|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
70.184.212.128
I got a failed first generation Prius dc-dc converter/inverter in. It has three good identical 2700uF 450V 85 degree Panasonic electrolytics. They're still good, but they're about 15 years old.
Interestingly the module has a liquid cooling system, so the temp may never have exceeded its rating.
Has anyone tried these in an audio tubed amp?
Follow Ups:
Lee
I have 40 or more of those caps I bought brand new in Dayton about 12 years ago and they are 2700uf at 500vdc and I have some that are rated at 450vdc.The problem with those caps is that the ESR is out of sight and it dramatically affects the audio path being the audio rides on the DC rail that they supply.Since Wheezer introduced me to power supply film caps,I never had an inkling to use those monstrous pigs..
Honest amplification is better than excessive 2nd order distortion anytime.
"the ESR is out of sight and it dramatically affects the audio path"
I'm a little skeptical about this, primarily because the the ratio of load Z to ESR is so high. My money's on hysteresis. :)
--------------------------
Buy Chinese. Bury freedom.
Jack
When the 2700uf caps were used in my ALtec 1530s, the mids were compressed and the high end extension was 1300cps lower on measurement until I took those out and had gone to film caps..I understand that bypassing changes that narrative but at the voltages we are dealing with and these were series connected,1350mf is an overkill that most likley changed the output impedance of the power supply. The original capacitance after the huge choke was a total 20uf at 900v.I had put in a pair of 51uf at 1500v solens in parallel and the sonic character changed dramatically just from those two caps as opposed to the series connected 2700uf caps/
I had the same thing with my Cary/AES 811 SET amps..Dennis had used his typical 1200uf series connected caps in the supply rail and I had taken those out and put in a pair of 80uf motorrun caps in parallel and just that upgrade helped the high end extension and it made the amplifier quicker because it was inherently slow..I did a few other upgrades but the motor run caps did the most for it.
I cannot understand why anyone would put that kind of capacitance in an amp at 600 to 800vdc because we aren't talking massive amounts of current at those voltages where you would justify it... YMMV tho.
Honest amplification is better than excessive 2nd order distortion anytime.
Edits: 08/23/14 08/23/14 08/23/14
"I cannot understand anyone would put that kind of capacitance in an amp at 600 to 800vdc because we aren't talking massive amounts of current at those voltages"
I agree with everything you've said, and I completely believe the improvements you heard from the caps you installed. I'm just skeptical that ESR is the culprit. 2,700 uF caps almost certainly exhibit a lot of hysteresis, not to mention their inductance at higher frequencies. As you say, it's just not an appropriate value for this application.
--------------------------
Buy Chinese. Bury freedom.
...Shouldn't the term "hysteresis" be used with some set of parameters such as hysteresis of the B-H curve or hysteresis of the force-displacement curve etc.
This...
--------------------------
Buy Chinese. Bury freedom.
...in that "test". It seems to be voltage coef of cap but he never states it outright. Do you know for sure what he's measuring?
I don't understand your statement, the article explicitly states what he's measuring and how he's measuring it.Short version: he's using an analogy to estimate the ratio of field strength to applied voltage.
Mark Kelly
Edits: 08/25/14
...and if it isn't, then surely the ratio you describe must have a common physics or industry designation or name, esp if it has any practical relevance in electronics. Then there's the question of the parameter's relevance to audio.
"Short version: he's using an analogy to estimate the ratio of field strength to applied voltage."
Yes, provided that by "field strength" you mean the displacement field, D, and not the electric field E. The current flowing through the capacitor is proportional to dD/dt rather than dE/dt, and that's why he gets a D vs. E curve.
I'm not sure why you say "analogy." Seems like the real thing here?
Chris
It's an analogy because he's not measuring D directly, he's measuring an analogue of it.
Mark Kelly
"It's an analogy because he's not measuring D directly, he's measuring an analogue of it."Hmmm...I suppose it comes down to semantics, really. By the same token, one could say that he's not measuring the current directly, but an analogue of it, since he is actually measuring the voltage drop across a resistor, and then using Ohm's law.
In one case Ohm's law is used, in the other case, a Maxwell equation.
If one starts on that path, then in the end all of life and existence is just an analogy, with the only "measurements" being our conscious perceptions in the mind...
Chris
Edits: 08/25/14
"If one starts on that path, then in the end all of life and existence is just an analogy"
You're lucky Ayn Rand is no longer with us. :)
--------------------------
Buy Chinese. Bury freedom.
"Do you know for sure what he's measuring?"
No, not in the terms you're asking about. I've always assumed this to be an artifact similar to the "memory" that large caps exhibit. It's a little out of my field though, so I just take Steve's work at face value as demonstration of a non-linearity in the caps. Maybe "hysteresis" isn't the best terminology, but it seems to describe the phenomenon being portrayed in the tests.
--------------------------
Buy Chinese. Bury freedom.
Pretty sure "memory" in a cap is more conventionally known as "dielectric absorption" and Bench is definitely not testing for that. Additionally, since DA is important in real world applications, there are well defined, standardized, objective test protocols to characterize it. No need to reinvent the wheel.
As an aside, I believe it's always wise to be a bit skeptical regarding claims for something when all the facts aren't clearly presented...or the claims haven't been verified independently.
It is not very clearly explained in that Steve Bench article in the link. But I suppose the essence of it is that he looks at the equation I= C dV/dt for the current passing through the capacitor, and relates that to the rate of change of the electric displacement D (as per the Maxwell equation curl H -dD/dt ~ J). So he has something proportional to dD/dt on the left-hand side of his equation, and something proportional to dE/dt on the right-hand side (since at constant plate separation, the electric field in the capacitor is proportional to the voltage). His "technical note" at the bottom of the article seems to suggest he is integrating the current to get D, and plotting versus V, which gives E. Thus his plots are supposed to be D versus E, I suppose, which would then be indicative of the constancy, or otherwise, of the dielectric constant as the applied voltage varies. A kind of analogue of the B versus H hysteresis curves in the magnetic case.Another way of putting it would be that by plotting the integral of I against the voltage V, his plot is showing how the capacitance C depends on the applied voltage (straight diagonal line = no dependence on voltage).
Chris
Edits: 08/24/14 08/24/14 08/24/14 08/24/14 08/24/14
"Another way of putting it would be that by plotting the integral of I against the voltage V, his plot is showing how the capacitance C depends on the applied voltage (straight diagonal line = no dependence on voltage)."
So yes, voltage coef of capacitance is what he's measuring (actually just displaying...no numbers involved).
The next question is how audible a large and variable V coef of C could be. Maybe spice has a C=f(V) capability so it could be simulated.
I just want to add that statements regarding the audible improvement of replacing power supply caps always bother me when there's no measurement data confirming a change. For example, what was the THD/IM before and after? Did frequency response change?
Just as important, if capacitors have been changed, did it result in a difference in the amplitude/frequency of signal components present at the output of the supply? It has always been my contention that when those components are zero (a theoretical concept only), the supply represents a perfect, zero-impedance voltage source, and in such a case, the capacitors can have no negative impact on sound quality of the amplifier. Derived from this, I also believe that the deficiencies of the supply, including whatever tonal inaccuracy is contributed by the capacitors, can be quantified by artifacts present at the top of the last capacitor.
Unfortunately, very few people take the time to make such a measurement, so there is virtually no data available with which to make comparisons or draw conclusions. Nevertheless, if the spectral content produced at the top of the last capacitor while the amplifier is in operation does not change when power supply components are replaced, I do not believe the amplifier can sound different.
--------------------------
Buy Chinese. Bury freedom.
That huge cap will make for essentially a short drawing a peak of 5.5 amperes (30 ohm transformer secondary) up to 1/2 a second and even after a second will still be drawing over 1 ampere while charged up to the operating voltage.
A soft-start circuit will remove stress from the power switch, rectifier and transformer.
I once had so much capacitance (I don't remember the number but way less than 5600uf or 2700uf) that I needed to use a series resistor that was shorted out after 45 sec. by a time delay relay otherwise the dual 5u4 rectifiers arced. And all the capacitance was after a large critical inductance input choke in a 450vdc LCLC supply.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/22/14
Aren't they going to draw a lot of current at start up?
I have some 5600uf 500vdc Hitachi HCG WA caps that took out the fuse on the Sencore trying to test them.
I charged them, through a series resistor, with my HP 712A tube regulated bench power supply and had to manually ramp up the voltage to keep the current draw below the HP's 200ma rating.
I only charged them to about 200vdc just to see if they held voltage.
I didn't what to charge them to 500vdc, they would be dangerous!
They did hold the voltage very well so I think they are OK.
I have 40 of them NOS.
Any ideas what I should do with them?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Those would be some great caps for a voltage doubler.
I like the implication that they are "safe" when charged to 200V, lol.
"I like the implication that they are "safe" when charged to 200V, lol."
Well.........safer? :-)
You can do some nice portable spot welding.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"Any ideas what I should do with them?"
Use 1 as the B+ reservoir cap. in a high power PP amp. Some sort of film bypass seems indicated, for "speed" and ESR/ESL reduction. Don't forget a substantial choke between the "bad boy" and the I/P cap. of the PSU filter. Even SS diodes might "pop" if the turn on surge is not moderated.
Eli D.
That 85 o C. rating is enough to make me say NO.
Eli D.
Yeah, I was surprised at that. But they haven't vented and they still test good. Two of the MOSFETs (IGBTs?) have blown up, though.
I was thinking of temporatily connecting them in a tubed power amp and seeing how they sound.
On a related track, has anyone tried using supercaps in a solid state amp? You'd need a bunch in series, and it would be expensive, but you could get farads of capacitance fairly easily. I've seen individual caps rated at 2.7V and 3000F. So if you need, say, 100 volts, that would be 40 in series, with equalizers (which are available in kits). 40 would give you 75 Farads! The thing might run for hours or days after shut-off. You'd probably need to incorporate a discharge circuit. But it would be interesting, if expensive, to see what it sounds like.
Cary used some monster caps in the early amps. Giant caps and tube amps rarely make good friends.
Properly implemented they can be beneficial. One issue w Cary is that they use their steel oe stainless chassis as ground. Try tieing in ground points w copper, amps sound much "faster".
I don't know. I have a great preference for chokes which seem to work better with just enough capacitance, When you throw 1000's of uF at a problem you usually get mush.
I prefer 30uF to 4HY choke (50 ohms or less) and 100uF. Very fast recovery power supply with good regulation.
"I prefer 30uF to 4HY choke (50 ohms or less) and 100uF. Very fast recovery power supply with good regulation."Chokes act like current sources. When the load Z falls rapidly - as happens during amplifier current peaks - more of the choke's fixed current output is diverted into the amplifier, with less available to replenish the capacitor. This action allows the capacitor to more fully discharge into the amplifier during the peak than if the choke were not present. Put another way, the choke degrades dynamic regulation and makes it more likely the capacitor will be drained on peaks. If the supply as a whole works well, it is only because the final capacitor is sufficiently large to remain adequately charged during peaks even when replenishment is reduced by the choke.
--------------------------
Buy Chinese. Bury freedom.
Edits: 08/23/14 08/23/14
Lee,Those caps get used fairly hard, and while the temp may not have exceeded that rating (with the cooling system for the inverter working properly) they were in a warm environment nonetheless. On later generation units they had more open area and liquid cooling of the board itself.
I'm not sure they have much life left in them.
Jim McShane - former Toyota technical instructor
Edits: 08/22/14
The module is water cooled.
I know it's liquid cooled - but the coolant temp itself is well above ambient.Those inverters rarely fail, but if they do it is often because the coolant pump quit.
Edits: 08/22/14
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: