PSU problem on Server Intel se7320vp2

[idata]

Hi all,

I have a Intel Server se7320vp2

As of lately it doesn't switch on any more, eventhough the power led on the motherboard is on and the NIC leds are flashing.

When I hit the power button, nothing comes on at all, not even the PSU fans.

I removed the PSU for testing and when I short the pins outside the chassis it comes on.

Trying to short pins after plugging in the connectors one by one to the mobo I have come to the conclusion that as soon as I plug in the auxiliary power signal connector it stops coming on.

I am not very knowledgable in electronics and would appreciate help on understanding

1. What is the auxiliary power signal for
2. What could cause the above problem
3. How can I further troubleshoot or fix the problem

Thanks a lot

Mayer

[Edward_Z_Intel]

The 5-pin power signal connector provides I2C monitoring of the power supply. Which chassis/PSU are you using?

[idata]

Hi Edward and thank you for the prompt reply

As I said, I am not very knowledgable in electroniics.

I am an experienced PC technician, but have near zero experience with servers.

From what I gathered "I2C monitoring" on a Intel mobo is called "SMBus" and serves to verify all kind of things.

I would appreciate if you could explain in layman's language what exactly it is used for on the motherboard.

My research has lead me to believe that it might be communicating with the 1st processor and deside according to the feedback it receives whether to disable it and use the 2nd processor only. I might be totally off there, but if you could explain in to be in this type of language it will make more sence to me.

The chassis is an Intel SR2400

the power supply is a HIPRO HP-U450UC3 REV:01

Thanks a lot,

Mayer

[Edward_Z_Intel]

The I2C bus here is used for the onboard mBMC (mini-Baseboard Management Controller) to monitor the status of the PSU. the mBMC is an independent ASIC that monitor the sensors (voltage, temperature, fans, etc.) and controll the system (fan speed control, power on/off, etc.) For more details, you can see section 5.2 of the http://www.intel.com/support/motherboards/server/sb/cs-012666.htm Technical product specification.

In your case for some reason when the aux power signal cable is connected, it seems the PSU is not generating proper Power Good signal, and thus the BMC won't power the system on. Did you check the http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=7470 System Event Log and see if there is any errors related to power supply?

[idata]

Thanks again for the information supplied so far.

I wish I could check the event log, but at the moment I can't boot at all, so no way of getting to the event log.

I have measured the current in the 5 pin connector (outside the chassis).

Pin 1 white-green 4.94 V

Pin 2 white-yellow 4.94 V

Pin 3 empty (I'm prety sure it always was and worked in the past)

Pin 4 black earth\grounding

Pin 5 white-brown -50 mV when not switched on. When switching the PSU on (shorting pins) it goes to 3.32V

I couldn't find a reference to the correct current, but the figures seem to make sence as those are the standard figures.

I would like to clarify again what happened during the initial test (described in my first post).

I removed the PSU from the chassis, plugged it in to the mains, and shorted its pins.

The fans of the PSU came on and remained on.

I plugged in the 6 pin connector to the backpane and shorted the pins.

The fans of the PSU came on and remained on, both, with the ribbon connecting the mobo and the backpane pluged and unpluged.

I plugged in the 8 pin connector to the mobo and shorted the pins

The fans of the PSU came on briefly, but shut down straight away, both, with the 6 pin plugged in the backpane and without it.

I plugged in the 5 pin auxiliary connector and shorted the pins.

No fans came on, but the lights on the backpane came on very briefly.

All that was done without the 24 pin connector inserted, as I was using it to short the pins. I did make sure to push down the little lever which indicates the chassis is closed.

Yesterday I made some more trials.

This time I pushed in some metal wire to the back of the 24 pin connector to get it shorted permanently. This allowed me to plug that connector in too, even though it is totally covered by the ribbon connecting the backpane to the mobo.

The second I connected it to the mains, everything came on and the computer started booting, even without the chassis intrusion button pushed down.

What I gather from that, and I'd appreciate your input, is that either the actual power button is gone.

Or, and I tend to believe this more, one or more of the checks which the system runs before boot, comes back negative and interupts the boot.

Now I have following questions:

1. Should I allow the system to boot into the Bios (by shorting the pins as above) in order to see if there is any info in the logs? Or is that to risky, as whatever it is that interrupts the system could cause permanent damage?
2. Could someone here give me a run-down on all the cross-checks the system goes through before booting (ideally in chronological order) so that I can go through them one by one and put my finger on the problem?
3. Is there any additional explanation that someone can think of, to explain the behaviour of the system?

All your help is really appreciated

Mayer

[idata]

bump