|'); } // End -->|
Hi you all,
My friend Dennis sent me a couple emails today that I 'd like to share up here - with any of those DIYers who have an interest :
- - - - - - - - - - - - - - - - -
You're right on your latest Forum stuff-- but you needed to point out that low-DCR is MOST CRITICAL in DRIVEN stuff.
ACTIVE stuff-- (That is-- the SOURCE for the power in the Power Supply (transformer & rectifier) NEED NOT be low-DCR-- as it is electrically DRIVING the circuit-- and can instantly change to accommodate current demand-- its own internal resistance is MUCH LESS a factor-- it is still there-- but it doesn't count for much.
Low-DCR is necessary in anything downstream from this-- all the devices which work on the power supplied must be low-DCR-- or the current/voltage losses will result in timing errors-- when current is tried to be re-supplied..... it will get there TOO LATE.
- - - - - - - - - - - - - - -
That last email is a bit muddy.
Here's the goodies:
(1) The Power Transformer can benefit a little bit by being low-DCR-- but the largest
benefit will come from extra current capability, regardless of DCR.
(2) The rectifier doesn't matter at all-- as to DCR. It is super-active, and can address current/voltage changes rather instantaneously. BUT-- just as in the transformer-- extra current capability here is a winner, regardless of DCR.
(3) The input choke should be low-DCR. This is because we want a TIGHT COUPLE to the rectifier output. Once the power leaves the rectifier, it is no longer adjustable-- actively. Now, we have passive components, unless we install voltage/current regulation-- but we don't want to listen to those things struggling to "keep-up" with music.... so they're best left out.
(4) Capacitors should be low-capacitance. This is because we want any "tuned" circuits to stay above human hearing--- by several orders. You can say here that we want SPEED. That simplifies it.
(5) Rules-of-thumb in designing the total amplifier:
(A) High-Current power transformer.
(B) High-Transparency rectifier.
(D) low-DCR chokes.
(E) SMALL capacitors.
(F) Low-DCR wiring-- everywhere.
(G) Anything that is in-series with power or signal-- low-DCR, and physically short.
(H) Anything that LOADS power-- or signal---- (plate & cathode resistors, input impedance, etc.) , ---- VERY high DCR.
OK-- Now they can get to arguing again. ----D----
- - - - - - - - - - - - - - - - - - -
In earlier posts I reported I had not yet experimented with bleed resistors on the two supplies in Greg's 245 amp. Well, that has been resolved somewhat this evening.
A high wattage bleed resistor, across the B+ to ground, inserted at the location after C2, in a L1/C1/L2/C2 filter chain, can be helpful in drawing more current through the chokes, and it is beneficially, a stable constant current. The main downside is the sonic signature of the bleed resistor itself, infringing upon the transparency and purity of the music's presentation. ( You can hear the R !! )
We are very fortunate today, because the Mills MRA-12 line of resistors is available from Mike Percy, and THAT specific type R has the least negative artifacts, of any I've tried, in this type of service. The MRA-12 is, IMHO, indespensible !!
We tried bleeder R's across Greg's front end supply and also, across the 245 output stage, a separate B+ supply. The front end supply powers a Type 76, one for each channel, two total, at 7 mA. each tube as I recall. The output stage draws only about 30 or 31 mA. per globe 45, so its a total draw of about 60 mA, versus 14 mA. total draw for the Type 76 input stage.
Bottom line, the bleeder R was VERY helpful to the Type 76 stages, normally drawing only 14 mA., and not too noticeable with the output stage B+ , when it was drawing 60-62 mA.
I experimentally listened to different values of bleed R's across the type 76 stage, but I didn't have enough " non -grungy sounding Rs " to pull all the current I would have liked to pull. For now, we have a 13.33 K ohm power resistor added across the Type 76 supply, drawing an aditional 12.4 mA. So, I've increased this stage's total draw from 14 mA. to 26.4 mA. This added draw decreased the B+ voltage a bit, and I simply addded another 10 VAC to the power trandformer, ( those Signal 500 VA DU-1/2s are flexible !! ) to get back my original B+ voltage, which we liked the best, sonically as a transformer-loaded Type 76 op point. The added bleed resistor sounds better to us, gets the choke drawing more current ( its rated at 600 mA. ) and the resulting music's presentation was both far more dynamic and more subtle in nuance, pretty cool stuff to listen to.
When going to the Triad C-40Xs as L1 and C1, I'd previosly estimated your final B+ will become : VAC on the PT secondary under load times about 1.35 , minus the voltage drop in the rectifier. I also promised I'd measure this supply in AC and DC terms, here goes, using Fluke 8060A:
Line was 120.9 VAC
Output Type 245 stage, two tubes drawing 31 mA each is 62 mA total draw. Power tansformer secondary, under load, 249.6 VAC. Final stage's B+ was 269.7 VDC with 0.43 VAC of ripple.
Input Type 76 stage, 14 mA. from two tubes and 12.4 mA. from an added shunt / bleed resistor, or 26.4 mA. total draw. Power transformer secondary, under load, is 156.3 VAC. Input stage's B+ was 165.7 VDC after temp stabilizing, with 19 mA. ripple. Rectifiers in both supplies were 5U4Gs as a Graetz ( sp? ) bridge. C1 and C2 were 40 uF ASCs in the output stage, and 50 uF oils, in round cans, in the Type 76 stage. Use nothing larger than 50 uF BTW.
I have posted this in an effort to be helpful to any who may want to experimentally duplicate a similar supply in their own gear, which I believe you will really enjoy hearing., Do NOT get hung up on ripple "numbers", but rather, listen to how this supply sounds in dynamic terms - playing music - which is its forte.
Oh, Dennis and I mentioned using heavy gauge wire as being critical. In my experiments with Greg's 245 SEer, I had to use three Radio Shack clip leads in parallel for each wire run in the supply, from the rectifier forward through C2. I can only imagine what we will next hear with good wire, shorter lengths, and Wonder soldered ........ Greg's next DIY project, two mono amps !!
Hope this helps someone out there, let me know please. Thanks a lot.
The intended function of a bleeder resistor is to discharge capacitors after the equipment is turned off. Good caps can hold their charge for a LONG time. So bleeder resistors are a saftey requirement, and not really an option. The secondary function, in a 'L' first supply, is to sink the minimum amount of current necessary to reach 'critical current'. But this isn't an issue in your low L designs, since 'critical current', is never reached.
By the way. I have a hard time believeing that a resitor of a value in the K ohms, shunted by a cap of at least 10 uFd, has any 'sound' at all. It's impedance, relative to that of the caps, is so high as to be insignificant.
leads to non-imaginative thinking and problem solving.
What exactly is 'bad EE trainig'? What is 'good EE traing'? And what does too little EE training lead to?
Strictly IMHO !!!
1) Bad EE training is training of too narrow a focus, such that one cannot get past rigid theories and think out of the box.
2) Good EE training allows you to think and successfully tackle a variety of engineering problems, not just electrical.
3) Too little EE training frees-up ones mind so they will try different posibilities and find out what works best, with no preconceived notions as to " how it should be according to theory ", to a point where good approches and effective practices are not even persued.
1) I'd feel sorry for myself, having 6 years of college EE training, except that my employers over the last 30 years do not share your view. Well, O.K., it's really only 5 years of EE education - the first year was all humanities.
I was not ever trained in EE. I have no limits to my thinking and I am prepared to try anything, not knowing why it may not work. You can say I know no better.
"We are very fortunate today, because the Mills MRA-12 line of resistors is available from Mike Percy, and THAT specific type R has the least negative artifacts, of any I've tried, in this type of service."
Rhoderstein 2 watters in parallel, and various high powered wire wounds.
The Mills MRA-12s were a tough act to follow. Do you know of something better ??
Hey, I'm not bashing the Mills. I just wondered what else you had tried ...
A young woman recently told me that her new Mazda Protege was the best handling car she had ever driven. I really wanted to ask what other cars she had driven ... (Not that I'm comparing the Mills to a Protege, you get the point I hope.)
Generally, I don't post here too much. Particularly on this topic, meaing the low dcr systematic approach, as my mind is pretty much made up. I know it works. And, I've got better things to do than to try and convience someone else here that already knows is doesn't, that it does. This in spite of never having tried it. I suspect that those others that know it works too are like me, off building, experiementing, exploring further, and enjoying music, trying to push the envelope even further still.
Why try to convert someone here when they seem to be dead set against it, believing and being convienced that what they are now doing is right? After all, some people are satisfied with music in and sound out, when it comes to their amplifiers. Some get it, some don't. It's pretty obvious after a while who falls were.
I try and convert some folks privately, but that's about it. I also use it my amps and now use it the first Ultra Fi standard production product, the I'O. I need to change my designation here finally. Is there one for a writer and builder BTW?
Maybe that makes me one of the polite first people to try the systemic approach that you are attempting to discribe here. I'd like to think so. I easily can remember back as far as the 80's when, for example, Dave Slagle and I were messing with this.
At any rate, keep posting!!!!! I enjoy the read, even if I don't chime in as often as maybe I should.
Let me defend myself.
1) I want to understand the reasons for what I am doing. Simply stating 'it sounds better' doesn't hack it. Why does is sound better?
2) 'Sounding better' is a pretty illdefined term. What the heck do terms like 'faster' and 'slugish' and 'mushy' really mean? And to whom does it sound better? I have been using a NAD non tube based preamp for many years. I tried to replace it with a tube preamp, which to me sounded 'better'. But you know what? My wife didn't like it. Without some kind of verifiable criteria, sounding 'better' is purely a matter of individual taste.
I'ld be more comfortable with simple descriptions of the perceived changes. The misguided attempts at technical justification, usually accompanied by excessive punctuation, capitalization and uncritical high-handing are an embarrassment. It gives the whole affair an unpleasant messianic vibe (especially the mid/hi/ultra-fi BS.)
Have to agree on the second, there’s no question in my mind what’s described as ‘faster’ is often ‘reams of higher order distortion components.’ Not saying it’s the case here though I also have to wonder just how much PS modulation is occurring and how it manifests on the output.
All that said, I have benched and measured a low-DCR test mule and tuning the first LC section – in this case 0.07H @ 3 ohms into 6.8uF - easily doubles the supply stiffness. It’s an approach worth investigating.
"All that said, I have benched and measured a low-DCR test mule and tuning the first LC section – in this case 0.07H @ 3 ohms into 6.8uF - easily doubles the supply stiffness. It’s an approach worth investigating."
Some questions, please.
Doubles the supply stiffness compared to what?
What is "worth investigating"? Tuning a supply? You do understand that that is nothing new, right?
"Doubles the supply stiffness compared to what?
Compared to replacing the 0.07H/6.8uF pair after a SS bridge (in this case) with the typical large value C. The supply is LCLCLC, last two L are 0.6H/11 ohms, last two C are 100uF oilers. The supply runs ~250 VDC loaded with a 5K resistor. To test stiffness another 10K is dropped across the output. The resultant drop in voltage displays an unambiguous minimum over a small range of first C values right around 6.8uF in this test mule. 50% divergence in either direction about doubles the voltage drop.
"What is "worth investigating"? Tuning a supply? You do understand that that is nothing new, right?"
Not sure what you mean by 'tuning' in this context. This first LC is 70 mH, 6.8 uF, well outside the range of any technical documentation I could find. Pointers to references covering this kind of supply appreciated. SMPS?
I get different results than you. With the 6.8uf in place there is a 25 volt drop increasing the current from 100ma to 150ma. With a 100uf in the C1 position there is an 8 volt drop.
Have Fun and Enjoy the Music
"Still Working the Problem"
PSUD is a godsend but has its limits. Oddball circuits like this push past. With a small L followed by sub- 0.1uf C it's possible to simulate kilovolts from a 200 VAC transformer. It's still useful to put you in the ballpark though, try 14uf. The transformer's secondary is 23 ohms per winding, 209 VAC. I get just over 6 volts from 70 ma to 130 ma (30 ma per SE output tube doubled, 10 ma constant for the two drivers), about 8 1/4 for 100uF. Again, in the ballpark of the measurements. My old EE prof used to tell us calculate for within an order of magnitude of the target and tune the rest by hand. I never bought a product he designed though. ;)
The problem is that the supply in your sim is not well tuned (sorry, Henry.) My guess is that rdf's transformer and first choke have DCR different from the ones you used.
For a given T1 and L1 (DCR included) you will find that there is a specific value of C1 that not only maximizes the output voltage, but also maximizes the supply stiffness (i.e. minimizes the step in voltage when the current is stepped.) Try playing around with the value of C1 in your sim. When you find that value I'll bet that the change in output voltage when the current is stepped will be smaller than 8V.
FWIW, according to the simulations, a second stage of low L low C filtering can improve stiffness further, though the optimum value of C2 is not the same as C1 even if L1 and L2 are identical. Around 3x in the cases that I've played with. I haven't seen much of an improvement with a third stage, but it could be I haven't found the sweet spot.
In any event, I'm talking about simulations. Finding the right tuning (sorry, Henry) in a real circuit might be a real challenge. rdf reported measured results. He wins. :)
Have you heard the supply ?? Do so !!
I'd personally would wanna add two more 100uF of oliers, and connect them to your existing 100 uFers, in SERIES, ( so you can HEAR 50 uF or less in C's ), and tell us how it sounds playing music, which is the goal of all this.
Use heavy wiring and / or multiple paralleled clip leads from the rectifier to the end of the filter chain.
I'd really prefer listening to the .32 HY Ls over your .6HY/11 ohms ones, due to its smaller core, having less of an effect dynamically upon the amplifier. This is another case where smaller is better. As a guide, over 1/2 a Hy is too big. Over 50 uF is too much C.
Great to hear people are OPEN to experiment and listen. Progress is being made. We will soon be out of the dark ages and using techniques that flat-out work well, quite well, which is all that matters IMHO.
The amp it's going into has 40uf oilers so that's easy. I can also parallel the chokes for ~0.3H at 5 ohms for a taste of both, then decide what direction to go in. Juggling it, two other amps and work at the moment. Not that I'm complaining about the first two.
Thanks for the update.
Well, in the case of paralleling chokes, the "numbers" you quote are correct, but we still have too much core and wire, twice as much !!!, hence the Triad C-40X suggestion to all. I also used 40 uF Cs on Greg's output stage !!
and what hes done is not the supply we are advocating. He was just doing some low L1/C1 experimenting, won't get ya to heaven that way I'm afraid. At least he mocked something up and tried ! Bravo for him.
It has to be a total approach to the supply parts, wiring, and to the amp. I've spelled out the supply here this month fairly specifically.
Jeff here. I got this nice email response ( to the Forum comments ) from Dennis today, to share with those interested in audio amplifiers.
- - - - - - - - - - - - - - - - -
Subject: --'ya done it now 1--
Date: Thu, 8 Mar 2007 11:54:49 -0800
It's all a matter of balance. I didn't say that the power transformer DCR doesn't matter-- it does. I just said it matters a bit LESS than the chokes & caps do.
I didn't say that the rectifier doesn't matter-- it does matter-- a lot. I said DCR of the rectifier, matters LESS than the DCR of chokes does, that's all.
There is NO change in my ideas here-- just some intelligent application discussions going on-- as to how this all can work together.
Remember also-- that transparency and speed matter most of all-- musically. You do whatever it takes to get that... but you don't quit there..... Let's get it all-- and that requires a total systems approach..
If the power transformer core gets too large, then THAT would make too large a choke out of it.-- but that will not hurt its performance as much as doing the same thing to a separate choke would.
We can agree that other improvements can be gained when the Power Source is "too large to pull-down".
In practice, we simply must get all of these things into perspective. Would you install a D-9 CAT Dozer flywheel into your Honda Civic? The engine would break before the "flywheel" got up to speed! If you did get the flywheel up to speed, could you stop the car if you applied the brakes?
Are power supplies flywheels? Not if I can help it! I don't want energy storage-- I want energy delivery-- instantaneously-- without it being impeded. That is correctly called impedance! Sure, it's a good filter. So is capacitance. Both must be used sparingly, carefully, and intelligently.
Are Batteries, then, the ideal power source? Well, they would be if they weren't also capacitors! Or diodes! Or have long leads!
There, you have it! We buy or custom wind those expensive copper-foil and tin-foil capacitors-- for Power Supplies! Why? They store very little, and they act very quickly.
They're low-capacitance, but highly energy-reactive.
That sounds a lot like music..... not so impeded, not so "canned" (stored somewhere), and not so resisted (high-DCR series transfer components), and not so loaded-down-- (low-DCR loads-to-ground). Those old terms for capacitance, impedance, and resistance really do tell you what they do.
Look again at that last-- loaded-down -- there are places where we do want HIGH DCR. On anything that is a load upon a signal.
- - - - - - - - - - - - - - - - - - - - -
I hope others will understand Dennis' intent a bit better by posting the above email. Hope it helps you.
"MUCH LESS" and "it doesn't count for much" has turned into "a bit LESS". Bull Shit!
"ACTIVE stuff-- (That is-- the SOURCE for the power in the Power Supply (transformer & rectifier) NEED NOT be low-DCR-- as it is electrically DRIVING the circuit-- and can instantly change to accommodate current demand-- its own internal resistance is MUCH LESS a factor-- it is still there-- but it doesn't count for much."
Come on Jeff, who do you think you are fooling? Do you think we are all morons?
Hate is a a terrible thing to harbor inside you, and it certainly shows up here loud and clear in your EVERY word and action! How sad and how rude.
Why don't you try to grasp the goodness of the information, and / or my good intent, rather than how well it may, or may not be presented.
What is being suggested works like gang-busters. Maybe you could learn how to build a much better, more satisfying audio amp if you opened your mind and lost the hate.!!
Instead of fully-digesting what we present, you knee-jerk react, and just look to discredit me.
However Tre', know this, I am correct in what I HEAR - what this type of simple supply does sonically, so in the end, you will be the discredited one, not I. Actually, you already are discredited, only the early adaptors of this technology are all polite, and do not get in your face up here publically.
You could be a big boy, and fork out 50 dollars for two Triad C-40X inductors from Allied Electronics and two Percy ASC 40 uF @ 440 VAC caps, and implement the supply, but instead, you wanna play hateful verbal games up here.
" with the solution to your problem "
When you think you are losing an argument or humiliated in some way, you suddnely turn vicious. Then we get to see what you are really like.
"What is being suggested works like gang-busters"
Which of the many different suggestions are you talking about?
"what this type of simple supply does sonically"
What supply are you talking about now?
The one you proposed 8 months ago?
The one you proposed 4 months ago?
The one Dennis proposed at this post http://www.audioasylum.com/forums/tubediy/messages/118189.html on 3-7-07 12:10?
Or the one Dennis proposed at the next post on 3-7-07 17:55?
They are all different. Can't you see how that shows that you don't know what you are talking about?
Jeff, I'm not full of hate but I don't like uneducated designers telling everyone they, and only they, have all the answers. I post so other, less knowledgeable persons don't buy into your unsubstantiated claims.
Jeff, it's wild, unsubstantiated claims like your's that give high end audio a bad name. Snake oil and magic dust.
I'm just standing up for what's right and not letting you get away with telling half-truths. If you want to call that hate, go ahead.
As time goes by, things get refined some. L1/C1/L2/C2 is as follows :
Ls are both Traid C-40Xs, available from Allied at only $10.79 each.
Cs are both ASC oils, the ones Mike Percy sells, 40 uF at 440 VAC, about $12.95 each as I recall.
Anyone can just build and LISTEN to that, with THOSE precise parts.
Its what we have listened to now in my latest report, on Greg's 245 SE amp, both for his input and output stages. Heavy gauge wire from the rectifiers forward is a MUST. We eliminated his damper diodes and went with 5U4GBs.
As an fair estimate, assume PT secondary VAC times 1.35 = B+ VDC, minus rectifier drop.
I'll look at Greg's amp notes later, on both supplies, and post actual measured AC and DC voltages and current load, for each supply, when I can. We have that information recorded in our notes taken.
With all due respect, Jeff, those Triad C-40X chokes have a DCR of 10 ohms. Wouldn't we be better using the Signal Transformer units that you mentioned earlier? The CH-1, while less than 1/3 the inductance (100mH vs. 320mH)it has only 1.5 ohm DCR.
Both of these inductors were intended for low voltage supplies. Using them at the AC voltages we're talking about worries me a little. Not for safety reasons necessarily, I'm just worried that the chokes will easily saturate even when run well under their maximum DC current rating. They are subject to MUCH higher AC flux in this application. VERY generously over rating them in terms of DC current might prevent saturation. OTOH, maybe you like the sound of the supply with saturating chokes ... :)
Also, if anyone is interested in 'tuning' the supply to achieve the 'flywheel' effect (as described by Henry,) it seems to be easier with lower inductance, depending on the PT.
Whoa.... written on the .32 HY L .... " 1500 VDC RMS insulation test " ... thats OK for tube amps of the 2A3 variety I'd think, thats surely not low voltage !!
It's not the DC voltage to ground that concerns me, it's the AC voltage *across* and SC currents through the choke. That's what might lead to saturation of the core.
I did a bit of reverse engineering on the Triad unit (not too hard given the dimensions of the core, the inductance and the DCR) and it seems that you might be OK in your application. I wouldn't try to push the DC current up to 600mA though. Keeping the DC current down leaves more headroom for the extra AC flux that the core is subject to in this application.
The Triad I'd prefer over the Signal, as the Triad posseses about three times the inductance, not too big and heavy a core, is well made, and rated for 600 mA current. I use under 60 mA through them, and have no worries with that part. The Triads and /or its similar "cousins" operate in amps known to me, in the L1 spot, at +490 VDC and never have failed, ever.
Two weeks ago, Greg and I listend to his vintage UTC L that was 1/2 HY, ONLY 2 ohms, rated at 1.2 A, and it sounded TERRIBLE. Too big, too heavy, too much core and too much wire inside it !! I mean, the $10.79 Triad smoked this 2 ohm UTC, the UTC lost everything we had gained, the music's presentation was ALL out of time and dynamically ruined. What else can I say ??!!
a few turns of heavy duty Litz wire wrapped around an engine block (steel, of course) nwould be perfect ...
Actually a dinky $10.79 Triad C-40X from Allied Electronics 1-800-433- 5700 will be way better bro'. C.Y. loved it in Korea last year...showed him the way.
Thanks, Jeff, for your honest report. It's a good data point.
It appears that there might be more to all of this than we have figured out yet, so I hope you won't be offended if I'm not too quick to draw any conclusions about what differences between the Triad and UTC were actually responsible for the differences in perceived sound.
No problem, it was a good question I thought. I'd go back and reread Dennis'posts of mid 2006, as its really HIS baby, I'm just the messenger who finally got around to implementing and hearing it last month on Greg's RCA field coil horn system.
I would have to disagree with the rectifier not mattering. It is fairly easy to see that the dynamic impedance of the power supply is much higher with normal tube rectifiers, and even damper diodes. Mercury vapor and solid state rectifiers have much lower forward conduction impedances. In simulations with PSUD II, I have found it easy to get down to around 65 ohms with mercury vapor or solid state rectifiers. However, I can't seem to find a combination to get normal tube rectifiers below about 280 ohms of dynamic impedance. To get the dynamic impedance of the power supply I do a current step of +20% and divide the voltage change by the current step amount.
Now, 280 ohms is not bad considering most choke input supplies have impedances around 400 ohms. Let's put everything into perspective with a typical 300B amplifier. The normal 300B will have a plate resistance of about 800 ohms. Now, which power supply do you think will allow the 300B to sing along with the music more freely, the 400 ohm choke input supply, the 280 ohm low DCR supply with normal tube rectifiers, or the 65 ohm low DCR mercury/solid state rectified power supply?
We are not using typical chokes JLH, 10 ohms DCR and .32 HY. as L1 and L2.
Just like YOU use 50 uF or less Cs, we wanna use 1/2 HY. or less Ls, without too large a core to impede the power supply dynamically upon music playback.
I've never disputed the validity of the "Low DCR" approach, within reason, and I suspect that also applies to many other people here. However, Dennis's new proposition of a division between "active" and "passive" devices seems to me to be somewhat arbitrary. In addition, the rationale behind the proposition that DCR matters far less with what Dennis labels "active" devices than it does with "passive" devices is not at all apparent to me.
It could be argued that everything is driven by the power tranny primary winding. Then everything in the PS is a passive device. The alternating current in the primary causes an alternating magnetic field in the transformer core, which in turn is passed on as an alternating current in the secondary winding. Downstream of that are the rectifiers and smoothing circuit.
I don't see how any of these things - the transformer core, secondary winding, rectifiers and smoothing circuit - can be considered "active". They get what they are given and the difference between what they pass on and what they absorb/waste/lose is governed by their efficiency. Magnetic inefficiency and DC resistance are unavoidable and unwanted burden, limiting performance and contributing to loss.
'(G) Anything that is in-series with power or signal-- low-DCR, and physically short.'
Dead right! - I changed the cathode resitors in my 845 SET to 10 ohm wirewounds, and the sound has been elevated to a whole new plane. NOS RCAs only last 40 seconds before self immolation, but after the experience I just can't listen to anything else....
So what does all this mean to the sound quality? There are contradictions here. Low capacitance combined with low DCR chokes (low H typically) equals low energy storage which is contrary to practice in high end audio products. Does this mean that better sound is achieved with less stored energy, or that it is difficult to store and release the energy in a manner that is conducive to good audio fidelity?
In your last sentence's questions !! "Yes and yes" for both questions.
And BOTH of the questions in your last sentence is what we are able to demonstrate with these new ( 1992 developed by Dennis, divulged onTube DIY 6-2006 ) supplies, so the answer is Yes, Yes to your last sentence !!
And as to sound............... wonderful, the best results known possible to me.
One other wonderful thing, Greg's amp's final L1/C1/L2/C2 filter parts retails for under $50.00, all four parts. Its SOTA.
- as oppoosed to researcher / investigator - passes personal judgement once again. What a surprise.
Most others will jump in and beat up; maybe Tre' will get some too. Now that's hangin' wit' da in crowd. Word.
Why would one listen to a guru who can't keep his own story straight?
A guru is a teacher or a guide. A real teacher does not teach the truth as one thing on one day and something else the next.
And you talk about lame....
> > > A guru is a teacher or a guide. A real teacher does not teach the truth as one thing on one day and something else the next. < < <
You make it real easy for me to make my point dude - thanks. Are you saying it is better for a teacher to stick to some arbitrary line even though experience, research, knowledge, etc. come to indicate something different? Y'know, progress an' all that stuff. Man, some of us are not cr#pping ourselves at the thought of change.
Yes, Dennis / Jeff should not present their hypothesis as an absolute truth, or Truth; but did they? It may have been *their* truth at the time. I am thinking they want people to give it go and come to their own conclusions. That said, I have not been following these threads in max-o-detail; this is just audio - time spent here is time away from da ladeeez.
Yes, their hypothesis may appear flawed, now as in the past. It appears that their approach and experiences have some merit though, as has also been shown (then better explained) as in the past. Y’know, something works but you lack the ability to explain it.
PS. Nice guru line - I can not recall Dennis / Jeff calling themselves ‘gurus’. I could be wrong though - I am more interested in his point than trivial crap like whether he is the guru. Sh!t. If he did not claim he is da guru, then kudos to your (mildly) inflammatory language. Yes, the commonly-used terms teacher, mentor, etc are less inflammatory than guru – but you know that.
Buzz, I understand your point -- that if someone changes their view and says so, that is entitled to a little respect. Drlowmu cannot be faulted for lack of interest in finding sonic nirvana, and he has certainly stimulated discussion.
But -- and I don't want to speak for Tre' or others -- I think the harsh reaction comes from the fierceness with which Drlowmu presented his views in the past -- views which he just recanted because his own guru said something different.
Here is a post that shows what I am talking about:
Only the HV secondary DCR matters. Bad news : 252 ohms DCR I must inform you, positively stinks, such high DCR is not accceptable for audio, guarantees NOT GOOD sound in a no NFB triode amp !! Buy a T amp!!
Even a crummy ST-70, which runs super warm to the touch, or likewise, a Dyna MK III power trannie has about 80 ohms DCR.
On Power transformers, FROM ALL MY LISTENIING EXPERIMENTS OVER 30 YEARS:
Eighty ohms is LOW FI.
Forty ohms is just MID FI.
Twenty ohms is good HI FI.
Under Ten ohms is ULTRA HI FI.
There you have it. About the same DCR parameters apply to inductors BTW.
No need to build your house on a weak foundation.
You can't get quality from left over TV parts. Don't mean to be hard on you.
Now, in fairness, low DCR to some extent goes with higher current capability. But, this does not hold up when making subtle distinctions below 100 ohms of secondary DCR. I have some honkin' power transformers with 50 ohm DCR secondaries rated at 310-0-310 and 500mA -- before this recent channelling from Dennis, I was under the impression that the transformers were not even mid-fi.
Most folks are like me -- they would never ever post something saying some other person's system or parts do not or cannot sound good, without hearing it. If one does not live by that rule, he or she must expect a little grief when abandoning the underlying rationale for making such bold assertions. And, when claiming to know something based on "LISTENING EXPERIMENTS", only to say "Never mind . . ." based on an email from a guru, seems a bit contradictory as well.
Let me be clear -- while other topics are worthy of the kind of focus low DCR has gotten, I think the forum and hobbyists have benefitted from the discussion, driven by Drlowmu's embrace and advocacy of particular ideas. A lively forum keeps us coming back.
My beef is with pronouncing things "mid-fi" or "low-fi."
" views which he just recanted because his own guru said something different. "
I don't think so at all !!
I recanted MAINLY because I just built and heard the Low C, Low HY, Low DCR supply on Greg's 245 amp - only several weeks ago, I trusted what I heard enough to bring it all up publically here on the 'ole Forum, so others may consider employing it in their amps, and gain the benefit.
is starting to unravel the emotional from the intended message. So much of these 'discussions' is a reaction to the way J & D presentation of their findings, their arbitrary categorisations, and it seems (now) to me their insistence on absolutes. Perhaps they should remember this is only an hypothesis, not yet theory nor reproduced with consistency in practice. Maybe their excitement over their discoveries and desire to share gets in the way of their moderation and use of more appropriate language?
That said, they encouoraged people to try it out for themselves - somewhat at adds with their claims of absolutes. Can be expensive for DIYers tho - all those TX, Lundahl chokes, 'n' stuff. At least the latest hypothesis is not so costly.
Oh dang, where did my stolen Buzzman persona go? Word. Ah, there it is.
'Not really just a hypothesis, Dennis' Serious Stereo SE 2A3 amp has been designed and replicated by him since 1992. Circuit hasn't changed any. It was and still is Low C, Low HY and Low DCR !!
as to what is tried and tested over time and what is been tired just once. And whether the specific formula has changed or not.
Oh yeah, if you have just a general description / understanding of why something works, you want people to experiment, and seek to refine the formula etc. it is an hypothesis.
"Yes, Dennis / Jeff should not present their hypothesis as an absolute truth, or Truth; but did they"
Yes they did. Jeff has said "If you don't have low DCR your system can be no better than lo-fi"
"Yes, their hypothesis may appear flawed"
Their original position is now flawed by their own admission.
"the commonly-used terms teacher, mentor, etc are less inflammatory than guru"
Both teacher and mentor are modern definitions of guru. Definition 2b.
"A trusted counselor and adviser; a mentor."
Now get back to "da ladeeez".
P.S. "Are you saying it is better for a teacher to stick to some arbitrary line even though experience, research, knowledge, etc. come to indicate something different?"
Your point is well taken but Jeff has been shoving this ever changing sh*t down everyone's throat for eight months or more now. If he really has something to say, he should have waited until he sorted it all out before telling everyone else that they were wrong.
pick out parts of what I write and leave out the stuff that puts it into some context. Oh, and the way you add my main point, which you could not fully refute, into a PS. Sweet. Then there is the use of technically 'correct' definitions whilst ignoring the connotations associated with the term. Some would think you are heavily left-brain dominated dude more interested in 'winning' some trivial debate than, well, being a more open and balanced kind of chappie.
That said, if Dennis and Jeff have claimed absolutes then they should be soundly spanked.
As for their *original* hypothesis being different to their current - hell, any researcher (tho perhaps not those that *apply* the results of research) understands that hypotheses change; that is the nature of improved understanding.
Got to move
Can't hang wit' da ladeeez at the moment - got to get me to the clinic... Damn
You have not followed this from the beginning. In the beginning I agreed with Jeff that low DCR in a power supply is a good thing. (I still think that) The voltage drop when the current demand goes up will be less so the voltage regulation of the supply will be improved.
Jeff said no "it's not about voltage regulation" He has also claimed that the improvement does not depend on the amount of current draw change. He stated that the full improvement would still be heard even under conditions of constant current draw.
Jeff also claimed to hear the full improvement even when the basic power supply is followed by two series regulators.
"Some would think you are heavily left-brain dominated dude more interested in 'winning' some trivial debate than, well, being a more open and balanced kind of chappie."
I don't see this as trivial. Some people come to the AA looking for instruction as to how audio electronics work. But instead they read that low DCR in a power supply will make their amp sound better even if the circuit is drawing constant current. They read that a proper sized input choke is a bad thing and are lead to build supplies with high current spikes that do cause audible problems. One reason choke input power supplies are better is the lack of high current spikes that cause noise. If the henry value of the choke is to small (as in Jeff's designs) the supply will act just like a cap input supply, high current spikes and all.
There may be something to what Jeff and Dennis are saying but they should get it straight before preaching it as gospel.
I remember three things:
- Henry P came with a ‘plausible’ explanation as to why low H (first) choke(s) might be beneficial (the flywheel concept);
- That the DCR of the PT played an important role in this explenation;
- That Henry P ordered some stuff and started building a low H / low DCR supply.
I didn’t hear from Henry P since. He might be too busy listening...
If you are out there Henry, any results yet?
IIRC the important thing was the relationship between the choke and the following cap (which is load dependent). The DCR of the transformer wasn't the biggie.
The results of the simulations were interesting. They allowed the first cap to be charged to more than the normal max of 1.4 times the secondary rms voltage. Some wild voltages in the choke troubled me but Henry felt they were a limitation of Duncan's simulation. I got sidetracked by speakers and dropped the matter. I would love to hear if Henry is building his amp.
Many of the things in Jeff's post are common practice. It is just a matter of degrees. Extra current in a PT is always desired and we do use caps that are small for the task at hand (because we want to avoid electrolytic capacitors). We perceive the results as a "faster" amp with "tighter" bass (my best attempt at audiospeak). I might beg to differ on rectifiers but I did just admit to liking the 5u4. And CCS, IT, and choke loading is in favor which reflects the "high dcr load" part.
So perhaps it is the low henry first choke that is deemed too radical around here? Small value line reactors are common place solutions in industrial circuits. Then you have those who report good results with having a choke (instead of a cap) as the last filter element (should we be returning at power supply caps?). Maybe the impedance of the power supply isn't the ticket?
The low DCR of the PT was indeed not the 'biggie', but I remember that Henry pointed out the a low DCR PT 'helps' to get the flyhweel spinning.
I can't find the exact post now, but look at his simulations where he uses low DCR PT's. If you increase the DCR of the PT, than the 'flywheel effect' ( C1 voltage rises above T1 voltage) dissapears.
or password protected. I'm not sure those posts were the REAL Henry. The one we dearly love and miss. =:-0
as I am sure others did. The posts were in keeping with the emails. The sims I ran were as well.
Ivan I have read enough of your posts to see that you have experience with better quality equipment. Not everyone has. When you get into more serious gear, details, which are basically unimportant in lesser designs, can become "make and break" things. For example Bud (O-Netic’s designer) posted down below how much difference the wire for the ground bus made. I can't say my stuff has reached the level where litz wire on the ground buss is a "make or break" item. So it would be easy to dismiss such a statement. But IMHO that is a recipe to insure my designs never reach that level of performance.
I suppose in the end it is all about whose opinions one puts trust in and if what they are saying has any basis of truth with regards to your own work. I think it is fair to say that many things, which we know effect sound, can’t be defended solely by electrical principles or normal measurements. However it sure is nice when you know “why” something sounds better.
he certainly could have benefited from one!
As one who both knows and likes Dennis, and has heard his amps a number of times and has actually met dr lowmu and found him to be quite a nice fellow, I try to stay out of all of this.
I’ve heard many of the amps that are supposed to be “legendary” and have met and know many of the better builders, having attended about a dozen CES’s, a couple VSAC’s and all of the RMAF’s plus a Stereophile show or two.
All that said doesn’t really mean much in the larger context, because my taste in audio might not be the same as anyone who posts here.
What I am sure of after all of this is that there is a shred of truth in many of these things from low DCR power supplies to silver capacitors.
I had the pleasure to meet Kondo San at CES many years ago and again shortly before his death at last years (2006) CES. I feel that musical reproduction is far more than just engineering theory but I also do not feel an obligation to defend my opinions to others who may or may not have had the good fortune to hear all that I have been blessed to hear.
The most sensitive measuring device is not a scope or a spectrum analyzer or even our ears. It’s our heart.
Whatever piece of audio gear allows the music to reach our soul and trigger an emotional response is Ultra Hi-Fi or whatever you wish to call it. If you do not agree it matters not a whit to me, as ones experiences in the area of a very personal nature to be shared with friends, if at all.
to expound a bit on your system and how you came to decide it was right for you?
and the decisions were easy.
It didn't cost much money, I could do most of the work myself, and it sounds pretty good to my ears.
But sadly, it's not Ultra-Fi. :-(
You want a low Z (output impedance) power supply that can supply lot's of instaneous current. Than why not use a large capactor for the last element in the filter? Caps are voltage storage devices, and can (if large enough) supply lots of instaneous current. They need to have low ESR, of course, to well above the audio frequency range. Most "electrolytic" capacitors are rather poor in this respect.
"The rectifier doesn't matter at all-- as to DCR. It is super-active, and can address current/voltage changes rather instantaneously. BUT-- just as in the transformer-- extra current capability here is a winner, regardless of DCR."
I don't see a rectifier as an active device, in the sense that there is no feedback from input to output. A rectifier is essentialy a non linear resistor (low ohms in one direction and very high ohms in the other) that appears in series with the transformer winding. Is a resistor an active device? No ...
The problem with large capacitors is they compete with your output tubes for the power supply’s current. A small capacitor has much less surface area between its elements. Because of this, it also has a lower electro-motive tendency to draw current to recharge itself. The weaker electro-motive tendency of the smaller capacitor also gives up its energy more freely than the large capacitor.
If a transit comes along and demands current, the small capacitor will deliver that current much faster and more easily than a large capacitor. In addition, once the transit is gone, the small capacitor only recharges to the extent that does not interfere with the output tube. However, the large capacitor with its stronger electro-motive tendency will compete with the output tube and be a current hog until it recharges. Therefore, amplifier overload or clipping conditions take longer to recover with larger capacitors in the power supply.
When thinking about power supplies, you must differentiate ripple reduction from energy delivery. While a large capacitor may appear to have better regulation and superior ripple reduction, it is very poor at delivering energy quickly.
I personally find that anymore than 50uF in a single spot tends to slow the amplifier down, and transparency decreases. I just use another LC section if more ripple reduction is needed to get things quite.
, , , use regulation. The input voltage to the regulator has to be greater than the required output voltage by a sufficient margin for this to work, of course. I'm aware that some people don't like to risk being able to hear the regulator and would rather go for the "brute force" approach; each to his own.
All PS capacitors 'compete' with the circuit for current. Small caps permit much larger changes in PS voltage with signal but return to static levels quickly, large caps maintain a more consistent voltage but require more time to stabilize. I find it unlikely - all else equal such as type, construction, quality, etc. - either is quicker than the other in delivering current at audio frequencies.
"While a large capacitor may appear to have better regulation and superior ripple reduction, it is very poor at delivering energy quickly."
Regulation is the ability to deliver energy quickly. The explanation also confuses 'large capacitor' and 'large capacity'. Would ten paralleled 10uF caps delivery energy slower than one 10uf cap? My sense is these power supplies are moving towards straight 'C' with inductive HF filtering (as opposed to using L as a regulator) coupled with a small first L as a pulse smoothing device.
It's not obvious to me why a large cap cannot deliver energy to the load as quickly as a small cap. A large cap has more stored energy than a small cap - why would it not be able to deliver more energy (current) to the load at the same speed as a small cap?
I agree that a larger cap will take more time to charge (draw current for a longer period of time), but it also will take more time to discharge than a smaller cap. That means, I contend, the the effective output impedance of the power supply at a given frequency decreases with increasing capacitance (of the last cap). Isn't that a good thing? Woudn't you want a power supply with the lowest possible output impedance?
> > It's not obvious to me why a large cap cannot deliver
> > energy to the load as quickly as a small cap.
Just listen to a large vs small cap in a coupling situation.
The larger cap does bass better but the smaller cap does
high frequency much better because it is quicker.
> > A large cap has more stored energy than a small cap -
> > why would it not be able to deliver more energy (current)
> > to the load at the same speed as a small cap?
Internal resistance would be my guess.
Why would a Corvette be slower off the line than a Honda?
The Corvette has higher power and much better specs but
also more vehicle weight and more engine mass to accelerate.
Choosing the right cap size is a matter of compromise.
Speed vs ripple reduction.
> > That means, I contend, the the effective output impedance
> > of the power supply at a given frequency decreases with
> > increasing capacitance (of the last cap).
Assuming that the caps have the same internal resistance.
Which they don't - unless one goes to a better cap for
bigger values which would get very expensive very quickly.
As in other posts related to this topic the difference in physical construction between large and small caps is being confused with the notion of large and small capacity. They're unrelated. If a 0.1 uF coupling cap is fast and a 1 uF coupling cap slow, what then is a 1 uF coupling made of ten 0.1 uF caps in parallel, and why? The latter unquestionably has 1/10 the effective and 'lossy' impedances of a single cap.
> > what then is a 1 uF coupling made of
> > ten 0.1 uF caps in parallel, and why?
It is as fast as (or faster than) the .1uFd cap
because the internal resistance is that of .1uFds caps
or maybe even quicker because they are in parallel.
But you have the problem of signal multi-paths
blurring as others have testified to before.
Again a compromise.
Why would a Corvette be slower off the line than a Honda?
The Corvette has higher power and much better specs but
also more vehicle weight and more engine mass to accelerate.
Sure, but how does this translate to a capacitor? I don't think this is a valid comparison. Electrons all have the same mass.
> > Electrons all have the same mass.
Yes but all caps don't have the same internal resistance.
Or internal inductance for that matter.
Nor do all caps have the same "speed".
> > I don't buy the 'big capacitor is slow' argument.
> > As of yet, I have not heard one single plausible
> > reasoning to support it.
I hear it and you don't - fine.
Much cheaper for you.
Or do you hear it and just haven't heard
an acceptable explanation for it?
Yes but all caps don't have the same internal resistance.
True, but is this a function of capacity or just a different parameter?
Or internal inductance for that matter.
I don't think induction really plays a role here. For a lot of (good) capacitors inductance is mainly determined by the length of the leads, that small.
I hear it and you don't - fine. Much cheaper for you. Or do you hear it and just haven't heard an acceptable explanation for it?
Actually I have no capacitor in the signal path anywhere. So I can't comment on that. Ofcourse I did experiment with different capacitors in the past, now I only use GE motor run capacitors. I use the size that I need, varying from 10uF to 80uF. I never noticed a slow or fast effect, but I'll pay some extra attention next time I'm on the breadboard again.
The only thing I said is that I don't buy the 'large capacitors are slow argument' based on the theory presented here. The arguments so far just don't make sense to me.
> > is this a function of capacity or just a different parameter?
Capacity is a definite part it.
As well as design and materials.
> > I don't think induction really plays a role here.
On most probably not but some cheaper ones ...
Especially big electrolytics which Need bypassing IMO.
> > now I only use GE motor run capacitors
Same here - convert as of late.
Especially in speaker crossovers.
I am getting to hate inductors there too.
Even Goertz 12gauge copper foil ones.
> > I don't buy the 'large capacitors are slow argument'
Fine with me, I am only telling what I have learned.
That is all I can do or even want to do -
I am NOT saying that if you don't use small caps or
don't bypass large caps you have low-fi equipment 8^D
I have heard the difference and am reporting such.
Take it or leave it.
Caveat - I haven't heard the need for any bypass on my
40 or 5uFd GE 97Fs used in my speaker crossover though.
....and we can all become good friends.
Spelling is "transient" BTW, but the thinking and post wins the day, so lets not carp.
For 30 plus years, I used to think, if 50 uF is good, 100 Uf is twice as good and I'd load it up. Ray Moth wants to execute it that way, with a large final cap.
But I always knew from this Forum that there was a fraction of builders who liked small capacitance, but I never figured out why. You see, with more uFs, the ripple measured lower on my fancy digital Fluke DMM !! That had to be better. But its worse DYNAMICALLY, and in terms of transparency - as you put it. The high C power supply impedes the music playback.
But my audio friends like Dowdylama prefer 20 uF or lower albeit with high HY Ls, and Dennis likes 50 uF or lower with Low HY Ls, and there are MANY hep others out there who were "tuned into" lower C's. But yours is the first Forum explaination I've seen on this. And, the fact that you tried to express this is appreciated in helpful personally to me, and hopefully others. It promotes understanding.
and not all that important in the greater scheme of things. Others can talk (argue) about it - you do not need to join in.
Don't read it, it really wasn't meant for you.
Low DCR is not of major importance in the power transformer, only high current sourcing capability?
I think there is some curve fitting going on. How can the DCR of the rectifiers not matter if Drlowmu discovered that he had to parallel rectifiers to lower the DCRs of the rectifiers to get ultra-fi sound?
This post is made possible by the generous support of people like you and our sponsors: