Thanks for buying a board from me! Have fun, and go slow, it’s faster.

See Also: Another Tour of a Hoppe’s Brain Adcom GFA-555 Restoration

This document is called “Assembly Notes” and not “Assembly Instructions” because every amp is a little different, and every technician’s approach to this work is going to vary. This is an advanced-level electronics project. It is assumed that you are experienced and comfortable working with audio amplifiers.

The board.

The schematic:

Click here to download the BFA-555 MK2 schematic.
The circuit is true to the original schematic, except for the addition of local power-supply bypass capacitors, consisting of four 47uF electrolytic capacitors with 0.1uF polypropylene capacitors in parallel.

Errors in GFA-555 MK2 Service Manual

  • R117 and R167 are listed as 381 Ohms, they should be 681 Ohms.
  • C103 and C153 are listed as 0.1uF, they should be 1.0uF. (Kit supplies Kemet F461 1.0uF/160V.)
  • C106 and C156 are listed as 50pF, they should be 33pF.

Parts:

Download the parts list here.

Soldering:

  • I recommend a regulated temperature soldering station or at least a powerful hand-held iron like the TS-80 or TS-100. This is a 2-layer board with plated through-holes, and large ground and power planes that wick away heat. No problem with a good iron, but a cheap pencil iron will be frustrating.
  • Getting the solder to flow to both sides of the joint: I use a cotton swab to apply a little paste flux to each and every solder pad on the top of the board. It’s messy, but it helps the solder wick through to the top sides of the solder pads. I touch-up from above the ones that did not flow through. Having both sides of the pads soldered not only looks nice, but is more reliable and serves as a double-check on each solder connection.
  • Solder: I do not recommend using a general-purpose “no-clean” solder, like Kester 44. It’s great stuff, but it’s very difficult to remove by design. Traditional rosin fluxes—like that contained in Kester 44—harden like amber, to keep out moisture and slow any chemical reactions. If you are going to be cleaning your board—and I strongly suggest you do—then this type of flux is very difficult to remove. Instead, I recommend solder wire with a water-soluble flux, like Chip Quik SMDSW.

Populating the board:

  • Start with the resistors and diodes, because they lie flat. The parts list is in recommended order of assembly.
    • Resistor values are printed on the board and use the convention of replacing the decimal point with the SI prefix, “M” K” or “R”. (Megaohms, Kiloohms and Ohms) For example: 47.5 Kilo-ohm = 47K5 or 68.1 OHMS = 68R1
    • Be careful with your decimal places. There are a lot of samey-samey values like 68R1 and 681R, or 47K5 and 475R.
    • Dale resistors are marked using at least two different conventions, that may differ from how they are printed on the board.
    • No shame; use your meter to check resistances if you are unsure.
    • Bend the leads of the Dale resistors so that their value can be read from the top when they are installed.
    • High-precision 0.1% tolerance resistors: There are (3) 22.1K resistors, and (4) 1K. Their locations on the board are marked with a “0.1%” next to the resistor value. The close tolerance assures near-perfect gain matching between channels.
  • Matched transistors:
    • Transistor gain varies with heat, so the two transistors that make up the differential long-tail-pair on the input should be thermally bonded. Copper foil tape and heat shrink is provided for this. Simply wrap the foil tape around the transistors so that it doesn’t touch any the leads, install them, and then install heatshrink. Do not use thermal compound as it may eventually cause oil to seep out and attract dust to the board. The copper tape does a great job keeping the temperatures even, and it doesn’t need help from thermal compound.


Transistors with Heat Sinks:

Populate the rest of the board:

From this point forward, it doesn’t really matter what order you install components, but it’s easier to install the components from smallest to largest.

Don’t forget to attach the thermal breakers, and wires for the RCA inputs.

Testing the input board:

I recommend testing the board before you hook up the output section, and risk blowing out eight expensive output transistors, or the board itself.

Conditions required to operate the board without the output section connected:

  1. Power supply: You can either use the amp’s power supply, preferably with a variac, or you can use two lab power supplies in a split-rail configuration. If using lab supplies, set both to 30V with a current limit of 50mA. Connect ground to both “Signal GND” and “Bypass GND”. Connect +V and -V rails to both channels.
  2. Connect your bias compensation transistors from the output modules.
  3. Install the thermal breakers or short their terminals on the board.
  4. Output and feedback hack: Solder a pair of standard 1/4W 1K resistors to the DRV+ and DRV- output and and tie them together on the other end. (Y-Connection) This is your output. From there, run a short wire to the feedback input.

At this point, the board should run without the output section connected. Inject a 100mV sine wave to the input and you should see a nice clean sine at the point where the two 1K resistors meet.

IMPORTANT: Testing your output modules

Please do not skip this step! You must be certain your output modules are good before applying power to the amp. This GFA-555 is direct-coupled from one stage to the next. If you try to power up the amp with blown transistors on the output modules, the “Domino Effect” may ripple backwards through your amp, blowing up everything all the way back to the input stage. It’s a big job to repair the damage.

All of the transistors on the output modules should be tested with the diode check function on your multimeter. Check out this excellent video on the subject by Mr. Carlson, and this discussion thread at AudioKarma, which contains a excellent guide written by user ‘Echowars’.

The four output transistors are wired in parallel, so if you want to test each one individually, they will need to be taken out of circuit. You don’t need to desolder the transistors entirely. Just remove their bolts, which disconnects their collectors, and then lift one leg of each emitter resistor.

If you find a bad transistor or transistors.: When power output transistors fail, they tend to do so very obviously, and will usually either blow out completely shorted, or completely open. If yours test OK on the diode check function, and show the usual 0.5-0.6V voltage drops, and no shorts, they are probably good. That is no guarantee. The most important thing is that none are shorted, because that’s how input boards get smoked.

If you find that you just have one or two blown transistors, it is recommended that you replace the whole set of four, as they need to match. I recommend On Semi MJ21193/94, but there are many good substitutes. If you order from a reputable source like Mouser or Digikey, you will probably receive a whole set with the same batch code. Manufacturing tolerances are so good these days that a set from the same batch code will likely be decent matches.

I recommend replacing the 2SD1047/B817 driver transistors with NJW3281/NJW1302, for better reverse voltage headroom.

Install into the amp

Connect all the wires.

  • V+ and V- supply lines
  • Drive signals DRV+ (NPN) and DRV- (PNP)
  • Thermal breakers
  • Bias compensation transistors.
  • Feedback wires. You may need to install longer wires for the feedback signals, as the connection points are further away than stock. These are thinner wires that go from the output binding posts to the “Feedback” terminals on the input board.
  • LEDs
  • Input wires
  • Ground wires, see below.

Star-Ground Connections:

The order in which the terminals are stacked onto the star-ground bolt makes a difference in noise! The most important thing, is that the signal ground does not share any conductive paths with charging currents between the rectifiers and the power supply and bypass capacitors.

Stack the star-ground bolt like this: From top to bottom

  • Signal Ground from input board
  • Speaker Grounds
  • Chassis ground wire
  • Bypass Capacitor ground from the input board. (Does not exist on original board. Hoppe’s Brain Input board adds local supply bypass capacitors, or you can add them to your original board.)
  • Bypass capacitor grounds from output modules. (MK2 has these capacitors installed, and they can be added to the MK1.)
  • Transformer center-tap
  • Buss-bar

Testing the amp:

Bring the amp up on a variac if you have one, or use a “Dim bulb tester“. Apply a 0.1VAC sine wave to the input. Monitor the output on a scope. If the amp is working, you should be looking at a nice clean sine wave of approximately 2.2VAC. If not, you have some troubleshooting to do.

DC Offset:

Short the RCA inputs to the amp. You should read less than 5mV at the speaker terminals. If you have more than this, something is wrong.

You should also read around 3-5V from pin 6 of the servo op-amps. This indicates good balance on the input transistors.

Bias Adjustment:

Bias spec is 5mV across the emitter resistors, or 10mV across the test points TP201 and TP301, which spans two emitter resistors in series. I prefer not to use TP201 and 301, and measure each emitter resistor separately.

I recommend checking all 8 resistors on each side. Take an average, it doesn’t need to be exact. Connect the negative probe of your meter to the speaker output and the probe each emitter resistor. Beware of probe slippage! I’ve blown way too many expensive output transistors that way. Now I use these Fluke extended fine point tips, and it hasn’t happened since.

What I do is let the amp sit on the bench, with no clutter around it, with the cover on. From stone cold, I power-on and adjust bias to about 5mV. This may climb as the amp warms up a bit. After 15 minutes or so, I’ll then re-adjust to 5mV. Then I put the cover back on and let it sit for about an hour. Check, adjust, check, adjust…. lather, rinse, and repeat as many times as needed.

What I’m looking for is that the bias hangs around the target, at room temperature with good ventilation.

Once adjusted properly, the heatsinks should feel just a little warm to the touch, about 100F.

That should be it!