caps question
Aug 21, 2002 at 9:02 PM Thread Starter Post #1 of 17

HD-5000

1000+ Head-Fier
Joined
Sep 16, 2001
Posts
1,400
Likes
10
My meta42 is still on the drawing board, but I have a dilemma on which caps to choose for C2 and C3.

Option 1
4 Black Gate 470uf 16V caps

Option 2
2 Black Gate 1000uf 25V caps

Option 3
2 Black Gate 2200uf 16V caps

I really can't fit much more than that because the black gates are a little larger than normal.

My amp will be powered by a 24V elpac wall power supply just in case you need to know.


The reason why I want to use Black Gate is because they are commonly refered to as the best electrolytic capacitor, and right now I don't feel like using any second-rate parts. So there you have it.

thanks
 
Aug 21, 2002 at 9:17 PM Post #2 of 17
I recently ordered parts for a meta, and I went with option 1. They are all roughly equivalent price wise, but the 470's are a good 10mm shorter than either of the other two options. I prefer this since I am trying to put them in a slightly small enclosure.

since you are using an elpac supply, i assume its regulated and you wont see too many spikes and whatnot. also, you will be using the voltage splitter, which means each rail will see exactly half of the voltage. this being said, you could get away with 16v caps.
 
Aug 21, 2002 at 9:37 PM Post #5 of 17
Quote:

Originally posted by HD-5000
would the 16V rating pose any risk of a problem?


As you are using a 24v psu, after the rail splitter, it will be 12v, well within the 16v rating of the BG's. But check the output of the elpac to make sure it is putting out 24v. I assume you are using the ad8610/20's? The ad8610/20 is limited to +-13V.
 
Aug 21, 2002 at 9:38 PM Post #6 of 17
as i said earlier.. *probably* not. well.. in ideal situations, no. if you happen to screw something up and one rail gets all the voltage.. the corresponding C2 and C3 would pop. people often say that for safety reasons, use a voltage rating at two times the rail. so if your rail is 12v, then use 25v. this way you do not have to worry about the situation i just described.

another thing you have to worry about with headphone amps is that they draw very low current. a lot of power supplies will drift up a bit when there is a small load. for example, if you have a 24vdc ps, each rail might actually see something like 13 or 14vdc. you are find there, though. first, your elpac is higher quality, so this probably wont happen. secondly, the 16v caps can handle the drift from 12.

so.. in conclusion.. many people will stay true to the safer route of 25v rated caps in this situation. some people, maybe those who are willing to take the risk, will go for the 16vdc.

just food for thought.. maybe you want to build the meta with some cheesy rat shack caps, just to get everythign right.. then once you know everything is kosher, throw in your black gates. this way you dont risk blowing them during the assembly process (this would be when you screw something up and one rail could see 24vdc.)
 
Aug 21, 2002 at 11:33 PM Post #7 of 17
I have hte regulated 24V elpac supply, and it stays true to 24V regardless of the load or AC voltage, so you shouldn't worry about that.

However, there are some "what ifs"...

what if you for a time you have to use a PS that isn't so well regulated?

what if you want to use a higher voltage opamp and want to increase the voltages to or past the 16V limits of the capacitors?

25V is a better bet for use under all circumstances. Also, if enclosure space is not an issue, you don't necessarily have to board mount the capacitors.
 
Aug 23, 2002 at 9:47 AM Post #8 of 17
You missed a few "what if"s, antness:

1. If you short V- to GND in the circuit, the distance from V+ to GND will be 24V, so you'll blow all your + side caps.

2. If you have more buffers on the outputs than on the virtual ground and you manage to short the output, you can drag vground to one rail or the other...same result, *poof* go your caps.

There may be other possible ways to blow the caps. Certainly these possibilities are enough to stop me from trying it. At least, I never test a META42 with more voltage than any of my capacitors can handle. Later, if I have a working META42 and I need to use a higher voltage temporarily, I have done it, but never permanently. It's nice to know you can get away with it, but I'm not comfortable leaving the amp in this potentially dangerous state.
 
Aug 23, 2002 at 4:20 PM Post #10 of 17
I say 25 volts. . . what happens if 1 rail get's shot with all 24 volts? Poof, and that makes me cry.
smily_headphones1.gif
 
Aug 23, 2002 at 4:25 PM Post #11 of 17
Yeah, with caps as expensive as the Black Gates, I'd go with the 25V ones, Just In Case. Especially if you're new/relatively new to amp building...

In general, there's only two reasons to use 16V caps:
1. You can't get any other ones.
2. You can't fit any other ones.

And in both cases, you NEED to make sure that the end user (you) will not ever use a different power supply than the one you design for (Elpacs are good for this, since the DIN connector isn't availble for Rat Shack p/ses).

Another option for the caps (one that I like to promote the heck out of) would be 2 Black Gate 2200uF and 2 Black Gate 220uF caps. (one from each value per channel)
 
Aug 23, 2002 at 8:01 PM Post #12 of 17
In general, higher voltage rated caps has lower esr, too. If space is not an issue use the higher rated caps. If space is an issue, then you have to make a decision: more uf's (better amp performance?) or a greater safety margin.
 
Aug 23, 2002 at 10:09 PM Post #15 of 17
There is one option you seem to be overlooking: save the Elpac power supply for another project and get a 15V model. I'm not aware of any good chips that need more than 15V to sound good. (I'm ignoring the 600 ohm headphone crowd...)
 

Users who are viewing this thread

Back
Top