I have something you can check quick.. With the unit powered down and unplugged, bottom cover off, set your multimeter to ohms or continuity, doesn't matter which, and see if you have continuity or very low DC resistance between any of the E, B, or C pads of Q6, the metal can 2n3053. There's have been a few cases of builders setting that component right down against the board and shorting it out against the pads.
Do the same for the three pads under Q5 as well. A failed-short transistor could also be a cause here, although I don't think we've seen that come up yet on these kits, you never know. Only takes a minute to check.
While you're down there you could even flip the meter over to diode mode and check each junction individually, if you read the link Mike posted and understand it.
You shouldn't have +30 rail voltage all over the place like that! I'm surprised it's not blowing R32. That's what usually happens.
JingleDjango said:
That's a super lucid explanation, thanks! I have to wonder, if the violet transformer wire is in place to ground the PCB, why do we also run a wire from C25 to the chassis? Redundancy? A thicker gauge for greater current?
The violet wire doesn't ground the PCB. It's the signal return path for that transformer winding. Just like you're guitar cable doesn't work if the shield wire comes undone, or a lightbulb stops working if someone disconnected the neutral wire, sometimes transformers need to be grounded. Transformers are actually (usually) the exception to the rule, being that they are often left "floating" (see Rev D schematic, the blue/red output winding doesn't reference ground.. it's referenced to itself!), but this winding is referenced to ground since the rest of the output amp is also referenced to ground.
If built correctly the PCB should only be earthed at one point and one point only. PCB power supply negative should have one wire flying out to the chassis.
The output transformer body also has a shield wire to be earthed. The bare wire. This wire by itself does not have continuity to any of the windings on it's own, but if the kit is built correctly it will still have continuity with everything else that's grounded, including the PCB's power supply negative rail and the rest of the ground plane, and the purple wire. The bare wire is intended only to help shield the transformer body from interference.
I know it's a lot to take in, but probe around if you have a minute, meter set to ohms, unit powered down and unplugged, and check all your ground/earth connections by connecting the black lead to PSU negative (power transformer center tap) and following along with the schematic probing with the red lead see just how many things are tied together there! It's crazy. The ground prong on the mains, every panel of the chassis, all of that should show no more than an ohm.
JingleDjango said:
Okay, that's useful information. I had assumed (i guess wrongly) that the same part number would mean the same pinout, regardless of manufacturer. I can't find datasheets for these because I bought them from an electronics parts shop that sorts them by part number but doesn't list any other info. The 2N3707 has literally no discernible logo, and neither does the 2N3053. The 3391 looks like this picture. Not sure what these colour bands are supposed to mean.
I'll swap the brand name transistors in for Q5 and Q6 when they show up.
Usually the same part number will have the same pinout, but not always. Often times you can google up the specific manufacturer's datasheet. Central is pretty good about having theirs' up and available for free. It's never a bad idea to try and find the correct manu just to be certain.
Those painted stripes on old parts from the 70's and 80's, as I understand it, is what's called having been "gain binned", where some poor bastard actually measured them (for hfe/beta I guess) and sorted them and or marked them according to their transfer characteristics.
I've done some light experimenting with several types I've bought like this, and it seems to be true that's what it is, I've been able to roughly group parts by hfe just by sorting the colors first and then checking with a Peak Atlas DCA55, but there's always a few that just don't fit in the groups so I suspect there may be more to it that just "gain binning".
Haven't been able to find much info online about the phenomenon. I'd like to know who did it and why, and how can we and hobbyists and technicians use them to help us build and repair shit.
Seems normal up to that point. For whatever it's worth, Q6 generates no heat. This was worrying me so I found a replacement transistor and swapped it in, but still no heat. On my working compressor, Q6 gets hot to the touch (obviously there's a heatsink for a reason).
