Summer / Consolation Amplifier: 1995 Adcom GFA-5200

[Andersonix]

Since I didn't want to go back inside again, I decided to throw almost everything at my 5200: IEC connector, RCA sockets, caps, resistors, wiring, jumpers....

It now has 2x 15000uF in place of the original 2x 6800uF per rail, as well as 0R25 in a CRC (68000uF + 0R25 + 15000uF + 15000uF per rail) for over 7 times the original PS capacitance. For inrush protection, it already had a 2R5 NTC on the switch pcb, which is enough to prevent the stock fuse from blowing at turn on.

Since I didn't have the schematic, I decided to change only the emitter resistors from 0R1 to 0R07 and leave the circuit otherwise mostly unchanged (except for different type caps and resistors here and there).

Next I would like to add some inductance in a CLC for the PS and modify the cover so it fits over the new heat sinks.

With the extra heat sinking added, I have the bias set as high as I dare: 0.4A (5W of class A into 8 Ohms), but it takes over an hour to reach a stable temp and sound its best. It keeps sounding better with more and more bias, so I suppose it's time to move on to real class A amps....

[jsfinusa]

Andersonix

Thank you for your comments about the service info and biasing. My 5200 is working well, runs reasonably cool, and I really like the way is sounds. So much so that I swapped it in place of my ST70 tube amp and my whole system sounds better. I was amazed! Who would have thought a $25 broken beater amp from eBay would end up replacing my main system amp.

I did add a small PC type fan under the 5200 which forces air up through the heat sink fins. It is just glued on with some RTV and it works very well. Anyway, I am enjoying the amp and really appreciate folks like you sharing your experiences.

Best regards

[Andersonix]

jsfinusa: Indeed there is some potential bottled up inside the 5200!
If you're confident in the state of your thermal insulators (and its location is cool enough) I suggest aiming for about 0.25-0.30A of bias with the stock heat sink. And since the HS is mounted to the bottom panel for extra heat rejection, raising the amp with some taller feet (like the 1.5 inch tall ones from ApexJr that I'm using) helps air get underneath. I had left myself the option to add two small fans (stuck on with double-sided foam tape to prevent noise), but I don't think I need them for now.

Updates to my 5200: As cooler weather descended, I adjusted the bias to 0.5A (that's 8W of class A), and it improved even further (with possibly some margin for even more bias -- I will aim for 0.65A next time).

Finally, as planned, I connected a couple of 9mH (~0R6 DCR, 19AWG) inductors from ApexJr in place of the 0R25 resistors to make a CLC: Ripple on the rails went from about 4mV to about 2mV, and rails went from +/-40.6V to 39.1V. CLC (CLcc, really) is 68mF + 9mH (0.6DCR, 19AWG) + 15mF + 15mF. The improvement in the highs and low level resolution from this quiet PS was beyond my expectations. Let's just say this amp is no longer the bottleneck in the system...

But it's clear this will never be a great bass amp: Bass is slightly mushy and impact-less compared to my hot-rodded Acurus with paralleled output transistors and similar PS. But that's alright, as it was primarily intended to drive just the ESL panel in my hybrids, which it does really well -- without the 'distortion alert' lights flashing too much!

More pics to follow when a new cover is completed.

[Andersonix]

For you DIY cheapskates following along at home and working up the nerve to re-cap your own little Adcom, the perfectly fitting 15mF 50V Panasonic caps that I used for my 5200 are on sale for about $3 at AllElectronics (see attached), which is less than half the 'retail' price. If memory serves, the 535 also takes this size caps (30mm Dia, 50mm H).

I have now dialed up the bias to 0.7A (over 15W of class A into 8R), and now the HS runs hot, so hot that I can barely hold it for 10 seconds. And after about an hour it sounds glorious, despite the PS ripple going up again to ~4mV. So now I'm battling to get the ripple down again, and thinking about working in some more caps into the new cover (see attached), so that it will have a total of 128mF per rail -- 9.4 times the original! Also I want to find a tin can to put the transformer into.