807 Audio Amp Progress Report

[Miles Prower]

The driver for the 807s is complete, finishing the gain stages completely. So far, the voltage amp doesn't show any signs of clipping until it is well beyond the required 45Vp-p that the 807s need. Bias adjust is smooth with a reasonable adjust rate, so that hitting the correct bias shouldn't be a problem.

The Sovtek 6SN7s and 6SL7s show no signs of microphonics, despite some insistant rapping. The 6J5 shows slight disturbances when rapped rather vigorously. The high frequency -3.0db point is 45.5KHz; low frequency -3.0db is around 5.0Hz. No sign of oscillation or other instabilities.

Next: complete screen voltage regulator, wait for the OPTs, prelimn test with a suboptimal OPT that came from junk box. Then complete other channel, then done!

[Miles Prower]

One channel is finally complete after the new xfmrs arrived today. Even without NFB, this really works great. No red glow of death, no smoke, nothing overheating even after playing three CDs through it. Really does Ozzy justice that I'm not getting from the store-bought SS amp. Bass responce is awesome, and there does seem to be a bit of "stridency" to the highs. So far, it doesn't look like this is going to require the full 20db of NFB.

The only problem that needs fixing is overvoltage on the negative rail. I'm getting somewhat less plate current to the finals: 30mA, as opposed to the expected 45mA. Even with that, cranking to full volume makes the overdrive LED on the speeks glow awfully bright. Running full out puts the 807s into class AB2 cleanly, with no hint of audible distortion. (Good deal, including the common plate drivers.) Not getting any audible microphonics from either the new Sovtek's or the NOS 6J5's, 807's, 6AQ5A, 6CB6A. Just that warm, orange glow plus that funky 0C2.

[sorenj07]

sounds like a really cool project! got any pics of the final build?

[Miles Prower]

Originally posted by sorenj07
sounds like a really cool project! got any pics of the final build?

Yep.

[img=http://img204.imageshack.us/img204/5956/807amplc7.th.jpg]

The finalized design includes 6.9db(v) of local feedback to linearize the 807s and reduce the effective r(p) to clear up that nasty high end "stridency", and improve woofer damping, and 4.0db(v) of gNFB to further reduce Zo and compensate for xfmr and general cirucit nonlinearities. Adding more gNFB makes this sound as dead as any SS amp.

As it stands now, the bass packs plenty of "authority" and clarity of detail without any apparant underdamped sloppiness. Everyone who's heard it says it sounds very good indeed.

[Original Burnedfingers]

Do you have a schematic that you might post?

Thanks,
Joe

[Miles Prower]

Originally posted by burnedfingers
Do you have a schematic that you might post?

Thanks,
Joe

I got 'em, but you'll have to e-mail for the details. Can't attach here.

[sorenj07]

well in the first few posts, you post the schem's right?

[Miles Prower]

^^^^

Preliminary schemos, not the finalized design.

[sorenj07]

OK, an email it is then one question though - what is the final wattage of this amp?

[Miles Prower]

Originally posted by sorenj07
OK, an email it is then one question though - what is the final wattage of this amp?

According to the application report, 26.5W. This is with the specified Q-Point: Ip= 88mA; Vgk= -22.5Vdc; Vi= 45Vp-p. When keeping to that, the output comes to about 25W, considering xfmr losses. However, this design includes a rather stiff (some would say overkill) grid driver. With this, the 807s can go into Class AB2 transparantly, and so the full power out is somewhat less than 30W (what the Hammonds are rated for) before the OPTs core saturate. Given the output capabilities of the front end, it's core saturation, rather than actual clipping, that limits the power out.

There isn't any problem getting considerable volume out of the Bose 201 speeks I'm running now.