Capacitor voltage rating for Dual power supply

[ruleworld]

Hello,
This should be a very noob question. I want to make a 35-0-35v dual supply like this -

I have four 4700 MFD 50v capacitors. Is it safe to use them in this dual supply or do I need 100v capacitors? The total voltage is 70v so I am not sure.


I opened up a sony cd player today and found 10000 MFD capacitors inside. Does 10000 MFD provide significant improvement over 4700 MFD? I will be powering a LM3886. Thanks.

 

[Einric]

If you're saying that the PSU Rail Voltage will be +-35VDC then 50V capacitors will work just fine.
If you are saying that your transformer secondaries are 35V then I would consider higher voltage caps unless the surge voltage rating on your caps is 60V or higher.

 

[ruleworld]

If you're saying that the PSU Rail Voltage will be +-35VDC then 50V capacitors will work just fine.
If you are saying that your transformer secondaries are 35V then I would consider higher voltage caps unless the surge voltage rating on your caps is 60V or higher.

Transformer output is 24-0-24v so rail voltage should be like 33-0-33v.

 

[Einric]

Then 50V caps are perfectly fine.
Each cap will only see the rail to ground voltage that they are connected to.

 

[ruleworld]

Thanks. Anything about the 4700 vs 10000 MFD capacitors?

 

[Einric]

It depends....
Almost always when the capacitance goes up so does the ESR of the cap.
Low ESR is almost always better.
I would prefer to parallel 2 4700uF caps for the 10kuF value since paralleling the 4700 caps will halve their already lower ESR.
However, if you are paralleling in that little 100nF cap its ESR is already super low.....so.....
YOLO, give it a rip!

I prefer Panasonic FR caps for PSU's.
FM's are good too.....

 

[nigelwright7557]

I always use 10,000uF 63V capacitors in my amps.

 

[havefun]

A new 4700uf may be better than an old 10000uF.
I used 2x10000uf (new, Samwha) with 4 LM3886 and it worked very well.

 

[PRR]

....Almost always when the capacitance goes up so does the ESR of the cap.
I would prefer to parallel 2 4700uF caps for the 10kuF.....

No, ESR tends to go down as uFd goes up. See attached clip from CDE SLPX data.

(Any other result is not natural.)

I have seen it that two smaller caps parallel give a lower ESR than one big cap. This does not show in SLPX data; you should of course read the specs for the caps you favor.

What *does* scale poorly is max ripple current. A cap with 3X the uFd is not even 2X the max ripple. My impression is that this is "never" a problem in audio, particularly hi-fi speech/music. Max ripple is mostly about outside surface, skin on the can. We audio-dudes aim for low ripple which means BIG uFd which means a can which is big for the job. Also we may burst to 100 Watts, even cook for an hour on the bench, but real use is under 10W even under 1W. In an AB amplifier the average current is quite low, and the ripple current also tolerable. For an amp drawing 50V 1 Amp, 10,000uFd, the ripple current is about 1.44 Amps. (Varying VERY little with uFd.) These "best value" caps are rated 5+A in this size. Even at 3,300uFd we have 2:1 safety margin. We get in trouble below 1,500uFd (not on SLPX data) where the can is very small (<1"x1"!) and max rip falls below 1.5A. But also ripple voltage is 16V, which is clearly awful for 50V audio supply.

And this 50V 1 Amp is Full Output. In listening we would be down near 0.2 Amp or less. Cap ripple current is near 0.5A, 7:1 safety factor.