Exciting news! Hoppe’s Brain power supply boards for the GFA-555 are here!
These boards feature a number of improvements over the stock arrangement with the old-style chassis-mount components:
- Soft-start circuit reduces stress on components like switches and capacitors.
- Reduced hum, buzz and noise from the amplifier’s speaker outputs.
- Capacitance upgrade from 4×15,000uF to 4×27,000uF. Nearly double!
- Earth-loop breaker with diode clamp safety shunt.
- Amp is converted to a 3-wire, earth-grounded chassis appliance, for safety and better noise rejection.
- Surge protection MOV
- Snubber capacitors across AC side of bridge rectifiers
- Compact layout. Overall wiring lengths reduced, fewer mechanical connections, and a much cleaner internal layout.
The power supply is available as a complete kit, supplied mostly assembled, for $375, or as an installed option with a GFA-555 refurbishing for the same price. The board comes tested and assembled except for the big power supply capacitors and one of the AC snubber caps. (The bridge rectifiers need to be soldered in place after the standoffs are drilled, and the board is test-fitted.) Web documentation provided.
I was really pleased with my improved power supplies for the GFA-535 and GFA-545, so next I turned my attention to the GFA-555! I wanted to take everything I’ve learned about the GFA-555 power supply, and integrate it into one circuit board.
Like the GFA-535 MKI, the GFA-555 MKI is not the quietest-running amplifier in the world. There’s a buzz if you put your ear close to the speaker, or if the room is very quiet. This is more noticeable with efficient speakers, or for people using the GFA-555 MKI for near-field studio monitoring. I know this is pretty nit-picky, but I suspect the noise has an effect on the amp’s overall sound quality, so I think worth picking the nits.
The source of the noise is no secret, and not unique to Adcom. It’s due to the proximity of the power transformer’s magnetic field, to the sensitive circuitry on the input board. Here’s the original layout, with the old-style chassis-mount capacitors and bus-bars.
Interestingly, the transformer can be rotated in such a way that one channel can be made dead quiet, while the other channel gets really buzzy. Rotate the transformer the other way, and the effect is reversed. This is due to a null point in the transformer’s magnetic field, and its alignment with sensitive components on the input board.
So, the transformer must be rotated into a compromised position, where each channel buzzes just a little. It comes adjusted this way from the factory. (Sometimes not so well adjusted.)
It’s always bugged me that I couldn’t make both channels nice and quiet.
With the compact layout of the new board, I can move the transformer to the right, and further away from the input board. And both channels go dead quiet!
Check out these FFT’s of the noise coming out the speaker terminals. These are the “before and after” results of the the same amp; First with the OEM power supply layout, (But with new capacitors) versus the Hoppe’s Brain power supply.
The Hoppe’s Brain power supply is shown in green, overlayed and shifted to the right of the OEM supply shown in white.
Caveat: These measurements were taken two days apart. Even though the only variable in these measurements is the OEM power supply versus Hoppe’s Brain power supply, there is some variability in the AC power utility to the house, so values should not be interpreted as highly precise. However, the improvement is so vast and obvious, even to the naked ear, that it swamps any possible deviations in my power line.
It’s a little hard to see detail with my messy overlay, so here’s the individual graphs.
Noise is way down across the board! At 60Hz it’s down 10db in the left channel and 14db in the right. The improvement is even more marked at the higher harmonics, well into the high 20db’s!
Hooked to speakers, the difference is obvious. This amp is dead silent! I have to put my ear right up against the cone of the B&W 602’s at my workbench to hear even the faintest noise.
Note: The GFA-555 MKII has a metal can shielded toroidal power transformer, and is super quiet to begin with, but all the other benefits of this board apply.
I get a lot of people asking about soft-start power supply options for the GFA-555. The 555 has a very generous power supply, with a big toroidal transformer, and 60,000uF total capacitance. It draws a huge surge of current when first switched on. You may notice your room lights dip out for a moment. This in-rush current causes arcing across the power switch contacts, and I often get 555’s in the shop with their power switches arc-welded shut. I always replace the power switch, but it would be better if it the amp didn’t burn up the switch! Most large amps have a soft-start circuit these days. I’ve implemented mine in a time-honored fashion; with a 555 timer and a relay. The soft-start circuit has it’s own linear regulated power supply. I considered using an off-the-shelf PCB-mount switching power supply, instead of this old-fashioned transformer, caps and regulator arrangement, but I was concerned about the longevity of such devices. I’m aiming to make these amps last another 30 years or more.
Which leads to another benefit…
Since the power switch isn’t under so much stress anymore, we can nearly double the size of the capacitors from 4×15,000uF to 4×27,000uF. Also, I get to use the excellent Cornell Dubilier 382LX capacitors.
I convert all the amps I work on to properly safety-earth-grounded-chassis appliances with a three-prong plug. Some isolation of the amp’s ground reference and the chassis is needed to prevent noise from ground loops. A common method is to insert a small resistance between the amp’s ground reference and the chassis, with a small capacitor in parallel to conduct RF. This works great, but gives no protection from power supply faults, such as a power supply wire touching the chassis. So I’ve included a bridge rectifier, also in parallel, acting as a diode clamp for safety. It does nothing in normal operation, but clamps any faults above 1V to the chassis.
More on the earth-loop breaker here…
The original bridge rectifiers were the old-style square types, bolted to the chassis for heat-sinking, and connected to the power supply with wire, spade and ring terminals. The newer, low-profile, flat-pack style bridge rectifiers allow me to mount them directly beneath the circuit board, bolted to the chassis for heat-sinking. Wire lengths are shortened, and 8 mechanical connections eliminated.
The soft-start circuitry is designed to cope with fault conditions such as a brown-out. The circuit will operate correctly down to 75VAC without relay chattering. Chattering is limited to 0.5Hz, the time-constant of the 555. If the relay should fail, then the thermistor will simply heat up and become nearly zero ohms, and the amp will continue to operate, though with an undesirable, variable, small resistance in series with the mains. If the thermistor should fail open-circuit, the soft-start circuit will not operate, and the relay will not attempt to close the connection, causing an arc. If the thermistor should fail short-circuit, then the power switch will take the brunt of the switching, and will arc and probably burn out after some number of cycles. But it will be obvious that the soft-start is not working, as your lights will be dimming much more than when the soft-start was operating. And if by some weird chance the thermistor were to fail half-way and become high-resistance but still conducting, then the soft-start circuit might still receive enough current to operate the relay, which would then be asked to take the full brunt of the in-rush current. It’s a 30A relay, so it should be able to tolerate that for some time.
So far, I haven’t heard of any failures in the field.
The Hoppe’s Brain GFA-555 power supply works in both versions of the GFA-555; Mark I or Mark II. (But not the current “SE” models.) Price is $375, either for a DIY kit, or installed in an amp along with a refurbishing.
Thanks for reading!
Here’s a board installed in a GFA-555 II. I’ve switched to WAGO cage-clamp terminal blocks. These things are awesome! It makes a super tight spring-loaded connection that will not loosen over time as the typical rising-cage clamp types can. These give a high degree of confidence in the connection, and are easier to install as well. There are little levers you push to open the clamp and insert the wire. At first glance, a spring-loaded mechanism might seem like it would be weak, but actually, these things pinch really, really tight, and are designed in such a way that they self-tighten even further if you pull on them.