low voltage, high capacitance design

[prof.anderson]

So I'm looking at the pictures page of the thread on the new Ray Samuels tiny amp, and I'm thinking "9V with a 10000mF capacitor?"

If this design is sound, why don't I take my 2x9v MINT, take out one battery and use the space for a BIG capacitor, then put WIMA caps in place of the little power caps on the board?

Any thoughts, even general thoughts on the merits of low voltage, high capacitance design?

 

[bg4533]

I think it has more to do with the opamp than anything else.

Don't just assume the Hornet sounds better than your Mint. Many people seem afraid to admit that DIY amps can compare to commercial amps.

 

[ppl]

Quote:

Originally Posted by prof.anderson So I'm looking at the pictures page of the thread on the new Ray Samuels tiny amp, and I'm thinking "9V with a 10000mF capacitor?" If this design is sound, why don't I take my 2x9v MINT, take out one battery and use the space for a BIG capacitor, then put WIMA caps in place of the little power caps on the board? Any thoughts, even general thoughts on the merits of low voltage, high capacitance design?

I would use the 9 volt Battery with the 10K Cap myself

 

[fewtch]

Well... I'm not really a DIYer (just some electronics classes at a tech college), but doesn't the Go-Vibe sort of do this? Single 9v battery, 2 x 1800uF power supply caps. Not that much capacitance I guess, but at least with AD8620 I can't hear much sonic benefit upping the voltage, which I've tried.

 

[kin0kin]

actually, I believe the 10k uF cap is mainly for filtering purpose(> uF = better filter), beside being a reservoir. RS doesn't like to compromise on PSU, he also mentioned that thanks to the cap, the amp is dead silent.

 

[nikongod]

i highly doubt that the cap is the reason that the amp is totally silent. as mentioned in other places, many peoople have found that a slightly smaller higher quality cap sounds better...

chances are better (good....) that the reason that the amp is silent is one of good design, and proper component matching where appropriate.

further i am willing to bet that shure e5's could pick up hiss on a hornett. never tried it though. there is noise, you just cant hear it. there is always noise.....

 

[peranders]

The size of this cap is a typical choice of the designer and there may be strong technical reasons but also emotional. 4 oz copper in a mA device is a typical emotional thing but not unimportant if you incorporate your soul into the design.

I think we all can agree on that 470-1000 uF is a good technical value and over that is esoteric...

Anybody who knows which opamps he has used? AD8620 maybe?

 

[bg4533]

Quote:

Originally Posted by peranders Anybody who knows which opamps he has used? AD8620 maybe?

AD8397 if the early reports are correct.

 

[Teerawit]

Quote:

Originally Posted by nikongod further i am willing to bet that shure e5's could pick up hiss on a hornett. never tried it though. there is noise, you just cant hear it. there is always noise.....

I listened to the prototype Hornet back in NY with the ER-6i and there wasn't any hiss. I forgot what gain setting it was on but, no hiss, it wouldn't matter.

 

[bg4533]

Quote:

Originally Posted by Teerawit I listened to the prototype Hornet back in NY with the ER-6i and there wasn't any hiss. I forgot what gain setting it was on but, no hiss, it wouldn't matter.

My E5s pick up hiss where the Etys I have tried were dead silent. I am not saying the Hornet will hiss with the E5, just that because it doesn't with the ER6i doesnt mean it won't with the E5.

 

[prof.anderson]

I've listened to the SR-71, and that is indeed a very quiet design. But I put it down to good board design, not the big caps (the SR-71 has two 4700 uF caps). It could also have something to do with the use of SMTs.

My own comparisons between battery supplies and a TREAD lead me to think that ripple from the power supply is NOT the significant factor in noise in these simple amps.

 

[Dougigs]

I agree - -the op amps and buffers used in these circuits all have enormous PSRR (power supply rejection ratio). The MINT amp, for example, has something like 16 constant-current sources inside its op amps and buffers (plus another two outside), all of which cancel/regulate PSU variations better than any capacitor. I would bet that you could hook an unfiltered/rectified supply up to such a circuit and not hear any noise.

But a big capacitor would give you a much lower impedance supply than a battery -- batteries have very high impedance, while modern electrolytics have tiny impedances. So the huge caps might provide some benefit for bass response. On paper, a <1W amp shouldn't need more than a few hundred microfarads of capacitance, but maybe more is better.

Has anyone played with big caps in tiny amps? I'd like to hear.

 

[cetoole]

Quote:

Originally Posted by Dougigs I agree - -the op amps and buffers used in these circuits all have enormous PSRR (power supply rejection ratio). The MINT amp, for example, has something like 16 constant-current sources inside its op amps and buffers (plus another two outside), all of which cancel/regulate PSU variations better than any capacitor. I would bet that you could hook an unfiltered/rectified supply up to such a circuit and not hear any noise. But a big capacitor would give you a much lower impedance supply than a battery -- batteries have very high impedance, while modern electrolytics have tiny impedances. So the huge caps might provide some benefit for bass response. On paper, a <1W amp shouldn't need more than a few hundred microfarads of capacitance, but maybe more is better. Has anyone played with big caps in tiny amps? I'd like to hear.

Above picture was me running my mint off of the 12v rail of a computer switching power supply, no other load, and no extra filtering. Granted, the Shure E2c I used dont sound great, but this combination didnt actually sound bad, not noisy. I havnt really played with high capacitance ever, that mint has a pair of 330uf Panasonic FC caps in it.

 

[prof.anderson]

Quote:

Originally Posted by Dougigs I would bet that you could hook an unfiltered/rectified supply up to such a circuit and not hear any noise.

Tangent's page has a discussion of PSRR for a typical op-amp that suggests otherwise.
But the noise has to be pretty bad, on the order of 100mV.