Some notes about restoring a GFA-5802

…just some thoughts and pictures from this Adcom GFA-5802 I’ve just restored for Jeff in Washington.

This was Adcom’s biggest and most expensive amplifier. A 300WPC monster, with 16 HEXFET output devices per channel. Minimalist in circuit design, with only three gain stages, but maximalist in execution.

wpid-20151005_150129.jpg

These HEXFET output transistors are interesting beasts, and have a bit of cult following. They have a wonderfully flat curve if they are biased heavily enough. Indeed, the GFA-5802 sits idle at about 450VA, at temperatures around 130F on the outside of the heatsinks. The heat ages the capacitors prematurely, so I have installed high-endurance 105C types throughout. In particular, the 4700uF 16V power-supply filters for the high-voltage supply are known to die and cause a loss of bass response. I’ve replaced them with 105C type Panasonic TS-HA’s with a slightly higher capacitance of 5600uF, and higher voltage rating of 25V.  These should hold up much better.

wpid-20151005_150042.jpg

A new power relay, rated at 16A versus the original 12A. I always replace power switches and relays just in case they are corroded over time.

Here are the input boards, which contain the RCA and XLR input connectors, along with the circuitry for the first stage of the amp. It’s a long-tail pair, with cascodes to minimize voltage distortion.

wpid-20151005_150108.jpg

There are four MOSFETs on the input board (IRFD210) and they need to be matched for VGS of less than 10mV between them. These were found to be off by quite a lot, perhaps due to heat stress.
Fortunately, these devices are still available, so I bought 200 of them, built a jig to match them, and sorted out pairs that match to less than 1mV. (Service manual says 10mV) There’s very little trimming of the DC offset when they are matched well.

I’ve replaced the input capacitors with Nichicon FGs, and bypassed them with WIMA MKS4. (Update; a polypropylene cap might be slightly better) There are four small mosfets on this board, which run hot enough to burn your finger, so I’ve equipped them with aluminum heatsinks. Two of these mosfets had actually lost some gain, possibly due to heat stress, so I replaced them. The symptom was that I needed to turn the DC bias pot all the way to one side to get it to zero.

Here’s the un-modified board. I don’t like all these unreliable pin-connectors, and have removed most of them from the amp and soldered the wires instead. Notice the pin-connector on the RCA input!

wpid-20150905_141553.jpg

It would be ironic to use fancy, expensive RCA cables, when there is this weedy little push-pin connector on the inside. I’ve soldered it permanently.

The amp boards.

wpid-20151005_150206.jpg

It sounds amazing! I’d say it has kind of effortless quality to it. Like a motor with a huge flywheel. Smooth sounds came out of the studio monitors. Hooked to my subwoofer, it was utterly unconcerned with the large current demands, and delivered tight, effortless bass as a matter-of-fact.

I measured 335WPC with both channels driven into 8 ohms, just before clipping.

UPDATE: 2021.04.26:

I get a lot of emails about the GFA-5802, from people looking to repair one that has cooked itself to death. I have no parts for this model, and generally speaking, I don’t sell parts, except for those that come with my PCB kits.

I stopped working on this model a long time ago; it’s just too problematic. The GFA-5802 is essentially a budget version of a Nelson Pass-esque amplifier—though he did not design it—and IMO opinion, too many corners were cut to meet a price point. For one thing, there is not nearly enough heat-sink for an amp with this much idle current consumption.

My issues with this amp are outlined here in this thread on DIYAUDIO.COM. I am forum user “Phloodpants”.

It’s a shame because the 5802 sounds great and performs great, but there are good reasons that Threshold and Pass amps of this type cost so much.

Leave a Reply