|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.22.235.48
In Reply to: RE: Are Direct Coupled amps reliable? posted by Ray Moth on December 04, 2014 at 13:08:22
... direct coupled amps will drift out of alignment as tubes age or power-line voltage drifts, so unless you re-trim them on a regular basis, or use servo bias circuits, regulators, or other complications, they will usually be operating at off-optimum operating points.
As Mr. Steady posted on the high efficiency forum recently, "To paraphrase Winston Spencer Churchill, 'direct coupled SETs are the worst amps in the world, except for everything else.'" There IS a magic to them, but they are high-maintenance. I currently prefer to spend bucks on a really really good coupling capacitor, and at the highest level on a shunt regulator. YMMV though :^)
Incidentally, the drift problem is IMHO one of the reasons that Jeff (drlowmu) and Dennis (tube wrangler) are so adamant about premium tubes operated at conservative dissipations.
Follow Ups:
That is right. Experience can be of use here.
I and others have operated this kind of amp hard & heavy over many years. Right now I'm getting 8 of them ready (new ones).
Since about 1989, not one output tube has ever failed. Only two driver tubes have ever failed.
Those were in my own amps at an RMAF show in Denver. I had hooked-up a signal attenuator backwards-- it was running full output at the "0" position-- about 7 volts.
I turned on the DAC, and of course, the full output (computer was playing music into the DAC) went into the amps..
What happened? NOTHING happened. And I really mean nothing. As Kevin Bacon would say-- (The movie "Tremors") "there's nothing but nothing between here and Bixby." True.
No matter what I did, no signal!
Both amps were happily lit up, the speakers were silent. After about an hour of re-checking cables & hookups (I did finally hook up the attenuator correctly), by Brother said "change the input tubes". Why? None had ever failed.
Well, I did it, and presto! Music. What had happened? Each grid in each driver tube was simply gone. That's right-- gone!
Did any other damage occur, even though both amps had run that way for over an hour?? Nope. Not even a fuse.
---Dennis---
Thank you for sharing your experiences, it eases me.
"We plan ahead, that way we don't have to do anything right now, Earl explained it to me." (Tremors)
"direct coupled amps will drift out of alignment as tubes age or power-line voltage drifts, so unless you re-trim them on a regular basis"
Depending on the time frame the "regular basis" consists of, and how much they drift by, would be a determining factor.
Maybe it's a good excuse to add a front panel and some gauges?
"I currently prefer to spend bucks on a really really good coupling capacitor"
Thus my dilemma, I would assume that a cap coupled 2A3 would sound better then my RH84 with 6BQ5's, primarily due to OPTs, not faulting Mr. Kitic's design.
"Incidentally, the drift problem is IMHO one of the reasons that Jeff (drlowmu) and Dennis (tube wrangler) are so adamant about premium tubes operated at conservative dissipations."
There is certainly merit in that. That would maybe make the trimming a little less critical as well.
I keep tossing around which design to try among the usual suspects: Shishido, Micro 3.5, Asano, DRDs and yes the Paramounts, although it is unclear to me if your schematics are intended to be public.
I even debating starting a thread looking for suggestions, hopefully would be a little more current results then my google search results.
For example someone in another thread showed some displeasure in regards to the JELabs 2A3, which was trendy at one point in time.
I am open to suggestions.
My schematics for Bottlehead are not intended to be public. We do not publish them or offer them to anyone but customers. I do occasionally put designs in the public domain (build for yourself, not for profit) but Bottlehead pays a license fee to me for the ones I do for them.
Stability depends a lot on the driver tube and how close to maximum current it draws. FWIW, we've seen the GE 5-star 5670s behave very stably, after the first 100 hours or so - re-trimming annually is probably enough, though it depends on how much usage the amp gets. I run it at about 25% of the maximum current.
in the cathode of the driver tube, suitably bypassed of course. Would this force a constant draw through the tube and therefore constant voltage drop across plate load R?
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
"My schematics for Bottlehead are not intended to be public. We do not publish them or offer them to anyone but customers. I do occasionally put designs in the public domain (build for yourself, not for profit) but Bottlehead pays a license fee to me for the ones I do for them."
Thanks for clarifying that for me.
"Stability depends a lot on the driver tube and how close to maximum current it draws. FWIW, we've seen the GE 5-star 5670s behave very stably, after the first 100 hours or so - re-trimming annually is probably enough, though it depends on how much usage the amp gets. I run it at about 25% of the maximum current."
I could handle that, or even trimming once every couple of months wouldn't be so bad.
I have a couple pf fish tanks, so I wasn't looking for any more bi-weekly maintenance. If things didn't work out I could always change it to a cap coupled design.
So I am going to put you on the spot while I have your attention, which 2A3 schematics mentioned would you recommend as a starting point?
I already have the 2A3s and have punched the input tube as octal as the 6SQ7 seemed like the candidate when I first started gathering parts.
Worst case scenario I could drill some new chassis as I have hundreds invested in cake pans :P, but I do like the ones I have now with the octal 1st tube.
Thanks.
I don't really like any of what's out there for direct coupled amps, for the reasons cited in this thread. But I'll work one up for you, if you like. Public domain type. Hey, it's my hobby as well as my business!
Let's see, I think these are suitable parameters:
Octal driver
2A3 output
Tube rectifier
Direct coupled
Readily-available (Hammond) iron
Relatively reliable (BB-proof?)
Uncomplicated - no special tricks
No solid-state components
Monoblocks
Of course this would be an untried design, but if we keep a thread going for adjustments that should work out. Let me know if you are interested, and adjust the parameters if you feel the need. Should be fun.
That would be great.
Your parameters sound perfect.
I do have iron that is a shelf above Hammond if that matters.
Also "(BB-proof?)"?
Does this mean I have to get rid of my Red Ryder carbine-action, two hundred shot Range Model air rifle with a compass in the stock and this thing which tells time?
"Of course this would be an untried design, but if we keep a thread going for adjustments that should work out."
I can do that.
BB-proof = tough but not bullet-proof
:^)
I will gather up the specs of my parts and start a new thread later this evening!
"I can do that."
Right on.
Big speakers and little amps blew my mind!
I'll observe with great interest.
Jamie
Big speakers and little amps blew my mind!
So all you need is one meter and a trim pot on the top-side? That's not asking a lot.
I have an amp that uses an LED as the bias meter. It goes, green, yellow, red. Very easy to trim. Then when the music plays the light jumps around really fast, and looks pretty cool.
Jamie
Big speakers and little amps blew my mind!
I could live with that.
A well designed DC amp will require NO trim pots of any sort. Trim pots are baloney, junk design. Amps should really be engineered to have NO pots in them, good amps !!The op points should be thoughtfully centered ( and conservative ) such that changes in AC line, and tube aging, are FULLY taken into account at the onset. Trim pots are band aides for incorrect designing.
Me personally, I have been working with DC SE amps since 2007, and I have never had a failure of any sort.
GarGoyle, you have mail. Punching a hole for an Octal high mu driver tube is a mistake of sorts. A 9 pin miniature will sound better, it does not have all the soldered connections to the octal's socket pins, and is "less hazy" more distinct sounding since the wires come directly out of the glass envelope, and do NOT terminate into solder connections (eight) on an Octal tube socket. Maybe its not much of a factor on a 6SN7 with a mu of 20, but it certainly IS audible ( as a degrade ) on a mu of 70 to 100 tube.
LMAO, it will be a cold day in Hell when anyone beats half a Telefunken etc. 12AX7, under one milliamp, DCed into a 2A3, IF you know how to do it. Proof .... is in the pudding !!! I am working on a stereo 1/2 12AX7 - JJ 2A3 amp for myself as of this very moment, applying new things I've learned. Installed L1/L2 for each channel - this afternoon. No pots in my amps !!
Gotta go now, tonight is the Kansas City Symphony's performance of G.F. Handel's "Messiah" , and I can't wait to hear / experience it !!
Jeff Medwin
Edits: 12/05/14 12/05/14 12/05/14 12/05/14
"A well designed DC amp will require NO trim pots of any sort. Trim pots are baloney, junk design. Amps should really be engineered to have NO pots in them, good amps !!"
To be fair trim-pots vs soldered in components is really a matter of semantics, but I get your point. If the drift stays within ones design parameters then no harm done.
"LMAO, it will be a cold day in Hell when anyone beats half a Telefunken etc. 12AX7, under one milliamp, DCed into a 2A3, IF you know how to do it. Proof .... is in the pudding !!! I am working on a stereo 1/2 12AX7 - JJ 2A3 amp for myself as of this very moment, applying new things I've learned. Installed L1/L2 for each channel - this afternoon. No pots in my amps !!"
Does the Rp of the 12AX7 limit the bandwidth in your application?
"Punching a hole for an Octal high mu driver tube is a mistake of sorts."
To quote Bart Simpson, "you're damned if you do, you're damned if you don't."
I will have a look at your paper, enjoy the show.
"A well designed DC amp will require NO trim pots of any sort. Trim pots are baloney, junk design. Amps should really be engineered to have NO pots in them, good amps !!"
Well, in a cathode circuit, maybe unnecessary. But, there are some nice, low wattage pots available - probably better than the metal film resistors some folks like to use.
As for trim pots in general, I use them sparingly, but they do have their place. Every amp I coceptualise, I ask: in a DHT cathode, does a hum-nulling circuit using a high quality, high-turns, low value bulk-foil trim-pot provide better sonic and practical results than the alternatives. To my mind, and how I use it, with my current level of knowledge, in my system, yes. I don't want hum pushing through the quiet music I enjoy in the wee hours (in a quiet neighbourhood).
Now, I know, the counter-argument is that am not as enlightened as those who might take a different POV. I'd be curious as to what noise levels folks are achieving when trying different, commercially available DHTs and running a CT filament transformer with no hum pot. There is a fair bit of variation amongst even the highest quality tubes.
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Interesting discussion, and one we all have to consider.
Variations in DHT filament balance are all over the place. In the case of the 2A3's, I once had an AVVT mesh-plate that delivered .8 MV at the amp output. That was with A Hammond filament trans, and A.C. direct-heating.
That .8 mv was rare, it's mate delivered 2.3 mv. The usual was about 4-6 mv, and I'd try to arrange these into nearly alike pairs, preferring to leave the cathode with nothing but a resistor and cap on it.
I'm still pretty hung up on this, and am trying to talk a trans-builder into making me a truly balanced filament trans. It will not, of course, balance-out wayward tubes, but at least the transformer as a hum source can be minimized!
---Dennis---
The experience here is consistent with mine as well.
In general, I think one can count on 2mV from an AC-heated 2A3, allowing for swapping in a variety of tubes. I know Gordon Rankin does a lot better, with careful attention to the details and I presume some willingness to cull the noisy tubes. Most of the time that's just fine, you don't hear it.
I did experiment more with 300Bs, which have twice the filament voltage and generally produce around 5mV hum with AC filaments. Using the winding centertap, I found anywhere from 4mV to 25mV depending on the tube (and culling one with visibly warped filaments). Too many speakers will make 5mV audible.
I prefer DC for 300Bs, and hum pots for both 2A3s and 300Bs even with DC. One of these days I'd love to play with one of those EML tubes with the filament centertap pulled out for connection!
Hi Paul,
My EML V4 centre tapped tubes are being made right now for Feb. Shipping.
I'll compare them to the standard SPs using CT hammond 266m2 and report my findings re. sound differences and hum performance.
Only problem is as these are specially made I will have to accept the matching of the pair 'as is'.
Also re. The Ultrapath cap, I found more important is the C2 bypass in the power supply to improve the sound.
Cheers
Smart
I'd be curious to know how accurate the filaments can be centre-tapped... maybe enough to get the noise down to acceptable levels and not need a pot in the cathode circuit (and able to keep the filament transformers out of the local current loop).
I'm temped to try it for my next build, but for comparative reasons, I'd like to be able to use the tubes I have invested in. Would be easy enough to convert over at some later stage though.
As for expected noise levels, I don't really know. Some folks say 3-4mA is acceptable when using a pot with AC fils on a 2A3, but I am with you, 2mA is an achievable target. Whether there are real sonic benefits with the lower hum, beyond hearing the noise, I don't know. Tradeoffs: to pot or not to pot?
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Can't wait to get my hands on those!
Will try to report what happens....
---Dennis---
Not knowing how your filament transformers are constructed, is it at all possible to partially unwind the long half to achieve a closer balance?
> > > I'm still pretty hung up on this, and am trying to talk a trans-builder into making me a truly balanced filament trans. < < <Sounds interesting Dennis; if you don't mind, keep us informed of your progress.
I'd like my amps to be as simple and needing as few adjustments as possible... and I'd like them to be quiet, stable and reliable... oh, and sound just as I like them! Seems like we have similar priorities, even if our methods differ. Of course, my implementation is not as evolved as yours.For my first amp build I used a 27-turn, 10 Ohm Vishay, bulk foil pot and even with less than ideal grounding (to my mind) I have been able to get noise down to ~1mV on one channel, and somewhere between 1mV - 2mV on the other using the EML solid plates (noise with the mesh plates was a touch higher). This is okay I think; I can't hear hum more than about a foot away from each of the 99dB+ efficient speakers. The difference between 2mV and 6mV hum is noticeable, I think, but it might just be "expectation" effect. As an aside, I find it interesting that my amp seems quieter than the two solid state amps it has replaced, despite using ac fils... maybe that slight hum is less offensive than whatever noise some solid state amps produce?
I am looking forward to my next build, but still have not settled on how I will manage hum...
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Edits: 12/06/14 12/06/14
Not all hum is alike-- like everything else, it has its own internal components.
A "little bit" seems to aid both interconnects and speaker cables, keeping a bias signal on them so that they really respond quickly to musical transients-- capture the "leading edge" as musicians say.
So, how much is too much, then, and will cause musical capture to start going "South"?
How much is too much depends on what components are in the hum, and what your speakers will tolerate.
In GENERAL, most of us like to see less than 4 mv-- if you're going to measure things. That's just a rough clue-- not an absolute.
Solid State amps may use center-tap (2 diode) circuits for H.V., and others will use a Full-Wave Bridge-- (4 diodes). Not all Solid-State amps use chokes-- some will just load-up the caps & rectifiers--only-- supply and then try to "regulate" it into submission!
Choke-Input supplies are used in the BEST Solid-State amps, just as in the best tube types.
In any case, many S/S amp's hum artifacts will often be very different sounding than the same "hum" levels from tube amps.
Interestingly, when the best possible tube amps are compared to the best possible Solid-State amps-- the two not only sound identical, but so do their hum components.
Small world!
---Dennis---
" Does the Rp of the 12AX7 limit the bandwidth in your application? "Only if you are able to think conventionally !! But in REALITY :
Not really, if you consider that the 12AX7 is just ONE part of a totally optimized amp we execute in modern times.
The 2A3 is very EASY to drive in a Direct Couple !! And of course, we have an actual, not theoretical, historical record of this : 1947 used a mu of 100 high Rp tube ( 6SF5 - Lipman and Robin, later "FI" amps ) as did the DC amps of Isamu Asano ( 2A6 at 0.43mA.) and Nobukazu Shishido ( 12AX7 at 1 mA. ). About 25% of the DIY Japanese amps were L-W variants, according to Jean Hiraga if I recall it properly.
If you wanna do it on a "proof by listening" , we have on A.A., Smart845's SupaNova Type 45 DC amp review, which I helped him build. I can refer you to various amplifier reviews found on Dennis Fraker's "Serious Stereo" web site.
In late 2014, we are able to do it even better, new parts and technologies are applied - positively.
Jeff Medwin
Edits: 12/06/14
"Gotta go now, tonight is the Kansas City Symphony's performance of G.F. Handel's "Messiah" , and I can't wait to hear / experience it !!"
Hear / experience it huh?
What kind of power amps are they using in the PA system?
What about crossovers? I'll bet DSP based feeding bi/tri-amped arrays!
What kind of wire are they using, particularly on the low level microphone feeds?
How many high feedback OPAMPS are in that audio board the signals have to transverse through?
Hi Gusser,It was a live unamplified concert, four soloists, a powerful 250 member mixed choir, and a Harvard ( summa cum laude ) music conductor. Not a single microphone, op amp, speaker, or electronic device was needed.
So you have struck-out on every account Gusser. Way to go.
Its the new multi-zillion dollar Helzberg Hall, about a three story pair of clamshell like buildings, with seating for several hundreds, and no microphone was needed. No prosentium arch either, Mr. Fulton would approve. Almost like a Greek ampi-theatre, but in 2014.
See the photos of the hall, especially the side-view shots down the page. Drool. It was fabulous. Everything is "up to date" in Kansas City.
Jeff Medwin
Edits: 12/06/14
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: