Another Indigo2 psu died...

marmotta

Member
Feb 22, 2021
60
6
8
I have changed all capacitor on 8001 LV (except the six 6800uf of two rails) the 8002 HV already have new capacitors.... all ok! The voltages is incredibly perfect: 12v is 12v, 3,5v is 3,5v ecc... but it continue to work only with a resistor on 3,5v graphic rail! If I remove it, the psu shutdown. I don’t think is normal, but I can’t explain why.

P.S. I can powerup the psu for only few seconds, the resistor on graphic rail is to small and burn! :-D
 

Attachments

weblacky

Member
Jan 13, 2020
91
23
8
Seattle, WA
I think you have a sense circuit issue. Trace back the sense wire, look for high value resistors. I think you have one more resistors that are no longer in tolerance of their marked values. Start checking your resistors, I bet you find a few that aren’t right anymore (or burnt).
 

marmotta

Member
Feb 22, 2021
60
6
8
I think you have a sense circuit issue. Trace back the sense wire, look for high value resistors. I think you have one more resistors that are no longer in tolerance of their marked values. Start checking your resistors, I bet you find a few that aren’t right anymore (or burnt).
I have checked the sense line resistance and on 8001 and on 8002 it's the same, instead the resistance of 3,5v graphic rail is slightly different (46,5 vs 47,9).

Exist another possibility: this 8001 is a very first revision (Rev.01), I have another 8001 but is Rev.Z and work good without graphic board.... can be a revision bug?
 

weblacky

Member
Jan 13, 2020
91
23
8
Seattle, WA
Please don’t make that mistake. The resistance of the sense pin does in no way check the resistors involved in the circuit. Anything after the first semiconductor on the sense line won’t show up on a ohm meter! So no, I don’t care what value you registered, you’ve already proven you have an overdraw on that sense line. If you’re asking for further advice, check you PSU for out of tolerance resistors.
 

marmotta

Member
Feb 22, 2021
60
6
8
Please don’t make that mistake. The resistance of the sense pin does in no way check the resistors involved in the circuit. Anything after the first semiconductor on the sense line won’t show up on a ohm meter! So no, I don’t care what value you registered, you’ve already proven you have an overdraw on that sense line. If you’re asking for further advice, check you PSU for out of tolerance resistors.
Yes you say right, but without schematics is impossible to track the line... is not a simple pcb, is smd multilayer, is not easy.
 

marmotta

Member
Feb 22, 2021
60
6
8
I do the external source test.... yes! The sense line is ok! I put the external 3,5v to sense line and psu start and stay up! The issue is the 3,5v graphic rail, probably without a load do not provide the correct voltage
 

marmotta

Member
Feb 22, 2021
60
6
8
Now I wait a new 6800 capacitors for better testing ;-)

In the meantime I try to clean the pcb corroded from electrolytic
 
  • Like
Reactions: Elf

marmotta

Member
Feb 22, 2021
60
6
8
All new capacitors arrived! Now the two psu (8001 and 8002) have all new and same capacitors!

First I have reassembled the 8002 un put on Indigo2…. Press start and… start! The 8002! Go!!!! Turn if off, disconnect the power (some time, I need to have dinner :) ) return on my room, reconnect the power cord, press start and… no any sign of life!!! it hasn't turned on anymore :cry:

Now the 8001: reassembled, put on Indigo2, press start and go perfectly! Some test with power cord, all ok! I remove the graphic board…. It no start, I reinserted the graphic board and it start. Like before… but with new capacitors :cautious:

I give up! I have tried everything in my possibilities.

The 8001 is not a problem (I never use the Indigo2 without gfx). The 8002 remain a mystery! Why is powered up the first time and not the second time? :unsure:
 

weblacky

Member
Jan 13, 2020
91
23
8
Seattle, WA
You may have lost a secondary diode on shutdown. It’s extremely common on switching PSUs to damage the unit on shutdown and not at startup.
I’d check your output line resistance, but likely you’ve lost a voltage rail on shutdown. First check your diodes on the LV out out side.
 

marmotta

Member
Feb 22, 2021
60
6
8
You may have lost a secondary diode on shutdown. It’s extremely common on switching PSUs to damage the unit on shutdown and not at startup.
I’d check your output line resistance, but likely you’ve lost a voltage rail on shutdown. First check your diodes on the LV out out side.
The diodes are ok, and resistance are ok. For me the question is: because is started? Is bewitched, or…. The only thing I have changed is the six 6800 capacitors on 5v/3,5v output rails. Is not resolved the previous issue, but it started one time, only first time, when the capacitors are new and completely discharged. The logic of switching psu is very strange and complex, and in this psu is more then others power supplies! This is a possible reason for first start, what it means I don't know.

P.s. To make understand how weird switching is: today I have repaired one psu of Sun IPX, changed all caps but not sign of life… Casually I have changed a brand of one capacitor (both low esr) and is started and now work perfectly.
 

weblacky

Member
Jan 13, 2020
91
23
8
Seattle, WA
At least your seeing the importance of following the cap series and family upgrade path. Different ESR than original spec isn’t always beneficial.
 

marmotta

Member
Feb 22, 2021
60
6
8
At least your seeing the importance of following the cap series and family upgrade path. Different ESR than original spec isn’t always beneficial.
Yes! Some psu has tolerant, others not. But now the 8002 have the same caps of 8001 (and 8001 work) there is something that alters the functioning, is very hard to find without original specs and schematics. And advanced diagnostics tools… I can do only tricks and experiment but is a Russian roulette :confused:

It would take a system to bypass the feedback circuit on LV for find where is the problem.
 

marmotta

Member
Feb 22, 2021
60
6
8
Can be helpful to know the pinout of brown connector, this sent the feedback to HV board and give consent to start. I have tried to remove it completely, the HV start but is dangerous… the voltages are completely wrong and can be damage some components.
 

marmotta

Member
Feb 22, 2021
60
6
8
The 6-pin brown connector:

-3 black wires are ground and connected together
-red wire is 18v with good amp
-first yellow have 0,9v with low amp
-second yellow have 15v with very low amp

I have tried to remove some pins without success (only if I remove the first yellow the psu start with reduced voltage)

I have tried to remove completely the brown connector and psu start! But… I turned it off after two seconds, smoke on resistors on 3,5v zone.

Voltage without brown connector:

12v perfect and stable
5v perfect and stable
3,5…. Raise very fast to higher value (I see 18v and immediately removed the power)

The question is: the 3,5v raise to high values because the psu can’t apply your voltage regulation (brown connector is removed) or because the 3,5v rail have effectively a problem?

p.s. the smoke is only why the resistors are near white glue… the hot burn the glue. The resistors are still ok ;)
 
Last edited:

marmotta

Member
Feb 22, 2021
60
6
8
I tried with working 8001 to remove brown connector… the 3,5v still out of control, the connector is needed to control the 3,5v rail voltage.
 

marmotta

Member
Feb 22, 2021
60
6
8
I have a crazy idea :cool:

Without the first yellow wire the psu start with very low voltages (for example 12v is 4v). With this wire not start… is possibly the voltages is too high (or psu think as) and it stop. I can try to put a trimmer on this wire to adjust the output, and in theory adjust the output voltages of the rails ;)
 
Last edited:

Elf

Storybook
Feb 4, 2019
469
118
43
Interesting, I have never investigated the connectors between the two boards. If I were in your position I would probably investigate more about how those connections figure into the topology of the supply (e.g. are they feedback to the high voltage side, are they some sort of control lines, etc.).

I think aside from getting lucky you may need to dig more into how exactly things are implemented here. Presumably a forward converter, but from there, what are the various blocks on the board and so on, which would assist in troubleshooting. If you don't mind the potential of dramatic failures though I guess you can certainly experiment. I wouldn't call it safe, but it could be interesting :p
 

marmotta

Member
Feb 22, 2021
60
6
8
Interesting, I have never investigated the connectors between the two boards. If I were in your position I would probably investigate more about how those connections figure into the topology of the supply (e.g. are they feedback to the high voltage side, are they some sort of control lines, etc.).

I think aside from getting lucky you may need to dig more into how exactly things are implemented here. Presumably a forward converter, but from there, what are the various blocks on the board and so on, which would assist in troubleshooting. If you don't mind the potential of dramatic failures though I guess you can certainly experiment. I wouldn't call it safe, but it could be interesting :p
Yes! Is a spare not working psu, if I repair it i’m happy, if not… I have tried ;)

The big square connector is surely the power line (is connected to transformer). In the brown connector the three black wires is surely a negative and red wire a power for LV board. The wires to investigate is the two yellow wires.

In the image is the connectors on HV board
 

Attachments

Last edited:
  • Like
Reactions: Elf

About us

  • Silicon Graphics User Group (SGUG) is a community for users, developers, and admirers of Silicon Graphics (SGI) products. We aim to be a friendly hobbyist community for discussing all aspects of SGIs, including use, software development, the IRIX Operating System, and troubleshooting, as well as facilitating hardware exchange.

User Menu