mrmanse
Member
Hi everyone,
So I bought a known non-working 5150 last week. As I drove home I really looked forward to some quality time trouble shooting, and man did I get just that!
Started off with the usual visual inspection, and apart from an after market Six Pack Plus equipped with 256kB the machine looked untouched and original. Disconnected everything, connected a home made dummy load (21W on 5V rail, 13W on 12V) and powered up. Good fan and good voltages! Connected mobo and tried again. Good voltages on +5V, +12V but 0V on -5V rail. Ouch, power supply... Took it out and apart. Immediately identified a burnt power resistor and replaced it. New test with dummy load on my work bench checked fine, so power supply went back into the machine for further checks. No POST. So I ran the supersoft ROM, and most motherboard parameters were actually fine, but the floppy disk read failed. So I turned off, swapped drives and turned on - but this time nothing! Checked everything again and again, and soon realized that the PSU output was 0V on all rails. OUCH, intermittent PSU fail... Did a very thorough examination of the PSU componens, all resistors and diodes were checked in place, de-soldered where needed. All electrolytes and tantalum capacitors were desoldered and checked. Some zener diodes were slightly out of value, but no more than a like 5 percent or so. Did some Kamikaze testing with everything still taken apart and the PSU worked again. Could it be a bad soldering? After 37 years?? Back into the machine again, but soon same problem. And with repeated testing I discovered that all the rails were fine or not. My initial assumption that only some rails were bad was wrong, the problem just happened to appear as I went from one test point to another.
So I took the PSU apart again (my otherwise very enthusiastic 8 year old son slowly started to lose interest in the project), but well, I did what had to be done - I made a complete schematic of the PSU in PCB software. After a few long nights I started to really understand the inner workings, and luckily the problem seemed to occur more and more frequently. I did some tests with either or both of the opto couplers disconnected (on the low voltage side), and when OK2 was disconnected the PSU ran just fine, and with it, not well at all. So with the OK2 disconnected I measured the outputs of the eight comparator circuits, expecting to find one of them signalling over- or undervoltage, but no, all were fine. So - if you read this far - pause and take a look at the schematic! Do you see it? Remember - i DID check all the tantalum capacitors. Well, if you thought C30 you are absolutely right. It's there to make sure that if there is a transient error in output voltages, the PSU is to stop outputting immediately AND then not to power up again but to stay offline. I desoldered it again, and now it was acting more like a resistor (140 Ohms), than a 4u7 tantalum capacitor. This was just enough to raise the non-inverting input of IC5 above 0V and activate the fail-safe cascade. A new capacitor and my PSU got a new life. And this wonderful community got a new schematic which I hope can be of use! I'll edit the post and add some pictures too when I have the time, for now only the schematic.
The PSU is a UK made Glesham Lion PLC type 32075. It's also labelled 8529342, and manufactured week 40 of 1983.
View attachment 5150 Power Supply.pdf
So I bought a known non-working 5150 last week. As I drove home I really looked forward to some quality time trouble shooting, and man did I get just that!
Started off with the usual visual inspection, and apart from an after market Six Pack Plus equipped with 256kB the machine looked untouched and original. Disconnected everything, connected a home made dummy load (21W on 5V rail, 13W on 12V) and powered up. Good fan and good voltages! Connected mobo and tried again. Good voltages on +5V, +12V but 0V on -5V rail. Ouch, power supply... Took it out and apart. Immediately identified a burnt power resistor and replaced it. New test with dummy load on my work bench checked fine, so power supply went back into the machine for further checks. No POST. So I ran the supersoft ROM, and most motherboard parameters were actually fine, but the floppy disk read failed. So I turned off, swapped drives and turned on - but this time nothing! Checked everything again and again, and soon realized that the PSU output was 0V on all rails. OUCH, intermittent PSU fail... Did a very thorough examination of the PSU componens, all resistors and diodes were checked in place, de-soldered where needed. All electrolytes and tantalum capacitors were desoldered and checked. Some zener diodes were slightly out of value, but no more than a like 5 percent or so. Did some Kamikaze testing with everything still taken apart and the PSU worked again. Could it be a bad soldering? After 37 years?? Back into the machine again, but soon same problem. And with repeated testing I discovered that all the rails were fine or not. My initial assumption that only some rails were bad was wrong, the problem just happened to appear as I went from one test point to another.
So I took the PSU apart again (my otherwise very enthusiastic 8 year old son slowly started to lose interest in the project), but well, I did what had to be done - I made a complete schematic of the PSU in PCB software. After a few long nights I started to really understand the inner workings, and luckily the problem seemed to occur more and more frequently. I did some tests with either or both of the opto couplers disconnected (on the low voltage side), and when OK2 was disconnected the PSU ran just fine, and with it, not well at all. So with the OK2 disconnected I measured the outputs of the eight comparator circuits, expecting to find one of them signalling over- or undervoltage, but no, all were fine. So - if you read this far - pause and take a look at the schematic! Do you see it? Remember - i DID check all the tantalum capacitors. Well, if you thought C30 you are absolutely right. It's there to make sure that if there is a transient error in output voltages, the PSU is to stop outputting immediately AND then not to power up again but to stay offline. I desoldered it again, and now it was acting more like a resistor (140 Ohms), than a 4u7 tantalum capacitor. This was just enough to raise the non-inverting input of IC5 above 0V and activate the fail-safe cascade. A new capacitor and my PSU got a new life. And this wonderful community got a new schematic which I hope can be of use! I'll edit the post and add some pictures too when I have the time, for now only the schematic.
The PSU is a UK made Glesham Lion PLC type 32075. It's also labelled 8529342, and manufactured week 40 of 1983.
View attachment 5150 Power Supply.pdf
