|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
So is it Lynn's Amity design or Allen Wright's DPA-300B type design or a combination of both. Take the Amity and put a CCS on the cathodes?When will a kit, book or schematic of Allen Wright's new design be publically available? This is exciting stuff wish I was at VSAC where Allen was handing out schematics.
Any idea when some of his stuff will be more available?
Follow Ups:
Having built Allens PP2CS (Kt88 based diff amp), I'll try to offer you my take on it.Firstly, I didnt want to give my impressions at this early stage as I consider it to still be running-in. Also Im yet to do any tube rolling and major voicing. So I'll keep this short and do an indepth review a few months down the track.
Now I've heard KT88's before, Ive heard em in PushPull + Class A + no GFB before. Ive also been acquainted with a few high quality SE amps, albiet for short periods during in home auditions.
To sum it up quickly, the PP2CS is giving me SE type ease and delicacy, with PP type dynamics. On top of it, the tonality, detail and pure listening joy is jaw dropping. And just to top it off even more, its doing all this on a pair of speakers that are murder to drive (VAF i93s - seas excel based, 5 driver arrangement, impedence that goes down to 2.5 ohms at points, and can push serious air down below 20Hz).
Its probably too early for me to find any annoyances etc, but its looking good so far. So much so that I can *somewhat* comfortly say it will be the last amp I spend money on, or at least some incarnation of it.
Oh and on a totally side note. I replaced one of the monoblock chassis (hammond aluminium) with one made of Carbon Fibre. VERY nice improvement (im not talking about just looks). Cant wait for the other to be completed.
and the carbon fiber chassis sounds way cool, but where are the promised photos?
The Compact do sound good. I made one with help from the nice guys her on AA.Regards puff
- http://db.audioasylum.com/cgi/m.mpl?forum=tubediy&n=42896&highlight=pufff&r=&session= (Open in New Window)
I see that the two designs have more in common than they do in differences. Sure, maybe one is in favor of transformer splitting and the other guy is opposed to that. But really they are both wonderfull designs IMHO and I would be happy to have either one. I suppose at this time I lean towards Allen's but mostly this is because I feel it would be easier for me to build using triode connected KT90's. I am just not familar enough with IDHT's and all the tricks associated with them. If uncle sam is kind to me in regards to an income tax refund I hope to get started on one in the next month or two. I will certainly be looking for lots of help on the CCS end as I have no prior experience. I also admit to having the preconcieved notion that the end result will so surpass anything I have previously owned that I may not be able to do it justice in a review. I am hoping that some asylum members in my area will be willing to give it a listen and honestly critique and compare it, especially to a SET as I don't have one. In short I don't feel one could go wrong with either design and that's it more a matter of personal taste. In closing I would also like to say how greatfull I am to these gentlemen for sharing thier ideas and work.
> So is it Lynn's Amity design or Allen Wright's DPA-300B type design or a combination of both. Take the Amity and put a CCS on the cathodes? <Why not build both amd find out - then tell us!
> When will a kit, book or schematic of Allen Wright's new design be publically available? <
Although not (yet) mentioned on the site, kits of the dpa300B are available and are shipping right now. Previous clients have had a private offering, and I need to get my flow lines optimised before making any publicity about it.
The TubePoweramp CookBook will be some time. Schematics will not be published outside of the book (at least not by me) - but the quasi differential PP-1C is on my site and it doesn't take too many smarts to bring it to a full differential. And Gary P's amp map is on his site.
> This is exciting stuff wish I was at VSAC where Allen was handing out schematics <
Such events are always worth attending if you want to get close to the cutting edge - who knows what I, Lynn and/or many other designers might have let loose over a beer or two if asked the right questions...
Having the pleasure of Allen Wright as a guest last fall, I got to see the schematic of his new DPA-300B, and hear it compared to the Aurora and Gary Pimm's PP47 amplifier - a trio of interesting PP amplifiers.
The biggest difference between what Allen is doing and my work is a matter of personal style - Allen likes balanced cascode front end/driver circuits and RC-coupling, while I avoid cascodes and use only (wide-bandwidth) transformer coupling. This gives a different presentation to the sound, hard to describe in the usual hifi terms, but evident on direct comparison.Probably the biggest difference (and some Asylum members may disagree with me here) isn't the circuit per se, but the sonic character of RC-coupling vs interstage-transformer coupling. To my ears, these sound quite different.
When I started the Amity project in 1995, I had never heard an interstage-coupled amplifier before - hearing the Sakuma amps was still two years in the future. (I was quite surprised to discover the Sakuma amps actually sounded totally different than the Amity - something I would not have expected.)
The Amity was unusual because it was both my first vacuum-tube design project and also because I wanted to create an amplifier with the lowest possible midband distortion in each stage of amplification. DC coupling with active loads could also do that, but my previous experience at Audionics with direct-coupled transistor amps had made me wary of cascading failure modes which can be dangerous with DC-coupled designs. So the Amity was very much a "thought experiment" - the math said it would work, but I had no idea how it would sound. The Ariel was a thought experiment as well, but combining an MTM with a transmission-line was less radical than reviving a mid-Thirties WE design with modern parts.
Listening to the Aurora has been educational - I can now easily hear the character of the SE stage compared to a balanced circuit, which is why I'm going forward with a new amplifier.
Hi Lynn,Is your new amplifier a new design? If yes, can I get a sneak peek?
By the way, how good is Amity in the 20Hz - 200Hz range? I will need something for bass horns with 15" Altec drivers.
Regards,
Here's the schematic - being built by Gary Pimm as we speak. There is a separate B+ supply chassis for the quartet of damper diodes, both B+ chokes, and the main B+ transformer. The relatively small and low-emission filament transformers are on the main audio chassis.This limits connections between the two chassis to a pair of HV B+ connections (think RG-6 coax) and 120VAC to power the filament trans. Power switch, of course, is on the B+ chassis. This greatly simplifies chassis layout for the main audio chassis, since noisy high-switching is kept away from audio circuits.
nt
Lynn,how did you arrive at the value of the 40 uF bypass capacitor on the OPT primary centre tap?
Is this intended as the equivalent of the harmonic equaliser cap in the WE86?
If so, did you solve the second order intermodulation component of the Taylor series expansion of the gain equation in the general case or just for the parameters used? Or was it cut and try?
Just a guess, really. Both of the Gary's (Dahl and Pimm) have fooled around with the value of the CT-cathode bypass, and you know what, the value isn't critical at all. This is nothing like a parafeed cap (even though it looks like one), where tuning is a critical value dependent on trans inductance and Rp of the power tube.In this case, all that's happening is the bypass cap is actually *preventing* the amp from becoming a differential amp - omit it entirely, and yup, you have a diff amp - maybe not a great diff amp, no current source, just resistors. Why would I do that? Well, this is part of the style difference between me and Allen - I want the extra surge power that momentary Class AB offers, Allen doesn't. And yes, it sounds different. It can even be auditioned in an Allen-style amp - just have a switchable cap bypass for the current source, and listen for yourself.
So the bypass cap is there for flavor as much as anything - there's not much effect on the rest of the amp operation except peak current delivery capability, which is improved by the option of momentary Class AB operation.
Just out of curiosity, what's the difference between connecting the 40uF capacitor as shown on the current schematic and just running it to ground as in a conventional self-biased amplifier? Due to the large (120uF) filter cap used, the OT primary center tap is sitting pretty much at AC ground anyway.I'm a bit surprised, by the way, to see that you've dropped the "forced symmetry" approach. But, of course, if you prefer the sound this way, all the more power power to you (literally and figuratively).
Two-part answer: it's easy to come up with periodic tables of circuit topologies, but preferences emerge when you listen to them. I tried the forced balance in both driver and output stages, and didn't care for it. I do like the sound of the various WE bypasses, and am curious what Mark turns up in his research on the WE Harmonic Balancer. Nice doctoral paper there for an enterprising grad student (hey, it worked for Richard Small at the University of New South Wales).
As I see it, the common-mode loop path for a PP circuit is from paired plates to paired cathodes (see Loop Distortion Talk ). You can force this current to traverse the PS caps and the cathode-bypass cap, but this puts a severe quality demand on the main B+ caps - any coloration present in these caps will emerge in the sound, due to residual imbalance and the occasional excursion into Class AB. By omitting the cathode-bypass cap, using a very high-quality cathode-to-CT bypass, this current avoids the main B+ caps, and flows through the bypass cap instead.I admit I have a very strong aversion to cap coloration, while other audio designers see caps as sonically benign, or at least more benign than transformers. I don't - but this could be my years of messing with speaker crossovers and trying to find sonically acceptable caps. If I could buy 10 to 50uF Teflon caps, I would probably tolerate caps better than I do. Since I can't, I look at various topologies with cap-coloration in mind.
I buy the argument about bypassing the power supply capacitors, but can you tell me how much trouble there is due to fluctuating bias when the amplifier is driven into Class AB?
Dunno the answer to that one. Will measure, listen, and find out. There is also the option of adding a (traditional) cathode-bypass cap and adjusting the ratios between the CT-K bypass and cathode-bypass caps for best dynamic properties (pulse response, overload recovery, etc.)There is probably an advantage in keeping cap values modest, for faster recovery from momentary-overload Class AB events. These amps aren't intended for subwoofer use, where extended pipe-organ passages are a consideration. That realm is better served by fixed-bias Class AB (tube or transistor).
> I want the extra surge power that momentary Class AB offers, Allen doesn't <It's not that I don't want it - just that going full differential with a very tight CCS made such a HUGE improvement to the sound that I'm willing to sacrifice the occasional peaks that can be had from AB to maintain this sonic purity.
But I had a "beta tester" who made both the PP-1C (on my site) that does have this peak capability and the pure differential - and overall he preferred the non diff - loved it even - but NOT me!
And now I've gone to electrostatics the difference is even more pronounced, to me.
I recalled that you had looked at this issue as well and I was hoping you would be able to save me from wading through the maths.
Working out the distortion contribution of a bypassed PP amplifier should be interesting - all I know is that I prefer the sound of a bypass cap compared to pure differential operation, so I'd surmise the distortion spectra is indeed different.
The key appears to be harnessing the intermodulation between evens and fundamental to provide a cancelling effect for odds.Perfectly applied it should remove third harmonic. Even imperfectly applied it should bring the relative proportion of odds to evens down.
I've never really bought Hiraga's argument that a consistent distribution was more important than the amount of distortion as there is no "control" in his experiment. This technique holds out some promise of providing something close to a control.
Well, let's see, to really approximate the fabled WE Harmonic Balancer you need a smallish cap going from the CT of the input trans to the paired cathodes, and a several-times-bigger cap going from the paired cathodes to the CT of the OPT (or interstage).This lets you set the ratios of the desired common-mode modulations, or distortion-cancellation if you are clever - and I bet the WE guys worked out the math pretty exactly - this was the early days of feedback, after all, and the theater amps had the best engineering of the day. What's not so clear if all this sophistication was merely to reduce hum, or the more exalted goal of lower 3rd-harmonic distortion.
Another really subtle refinement would be separate bypass caps for each side of the filament circuit, thus removing the small inductive/resistive loss of the filament trans CT connection. You'd have to match the caps pretty exactly, though, otherwise you'd get hum from AC heating filament circuit. Then again a small trimmer cap could be switched into one side or the other to minimize hum.
Yes, Mark, you've stumbled onto something really quite deep and remarkable - a very interesting form of common-mode feedback that only acts on trans-coupled balanced circuits. Fun fun fun (to the closing tune of Red Dwarf).
Or at least I'll probably feel that way by the time I get to the bottom of this.The WE engineers' description of the Harmonic balancer says explicitly that its purpose is to re-inject even harmonics into the cathode so that the intermods with the fundamental form an odd harmonic cancellation signal.
I struggled for ages to get to the bottom of this and then realised that what they are effectivley doing is positive feedback of the "unbalance" component present at the OPT CT to the cathodes. The increase in even harmonics in each half is not so important because of the differential nature of the stage, so they calculated how to get a voltage divider so that the increased intermodulation products were of equal amplitude with the existing odd harmonics and out of phase.
I think that to get this right we need not only to take account of such things as the reactance of the heater windings as you said but also the effective output impedance of the OPT CT connection and the input impedances to the cathodes.
I have an idea that the cap connections to the centre taps of the input trans and the heater trans are about shifting the contributions of the current and voltage relationships of the harmonics - the signal fed to the cathode changes the total current through the diff stage while a divided signal goes to the input transformer centre tap so that the grid to cathode voltage change doesn't necessarily track this. This is as far as I've got, I need more time and clear space to worry it through.
Perhaps that shipwreck would be a good idea after all, and I quite like mango juice.
The Harmonic Balancer technology has lain dormant for seventy years - I'd say it's high time we brought it back.(rant on) It seems more than passing strange that PP has been done "by the book" for most of its history - one engineer mindlessly passing on stale Williamson circuits to another, without any understanding of what's really going on at all. At least by going back to the dawn of PP, we can do some audio-archeology and re-discover some of the fundamentals of amplification, before it was washed away in the deluge of feedback, "ultralinear" pentodes, Class AB, and then transistors. (rant off)
Pant pant, OK, I got that out of my system. Seriously, all-trans-coupled PP topologies lets us have some interesting degrees of freedom in circuit design, such as exploring the relationship of AC loop currents in both differential and common-mode terms (see Loop Distortion Talk . I'll probably be doing an expanded version of this talk at the European Triode Festival this December - the concept of loop currents still seems to fly over the heads of most audio engineers.
This design freedom is actually one of the main reasons I designed the Amity in the first place - I wanted to build the fabled "textbook" PP amplifier and listen to it, instead of the usual box of clever engineering tricks that was common in the mid-Fifties.
I had read some stuff about the harmonic balancer about ten years ago and not understood it at all. I've revisited it on and off over the intervening period, the biggest push coming after I read your stuff on differential and current mode loops about a year ago I guess.My major thrust is an attempt to understand the sources and effects of the non-linearities of triode devices. I think we've accepted approximations that are "good enough for Jazz" (as my classically trained wife is wont to say) for too long. Chief amongst these is the assumption that mu is a constant.
...is descended from the Amity, Raven and Aurora.The Amity is a 2-stage amp, all push-pull, which must be driven by a line stage.
The Raven is a push-pull line stage, with transformer input and output.
The Aurora, an integrated amp, is essentially a single-ended line stage grafted onto an Amity.
The new amplifier, Karna, is all push-pull. The first stage is essentially a Raven. The second stage (driver) uses 2A3's, and the output stage is like the Amity.
The Karna has had a number of working titles along the way (Aurora DHT, Concordance). It started out as an Aurora with a Raven-type line stage. The change to DHT drivers required power supply revisions, and eventually Lynn decided it was time to give the amplifier its own name, since it was really no longer an Aurora.
Karna, the amplifier, is named after Lynn's charming and brilliant wife. The amps are currently under construction. Living up to their namesake will be a tall order indeed!
If you have any pictures, more info, etc., you can pass it my way.
... Lynn is probably already working on a merging of his and Allen's designs. Which of them will get their circuit posted to the internet is kind of a moot point. I doubt very much whether either of them would make any claim about it being the final last design. Where's the fun in that?The idea of using a large shared cathode impedance to force differential currents is very old; I don't know when the long-tailed pair was first described but I'd expect it to be in the twenties. It can't have been very long before someone used it in an output stage, combining phase splitting and P-P power amplification.
The debate about shared cathode impedance in push-pull output stages has also been going on for many many years. I first heard it in the context of the original Williamson amp (the one Williamson designed, not the later Hafler/Keroes ultralinear version!). At the time, the debate was in the letters column of the late lamented Wireless World magazine - this before the internet of course :^) The original amp had (IIRC) unbypassed cathode resistances, on the grounds that it improved AC balance - a theoretical argument. The subjective counter-argument was that it sounded better with a big bypass cap. I don't recall that any real consensus emerged. But after all, the Williamson amp is derived from an older P-P triode amp, and I imagine that the debate started as soon as there were capacitors large enough to bypass the cathode resistors.
This is just an extension of the Class A vs Class B arguments. Today, with very high impedance current sources, it is possible to force differential behavior to great precision, probably 0.1% or better. Whether you call it super-deep Class A or differential or whatever, the current claim is that it is audibly a different animal that "ordinary" class A, with or without bypasses. The other end of the argument is that by using Class AB you can get twice the power from the same tubes and power supply, and four times if you go to Class B. Of course, in many systems, more power sounds better. Hard to tell the apples from the oranges here.
As a perceptive friend of mine often says, "Ho hum, another major breakthrough."
off of Ned's site?
http://www.triodeel.com/compact.htmland then look at the work/mods by Yeo at
http://diyparadise.com/simpleel84.html
what strikes me as utterly elegant about this cicuit is it's simplicity....
low parts count... singe ended voltage amp up front....
a friend of mine was showing me a McIntosh 240 amp schematic where they used the same "long tail" phase inversion (but not in output stage)...
looks like the big trick is tying down the cathode sufficiently...
and of course you could go wild and do parafeed PP output stage to get the benefits of much improved PSRR and etc....
but I think conceptually it's a competitor to the very fine work of both allen and Lynn.
MSL
I have a few reservations about using a multigrid valve sitting on a common cathode CCS. THe g2 curents will do some screwey things. Rig the valves up as triodes and it goes away. Or use single-grid triodes.
I raised this issue with Allen Wright( I think, memory not what I remember it being a while ago...). He had a KT88 amp with a CCS in the cathode circuit, and noticed the same thing, g2 stuff doing screwey things.
It is a PP amp, make no mistage about it. Rigged with the high gain 6336( 6528?), I suspect a really neat amp, capable of 15 watts would be possible. It would take a bit of drive voltage, and 10x the g1-a capacitance for that valve is no joke...
regards,
Douglas
I picked up some 6528s based on the massive construction and specs but find linearity is poor. A triangle wave has a nice curve to both rising and falling slopes.I've read that with a more complicated circuit and several feedback loops this tube sounds great. This isn't a good tube for simple no feedback amps.
The compact works because of the sensativity of and ease of drive of the 6BQ5. Other tubes that might work well are the 7591 and 8417.
I'd be interested in seeing how ballanced the UL design would be. Having a signal on the G2 that is effected by the other output tube might make it more ballanced than pentode operation.
Of course not bypassing the cathode resistor or CCS would increase output Z.
Hey-Hey!!!,
I have seen this arguement, re. the cathode resistance increasing output z. for a resistor it is a closed form sol'n. If the CCS is near infinite z then output z would also be ridiculous. This is not the case. I suspect that the contribution to the output z is mostly the input z of the other cathode. WHich would yeild an output z for the circuit much more in line with what would be measured. As far as measurements go, the freuency rsponse of one of these driving another triode( read: Miller ), the output z is in line with a single grounded cathode with small unbypassed resistance.
regards,
Douglas
I guess I'm regarding output Z like SE preamp designs from the RCA manual. I don't think I've read anything about P-P or diff output Z.
In pure differential amps the signal doesn't pass over the cathode to ground route at all, it circulates around both valves instead.
This is what I understood for diff amps driven with both grids. But in the case of the Compact one grid is grounded. The cathode current throught the resistor then creates the drive for the grounded grid tube.How is output Z determined with grounded grid diff amps? Is this even a true diff amps?
With a true current source in the tail, balanced drive, and ideally linear tubes, the common cathode connection sits at virtual AC ground halfway between the DC bias on the grids. With unbalanced inputs the cathodes go up and down with the common-mode signal. Because the tail current is invariant, there can be no common AC current. That means any change in plate current in the first tube flows directly and fully into the cathode of the second tube. This is true regardless of how you drive the grids (provided no grid current flows).If you change the current source to a resistor, some of the first tube's signal current gets diverted into the tail resistor. Looking into the second tube's cathode you see about (rp + Rp) / (u + 1). You can use the current divider equation to determine what fraction of the first tube's plate current winds up as cathode current (and hence plate current) in the second tube.
The impedance looking into the plate of the second tube should be the cathode impedance of the first tube (given by the same equation above) in parallel with Rk (the fixed cathode resistor), all times mu plus one.
It's helpful to think of this output section as a two-stage circuit consisting of a cathode follower driving a common-grid amplifier. A common-grid amplifier has the characteristic that the output current is equal to the input current. It provides voltage (and hence power) gain, but not current gain. You can also think of it as a factor-of-mu impedance transformer. It's less intuitive to try to figure the voltage across the cathode resistor and then to compute the second-stage output current using gm times Vgk.
I agree, this circuit doesn't really qualify as a differential amplifier.
Thanks,It sounds like the output Z of the Compact is missmatched between the tubes.
Actually, I think the output impedance is the same from both plates. The output stage is symmetrical, even if the drive isn't.
Henry wrote::::The output stage is symmetrical, even if the drive isn't.:::
if the "drive" (depending on how we define "drive") is not symmetrical... how pray tell would the output be "symmetrical"....
and everyone is using the term "symmetrical" as if it's an "all or nothing" proposition....
for istance the basic concepts of the 'compact' amp (advanced versions NOT.... NOT..... NOT.... the favorite whipping boy that was publshed in 1961) are practically the same as Henry's fave design he referred to earlier with the exception that his favorite is much more complex than an advanced compact design...
but back to symmetrical.... read the words written by Paul Joppa earlier in this thread..... in the box below....
*********************************************************************
The idea of using a large shared cathode impedance to force differential currents is very old; I don't know when the long-tailed pair was first described but I'd expect it to be in the twenties.
Today, with very high impedance current sources, it is possible to force differential behavior to great precision, probably 0.1% or better. Whether you call it super-deep Class A or differential or whatever....
*********************************************************************
the compact with a stiff impedance source in the cathode circuit does promote or acheive a very high degree of balance in the phase inversion....perhaps we should measure the phase inversion and the balance instead of hypothesizing...
again... anyone interested I would go to Yeo's site and check out ALL the compact pages... or check out several of the other posts in this thread where folks have built the amp....
and argue the semantics of what is or isn't a differential... but... such arguments appear to me to be semantical and to quote Joppa once again,
:::Hard to tell the apples from the oranges here.::::
too bad Norman crowhurst isn't around... or maybe he did write something on this... as he had one piece as I recall that was devoted to exploring several of the various phase splitting schemas...sorry, henry, but there is no clarity being added or obtained in this discussion at this point....
MSL
(1) The question was whether or not the output impedances (presumably at the plates of the two power tubes) are mismatched. The standard procedure for measuring output impedance is to ground the inputs of the unit under test (do you know why?) and to observe the voltage that results when a signal current is driven back into the output. If we ground the input of the amplifier there will be essentially no signal voltage on the first output tube's grid regardless of what is going on at the plates. From an AC point of view, the first grid is, therefore, at ground potential. With both output grids effectively grounded, there is physical and electrical symmetry in the output stage. It follows that the two impedances looking into the plates are equal.(2) Symmetry is a pretty clearly defined concept. Either a circuit is symmetrical or it isn't. With balanced drive, the unadorned Compact's output stage is symmetrical. Without balanced drive, the question becomes how much imbalance we are willing to tolerate (as a consequence of finite tail resistance) before we say the thing is asymmetrical. I'll leave it to you to answer that question.
(3) With a constant current source in the tail, the amplifier should have reasonably good balance and performance. There are some second-order effects that will impact the balance, but theory and practice say the amplifier should perform acceptably this way. In my opinion, this is a sufficiently evolved version of the circuit that it's no longer fair to call it the "Compact" (especially if you stick to the technology of the era and build the current source with tubes). Others may disagree.
(4) What you call "hypothesizing" is actually extrapolation from extensive practical experience. I have the confidence to say that theoretical predictions will closely match real-world performance as measured on the test bench. And I have some sense, within very broad limits, of what is likely to sound good, and what probably won't.
(5) Poindexter's amplifier avoids the symmetry and performance problems of the "Compact" with the addition of one tube stage, which I would hardly call "much" more complex -- certainly not compared to the complexity of a constant current source. Thinking about the two solutions, we might want to reconsider our priorities. If the goal is true simplicity and all-tube design, then Poindexter's circuit comes out ahead. If we can deal, conceptually, with lumping the constant current source into a black box, then maybe the modified "Compact" is more appealing.
I think we all agree that the basic "Compact" amplifier sucks. The only way to make it work properly is to complicate it. And then it boils down to choosing between several more advanced circuit possibilities. There is no one right answer. Personally, if I'm going to bother with solid-state current sources, I'd rather go all the way and adopt Allen's solution.
(6) This is not just semantic posturing. The technical issues are straightforward.
(5) With respect to Norman Crowhurst, frankly, many of his articles are less than lucid. I can do as good or better a job of explaining any of these subjects as Mr. Crowhurst, no offense intended to the memory of the dear old fellow. Obviously you feel my comments have been unclear or inaccurate, but reading your replies I don't find anything I want to revise, aside from my initial confusion (now resolved) due to your linking to the wrong schematic on Yeo's website.
If I've actually made any other mistakes, I will gladly correct them if someone will point them out to me. But, of course, just because someone says I'm mistaken, it doesn't mean I actually am.
Henry wrote:
::::(2) Symmetry is a pretty clearly defined concept. Either a circuit is symmetrical or it isn't.::::
Well, that certainly clarifies the concept with precision. In other words your saying it's 100% or nothing.
so that, if we have balanced output windings on a transformer... it's only balanced if it's 100% symmmetrical btwn the two halves from dc to light (frequency)... the cmrr would have to be, what? Infinity?
::::With balanced drive, the unadorned Compact's output stage is symmetrical. Without balanced drive, the question becomes how much imbalance we are willing to tolerate (as a consequence of finite tail resistance) before we say the thing is asymmetrical.:::::
Hello, Henry.... are you there.... no one has been talking about the "unadorned" compact (as published fifty years ago) as the archetype of high performance.... earth to Henry...
and... if we want to be more exact... who determines at what degree of imbalance a circuit is said to be "unbalanced"? Is this the best precision we can get?
are all interstage phase splitting transformers achieve perfect balance on the outputs? at what degree of error or what percentage of error do we then label it as "unbalanced"?
::::(3) With a constant current source in the tail, the amplifier should have reasonably good balance and performance. There are some second-order effects that will impact the balance, but theory and practice say the amplifier should perform acceptably this way.::::thank-you. so there MIGHT be hope for the patient. I sort of thought so myself...
::::In my opinion, this is a sufficiently evolved version of the circuit that it's no longer fair to call it the "Compact" (especially if you stick to the technology of the era and build the current source with tubes).::::henry.... how long ago were current sources and/or regulators used or developed? what about chokes? we could put a choke in the cathodes to increase the impedance as well.... chokes been around before 1961.... so does that keep the design authentic to the era?
your opinion as presented is laughable. so we shouldn't call any transformer made past the first or second decade of twentieth century a "transformer" since they evolved quite a bit over the next several decades in terms of design, materials, and winding practices...and a modern dry film teflon capacitor... since it is very different than the bathtub capacitors made in the twenties and thirties.... then the teflon ________ (blank) should not be called a "capacitor"....
and speakers of course to be true speakers would have to look like them cute projected trumpet speakers that pre-date the permament magnet type of drivers made several decades later.... so even an altec 802 cannot or should not be called a "speaker" or "driver" since it is an evolution of an earlier Lansing field coil type device.
nonsense....
::::(4) What you call "hypothesizing" is actually extrapolation from extensive practical experience. I have the confidence to say that theoretical predictions will closely match real-world performance as measured on the test bench. And I have some sense, within very broad limits, of what is likely to sound good, and what probably won't.:::::well, Henry.... if we're speaking of "confidences in our predictions" then perhaps I too could toot my horn just a little bit... but that gets into ungentlemenly public displays of self-affection...
but suffice it to say in this context.... that I had recommended to John Tucker the "compact" circuit initially and showed him where to get the schematic... and I showed this to John because the very first time I looked at it.... I realized that with the work that John has done in active loads, current regs, current sources, shunt regs and etc... I knew that John could bring this guy to life.... and I was right... John intuitively knew what was needed and saw also the beauty and elegance of the "compact" as an architecture and then set about building it and refining it.... to the point that one of the premiere recording studios in the US now has a pair in their studios....
::::(5) Poindexter's amplifier avoids the symmetry and performance problems of the "Compact" with the addition of one tube stage,::::nope.. has the exact same challenges... the circuit values will be different because they are different tubes and etc... but the level of refinement is no more or no less to make the phase splitter work at the front door than at the output stage....
:::which I would hardly call "much" more complex -- certainly not compared to the complexity of a constant current source.:::::
think outside your box, Henry. OK.... don't like a current source use a choke.... performs the same function in the sense of provides a high ac impedance....
what we need is a certain amount of impedance in the cathode circuit.... different devices will generate differing amounts of ac impedance.... one might use a resistor, a choke, or a current source... you pick the guy that best performs the function you need.
the rest is prostylsizing and religious dogma...
::::Thinking about the two solutions, we might want to reconsider our priorities. If the goal is true simplicity and all-tube design, then Poindexter's circuit comes out ahead.:::::
OK... Henry... we'll make the "ultra compact" all-tube... use another el-84 for your current source... siple enough....
again, you pick the right tool for the right job... and leave religion for your day off on sunday....
as much as I like Poindexter's.... it is more complex than the "compact" not less....
:::::If we can deal, conceptually, with lumping the constant current source into a black box, then maybe the modified "Compact" is more appealing.:::::Ok... this is a bit of what I have been aiming at in this discourse...
think of your "black box" as your "tool box".... we need to figure out which tools might work... a tube, a resistor, a ss current source, a choke, etc...
you keep on presenting one or two choices as though it exhausts the possibiity of outfitting our toolbox....
:::::I think we all agree that the basic "Compact" amplifier sucks.::::
well... how much imbalance does the stock compact have? what is our criteria on technical grounds for adjudicating a design as "sucks" because the phase inversion is not sufficiently good?
since you have never heard even the stock amp which you condemn then it must be on technical grounds that you have reached your "sucks" conclusion.... I'm just curious as to the obective criterion applied so that we might know when an amplifier does or does not suck...
::::The only way to make it work properly is to complicate it. And then it boils down to choosing between several more advanced circuit possibilities. There is no one right answer. Personally, if I'm going to bother with solid-state current sources, I'd rather go all the way and adopt Allen's solution.:::::go for it... it's an option available.... as is Lynn and Gary[s work... the compact is just another alternative...
and, again, you have blinders on.... you can only see in front of your nose.... how 'complex' is a choke for instance? what about a tube based current source? what.... throw gary Pimm's work out the window as to complex?
:::(6) This is not just semantic posturing. The technical issues are straightforward.;;;I believe I have addressed this contention already.
::::(5) With respect to Norman Crowhurst, frankly, many of his articles are less than lucid. I can do as good or better a job of explaining any of these subjects as Mr. Crowhurst, no offense intended to the memory of the dear old fellow. Obviously you feel my comments have been unclear or inaccurate, but reading your replies I don't find anything I want to revise, aside from my initial confusion (now resolved) due to your linking to the wrong schematic on Yeo's website.::::
first off... I did not link to the "wrong" schematic on Yeo's site.... the link provided was the starting point.... not hte ending point.... on the very same page whose url I posted were links to and an invitation to explore the 'refinements' that Yeo and Co. had experimented with.... so don't try to blame me that you could not have been bothered to read or simply peruse the whole page which I referenced....
regarding crowhurst... and your favorable alignment with the late crowhurst.... sorry, henry, as much as I like you... in terms of writing style and clarity your no crowhurst...
we haven't even gotten a good definition of "balanced" or "symmetry" yet... just some vague supreme court like enunciations like "I know pornography when I see it"....
:::If I've actually made any other mistakes, I will gladly correct them if someone will point them out to me. But, of course, just because someone says I'm mistaken, it doesn't mean I actually am.::::
good evening Henry.
-Mike
Your basic point -- that the "Compact" circuit, updated with a current source in the output stage, is a potentially worthwhile design -- is valid and defensible. Some of the things you've said to defend your position, once you felt it was under attack, are not. Out of respect for you, I've tried not to come back at you with my characteristic... vehemence.My original posting in this thread (look at it now if you forget what I said) was inoffensive, and based on a misunderstanding of what circuit we were discussing. Although I didn't have to, I went out of my way to post an apology to you and a clarification to Yeo. I don't see what it is that I've said that is so problematic, except that I've apparently bruised your ego. I've definitely added a lot of value to this discussion, whether or not you are savvy and discriminating enough to realize it.
Finally, regarding Norman Crowhurst, on more than one occasion I've stared at articles of his and tried to make heads or tails out of what he's saying. Crowhurst writes like an engineer. He's OK, but he doesn't have the gift for explaning things in a way that leaves you breathless with admiration for his writing skill. He could have used my help reviewing his articles. On the other hand, if you want an example of truly brilliant writing for tube hobbyists, it's got to be John Broskie's TubeCad website...
> > I've definitely added a lot of value to this discussion ...And I most definitely agree. Thanks a lot for your posts, I certainly learn a lot from them. Probably because of your clear, factual and technical style. Great stuff.
but I think he was still talking about driveing both grids not grounding one.I'm all in favor of jumping in and building circuits. Audio design is 30% ideas, 30% figuring out your idea, 30% building and tweeking and 10% luck. Sometimes building, tweeking and luck are 90%.
I have a Heathkit AA151s that I'll try this with. I'd start with the direct coupled Mullard design.
scott wrote:::::but I think he was still talking about driveing both grids not grounding one.::::
well... since he posted in a thread on the "compact" and made some comments about the trick being getting the impedance in the cathode circuit stiff enough....
and since I talked to him on the phone about this very same subject....
and since he has seen John Tucker's advanced "compact" design...and has heard it...
and has worked with the same recording studio customer of Mr. Tucker's.... alongside Mr. tucker on various projects for this same customer....
me thinks.... his commments were probably meant to apply to or include circuit genus's such as the "compact"...in other conversations with Paul (IIRC) he places the "compact" within a family of circuits or circuit variants that he would perhaps label "long tail pair"....
::::I'm all in favor of jumping in and building circuits. Audio design is 30% ideas, 30% figuring out your idea, 30% building and tweeking and 10% luck. Sometimes building, tweeking and luck are 90%.::::
interestingly, the more I look around the more examples of this strategy I see.....for instance, Cary Audio Design..... uses this same "LTP" ... with the grid of the bottom tube grounded and cathodes tied together.... with a current source in the cathode.... in several of their own PP amps.... for the phase inversion.... only diff is that they are doing it at the front door instead of in the output stage... and dennis uses a tube as the current source....
I have a nearly ten year old catalog from P&C Electronics (Japan) that has several PP amps wherein a "LTP" is used in the front end as the phase splitter and then coupled to the output stage.... again cathodes being tied together and the grid of one tube being at ac ground.... with resistor loading in the cathode circuit...
so like Paul Joppa was saying the concept isn't new at all...the goal or idea here being.... what potential does this circuit have if we study it, refine it, and work with it.... not.... does one instanitation of this concept as presented nearly fifty plus years ago exhaust the universe of possibility and development for this idea...
and whether or not you do this at the front door or in the output stage (invert the phase).... either the phase splitting is a "good" (however we might qualify good) or it isn't... the location isn't critical in and of itself....
::::I have a Heathkit AA151s that I'll try this with. I'd start with the direct coupled Mullard design.::::I liked your idea of direct coupling the single ended voltage amp to the "compact" output stage that you mentioned on this thread two or three days ago... it sounded like a really neat idea... could you possibly post up your ideas drawn out schematically (even if the schema is has black boxes (unspecified values in it)).
If your able to post attachments.... I would also like to see your proposed Mullard design....
thanks and carry on,
Mike
I did a quick and dirty rewire to use just the 6BQ5s UL connected and bias a little low 26ma at about 390v. Ballance was very close to equal up to over 30Khz where it started to shift phase but amplituded stayed good.The problem was power output power was very low at 1.2 watts or 9vp-p
and when I disconnected the 8 ohm load the output voltage shot up to over 23vp-p. Looks like output Z is high.
Hmm, see what you mean.
Not in any definition of a differential amp I've ever read.I would consider ta better description to be a "cathode coupled push-pull output stage".
The "Compact" could be quite interesting with a few mods. As it is, the output impedance with pentode output stage and no global feedback will be too high and distorted. So, needs either ultralinearizing V2, or trioded V2 or global feedback or some combo. Putting a CCS into the cathodes would change it to a real P-P from its present near SE operation and would increase power output back to two tubes. Its a neat idea for trying SE with a P-P transformer.
I Love Lynn's writing. And apparently Allan's stuff is super...(Am collecting parts for the Amity) However purely on a cost basis...I think I might build something like this ...before doing Lynn Olson's Amity.
I've seen Ralph's schematic before and overall I like it...Do I think it's the ultimate conceptually and practically? Probably not.
Is it elegant and simple.... in many ways yes....
one of the questions we should ask is.... does a transformer (or "all" interstage tranneys) split phase *better* than alternative strategies like a long tail pair or other tube based phase spitters?
to answer the above question.... we need to define what would make for a good "phase inversion".... what properties or technical grounds would we evaluate the different candidate's on?
and then... how does the interstage interact with it's source (the anode of the tube driving it)? does the interstage draw any exciting current? Is this exciting current itself a source of distortion in the iron? what is the "actual" impedance presented to the anode of the tube? Is it a steady or stable impedance? Or is it reactive? does the primary "show" a different impedance at different frequencies?
then.... look at the output side of the IT phase splitter.... what is it's bandwidth? How much disotortion? does the secondary have a high degree of AC balance to ground? Over what frequency range does this balance hold? what is the degree of balance? Any phase shifting at either end of the audio spectrum?
I'm not necessarily saying that IT phase splitters can't bee good... but SIMPLY because it's a tranney is no assurance that it is a good quality phase splitter.... which seems to be the mode of thinking amongst some in the diy community...
again... I like ralph's concept/architecture.... I would parafeed the IT (to get the benefits of having the AC signal and the dc supplies on separate paths and to enhance the PSRR) and I would use a grid choke in place of the 47kohm grid resistor of the 417A.And if I was going to use a trans for phase splitting I would be darn tempted to look at using a ct autoformer phase splitter with adequate interleaving btwn the driving (source) and the driven (load) halves of the autoformer... sort of like Peerless did in the tl-404 autoformer output that they designed for MGM recording studios back in 1946...
and if I were on a budget.... I would doubly look at the 'compact' architecture.... so that I could take my finite dollars and put them into an absolutely dynamite pair of output tranneys.... as opposed to watering down my iron budget and having to purchase two sets of lesser quality ironware... so that I could say that I have an interstage tranny...
I'm sort of from the old school... use as few pieces of iron as possible in the signal path but of the highest quality possible...
but again, I applaud ralph and admire his circuit on many different levels....
msl
I've done alot of work with this design in the 90s. The devil as we all know is in the interstage.Is this the best phase splitter? In a perfect world, yes it is. Here on Earth Williamson, Mullard, Wolcott and concertinas are much easier (maybe not Wolcott) to build and produce wide bandwidth phase splitting.
Which sounds better? Take 20 different speakers in 20 different rooms with 20 different audio nuts and you'll get no clear answer.
AS for me I liked the interstage better than a Sowter 8920 input phase splitter. I also abandonded transformer phase splitters for Williamson, Mullard and concertina.
Is there a Wolcott schematic for the public to see? I have read thier white papers and am curious to say the least. I think I get the basic idea but I do wonder how they carry it out. Any idea who makes thier output transformer?
Wolcott will send you one for a reasonable fee. Don't get your hopes up. The Wolcott is a crazy circuit with all kinds of cross coupling and looks to need a lot of tweaking.JC Morrison posted a good discription of phase splitters but I can't find it. If you find it link me there is one design that is close I'll tell you which one.
Van Scoyc?
Not exactly. The jist is similar but he uses 6BM8 with the triode in a diff like config with ccs and the pentode as a common cathode and there is crosscoupling and trim-caps etc. You have to see the schema. Call up Henry and ask him for Optimation 250M schematics. Start at www.wolcott.com for the phone number. I hate to say it but the ITT version was the most dissapointing amplifier I've ever heard. I got rid of all of my Optimations and documentation shortly after hearing a set for the first time. I hope the Wolcotts are much better, I've never heard them
Surely not :-))
This is the one...
NT
Hi Scholl,I agree, the problem is coming up with an interstage that can handle phase splitting *and* handle the voltage drive requirements of a DHT output stage. Finding none, we gave up on that approach.
Instead, we use the superb LL1676 input transformer from Lundahl, which gives beautifully symmetrical phase splitting and excellent bandwidth (~8 Hz - 140 kHz). This allows the use of push-pull interstage transformers in the subsequent higher-level stages. The Lundahls have also served us well in these applications.
The only trouble is that the LL1676 isn't easy to drive, so must be fed from a low impedance. Gotta get that new Raven done!
A whole 'nother layer of inspiration.Gary Dahl(of Amity-Aurora/PP47/and endless CCS fame) designed the very first version of this amp.
From Ralph Powers' on website:
I've always subscribed to the "less is more" and the "keep it simple stupid (KISS)" school of audio design. Luckily both of these prerequisites came together beautifully in a schematic that was given to me by a serious audio experimenter, Gary Dahl. This circuit is exceedingly simple and pure, with no capacitors or resistors in the signal path. And it requires only a hand full of parts to build. I knew the first time I laid my eyes on this schematic that I HAD to build this amplifier.
But, I bet there's something, somewhere, during 20's thru 40's that looks eerily similar.
Hi darkmeobius,While it is true that Ralph Powers' PP-6B4G was derived from a design I gave him, it is Gary Pimm who is the designer of the famed PP47 amp, and fountain of endless CCS's.
I am a business partner with Lynn Olson, and built the Aurora amps shown on our site, and also did the PP2A3 revision of the Amity to answer concerns I had about the Ralph Power PP-6B4G design.
And yes, you're absolutely right--these designs are straight out of the '30's, with a few niceties that weren't available at the time.
I think Peter is into ham stuff. I ask because I have a power tranny from an old heath scope that I have heard Peter gets a lot of money for. I am curious as to how to best use this iron and thought if perhaps you were related you might have an idea as to why it has worth and it's possible hi-fi applications.
.
"So many tubes, so little time..."
I had the other Gary's website open to investigate CCS's for Yeo's "Simple EL84" and somehow merged the two of you.Silly me, I should know you can't watch SciFi Channel, study amp designs, and type at the same time.
thanks darkmobieus for the post in which ralph is quoted as saying;
From Ralph Powers' on website:::::I've always subscribed to the "less is more" and the "keep it simple stupid (KISS)" school of audio design. Luckily both of these prerequisites came together beautifully in a schematic that was given to me by a serious audio experimenter, Gary Dahl. This circuit is exceedingly simple and pure, with no capacitors or resistors in the signal path. And it requires only a hand full of parts to build. I knew the first time I laid my eyes on this schematic that I HAD to build this amplifier.::::
allow me to extend kudos to gary dahl!!!! good job gary....
Just one note... though we strive for simplicity... and on first blush I too am or want to be a church member of the KISS congregation...
but least we never forget.... capacitances and resistances are a natural part of the electronic universe.... L's, C's, and R's...
what's hidden in the schematic (by hidden I mean not shown schematicaly) is dozens and dozens of capacitors... capacitances from turn to turn, from layer to layer, from each end of each layer to the core (ground), from primary to secondary, and etc.... the 'capacitances' are small in value.... but they are there none-the-less in any transformer...
as well as resistance.... most simply is the resistance of the copper windings.... this is a resistance.... then add in the resistances (losses) of the magnetic core...
so, even where we don't have any obvious R's and C's.... they are there... just not shown unless you see a schematic of the transformer with all of the details shown....
point I want to make is that capacitance and resistance can be as "pure" or "impure" as inductances... it's what can we do with L's, C's, and R's that counts.... and each used well can produce wonderful results....
and each of these three building blocks can and do have their own gremlins...
but.... point taken and admired that Gary's circuit has what looks to bee a fair amount of woo...
cheers,msl
I agree with Mikey--capacitances and resistances are most certainly present in any transformer, which is one of the reasons why the choice of transformer is such an important decision. The capacitances, especially, make a significant contribution to the "voice" of the circuit...for better or for worse.Speaking of which: Mikey, those custom S-240-A's you made for my Auroras are still singing beautifully!
I built this amp (original compact as on Neds site) over a year ago to resolve whether or not it was really a super elegant implementation of a single ended design using a more commonly available PP OPT. The "speculation" was that the undriven 6BQ5 was actually a current source designed to null out the DC from the driven 6BQ5 thus resulting in zero net flux in the OPT under idle conditions.Testing indicates that the output stage is nothing more than a phase inverter capable of delivering power. It is very poorly balanced WRT drive on each phase due in large part to the relatively miniscule "tail". Power output is also a bit low by 6BQ5 standards at about 7W @ 1KHz with very obvious clipping...this using a regulated PS for optimum measured performance. THD was not measured but I expect it to be high for a PP design because the "undriven" 6BQ5 operates at about 50% of the drive of the "driven" 6BQ5. One interesting improvement should be replacement of the common cathode resistor with a CCS but I never got around to trying it this way.
The amp sounded just OK but after a few weeks was dismantled because I needed the OPT for another project.Overall, I don't believe this amp will satisfy the discriminating listener over the long term. However, it's simple and low enough cost that nothing is lost if one wants to try it. IOW, YMMV.
Hi Steve:you wrote:
::::I built this amp (original compact as on Neds site) over a year ago to resolve whether or not it was really a super elegant implementation of a single ended design using a more commonly available PP OPT.::::
the "original" as published on Ned's site.... is a starting point only... it shows, imo, the 'architecture' of the concept... the 'black boxes' that then must be filled in and optimized...
::::The "speculation" was that the undriven 6BQ5 was actually a current source designed to null out the DC from the driven 6BQ5 thus resulting in zero net flux in the OPT under idle conditions.::::
and the speculation proffered was worth not much more than speculation, imo. very little insight and understanding into how the circuit actually works.... heck.... if I remember right... some folks were saying it was a single ended amp....
::::Testing indicates that the output stage is nothing more than a phase inverter capable of delivering power.:::::
OK.... so now we have a phase splitter (necessary for PP operation) and a stage capable of delivering power.... two necessary ingredients or functions electrically for a PP amp....:::It is very poorly balanced WRT drive on each phase due in large part to the relatively miniscule "tail".::::
ok... so now you know what needs to be 'fixed'..... are there any solutions/optimizations available?
:::::Power output is also a bit low by 6BQ5 standards at about 7W @ 1KHz with very obvious clipping...this using a regulated PS for optimum measured performance.::::::
regulated or not.... all amps will clip at some point... again, you need to optimize the design for the performance requirements/goals that you set...
there is nothing inherent or necessary in using an EL84 output stage.. you could use a 6550, 845, 2A3, 807, etc.... and vary the power output accdg to needs and tastes.....
::::THD was not measured but I expect it to be high for a PP design because the "undriven" 6BQ5 operates at about 50% of the drive of the "driven" 6BQ5.::::
second harmonic distortion should be largely eliminated in the output stage as per normal PP operation if your balance is good....
:::::One interesting improvement should be replacement of the common cathode resistor with a CCS but I never got around to trying it this way.::::
so you need to improve the balance..... look at yeo's site for some ideas and insight into how this can be accomplished.....
and there are actually many or several approaches... for example you could use a 'plate choke' in the tail... and get substantially higher ac impedance than the 130 resistor provided as in the original schematic.
get creative.. put on that thinkin' cap... beauty is there...
::::The amp sounded just OK but after a few weeks was dismantled because I needed the OPT for another project.::::in other words it was never fully optimized.....
:::::Overall, I don't believe this amp will satisfy the discriminating listener over the long term. However, it's simple and low enough cost that nothing is lost if one wants to try it. IOW, YMMV.::::::
sorry... but your conclusion is a laugher...
one of the biggest, best and well known recording engineers in the states commissioned an amplfier that was built on the basic architecture of the "compact"..... and I would guess his ears have some pedigree....
the designer of the studio's commmissioned amp... is also a seasoned builder.... and it more than satisfied his ears....
of course, he went beyond... the circuit as published on Ned's site.... and refined, refined, refined it... and used some pretty advanced current sources and shunt regs and etc...but... it's much to early and counter to the evidence that this puppy.... is incapable of being anything less than a stunning performer....
MSL
You wrote:
sorry... but your conclusion is a laugher...Sorrow not necessary! However, I do believe you dismiss my conclusion too readily.
My conclusion is based on what I built and listened to: the Compact Amp as posted on the Triode Electronics site. The conclusion is not based on an optimized version of the amp or on what a golden eared recording engineer designed and built. I stated this clearly and unambiguously. Regardless, it could be argued that the "optimized" versions are no longer the "Compact Amp" but that's for another day.
Bottom line: I stand by my conclusion that as described in the original schematic, the "Compact Amp" is anything but a stunning performer either technically (significant output stage drive imbalance) or sonically (my parts, my speaker my ears, my music).
However, in spite of my less-than-totally-positive experience with the original design, I'm still interested to hear what you liked so much about the original version you heard (or built?).
SO
Hi Steve:you wrote;
::::My conclusion is based on what I built and listened to: the Compact Amp as posted on the Triode Electronics site.::::one amp on a fifty plus year old design iplementation who had as an explicit goal "low cost" and "ease of building"....
na... before condeming the whole architecture you should really try to optimize the circuit and refine it to squeeze all the goodness out of it you can....
imagine if we had only built single ended amps with the exact same parts, methods and strategies that were used eighty years ago....
remember all the folks who did harebrained clumped together SE amps and then said they sucked?
Or like saying all single full range drivers stink because the 35 inch RS jobbie ya built wasn't so hot.....
ya gotta do more than copycat a fifty year old design without any common sense widely recognized improvements in parts, methods or strategies....
::::The conclusion is not based on an optimized version of the amp or on what a golden eared recording engineer designed and built. I stated this clearly and unambiguously. Regardless, it could be argued that the "optimized" versions are no longer the "Compact Amp" but that's for another day.::::::
it's the architecture the CONCEPT.... not the exact circuit... so all of today's single ended amps aren't REALLY the same thing because they are built differently than specimen's that might have been designed and built eighty years ago?and... no need for the namecalling.... "golden eared".... as though we should trust your ears moreso (or less or differently) than this fella's ears and aesthetic judgement?
or the OTHER professional who designed and built this evolved compact type architecture for the recording engineer? His experience and aesthetic judgement must also be dismissed?
that's what I find to be so laughable....
::::Bottom line: I stand by my conclusion that as described in the original schematic, the "Compact Amp" is anything but a stunning performer either technically (significant output stage drive imbalance) or sonically (my parts, my speaker my ears, my music).:::::now try to optimize the circuit... go visit Yeo's website for some good ideas.... get creative... solve problems... refine circuits.... work with something for more than a week and ONE implementation.....
study the phase splitter in the mac 240.... it's comparable to the phase splitter in the compact architecture.....
read Paul Joppa's post above..... this ain't new.... but maybe there are some 'newer' ideas and strategies that could be implemented in this architecture....
Not to belabor the issue but..1. I did want to point out that I'm more-or-less dismissing the original compact amp design, not the basic concept. However, to my way of thinking, it must be implemented with a CCS (more active components) or a good quality inductor (more expense) in the common cathode circuit. In either case the thing then deviates from the raison d’etre of the original design.
2. At a couple of places in your discussion you reference the phase inverter of the MC240 as another implementation of the same basic circuit. If you look at the circuit you'll notice that Mac used an 18K resistor for the common cathode tail (much larger than ~100 ohm) AND unequal plate resistors. The inverter is followed up with a differential amp of sorts and the whole thing is wrapped in lots of global inverse feedback with some local positve and inverse feedback thrown in for good measure. The point is that Mac realized that a simple implementation of a long tailed phase inverter was inherently unbalanced and they took special measures to balance it. Their implementation bears only a passing resemblance to the compact amp (IMO of course).
Has this discussion become too philsophical? Possibly.
steve writes:
::::1. I did want to point out that I'm more-or-less dismissing the original compact amp design, not the basic concept.::::
that was my only real sticking point with ya!!!!
to appreciate the context of the orginal circuit shown on Ned's website... recall that the author is pretty explicit about pointing out that his circuit is tailored for the "cheapstake"....
and I only wanted to point out that this is equivalent in a sense to reading an article done by Eric Barbour in glass audio... on how to build a $49 (or was it $99) singe ended amp... yeah... give it a try... but if it doesn't 'ring your musical chimes' don't conclude anyhing universal about single ended amps from it.... or take it a step further.... get some purpose built single ended outputs and optimize the published circuit till ya get it to sing....
that's what the COMPACT offers... it is a starting point....
:::However, to my way of thinking, it must be implemented with a CCS (more active components) or a good quality inductor (more expense) in the common cathode circuit.:::I suspect your right here... it does need to be refined a bit... but... hey.... what a great starting point....
a simple circuit... with a ton of flexibility.... again, see yeo's site.... you can UL the critter, run it as a pentode, strap it as a triode....
or if you take heed of sector7g's post.... you can use a pair of triodes in the output stage....
gotta run it (as Paul Joppa points out in his contribution in this thread) pure class A.... but what the hell...
it's still an architecture with considerable flexibility.... and simplicity...
::::In either case the thing then deviates from the raison d’etre of the original design.:::::
I don't see this as being the case any moreso than using a plate choke on top of 45 tube in a SE amp makes the amp any less "single ended"... these are simply circuit refinements and enhancements that if they make the amp funciton better and sound better... then they are worth doing... and should be done....
::::2. At a couple of places in your discussion you reference the phase inverter of the MC240 as another implementation of the same basic circuit. If you look at the circuit you'll notice that Mac used an 18K resistor for the common cathode tail (much larger than ~100 ohm) AND unequal plate resistors.::::I do remember seeing the 18K resitor in the cathode circuit...yep....
and that's the point... author A (compact design author) used a different resistor value due to the biasing requirement of his circuit.... the small value resistor doesn't optimize the phase splitter function of that circuit....
mac... on the other hand.. used a larger value resistor in the cahtode circuit of their phase splitter... and apparently it worked reasonably well... or so I assume
but I don't recall seeing the unequal plate load resistors in the mac circuit.... I will loook again at the circuit on monday when I am back in the shop....
but my reference to mac was not about all the nitty-gritty details and choices of component values.... nor that's it's implementation was exactly like the compact.... it was to simply show that the concept has been used by others with some degree of success...
::::The point is that Mac realized that a simple implementation of a long tailed phase inverter was inherently unbalanced and they took special measures to balance it.::::
if by simple you mean not optimized.... this seems to me to be getting to bee a semantical argument.....
you say that a "proper" or a "refined" or an "optimized" circuit is proof of something being fundamentally askew.... that the need to "optimize" is itself proof of some shortcoming.....
as a designer of tranneys.... I know that optimizing is the whole secret to success.... I imagine that circuits respond likewise to wise refinements and enhancements...
:::their implementation bears only a passing resemblance to the compact amp (IMO of course).:::::well.... let's see.... the top tube is driven by the grid... the cathode of the top tube is tied to the cathode of the bottom tube.... the grid of the bottom tube is tied to ground.... the output of each tube comes off the plate and is the source for an external load...
geez.... those points look exactly the same in the mac phase splitter and the compact's phase splitter...
guess it's all in the eye of the beholder....
best of,
Mike
I'm sorry to say it, Michael, but personally I don't see what's so great about this amplifier. A pure single-ended amplifier is simpler, and a proper push-pull amplifier has much better balance with hardly any more complexity. I can't speculate on the sound, but technically the "Compact" design is flawed. For simplicity, I think there is a lot to be said for Bas' circuit. And I think Allen's circuit is pretty nice, too, and it offers much better balance -- which seems to be the point of push-pull after all.(Does this mean you hate me now?)
Hi Henry:by now i've come to expect that if I say orange you will say blue... just for the thrill of disagreeing.....
I'll address a few of your points none-the-less;::::A pure single-ended amplifier is simpler::::
how so? especially when you consider an output stage w/ paralleled output tubes... i don't see that it's really much simpler at all...
::::...and a proper push-pull amplifier has much better balance with hardly any more complexity.:::::
hey... basically this arrangement worked reasonably well for McIntosh... look at the phase invertor in the 240 monoblock amp.... essentially the same thing except that it's done several stages earlier...
and how are we defining what constitute's "better balance"? and is your criticism of the "compact" architecture based on the ONE instantiation of the concept done nearly fifty years ago.... or a modern implementation of this architecture as say done by John tucker for stubblebine recording studios?
you guys crack me up sometimes.... I realize it's sport for some....
::::I can't speculate on the sound, but technically the "Compact" design is flawed.::::Really. so mac didn't know what they were doing either? OK.
::::For simplicity, I think there is a lot to be said for Bas' circuit.:::::do you mean ralph power's circuit that bas put up? I was going to make some comments under separate cover on that....
::::And I think Allen's circuit is pretty nice, too, and it offers much better balance -- which seems to be the point of push-pull after all.:::::
funny... they may be more similar than you think....
and... again... "offers much better balance" than what... the compact as drawn in ONE example fifty years ago.... or better than this concept executed with a current source in the cathode circuit? any proof for the above claims?
you guys are funny....
The MAC circuit and other LTPs phase splitters work a little better than the Compact because the LTP is DCed to the previous stage and floates on a larger cathode resistor.And this little tidbit of a refinement for the Compact could be all that's needed to get much better balance. Simply direct couple the 6BQ5s to the input\driver tube (my vote would be 6sn7 run low and hard) and increase the cathode resistor to get proper bias. Heck the larger B+ would be easier to come buy than the one speced in the artical. The grid on the non-driven tube would have to be cap coupled to ground and connected to the driver tube's plate via a large resistor. The only draw back would be the cathode resistor would have to be hefty 12watts min 20 watts typical. In this case the CCS would have more room to work, the choke wouldn't be the best choice. Another benny would be no coupling cap but the grounded grid cap usually is larger .47uf vs .1uf.
Back to the MAC. This design, as we all know, uses cathode windings in the output tubes plus an overall feedback winding that is directly in the input tubes cathode circuit plus sundrie feedback loops in other stages. With all of this NFB the dang thing can't be anything but ballanced. The low turns ratio and bifilar wound primary legs ain't no bummer either. With all of these great inovations some folks still like 1920s designs better, go figure.
A better example would be Mullard.
Later, Scott
Mikey, I'm not just disagreeing for the thrill of it. You asked me to have a look at this circuit and my comments reflect my sincere and carefully considered opinion.(1) With respect to simplicity, if you compare this circuit to a single-ended amplifier, the "Compact" is definitely more complex. I don't necessarily mean in parts count, but surely in terms of function. In the single-ended amplifier, the output tube grid voltage is converted in a straightforward manner to a current flowing in the transformer primary. In the "Compact", the current in the first output tube is split between a resistor and the cathode circuit of a second tube, connected as a grounded-grid stage. There is some cathode negative feedback applied to the first tube. The two dissimilar currents produced these tubes operating under very different AC drive conditions are summed by the output transformer to produce the final output.
Parallel single-ended isn't really comparable. In PSE, both tubes run under (theoretically) identical AC and DC conditions. From a conceptual point of view, they are a single tube, just split into two envelopes.
As I think everyone agrees, the output stage in the "Compact" serves both phase-splitting and amplifying duty. I believe it's fair to call this a bona-fide complication.
(2) The definition of balance is well-understood by circuit designers. In a perfectly balanced circuit, the sum of the two signals is zero. Any departure from perfect balance results in a non-zero sum when the phases are combined. We can represent the imbalance as a common-mode residual superimposed on the perfectly balanced phases. This is first-year electrical engineering curriculum.
If you look at the EL-84 data sheet, two volts change on the grid gets you about 20mA change in plate current. This means the cathode impedance is going to be no lower than about 100 Ohms. With a 135 Ohm common cathode resistor, about half the signal current flowing in the first output stage valve is going to be shunted to ground instead of flowing into the cathode of the second tube (and therefore into the other leg of the output transformer). This means the circuit is fundamentally unbalanced and it can't be improved without increasing the impedance of the "tail" circuit (or resorting to other, more extreme fixes).
An important measure of a phase splitter is its common-mode rejection ratio (CMRR), the extent to which the output remains balanced even when the drive signal is unbalanced. If we lift the grid of the second tube from ground and connect it in parallel with the first tube grid, we will observe considerable AC plate current in both tubes when a drive signal is applied. This is because the stage has a poor common-mode rejection ratio. Because of the asymmetric driver, the output stage of the "Compact" sees a large common-mode input signal, which is propagated to the plate circuit for the output transformer to deal with.
In the MC-240, there are three well-balanced differential stages following the phase splitter. Each stage (assuming matched tubes) has a high CMRR, so that by the time the signal reaches the outut transformer any imbalance due to the phase splitter has been thoroughly canceled out. As well, the 10K tail resistor more closely approximates a current source than the 135 Ohm resistor in the "Compact" (even taking into account the difference in cathode circuit impedances between the two circuits). And finally, the plate resistors in the phase splitter are deliberately mismatched to correct the imbalance. So, in my opinion, this part of your argument doesn't really apply. McIntosh knew what they were doing, and it was very, very, VERY different from the design of the "Compact". In fact, the McIntosh amplifier is arguably the very antithesis of the "Compact" -- just about as complex a push-pull topology as you can imagine.
With respect to John Tucker's implementation, I can't comment without seeing it. The version of the amplifier on diyparadise doesn't really make any changes to deal with this balance issue.
(4) There's no question that in spite of its other limitations, a circuit using a quality interstage transformer will yield orders of magnitude better balance at the output transformer primary than the "Compact" will.
(5) Philosophically, the "Compact" and Allen's design are quite different. The essence of Allen's philosophy is balanced signals at all points. The output stage does not serve phase-splitting duty in his circuit; this is handled by the cascode differential driver. The current source in the output stage is there for a number of reasons that have been discussed elsewhere. While it's true the "Compact" strives to have a differential output stage, the implementation (of a diff amp) is so crude, and the driver stage is so different, that it's hard to claim this circuit is similar to Allen's.
There's really nothing "funny" here. Just some straightforward and unemotional engineering observations.
There are about a thousand ways to design an amplifier that sounds good and communicates music well. Sound quality for the sake of audiophile bragging rights isn't my priority. On the other hand, I enjoy the theory and practice of circuit design. I view a circuit as a statement of a design concept. The concept of push-pull is balance. The "Compact" doesn't express that concept very well. Nor does it particularly express the single-ended concept. It's some kind of hybrid. I think the main point of this circuit is "cost savings". Otherwise, it doesn't say much to me conceptually. That's not to say that it's "bad" or "wrong", or that it doesn't play music well. I just don't find the concept as noteworthy as you do.
-Henry
OK, Mikey, I apologize for being hard on you. As I said below, I only looked at the page you linked, and not the "refined" version of Yeo's circuit.Given limited time and resources and an infinite number of circuits to build, I try to pick projects that are a clear statement of some design idea. If simplicity is the idea, then it should be very simple. If push-pull is the idea, then it should be very well balanced.
Sometimes you can make a circuit conceptually simpler by adding a bit of complexity to it. When you strip a circuit down to a minimum number of elements, each of those elements has more work to do and the non-ideal characteristics are more clearly exposed. This makes it harder to predict how the circuit will perform as a whole.
My feeling is, if you put a good phase inverter in front of a push-pull output stage, separating out the two functions, then it's actually clearer and simpler from a design standpoint how the circuit works. Other people may feel differently.
I hope it doesn't bother you too much that I sometimes challenge your ideas in public. It's not personal -- just trying to state my understanding of the subject as clearly and completely as possible.
but I get a lot out of your posts. So I hope you never stop challenging. IMHO one shouldn't feel attacked when asked to defend thier beliefs. Especially when asked to do so by one with knowledge of the subject at hand. Good debate is what makes this forum what it is.Your comment about a circuit appealing to you from a conceptual aspect rather than how it might actually sound struck a chord in me. I had thought I was trying to keep an open mind but saw that wasn't quite the case because when I looked at the compact I pretty much dismissed it without much thought in spite of the source.
Anyhow, not trying to blow smoke up your a** but I think you have a wonderfull writing style and clearly lay out and make your points. If you haven't written a book on tubes I think you should....and if you already have do let me know the name. I have only seen things along the line of short "interviews".
hi henry> A pure single-ended amplifier is simpler, and a proper push-pull amplifier has much better balance with hardly any more complexity. I can't speculate on the sound, but technically the "Compact" design is flawed. For simplicity, I think there is a lot to be said for Bas' circuit. And I think Allen's circuit is pretty nice, too, and it offers much better balance -- which seems to be the point of push-pull after all.
i think we are all aware that the limiting factor to a push pull amp's performance is none other than the phase splitter. almost all phase splitters have problems of their own.
the "compact" doesn't have such good phase splitting performance either but the evolved "simple el84" on my website has a much much better performance. and the simplicity of a single ended amp!
there is nothing flawed about this design. i have built it and numerous guys have built it and this amp has been running in my system for more than a year already.
> I can't speculate on the sound, but technically the "Compact" design is flawed.
don't speculate on the sound. don't dismiss the design. build it and listen to it.
all this talk without even listening to one in action (with constant current source at the cathode) is moot.
What's the URL? I'd like to take a look at your amp.Regards,
Poinz
Try www.diyparadise.com
Yeo, I didn't realize from looking briefly at the URL Mike posted that there were several subsequent pages of development of the amplifier on your site. Your listening impressions of the "uncomplicated" version (comparable to the original "Compact") are in line with my expectations for this circuit: "The sound is veiled, muddied. The highs are okay but the bass is quite rolled off. Also, you need to crank your volume pot quite high..."With a constant current source in the tail, the amplifier is quite a different animal, and, yes, now similar in one respect to Allen's design. (Of course, with all those active devices hidden inside the LM-317, one could argue the amplifier isn't really so simple anymore.)
Personally, I'd be inclined to try this amplifier with and without a differential input stage as well. My favorite design for simplicity and elegance is Poindexter's 6V6 amplifier. I might ask alternatively how his amp would sound with a CCS output stage...
-Henry
that are anywhere near a signal path\. And NO WAY use them for differential cathode CCS's - been there and it took some time to calm down after ripping them out with my bare hands - just horrible!
Poindexter's amp is really neat - but I expect it would jump up at least one major level with a JFET CCS on the front end, and again with a simple MOSFET CCS on the output stage. pretty easy to do...
Your assessment of Poindexter's amp makes sense to me (on paper at least since I haven't built or listened to it). It strikes me as a design that is as simple as it can be and still have any chance of operating the output stage in a balanced condition. Adding a CCS to the "Compact Amp" increases its complexity a little or greatly depending on how one views the CCS itself.I also perused Yeo's site and was surprised at the use of the LM317 as a CCS. The sonics of this device are routinely criticized when used as a conventional 3 terminal voltage regulator in the PS of audio equipment. Most of the criticism seems centered around poor performance above a few KHz IIRC. OTOH, my own experience with the 317 (and the 337) as a voltage regulator has been good so maybe it works well in Yeo's app in spite of what the PS gurus claim.
it's the very same mindset that first dismissed single ended....then said parafeed couldn't work....
then said the compact couldn't work...
no experience, no building, no listening, no creativity.....
FWIW Paul likes JTs amps enough (after I recommended them, ahem) that he ended up ordering a third pair this year for the subwoofers on his Magicos. You'll never get me to say they are better than SE, but they are pretty close, and if you need 20W on tap they are a super way to go.Hey Mikey, speaking of interesting new ideas I got one for ya-
Damascus steel bell ends.
Damascus steel bell ends might be too hard to form, press or shape. I've never seen it in anything but knife blades.
It's used for more than knife blades. Premium shotgun barrels were often made of Damascus steel in the 19th and early 20th centuries. In fact you can currently buy a Damascus steel slide for a Colt 1911 from Caspian Arms.Damascus steel is a general term covering a lot of different combinations of layered steel, both high carbon and stainless. There are even guys out there making Damascus steel from iron meteorites.
The idea was proposed to me by a machinist and fabricator with experience in all sorts of materials from aluminum to titanium to tungsten. He has made Damascus steel himself and has access to a very high pressure press. He feels pretty confident that the press can handle the material, and the Damascus steel maker is enthused about the project too.
Like Mikey always says, I'll try anything as long as it doesn't blow up the microwave and burn down the house.
how about damascus steel Lams! now that would be cool. still need the endbells too though since you couldn't see the lams... ;-)or maybe a damascus chassis? even better.....
"Damascus steel bell ends."
shoud I bee afraid, doc? Is it something that could be discussed in front of the young'uns? Should the ladies first leave the room?hey... hope your doing well.... been a while since we talked....
I think it's safe for the kids as long as you can skillfully screen the embarrasing stains you might make when you see this pic. We're talking hand forged Damascus steel, something like that used in the blade of this 6" Japanese style kitchen knife by Bob Kramer of Kramer Knives. Possibilities discussed include having bell ends drawn, control knobs turned, cabinet feet forged, etc.And yes, as a sucker for practical items done as beautiful art pieces I ordered a knife...actually two, in carbon Damascus with Thuya burl handles - 8" chef's knife and a 5" utility knife.
Hey, speaking of practical art, any truth to the rumor that a prototype run of MQ crossover inductors may be making their way west soon?
This post is made possible by the generous support of people like you and our sponsors: