|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.0.212.246
In Reply to: RE: ALT. 1. posted by rage on January 09, 2015 at 09:11:59
Thanks for responding. There is hardly any meaningful difference in dynamic impedances and ripple between the two supplies, Chris gave the impedance figures, ( 570 Ohms stock, 650 Ohms ALT.1, a 17% increase ) and I have originally posted, ALT 1. is 437 mVAC. versus 457 mVAC ripple with the original 10 HY choke, ( about a 5% decrease with ALT.1 ).However, there IS a 14 times reduction in B+ filter series resistances with ALT.1 to the Finals tube. Which B+ filter would sound better ? Why?
Jeff Medwin
Edits: 01/09/15Follow Ups:
the difference is 80 ohms. that's hardley a difference?
we used to shoot for ~ 200 ohms...we're talking 570 vs 650 ohms now. (when I say "we" I mean those of us that attempt to follow you on power supply design)
if you do believe in using the step response then surely the math that is used to evaluate it is relevenat?
or are we just looking at DCR and using step response to show the smoothness of the waveform?
all you're doing is comparing "under dampened, over dampened and ringing power supplies". (from the JLH documents you sent me)
also, do these 1H 9.5 ohm chokes really exist, or are you manipulating the inductance to get the results you are looking for? since you are now locked in at 30uf and 50uf, all you can vary is the chokes?
Rage,
You write : "the difference is 80 ohms. that's hardley a difference?"
Correct about 17%.
Correct, for all intensive purposes, the supplies are similar in Z and ripple. Chris was just posting the "usual" up here, anything to confuse and obfuscate my question of : which sounds better, one with 270 Ohms series resistances or 19 Ohms, a 14 fold difference !!!
We have another cat below that is talking about P-P amps, and this 91 amp is a SE er, and the question doesn't get answered by him either. Confuse, obfuscate, never answer which sounds better and why.
Rage, since early 2014, we do 2.0 HY at 9.5 Ohm chokes, under two pounds. It would not be any problem what so ever to do 1 HY, small in mass - under two pounds, at the same 9.5 Ohm DCR, or likely even less in DCR.
Jeff Medwin
"which sounds better, one with 270 Ohms series resistances or 19 Ohms, a 14 fold difference"
But you are presenting an incomplete picture here, by only including the DC resistances of the chokes. If you want to add up the DC resistances in that circuit, then you certainly need to add in the DC resistance of the secondary winding of the power transformer. And the rectifier is in the circuit too.
In your preferred PS, for example, any current that flows through L1 must necessarily also flow through the power transformer secondary and also through the rectifier. So just adding together the DC resistances of the two chokes is simply wrong.
Chris
You are dancing around again, we are using the same EXACT power transformer and rectifier tube, so there is nothing else to take into account. We are just looking at two different B+ supply filter topologies, in the same amp.
Now, answer my question Chris.
Which B+ supply filter will sound best? And WHY ??
Jeff Medwin
" We are just looking at two different B+ supply filter topologies, in the same amp...
Now, answer my question Chris. Which B+ supply filter will sound best? And WHY ??"Well, if Dave Slagle's approximation for the power supply is a reasonable one, then your preferred filter set-up gives that impedance resonance peak around 45Hz, in the audible range, which I showed in my previous post. The other filter set-up, disfavoured by you, gives this result instead.
Based on the appearance of the two plots, I would probably expect the one you don't like to give a better rendition of the audio signal. I don't think either of them is anything to write home about. But then, if it were me, I'd just use solid-state rectifiers, big capacitors and a small resistor, and avoid most of the LC resonance problems altogether.
Of course, an entirely different question is which one would give the sound that you personally prefer. That is really a psycho-acoustical and psychological question, which I would not attempt to answer.
Chris
(This is a corrected plot. Had a typo in the Mathematica routine before.)
Edits: 01/09/15 01/09/15 01/09/15 01/09/15 01/09/15 01/09/15
Hi Chris,
I am rather new to this (surprise!), but I am wondering: what are the technical and sonic implications of the impedance peak in the "audible" range. How does it affect what is produced at the speaker drivers?
Does the increased impedance affect the ability to deliver current or affect the voltage at a given frequency? How does this translate sonically?
Just trying to learn.
Regards,
91
"I am rather new to this (surprise!), but I am wondering: what are the technical and sonic implications of the impedance peak in the "audible" range. How does it affect what is produced at the speaker drivers?"
Here is my attempt at modelling in LTSpice what happens. I've set it up so that the power supply is turned on at t=0, there is a quiescent load of 60 mA, and then at t=0.3 sec, after the supply has settled down, a burst of 36 Hz sinewave, with amplitude 60mA, is superimposed, for a duration of 15 complete cycles of the sinewave. This represents the SE amplifier operating at full volume for 15 cycles of the 36 Hz audio signal, and then going quiet again. The trace shows the output voltage of the power supply, with the resulting modulation by the audio signal shown. 36 Hz is near the resonance peak. It can be seen that the resonance continues, decaying away, for a few cycles after the burst of audio signal itself has finished.
Below is an exaggerated version of the same phenomenon, achieved by dropping the output capacitor C2 to 20 uF. This shifts the resonance to about 44 Hz. The trace again shows the voltage at the output of the power supply, this time with a 15 cycle burst of 44 Hz audio signal played at full volume. The resonance is now very prominent, and there are at least a dozen or so cycles of decaying oscillation visible after the burst of audio signal has stopped.
The second simulation is, of course, deliberately chosen with a very small value of C2 in order to exaggerate the effect so that it becomes very obvious what is happening. But the same phenomenon is occurring, in a somewhat less dramatic way, in the first more realistic simulation.
(Of course I'm just using "off the shelf" solid-state rectifiers in my simulations, because I haven't mastered more sophisticated aspects of LTSpice yet. And my transformer construction might leave a little to be desired! But I think these simulations capture the essence of what is going on.)
I'm not sure what audible effects one might expect from this, but it is evident that in addition to the resonant audio ripple on the power supply output voltage, there is also a "ringing" phenomenon after the audio burst signal has stopped, which might or might not have audible consequences.
Chris
Myself, and Dennis, never use solid state rectifiers. Sub in a directly heated rectifier, either a 5V3, 5V3A, or a 5U4G, and see how that goes.
Q.: Have YOU ever heard one of Dennis' amps at RMAF, an owner's home, etc ??
Jeff Medwin
Interesting.... try putting a 5K resistor across the current source.
dave
"Interesting.... try putting a 5K resistor across the current source."
I think maybe that damps the ringing down just a little. But it's a fairly marginal difference, I think. Another thing I've just tried is adding a 100 ohm resistor in series, immediately after the silicon rectifiers, as maybe a slightly closer approximation to a vacuum tube rectifier. The resonant frequency seems to shift up a tiny bit, and the amplitude of the modulation on the power supply increases somewhat, of course. Here's a simulation with the 100 ohm series resistor and the 5K shunting the current source also.
Chris
(I'm not sure why some of the wires in the schematic aren't showing up in this jpg file, but they really are there! I don't seem to be able to find a way of directly saving the LTSpice window as a jpg or whatever, so I've been resorting to taking a screenshot and then cropping it. There must be a better way?)
..
Nice. Seems to confirm that tube rectifier or SS doesn't make a big difference.
Now...do we have a way to estimate what effect this would have on the sound?
Chris
using the same Triad parts posted back then. He indicated his 2A3 amp seemed to roll off the bass below 80hz. After less than an hour of listening he went back to the conventional supply. This was two years ago so maybe Jeff made some improvements since then?
This has been a lot of fun doing these sims.Inspired by the sims posted by cpotl and deafbykhorns, I did a little playing around with some power supplies.
It's cool to have a little more insight beyond PSUD.
It is also reassuring to see how well adding capacitance to the supply really stiffens things up with a "regular choke." It's pretty easy to make that squiggle disappear to a negligible amount ~1V with 300uF
I will also note that my power supplies did not exhibit any ringing, so I will pat myself on the back there too. lol.
My models could use more refinements to suit my specific needs but overall it is very insightful.
Cheers.
Edits: 01/11/15
but will usually require some tweaking when bread boarding. It will at least get you close and give you a visual indication of any potential problems. It's a great little tool and fun to build the amplifier in Spice before spending hundreds of dollars on parts. If anyone is interested in some sim files, shoot me an email and I'll send you some files to get started.
This is interesting stuff that you're all doing-- I suggest you involve Dave Slagle as much as you can.I have a suggestion for you-- perhaps more a question: what happens when your simmed power supply is connected thru an output transformer (inductance, resistance and reactance) to a tube's plate, and also, what happens to the supply when it is tapped and shunt-loaded after the tap-- to also run a driver stage?
What does the power supply do under the conditions of running distributed loading and filtering in the amplifier-- beyond the simple simmed power supply?
I humbly submit that the entire amplifier-- ALL of it-- is the power supply, and its overall performance must be considered as a whole power supply system operating under changing and mixing pulsing conditions.
The question therefore becomes: "how can this possibly be simmed or evaluated if you only consider the main power transformer, the rectifier, the chokes and two capacitors?". Those are but a few of the parts in the power supply!
---Dennis---
Edits: 01/12/15 01/12/15 01/12/15
we can't SIM a complex musical signal......
Wait a minute, has anyone injected several different freq. signals into a SIM yet?
sure you can. Just instruct your voltage source to play a .wav file. You can even save the output as a .wav and listen to how your circuit sounds.
dave
Wow, I'm going to try this one! If I can figure it out
Yes, some guitar amp guys play through spice to simulate various effects before they build stuff.
.
instead of incessant condescending philosophical BS....tell us about your new methods that your cheerleader says are 50% better then anyone else could possibly think of or apply...or just start with how you deal with the inherent distortion ( you admitted to when questioned by Tre). later in the design of your amp... throw the little people a few crumbs...
Dennis HAS shared with us all up here, over the years. Its been posted. Maybe I need to try to do a compilation of his posts, in chronological order, without the rebuttal type posts included.
I would be qualified to do that, because I know him for 30 years, and generally how he thinks. THAT would be a big project. But, I have done this for John Hasquin's supply posts, and Bill Swenson's supply posts....sweet stuff to read. Ask me and ye shall receive, anyone !!
You need to read Hasquin and Swenson first, to get a decent background that makes sense, cause they have done it, heard it, know from their direct personal experience.
Jeff Medwin
.
Love it! If it gives you good music, Dig it!
---Dennis---
BTW, the values you are simming are strictly "mine" as it applies to the Joppa filter, and it is NOT what DF and JM use for themselves exactly !!
Jeff Medwin
"I am rather new to this (surprise!), but I am wondering: what are the technical and sonic implications of the impedance peak in the "audible" range. How does it affect what is produced at the speaker drivers?
Does the increased impedance affect the ability to deliver current or affect the voltage at a given frequency? How does this translate sonically?"
A very good question. I've spent some time today learning some of the basics of how to use SPICE, and I think I've succeeded in modelling a reasonably realistic power supply, with an applied load corresponding to the kind of thing that would arise with a class A single ended amplifier. It seems to confirm more or less the same qualitative features discussed already, in the case of a drlowmu style power supply with small values for the chokes. Namely, some sort of a resonance in the low audible spectrum, resulting in quite a strong modulation of the power supply voltage. Whether it is enough to be audible, I really don't know. But since he says he hears a difference with his power supplies, and since this modulation of the supply voltage seems to be a fairly robust feature of the models, perhaps these two things are indeed correlated.
Chris
"Does the increased impedance affect the ability to deliver current or affect the voltage at a given frequency? How does this translate sonically?"
Weak bass.
I would question this : Does your friend know how to build an amp properly, or, is he like you ??
Jeff Medwin
Weak bass. This was a couple years ago so maybe there's some revisions since then?
It was powering a 2A3 amp that had perfectly good bass response with the traditional supply.
Edits: 01/11/15 01/11/15
Correct, and that is certainly not the case in Dennis' amps or mine, so, the simulation is inaccurate, looking at the wrong thing, or both !!!!
Jeff Medwin
"...the simulation is inaccurate, looking at the wrong thing, or both..."I think the simulation is looking at the right kind of thing, namely how the power supply responds to the kind of loading an SE amplifier might put on it. I'm not claiming that the simulation is completely accurate for the specific details of your rectifiers, etc. And the capacitor and inductor values may not be precisely those you would choose to use in your preferred supply. Those could easily be changed, of course.
But I think the broad features are going to remain qualitatively similar, even if some of those changes are made. If the supply has a resonance in the (low) audio spectrum, then it is going to exhibit some ringing also, in the low audio spectrum.
Chris
Edits: 01/11/15
I don't know why it's so hard for someone who likes that supply to just say they like the sound of a sloppy supply. It obviously sounds better to their ears.
"I don't know why it's so hard for someone who likes that supply to just say they like the sound of a sloppy supply. It obviously sounds better to their ears."
Yes, I agree. Jeff says he likes the sound of it. Basic EE theory and simulations say that it indeed has some characteristic features that quite possibly could account for there being audible differences between it and a more conventional power supply. It should not be too much of a stretch to make the link between these two statements.
Chris
It would be interesting to know how low Drlomu's speakers go.
If they are anything like the 9cu' delirious stereo speakers, they probably don't go low enough to be effected by the impedance hump anyways. :P
Chris,You only theorize.
Have you ever actually gone to a show, etc. and heard a Dennis Fraker 2A3 DC amplifier? Yes or No ?
Any idea why it got " Best Sound at Show " 2005 RMAF by Mr. S. Harrell, writing for Six Moons ??
Jeff Medwin
Edits: 01/11/15
See Jeff's link, first picture on the right.
Top right photo is the rear view of one of Dennis' amps. Notice the Cardas binding posts, and what appears to be an AVVT meshplate 2A3.
Jeff Medwin
I'm sure it was a damn good performing amp with those Slagle transformers.
Though I have experienced only 4 manufacturers in my DIY career, nothing has come close to reproducing the music so beautifully until I purchased a pair from Dave. Well worth the extra $$ and I cant say enough about them until you hear a pair yourself. And did I mention they test really well, I'll have to post a square wave of my amp one of these days. It almost looks like a good transistor amp, you almost wouldn't know a transformer was in the signal path.....
Edits: 01/12/15
For the record, Dennis does not use my transformers. From the looks of it I suspect those are MQ from mike (acrosound)
dave
Hi Dave.
Yes. The guys in the brass covers are mine (MagneQuest).
MSL
Builder of MagneQuest & Peerless transformers since 1989
Hiah Dave.....
Correct, Magnequest.
Hope you are well, long time no see !! My fault. Regards and best wishes.
Jeff
nickel core ;-) , my bad. I was lured by the end bells....
Yours are still better!
Edits: 01/12/15 01/12/15
"Chris, You only theorize."
So, if I understand correctly, you do believe that your power supply gives your amplifiers a special quality that makes them sound better. But you don't like any suggestion that this might be associated with any distinguishing features of the power supply that could be calculated, simulated or measured?
So you therefore dismiss any evidence based on theory or measurement that might support your claim that your power supply gives the sound a special quality, because you would prefer to believe that the special characteristics of your power supply must necessarily be unmeasurable and uncalculable by any methods known to science?
Chris
I don't recall reading somewhere that you said the supply necessarily sounds bad. Thanks for your diligence.
Thanks for contributing Chris.What about from 100 HZ to 20 KHZ, does that matter any ??
Jeff Medwin
Edits: 01/09/15
"What about from 100 HZ to 20 KHZ, does that matter any ??"
Nothing particular of interest, in the higher frequencies. Just the expected fall-off because of the final capacitor. All the "drama" is in the lower frequency part of the spectrum.
Chris
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: