I've done some more reading, and I think it's possibly better to split the signal ground between the IPS/VAS and OPS. Here's my thinking: it's a class A design, so it should be drawing a more-or-less constant current out of the PSU capacitor bank, and then switching between returning it through the speakers and returning it through the power resistors at the bottom of the OPS. So I'll get the most even ground current if I connect the speaker return to the bottom of the OPS, and then return a ground from there to the PSU board.
The amplifier in question (F3) is a single-stage topology, so that actually means separating the input signal ground from the power return resistor ground.
So still 3 grounds back to PSU, just subdivided differently. Conceptually:
(a) IPS/VAS signal
(b) speaker/OPS ground
(c) local filtering/regulation ground
For the F3, that nets out to:
(a') input signal ground
(b') speaker/signal ground
(c') Cmultiplier/Aleph current source ground
Just to test my hypothesis further, the F5 is a PP topology, which I believe means the idle current is returned through the opposing power rail. So there is no OPS ground, and it has no local filtering or regulation either. We're therefore left with:
(a'') IPS/VAS signal
(b'') speaker return
Comments encouraged (even of the "hey you idiot" variety).
The amplifier in question (F3) is a single-stage topology, so that actually means separating the input signal ground from the power return resistor ground.
So still 3 grounds back to PSU, just subdivided differently. Conceptually:
(a) IPS/VAS signal
(b) speaker/OPS ground
(c) local filtering/regulation ground
For the F3, that nets out to:
(a') input signal ground
(b') speaker/signal ground
(c') Cmultiplier/Aleph current source ground
Just to test my hypothesis further, the F5 is a PP topology, which I believe means the idle current is returned through the opposing power rail. So there is no OPS ground, and it has no local filtering or regulation either. We're therefore left with:
(a'') IPS/VAS signal
(b'') speaker return
Comments encouraged (even of the "hey you idiot" variety).
I have room to route transformer primaries straight under PSU board (6mm spacers) down the center line. This is neat, and keeps mains away from sides of an F6 build. But, it's a snug fit under there. And, it would involve wiring thermistors and line cap with primaries into disconnects that attach directly to the IEC blades. Are there strong reasons NOT to do this, and instead to route them around the PSU and connect to mains routed through a terminal block? Thanks for any input!
6mm seems plenty for a mains cable.
Are you using a 2core flex? or a twisted pair?
I would use a twisted pair with an extra layer of insulation to protect the main insulation from damage. I use red sticky insulation tape spirally wound with a 55% overlap along the whole length off all my internal mains cables. Instant recognition of the danger.
Put the soft start Power Thermistor at the end of the cable, either at the input socket end, or at the transformer end.
Are you using a 2core flex? or a twisted pair?
I would use a twisted pair with an extra layer of insulation to protect the main insulation from damage. I use red sticky insulation tape spirally wound with a 55% overlap along the whole length off all my internal mains cables. Instant recognition of the danger.
Put the soft start Power Thermistor at the end of the cable, either at the input socket end, or at the transformer end.
Thanks, AndrewT. Actually, it's an 8mm spacer. My idea (option 1) is to use the transformer primaries, as a twisted pair. I'd route them (2 pairs, twisted, side by side) directly from the transformer, underneath the PSU board (8mm spacers), and up to input socket. I'd place power thermistors at the end of that run, right at the input socket end. I could put some techflex on it and/or wrap in insulating tape as you suggest, for damage protection.
The alternative (option 2) is to use a separate twisted pair as the mains run, routed around the PSU board, joining up with the transformer primaries at a terminal block. Thermistors and line cap would be at the terminal block. This run is a little bit longer, includes a couple more joins, and is nearer to other wiring and one of the F6 amp channels. It does seem to have the advantage of clarity, however.
My main concern about option 1 is any interference. Otherwise, it seems cleaner.
Thanks again for your input.
The alternative (option 2) is to use a separate twisted pair as the mains run, routed around the PSU board, joining up with the transformer primaries at a terminal block. Thermistors and line cap would be at the terminal block. This run is a little bit longer, includes a couple more joins, and is nearer to other wiring and one of the F6 amp channels. It does seem to have the advantage of clarity, however.
My main concern about option 1 is any interference. Otherwise, it seems cleaner.
Thanks again for your input.
Do I need to screw (or solder) the radiators to the PCB? Mine came without a thread in the posts, and the material looks like aluminum, so soldering may be out of question. If I screw the diode to the radiator and just solder the diode to the PCB then it seems to be stable enough, but I thought I would ask the smart people anyway in case I am missing something. Thank you for your help!
Are you sure the pins are aluminium? They're usually tinned brass so they can be soldered.
https://www.aavid.com/sites/default/files/technical/rohs/Solderability_White_Paper.pdf
https://www.aavid.com/sites/default/files/technical/rohs/Solderability_White_Paper.pdf
No, I am not sure. Here are the heat sinks I have got:
513201B02500G Aavid Thermalloy | Fans, Thermal Management | DigiKey
I would think so but ideally you would want 2 to get dual rails, either dc or ac. I don't think that this is a particularly well suited application for this board though. Wall warts are usually low power so that leads me to think that you want to use the board for smoothing. I think there's probably much better ways of doing this. It's also not regulated.
What's your intended application?
What's your intended application?
High temperature rating (105 preferably, these amps do get warm), low esr and high ripple current rating (I think that's the term) . Do a search on these forums, there's definitely a post about this somewhere and how to work out what you need, this is really important. And then the usual like voltage/capacitance etc. Remember it's the rectified voltage and not the transformer output voltage. Multiply tx output by 1.4 and add a bit, that should keep you safe.