Ripple current Q for amp capacitors

@tester007   7mOhm ESR @ 120Hz is an internal impedance of the capacitor (Equivalent Series Resistance) measured at 120Hz - lower is better.  It represents losses in capacitor.  29.3A @ 120Hz is the max charging current at 120Hz - higher is better.  Value at 10kHz is likely for use in switchers (SMPS).
With higher breakdown voltage  in the same case size - something has to give.
Your amp can deliver a lot of current (375W per channel) while capacitors in power supply are charged in narrow pulses of high amplitude.  There is no way to know for sure if 23.4A in 100V version is enough.  I would stick with exactly the same as original or at least one with the same specifications.

 

Thanks.. I guess that's why they are twice as much. No other brand is close, if it even exists (-10/+75%). I was reading about reforming the large caps. I'm still not entirely clear on a rig to try it, and i've been rebuilding crt monitors and late 70's-early 80's arcade machines as a hobby for 3 decades. But a question I had, and maybe you'd know, is how is applying a low voltage for a while to reform a cap, different than just leaving it plugged in for a long time, then unplugging it and letting it draw down? Shouldn't they naturally reformat themselves during regular use? 

@tester007  Sorry, I've never done it.  When I replaced them once I just turn power on.@atmasphere would know, I'm sure, perhaps he can chime in.

Did replacing the other caps eliminate the slight hum? If so, then the best advice is to leave the big caps alone. If it works, don't fix it.

As far as which of the two caps to use: you'd have to do a little bit of research and see if both versions of the CDE caps were available at the time of the EAD1000 production. If so, then the designers made the choice. 

When I use new large can caps, I hook them up to a bench power supply and run it on 5 volts for an hour then quickly turn up the voltage to the cap's max rated voltage (or the limit of the power supply) and see the if the current rises and drops to zero fairly quickly -- which is what a normal healthy cap does.

Replacing vintage caps it’s often expensive to try to get a perfect match.  New caps are significantly smaller in physical size and have longer life spans. 

Ripple current for the main power supply caps is high, so you need caps with relatively large ripple handling capabilities.  The currents that come from the bridge rectifier are, by definition, high ripple.  A low ripple current cap here will overheat and die in an hour, even when V values are exceeded.  Trust me on that one.  

The DC current that comes OUT of the power supply (transformer, bridge, caps)  to the amp boards is not high ripple though, so small buffer caps (50-100uF or less) don’t need to worry so much.  

Also, all other things being equal, higher temp caps will last longer, even if neither is really temperature stressed, and also improves ripple current handling since  ripple currents cause heating.

Check the physical size, try to stay close in uF but feel free to go up in V and temperature ratings and make sure that you get caps designed to handle AC ripple. 

Of course, a physically smaller cap is also a good thing, so long as all conditions above are met.   If they are too much smaller you may need to apply some glue to ensure they aren’t rattling around on the board.  If your original caps are glued in and these are smaller, don’t rely on the tin leads alone for mounting. 

So it seemed to have a slight hum for a while. I moved out of a long term house, then rented for about 6 years and had the amp sitting in a box. When I built the theater, I bought the second PM1000. I have 11 channels in there (A Parasound Halo rounding it out). Everything is fed through a Tice Signature III power conditioner. I noticed the hum fairly soon I recall, but the build took a year so wasn't in the mood to hit something else and take things apart. But it got to the point I noticed it every time I went in the room.. and this is a treated space. But I think it might have seemed like more based on where I stood. If I put my ear against the wall where a speaker was, I could definitely hear it, and determined which amp it was.

As mentioned, i've rebuilt lots and lots of crt's and arcade pcb's as a hobby, and it is amazing what new caps will do to a crt. But I know those run far hotter. Still, I have always wondered if the same degredation happened in amps. The caps I took out, were as good of a reading as new ones were which completely surprised me. Did the hum go down? I would say it probably cut it in half.. but sadly the years of Orion amps and subs in my car in the 80's took a toll on my hearing so have a permanent slight ringing, so I can only judge by what range I can still hear. I can't stand in an open space and hear it easily like before.. I have to think about it and really try.. but if I go to a speaker I can definitely hear it. I used to have solid silver interconnects, but those started to get flaky at the connector, and I didn't want to have them in the system and not notice a channel  was out (like had happened before). I also thought the wires might be contributing to the hum, but I replaced 'everything' with Mogami wire to eliminate other sources before opening the amp.

The second PM1000 seems to be fine.. but it may have been taken out of a working environment right away and didn't have the 6-7 years of downtime my first one did. It's this downtime that has gotten me curious about reforming the caps. If I could get new date code ones I would just replace.

Thanks for all the help so far.. When I went this route I reached out to Noble Electronics who are the only authorized EAD service center to ask if there were any other possibilities or weak spots, and he only suggested the two large caps. I guess i'll focus on those, and maybe reach out and see if he'd suggest an alternative or what he uses.

I was reading about reforming the large caps. I'm still not entirely clear on a rig to try it, and i've been rebuilding crt monitors and late 70's-early 80's arcade machines as a hobby for 3 decades. But a question I had, and maybe you'd know, is how is applying a low voltage for a while to reform a cap, different than just leaving it plugged in for a long time, then unplugging it and letting it draw down? Shouldn't they naturally reformat themselves during regular use? 

If the caps are old enough, reforming them really isn't a thing unless you like the smell of a power transformer burning up. You might try this on a part only a few years old (5 on the outside). The easiest way is to put an incandescent lamp in series with the part when charging it. When the lamp goes out you're done. If it does not go out you're done also, except the cap goes in the waste bin.  

As far as which of the two caps to use: you’d have to do a little bit of research and see if both versions of the CDE caps were available at the time of the EAD1000 production. If so, then the designers made the choice. 

According to the specifications of EAD powermaster 1000, it is a 5-channel power amplifier with a power output of 200W RMS per channel into 8 ohms and 375W RMS per channel into 4 ohms. 

It looks like the 50V power supply filter capacitor might be a little too low for its duties, did you measure the rail voltage?

When I went this route I reached out to Noble Electronics who are the only authorized EAD service center to ask if there were any other possibilities or weak spots, and he only suggested the two large caps. I guess i’ll focus on those, and maybe reach out and see if he’d suggest an alternative or what he uses.

If budget and space allow, my recommendation would be the EPCOS / TDK B41456B8100M000, 100,000uF 63V capacitors, 12000 Hour life @ 85°C with 4mΩ ESR and 57A ripple current.

Lower ESR means lower internal operating temperature, and every 10°C difference will double or decrease the life of the capacitor! A capacitor rated for 12,000 hours at 85°C will have a useful life of 96,000 hours if operated at 55°C.

https://www.tdk-electronics.tdk.com/inf/20/30/db/aec/B41456_B41458.pdf

It’s half the price of the CD 100k cap so that would be great. Size seems to be about the same as whats in there. No I have never measured the voltage, only had it open for the first time recently. Noble does transformer upgrades and things for EAD parts, and didn’t mention going up. I believe he helped design these. What is the difference, of the original cap being -10/+75%, compared to the TDK cap that is a flat 20%? Can the CD store basically 175,000uf if needed? That seems to be the only parameter that is making this replacement really difficult.

Can the CD store basically 175,000uf if needed?

No, that’s not how it works, modern electrolytic capacitors typically have a tolerance range of ±20% being a common standard. This means that a capacitor labeled with a specific capacitance value can vary that much (either higher or lower) from the stated value when measured. The benefit of wider tolerances is lower production rejection rates for manufacturers, some older or lower quality capacitors may have a wider tolerance range, and for this Cornell Dubilier capacitor, that means the capacitance can vary a lot, you might get one that’s 90,000uF and another that’s 175,000uF.

However, in power supply filtering applications, the exact capacitance value is less critical. And the TDK is superior to the CD in every parameter!

Thanks a ’lot’ for all the help here.. I should have started this thread months ago. Everything I have done in the past has been max 1000uf, or the occasional 3300uf in the audio amps of old arcade games, and topping out at 470 everywhere else. I’ve always just bought 105’s and paid little attention to anything else as the machines after restoration would never be routed, or left on more than the occasional hour here and there and they weren’t exactly made for SQ. But with this amp, it literally changed a part of my life going from what I thought was ’high end’ at the time, a Yamaha rx-v992, to a Lexicon MC-1 and this amp. That first time turning it on and hearing that combo is something I’ve never forgotten so I’ve been incredibly paranoid of not getting the right part in there.

Question on that above though as that’s interesting.. I assume at the time EAD had capacitance testers, and that every cap in this size they got, would be tested first to see where it was at. Why would they pick one like this with potentially a large throw away range? Or did they pick it because on the low capacitance end they would by odds be closer to 100k which was important to not be below that, but on the top end of the error they didn’t care if it was 120 or 150? I’m sure package size would have played a part as well.. there are two of these in each amp.

All the other caps, I think were 100 and 470 on each side of each channel throughout and were mainly Panasonic. I say mainly because there was one, that was a different brand I can't recall right now that had initials written on top. I saved it just in case, but it wasn't special, just a different brand and metered about as well as all the others. Not sure why it was initialed unless it was just to signify who assembled or something.

So when I have read before that a lower ripple is better for SQ because it filters more, if that’s not really true, and that the esr and ripple ratings are ’only’ an indication of how well it will take abuse, and by that it’s lifespan, and that every 100k filter cap will pretty much ’sound’ the same? If people are picking these specific brands like Black Gate, Audio Note, I have to think there is something that makes a difference. Which spec is it? Or is it more what materials are in it.

I’ve been incredibly paranoid of not getting the right part in there.

Question on that above though as that’s interesting.. I assume at the time EAD had capacitance testers, and that every cap in this size they got, would be tested first to see where it was at. Why would they pick one like this with potentially a large throw away range? Or did they pick it because on the low capacitance end they would by odds be closer to 100k which was important to not be below that, but on the top end of the error they didn’t care if it was 120 or 150? I’m sure package size would have played a part as well.. there are two of these in each amp.

Capacitors produced in the same path usually have very similar capacitance and will not cause any problems if you replace both capacitors in the amplifier. Just for a peace of mind, the TDK also comes with 150,000uF 63V. However, I wouldn’t recommend anything over 180,000uF unless the power transformer is up to the task.

BTW, the TDK is a newer design with better technology, and the CD is a three decade old design.  

So when I have read before that a lower ripple is better for SQ because it filters more

I think you got it wrong!

Yes, a power supply with lower ripple is better for SQ. However, this does not mean that using a capacitor with a lower ripple current rating will help SQ, the fact is that you must choose a capacitor with a ripple current rating that exceeds the power supply requirements. There is an equation and I recommend you read this article to better understand how circuit designers select power supply filtering capacitors.

if that’s not really true, and that the esr and ripple ratings are ’only’ an indication of how well it will take abuse, and by that it’s lifespan, and that every 100k filter cap will pretty much ’sound’ the same?

Well, since you have two powermaster 1000 on hand, you can easily compare it, one with CD capacitors, and one with TDK capacitors. IMHO, lower ESR is better, theoretically.

BTW, I should have mentioned this before.  DIYaudio is a great place for this kind of question as many there touch this subject regularly.