Help for open frame dual output PSU - model FSP012-2P01

[Cavia]

Hello,

I'm trying to repair an old (retired) open frame dual output PSU: model FSP012-2P01.
This model is commonly found inside old Extron devices (i.e. 201Rxi, DVI-RGB 150, and more). Since Extron does not provide any schematics (as per company policy), I need some experienced soul that can help me debug the problem (I also contacted FSP asking for the schematics of this retired device, but I got no answer so far).

Let's jump into some details. The model FSP012-2P01 (produced by FSP) is a dual-output PSU with the following outputs: 5V at 1.5A and -5V at 1A.
Note: Regarding a possible substitution with a similar device, the only candidate I found is a Meanwell PD-2505 which unfortunately does not fit into the Extron case. I may keep the power supply outside the Extron in a 3D printed case and route the cables to the logic board, but I would like to keep this option as the last solution.

The faulty PSU is showing the usual ticking [tik tik tik] cyclic-reset behavior. I've already replaced all the capacitors, checked the diod+high resistor (measured 15kOhms) line next to the transformer and looked for shorts. Could not find the problem 
As I'm not an electrical engineer (or expert), I'm a bit lost now. Any help would be greatly appreciated. If needed, I can setup a video call with anyone who wants to help me live (I have a streaming position at my bench).

Here are some pictures of the PSU.

Thanks all.

[voltsandjolts]

Check as many passives and diodes as you can in circuit.
Then I would probably try replacing Q10 and U30 if you can get new ones. What are they?

[wraper]

You say replaced all caps, but did you replace small electrolytic cap on primary side, just in case? Check if diodes on secondary side are not shorted.

[wraper]

Check as many passives and diodes as you can in circuit.
Then I would probably try replacing Q10 and U30 if you can get new ones. What are they?

Q10 would simply short and blow the fuse when fails. Also does not look like UC3842 issue. I suggest checking its datasheet for typical schematic. This also could be an issue with resistor on it high voltage power input or faulty diode from auxiliary transformer winding.

[FIXITNOW2003]

some info here

FSP012-2P01

https://goughlui.com/2015/07/12/salvage-teardown-extron-dvi-rgb-150-iluv-usb-charger-nvidia-9600gt-oem/

[Cavia]

Hello

@wraper, @voltsandjolts

Diodes report:
All diodes work as intended.

@voltsandjolts
Q10 is a Fairchild SSS4N60B 600V N-Channel MOSFET
U30 is a STMicroelectronics UC3842B PWM Controller
Schematics are attached.

[Cavia]

Hello, some updates.

I've replaced Q10 and U30 with new parts. The PSU is still cycle-resetting.

Interesting behavior: I didn't work on this circuit for a long time. When I took it again, I immediately replaced Q10 and tested the PSU (first cold boot in a long time).
The PSU started perfectly (no reset) and it was outputting steady 5V DC and -4.7V DC. As soon as I turned it off, and on again, the PSU was showing the issue again and never restarted correctly.

Not sure what to do next. Any help?

Thanks.

[Cavia]

Updates:

Following some suggestions on Youtube, I measured the DC voltage across the diodes in the secondary part of the circuit: D60 is carrying 5V (ofc it's oscillating due to the circuit reset), while D50 measures 0.330V. The transformer is outputting on all 5 pins 87.1V AC.
In the feedback circuit, the DC voltage across the input pins (1,2) of the optocoupler PC20 shows a fluctuating 1.0-1.1V, while the output pins shows an unreadable DC voltage (random values).

If anyone can spot the problem, or a direction to debug it pls let me know.

Thanks.

[Cavia]

Updates:

I've replaced the Optocoupler PC20 (a classic model 817A), and the Programmable Shunt Regulator IC30 (model KA431A). Still same issue. I'm not sure if the problem is in the feedback area, or the secondary part (after the transformer).

This is challenging

[jd213]

I have the same PSU in my Extron RGB 160 xi that recently stopped working.
Replaced all the caps (all of which were cheap brands, around half of the caps were way out of spec when tested out of circuit).
5V rail seems to be working but not the -5V rail.  Some of the resistors and diodes didn't test right in-circuit, will try testing them out of circuit later and report back.

[jd213]

All of the resistors and diodes seem fine when tested out of circuit.  I believe the transistors (Q10 and IC30) are also fine since my TC1 component tester recognized them.

Would anyone have a guess as to what else could be the problem? Should I test the film/ceramic capacitors?

Not sure how to test U30 (UC3842B) or PC20 (marked L0421 817A, not sure what kind of component this is).  Are either of those responsible for the negative voltage?

Thanks in advance.

[Chris56000]

Hi!

. . .PC20 is a Sharp PC817 Opto–Coupler!

. . .The best way to debug any power supply based on the UC384x series is simply to measure the d.c. voltage between pins 5 and 8 – pin 5 is power ground and pin 8 is the Reference Output, connected internally to the protection circuits within the chip!

. . .If the voltage between pins 5 and 8 is 0.5 V or less, either the chip is locked out due to an under–voltage lockout, a short within the chip itself or across the supply decoupling between pins 5 and 7!

. . .A pulsing voltage indicates the chip is tripping due to either excess current, most likely a defective output rectifier diode or capacitor, (open–circuit or high e.s.r. output capacitors also cause tripping!), or  (rarely!) short–circuit turns in the power switching transformer!

. . .Loss of, inadequate negative feedback from the output side can be due to PC20 (Sharp PC817) going low–emission, will also cause the chip to persistently trip, substitution with a new replacement will prove this!

. . .If the voltage between pins 5 and 8 is 5V approximately and the board is not producing power, then the components between pin 6 (output) and the gate (usually the left–hand leg) of the power switching MOSFET may be defective – a defective MOSFET can destroy these components , commonly a resistor from 22–100 Ω and a protective zener diode of about 15 V 400mW between gate and source of the power MOSFET.

. . .Dry–joints where the transformer lead out wires are soldered to the bobbin/PCB pins is another common cause of persistent tripping or failure to start, particularly if that lead–free muck has been used to make the original board, as is very likely – removing the transformer from the PCB and resoldering the leadout wires where they connect to the bobbin pins is the cure for this!

. . . Rectifier diodes do sometimes short out as soon as power is applied and then promptly test OK when power is removed – binning the lot and fitting suitable replacement types will eliminate this possibility – the "UF" series is suitable as replacements in most cases!

Chris Williams

[Harry_22]

Hi!
The PaintNET diagram will help you.

[jd213]

Hi!

. . .PC20 is a Sharp PC817 Opto–Coupler!

. . .The best way to debug any power supply based on the UC384x series is simply to measure the d.c. voltage between pins 5 and 8 – pin 5 is power ground and pin 8 is the Reference Output, connected internally to the protection circuits within the chip!

. . .If the voltage between pins 5 and 8 is 0.5 V or less, either the chip is locked out due to an under–voltage lockout, a short within the chip itself or across the supply decoupling between pins 5 and 7!

Ah, thank you!!  Looks like it is under 0.5V, but not pulsing.  I'll try replacing UC3842B and PC817A then, doesn't look like they will cost much.

. . .Dry–joints where the transformer lead out wires are soldered to the bobbin/PCB pins is another common cause of persistent tripping or failure to start, particularly if that lead–free muck has been used to make the original board, as is very likely – removing the transformer from the PCB and resoldering the leadout wires where they connect to the bobbin pins is the cure for this!

Oh yeah, I've dealt with lead-free solder before... thankfully this board is from 2005 and seems to use leaded solder.

. . . Rectifier diodes do sometimes short out as soon as power is applied and then promptly test OK when power is removed – binning the lot and fitting suitable replacement types will eliminate this possibility – the "UF" series is suitable as replacements in most cases!

Chris Williams

Hmm, when read in-circuit, D21 read 150 both ways and D50 read at 43 both ways, but the readings were the same both with and without power to the board.  Should I assume they're bad?

Thanks again!

[jd213]

Hi!
The PaintNET diagram will help you.

Yes indeed, that came in quite handy when testing the board when powered on! Thanks a lot!