Newbie built a META42; sounds wonderful with OPA637's, but...

[ljordan]

BTW, Thanks everybody for advice and the tips, especially, ppl, Tangent, & PRR for the low resistance feedback advice.

These are now my favorite cans, SONY MDRF1's, (12 ohms). These cans never sounded good until I build this amp. I'm using three BUF634's per channel for the buffers biased with resistors. I'm doing 35ma per buffer. Running off two batteries + - 12 volts. Built it using a bread board. Played around with a lot of chips. The LM6172, AD8610 & 20's, BURRBROWN 2134's, 627's and finally, the 637. The 637 beats them all. The mid range and the highs are just wonderful. I can hear things I can't hear with any of the others. However, the base seems to be not as strong as the 8610. BTW, does anyone have any advice for bringing the base up for this chip?

I also am running two pairs of 637's in my DVPS7700 replacing the 2134's. I thought I would have trouble with stability, because I believe these are used as unity gain buffers for the DAC. You know, taking the positive and the negative out of the DAC with no feedback. I thought that I would try the 637's just as a lark because the 627's were boring compaired to the 8610's. But WOW! IT WORKS!! No stability problems.

Of couse, I'm bypassing the power supply at the chip with 10uf tandelumn's in series with some 2ohm resisters and then these in para with some 1uf ceramics. I'm waiting for the schematic on this player, so I can really know what I'm doing. I want to match the resistors in the filter network coming out of the DAC, but I'm just wondering why is it stable? The 637 in the meta42 is running at a gain of 10, but this, I'm pretty certain, is unity.

Can anybody tell me if I'm about to blow my ears out? At least it would be like dying during sex. Anyway, thanks again to everybody.

Happy tweaking!

 

[ppl]

if a scope is available you might want to look at the output since the op amp can be outputting a 50 MHz RF wave while playing music just fine with only tell tale signes like a slight hum of higher than normal DC voltage on the output . with no Disc inserted put a scope probe on the DCP output jack and see if anything other than digital garbage is comming out. look for wave forms less than 50 MHz and probaly around 10 MHz. this would indicate stability problems in the Analog section probaly from that OPA-637. Note some in fact most OPA637s will work at unity gain however this opa has almost no gain phase margin so if it is not unstable you are lucky.

 

[tangent]

Quote:

does anyone have any advice for bringing the base up for this chip?

You've done about all you can, short of adding a bass boost circuit.

Quote:

I'm just wondering why is it stable?

If Burr-Brown's SPICE model for the chip is to be believed, this chip doesn't become unstable until about 15MHz. If such a high frequency signal never enters the chip, it won't take off into oscillation. Now, just because audio only goes up to 20kHz doesn't mean that you can't have 15MHz+ on the signal line. You're in a digital environment; clock noise is just one of many sources of HF signals that could couple into the signal lines.

ppl's also right to note that oscillation isn't always a complete rail-to-rail wave running at the full slew rate of the chip. Sometimes it's just a small bit of extra noise. You should still be concerned about it.

Bottom line, just because it isn't in wild oscillation now doesn't mean it isn't oscillating or won't oscillate in the future. The whole point of those minimum stable gain numbers is that it is possible to lose stability under the rated gain level, but it doesn't guarantee instability below the rated gain level.