CBM-II/B700 PSU fail

[Witchy]

Hi folks,

Recently I dug out my B700 high profile, UK edition, and predictably it seemed dead. Probing with a 'scope showed activity though, and the CURSOR pin of the CRTC gave a trace of a flashing cursor so I checked over the video side of things. I could see a raster on the monitor plus H and V sync on the monitor connector but no video. Following the schematic backwards from the monitor pins I discovered a 74LS86 with a dead output; changed that and yay, a display!

Keyboard was good too so I added my PET MicroSD adapter and while I was testing that I heard a crack. I powered off immediately and pulled the MicroSD then inspected everything around the IEEE bus, all looked OK. Tested the PSU for a short and it was OK, so I put a DMM on the 5V rail and powered up again, turns out the crack was the MC1488 RS232 driver chip shorting out so it smoked and my 5V rail disappeared. Testing the PSU again showed +5V and +12V had collapsed to ~1.3V and -12V had disappeared altogether.

The datasheet for the MC1488 shows it takes in +12 and -12 so I suspected I'd had a short between those two rails and it's taken out...something. Testing the components on the output side however showed a failed rectifier diode pair on the 5V rail (marked BVW51-100) so I replaced that and tested all the diodes which were OK. The caps are below spec though so I'll replace those too.

Has anyone worked on these before? The B700/CBM-II 710 seems to be pretty rare these days. The PSU is made by a company called CEAG and thankfully a schematic is on zimmers - http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/b/cbm700-ceag-PSU.jpg

Cheers!

 

[Hugo Holden]

It would be pretty unusual for an IC like the MC1488 to be able to take out a switch-mode psu. Likely something else shorted out the supply rails (like a tantalum bypass cap) or more likely the psu failed on its own.

As far as switch-mode psu's go, it is a fairly simple one, with an NE5560 as the pulse width modulator, feeding back via an isolation transformer to the chopper transistor on the primary side of the transformer.

I'd suggest rigging up a dummy load for the supply to at least load it with close to 1/2 an amp on the 5V rail, so maybe a 10 Ohm 5W power resistor and then fault finding the supply with the scope. You might find the supply is ok and the fault loading the supply is elsewhere and it is in shutdown mode.

 

[Witchy]

Ah ok, so just coincidence then I hope, given the age of all the components. The reason I suspected a short was because the damage to the 1488 is at pin 12 which is the CS0 line of the ACIA at U7 as well as the Pullup Enable pin of the 75160A at U5, and that is pulled high so it was theoretically possible for +5/+12 and -12 to mix inside the 1488, indeed the only failed part I've found so far is the rectifier diodes on the 5V rail.

I'm sure when I first tested this PSU I did it with no load and got all the right voltages, but now when I probe the + rails I initially see around 3.5V before it collapses so I'll try with an old hard drive as a load and see what happens. I should also pull the 2812 on the -12V rail to see if it's failed too. It didn't test as shorted in circuit but... I wonder if it's also possible that I triggered this by attempting to use my MicroSD adapter on the IEEE bus.

 

[Witchy]

Update on this. With a bulb on the 12V line I still got nothing so I pulled all the electrolytics apart from the weird 4-terminal 100uF inrush cap (if that's the right name for it) and they all tested well below tolerance and/or high ESR. I knew the output rectifier diodes were OK because they're new and I tested them before fitting, the 2812 on the -12V rail is OK too as is the BUV48A chopper (tested out of circuit). Yesterday I replaced all the eletrolytics with new Multicomp types and still got nothing so I started probing various parts with the power off to map the DC side of things and discovered there's a dead short between +5 and +12 on the primary windings before the rectifier diodes. I can't see anything on the schematic that bridges both those rails if it fails so hopefully I'm missing something and it's not the primary transformer that's failed.

 

[iz8dwf]

Update on this. With a bulb on the 12V line I still got nothing so I pulled all the electrolytics apart from the weird 4-terminal 100uF inrush cap (if that's the right name for it) and they all tested well below tolerance and/or high ESR. I knew the output rectifier diodes were OK because they're new and I tested them before fitting, the 2812 on the -12V rail is OK too as is the BUV48A chopper (tested out of circuit). Yesterday I replaced all the eletrolytics with new Multicomp types and still got nothing so I started probing various parts with the power off to map the DC side of things and discovered there's a dead short between +5 and +12 on the primary windings before the rectifier diodes. I can't see anything on the schematic that bridges both those rails if it fails so hopefully I'm missing something and it's not the primary transformer that's failed.

The "dead short" might be intentional as probably one end of both +5V and +12V secondaries goes to a common negative rail (0V), you're measuring basicly the very low resistance of both secondaries in series.
What you shouldn't find is a dead short between any output rail and the common 0V rail.

Frank IZ8DWF

 

[Witchy]

Ah ok, yes, my measuring was from the transformer to the common 0V rail at the output, so the reason I don't get a short on the -12V rail too is because there's a resistor in the way whereas +5/+12 are directly connected. I tested between the outputs and common first which look ok, so I guess I'm safe to power up and probe.

Cheers!