LiteOn lit on fire in HP Scanjet 5

[clearchris]

So I haven't powered this machine up for some years, but when it powered it off last, there were no problems. Before powering it up, I opened up the machine and blew it out with compressed air. Everything was clean. So I plugged it in on turned it on. BZZZZZZZZ! Turned it off. Opened the machine, pulled the power supply, hooked up the PS to AC with no load. FIRE! Turned it off.

It looks like a diode shorted and burned.

Luckily I have another identical PS that I can identify the parts. So, get new parts and turn it back on? Should I also replace the same part on the non-burned power supply? Do diodes age? I haven't plugged in the other PS just in case.

Burned section:


Same section on a different power supply (same model)


Bird's eye view of power supply and my messy bench.


LiteOn Label

[stj]

Re: LiteOn lit on fire in HP Scanjet 5

clean up the mess and replace the diode - maybe it was that fucking conductive glue that caused it!

btw, when i saw the word "fire" in the title i thought "damn - now everybody will want one".
i must be watching too many video's from Big Clive and Diode Gone Wild.

[budm]

Re: LiteOn lit on fire in HP Scanjet 5

That diode looks to be part of RCD snubber circuit, check to see if that cap next to it has leakage resistance or not.

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

Looks to me like the resistor R206 burnt up, and you removed it, where is it in the circut?

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

clean up the mess and replace the diode - maybe it was that fucking conductive glue that caused it!

Could be, the other one didn't have glue near the diode though.

Budm: I'll be sure to replace the cap too.

R_J: It definitely wasn't the resistor, the resistor was covered in heat shrink tubing. Once I removed the heat shrink, the resistor was unscathed. I removed it to get a better access to the other components. Even the ceramic cap was relatively unscathed, but the diode was burned to a crisp.

Budm thinks it's part of RCD snubber circuit, and I'll have to defer to his expertise. I'm really weak on reparing power supplies. If this wasn't a very specific form factor, I'd probably just buy a new one.

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

It could have been a poor solder connection, I have seen them start arcing and burn up the pc board or maybe the diode just desided it was time to go.
As the resistor was remove I could not see what condition it was in.

[PeteS in CA]

Re: LiteOn lit on fire in HP Scanjet 5

Did I spot a "9817" date code on a magnetic component? That partly visible Samsung IC (KA384x?) places it no later than the early 2000s.

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

It could have been a poor solder connection, I have seen them start arcing and burn up the pc board or maybe the diode just desided it was time to go.
As the resistor was remove I could not see what condition it was in.

Yep, I initially thought it was the resistor too. I was rather surprised to see 2-3 inch flames and it end up being a tiny diode.

PeteS: Oh, yes, it was old when I bought it and that was around 10 years ago. It is a sheetfeed scanner that is driven by a 486. I installed gentoo linux on it, and it works great, when it's not on fire. This guy did the bulk of the work, I took his released product and tweaked it to do what I wanted. http://www.berklix.com/scanjet/ The really surprising thing is that I haven't found a better network scanner to replace it with. Old HP hardware was real quality. I really wish someone would make a scsi adapter (scanner is scsi) for a raspberry pi so I could chuck the 486 board and upgrade it.

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

Ok, I sacrificed my other PS in the name of science. The diode is marked "1NU 6.G".

The capacitor is marked "MPE103K 630V C". From my searches, this appears to be a Metallized Polyester Film Capacitor with a 0.01uf capacitance and 630v rating.

Would a panasonic ECQ-E6103KF be a good substitute for this? Is there something better to use here?

https://www.mouser.com/productdetail...V8RxLUrW1ls%3D

I'm coming up empty on the "1NU 6.G" diode. Anyone have any ideas about how to substitute these?

[CapLeaker]

Re: LiteOn lit on fire in HP Scanjet 5

replace resistor, cap (higher voltage rated than 630V since it is a snubber cap and the diode. Probably 1 = 1Amp, NU= something ultrafast and the 6 something like 60V?

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

I checked a schematic for a pa4131-2 and it has a similar circuit with a BYV26E in that position, 1 Amp, 1000 volt
Does it look like these 1NU41 diodes... https://ru.aliexpress.com/item/1NU41...517629040.html

[CapLeaker]

Re: LiteOn lit on fire in HP Scanjet 5

that sounds better.

[stj]

Re: LiteOn lit on fire in HP Scanjet 5

so maybe a UF4007 would work.

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

I checked a schematic for a pa4131-2 and it has a similar circuit with a BYV26E in that position, 1 Amp, 1000 volt
Does it look like these 1NU41 diodes... https://ru.aliexpress.com/item/1NU41...517629040.html

Yes, the diode in the PS looks exactly like the aliexpress link (same font and everything) , but instead of the "6.G" there is an "0.K"

The BYV26E sounds like an upgrade. I think I'll go with that. Thanks for your help!!

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

Glad I could help, and I hope it works for you.

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

Ok, I'm ordering the parts for the burned power supply today. So I have turned my attention to the non-burned power supply. I hooked it up to the scanner, and it didn't work. I removed the power supply, and checked the line voltages. I got nothing on the 24v, nothing on the -12v (blue) line, +12v (yellow) was at 11.5v and +5v (red) was pulsing between 4v and 0v about once a second with an audible click. Now I get nothing on any lines. But this might not mean anything because I didn't test with a load.

I found the schematic for the pa4131-2 and I'm currently reading through a switched ps tutorial.

I understand I need to construct a dummy load, so I'm going to get a 4.7ohm 10w resistor to put on the 5v line. Since the output of the 12v line is so small (0.8a) is a load required on this line?

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

I got snowed in today, or what passes for it in Texas. So I spent my day in front of the fireplace reading about switched mode power supplies and watching tutorials on youtube. I think I might have acquired an eastern european accent.

Armed with my newfound knowledge, I went over the backside of the ps with a magnifying glass and found a few suspect joints, one was nearly dry. I fixed the bad joints. I also rigged up a dummy load with a 12v auto bulb.

Now I'm getting good power on the 5v, 12v and 24v lines. I'm getting about -0.1v on the -12v line and nothing on the "power good" line.

So, I think that this now narrows the problem down to the -12v output rectifier and filtering section. Is this true?

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

If you have the other voltages, tracing the -12 volts should be easy
It looks like there is a 4 pin regulator ic located on that heatsink, gussing by that cap polarity around it. get its number, check if you have input, if you have input and no output it may use a control pin to turn it on and off?
If might also need a small load on the -12 volt line.

[clearchris]

Re: LiteOn lit on fire in HP Scanjet 5

If you have the other voltages, tracing the -12 volts should be easy
It looks like there is a 4 pin regulator ic located on that heatsink, gussing by that cap polarity around it. get its number, check if you have input, if you have input and no output it may use a control pin to turn it on and off?
If might also need a small load on the -12 volt line.

If you can see what I think you see, that's pretty amazing!

I'm looking out the -12v output regulator. It's getting 15v solid on pin 1 when the ps is on. Pin 3 (ground) is what is hooked up to the blue -12v wires. Pin 4 is control and voltage is low on this pin. Control pin on the regulator appears to be driven by an LTV817 optocoupler, which has 5v across the led pins (1 &2).

So, is there anything else that could be driving the voltage down, or are we looking at a bad optocoupler here? I have some spare PC123 optocouplers on hand and they seem to be pin-compatible.

[R_J]

Re: LiteOn lit on fire in HP Scanjet 5

Pin 4 needs a min. of 2 volts to turn on the regulator, if the opto has 5 volts across the led, the output pins 3&4 should be low connecting pin 1 to pin 4 via the 2.2k resistor.
you could short pins 3&4 of the opto and see if the -12 is at the output of the reg.