After this horrendous illness/flu/cold/cough thing, I’d not been feeling well enough to really do very much at all. Yesterday was a change to that so I decided to set about fixing the power supply again.
Replaced a couple NPN 2N2222A transistors, a couple tantalum capacitors that had cracked.
Tested the circuit, and the gate of the (new) chopper transistor was sitting at zero, then would rise to about 0.5V and then the crowbar would kick in. The rest of the gates were being modulated by the clock signal.
Replaced the chopper transistor, and up came the circuit, albeit operating at much lower frequency than expected. It was whining away so I dug back further and found another bad transistor. Replaced that and it was sitting with the correct sawtooth, and the 5V line stable.
That should hopefully be good.
I should really put a bunch of lights on the back to indicate the correct presence of all the various voltages.
Having a thought on this overnight
There are 4 choppers, independently controlling either a +27 or -27V supply, three are positive rail supplies, for +5, +12 and +15V
The other is -15V with a separate regulator that bleeds off the -15 supply to create -12.
+15 and -15 are only for the printer.
The +12 and -12 rails have to come up first, the +5 rail chopper is padded with a 100 Ohm resistance, presumably to cause it to come up after the +12 and -12 rails.
However, what’s happening is the +5 is wide open (as if the chopper were short) and the clock doesn’t come up to provide the chopper signal until the +5 has tripped out on the crowbar. I need to check see if the chopper is being held (or pulled) open by a false signal from somewhere because I don’t know why it’s rising up so quickly before the other circuits become active- the others don’t come on until the clock starts to operate.
The more I think about it the more it makes sense, but the less it does because I can’t see why the signal should be operational until the clock becomes active.
Result- the +5 comes on, rises to about 5.8V then kicks out, the rest of the board becomes active and the +5 chopper and crowbar get really rather warm.
I’m thinking this is a problem the board’s had for a while…
Forked out for a set of new tires on the Renault. the old ones were dangerously bad and split.
New ones are very much better. (They need to be at the price). Had an alignment done also.
Ride is significantly improved. Shall see how the fuel economy is.
After a relatively lengthy wait, the replacement chopper transistors came in.
Fitted the spare.
Wobbly photo, but that’s a PIC645, from 1997.
Net result, all voltages present on board. Put the thing back into the chassis finally and hooked it up.
I was presented by a raster. Short, but pressing reset made it expand out then come back. Pressing CAPS LOCK also made the light come on. It wouldn’t go off again, but that’s a sign of life.
The voltages were a bit off, 5V was 4.6, +12 was 11.6 and -12 was -11.7 so I got that all adjusted up.
Had a poke about in a few places on the CPU board. There’s signs of sensible life.
With the viewing hood on, heh..
We have strobe on the RAM.
We have something resembling data elsewhere.
It then took a dump on itself and went unresponsive with the ? light lit and the raster all stretched out. Further investigation required.