HP ML115 G5 shuts itself off a little while afrer pushing the power button

We have (quite an old, yeah) HP ML115 G5 AMD based server which shuts itself off 10-15s later (during fan test I suppose) after pushing the power button and before BIOS POST single beep appears.

We need some help with remote (200 km) hardware failure diagnosis. Our hardware specification is as follows:

root@linux:~/# dmidecode -t1
# dmidecode 2.12
SMBIOS 2.5 present.

System Information
        Manufacturer: HP
        Product Name: ProLiant ML115 G5
        Serial Number: CZC94743QJ
         SKU Number: 470064-894`

root@linux:~/# head -n 30 dmidecode.txt 
# dmidecode 2.12

Handle 0x0000, DMI type 0, 24 bytes
BIOS Information
        Vendor: HP
        Version: O18    
        Release Date: 07/06/2009

At this point it works stably. I've managed to turn it on by:

• turning the server off,
• removing the power cord for five minutes,
• putting it to the ground on the longer edge, CPU heatsink turned in the direction of ceiling.

If we put it into standard position as it is supposed to stand it doesn't turn on like I wrote at the beginning. Totally reproducible.

Voltage/Temp/Fans stats look okay to me:

root@linux:~/# ipmitool sdr
POST Error       | Not Readable      | ns
Memory ECC       | Not Readable      | ns
ACPI State       | 0x01              | ok
PCI Reset        | 0x00              | ok
CPU Fan          | 1048.88 RPM       | ok
Rear Fan         | 2107.04 RPM       | ok
CPU Diode        | 26.50 degrees C   | ok
Front Ambient    | 19 degrees C      | ok
System 12V       | 11.93 Volts       | ok
System 5V        | 5.12 Volts        | ok
System AUX 5V    | 4.98 Volts        | ok
System 3.3V      | 3.39 Volts        | ok
System AUX 3.3V  | 3.33 Volts        | ok
CPU Vcore        | 1.07 Volts        | ok
CPU 12V          | 11.82 Volts       | ok
HT 1.2V          | 1.20 Volts        | ok
Mem Vcore        | 1.81 Volts        | ok
MEM VTT          | 0.90 Volts        | ok
MCP55 1.5V       | 1.50 Volts        | ok
MCP55 1.4V       | 1.40 Volts        | ok
Therm-Trip       | 0x00              | ok
CPU Prochot      | 0x00              | ok
System Reset     | 0x00              | ok
NMI              | 0x00              | ok
PCI Error        | Not Readable      | ns
CPU Socket       | 0x01              | ok
LO100 Present    | 0x00              | ok
Watchdog         | Not Readable      | ns

IPMI events:

  18 | 03/18/2015 | 09:29:46 | Temperature #0x20 | Upper Non-critical going high | Asserted
  30 | 03/18/2015 | 09:30:08 | Temperature #0x20 | Upper Critical going high | Asserted
  48 | 03/18/2015 | 10:38:59 | Temperature #0x20 | Upper Non-critical going high | Asserted
  60 | 03/18/2015 | 10:39:20 | Temperature #0x20 | Upper Critical going high | Asserted
  78 | 03/18/2015 | 10:45:26 | Temperature #0x20 | Upper Non-critical going high | Asserted
  90 | 03/18/2015 | 10:45:30 | Temperature #0x20 | Upper Non-critical going high | Deasserted
  a8 | 03/18/2015 | 10:45:56 | Temperature #0x20 | Upper Non-critical going high | Asserted
  c0 | 03/18/2015 | 10:46:12 | Temperature #0x20 | Upper Critical going high | Asserted
  d8 | 03/18/2015 | 10:48:42 | Temperature #0x20 | Upper Non-critical going high | Asserted
  f0 | 03/18/2015 | 10:48:46 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 108 | 03/18/2015 | 10:49:04 | Temperature #0x20 | Upper Non-critical going high | Asserted
 120 | 03/18/2015 | 10:49:18 | Temperature #0x20 | Upper Critical going high | Asserted
 138 | 03/18/2015 | 10:50:24 | Temperature #0x20 | Upper Non-critical going high | Asserted
 150 | 03/18/2015 | 10:50:25 | Temperature #0x20 | Upper Critical going high | Asserted
 168 | 03/18/2015 | 10:57:53 | Temperature #0x20 | Upper Non-critical going high | Asserted
 180 | 03/18/2015 | 10:57:57 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 198 | 03/18/2015 | 10:58:24 | Temperature #0x20 | Upper Non-critical going high | Asserted
 1b0 | 03/18/2015 | 10:58:41 | Temperature #0x20 | Upper Critical going high | Asserted
 1c8 | 03/18/2015 | 11:14:23 | Temperature #0x20 | Upper Non-critical going high | Asserted
 1e0 | 03/18/2015 | 11:15:06 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 1f8 | 03/18/2015 | 11:16:33 | Temperature #0x20 | Upper Non-critical going high | Asserted
 210 | 03/18/2015 | 11:16:33 | Temperature #0x20 | Upper Critical going high | Asserted
 228 | 03/18/2015 | 11:49:12 | Temperature #0x20 | Upper Non-critical going high | Asserted
 240 | 03/18/2015 | 11:49:18 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 258 | 03/18/2015 | 11:55:45 | Temperature #0x20 | Upper Non-critical going high | Asserted
 270 | 03/18/2015 | 11:55:46 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 288 | 03/18/2015 | 11:56:32 | Temperature #0x20 | Upper Non-critical going high | Asserted
 2a0 | 03/18/2015 | 11:57:06 | Temperature #0x20 | Upper Critical going high | Asserted
 2b8 | 03/18/2015 | 12:00:11 | Temperature #0x20 | Upper Non-critical going high | Asserted
 2d0 | 03/18/2015 | 12:00:14 | Temperature #0x20 | Upper Non-critical going high | Deasserted
 2e8 | 03/18/2015 | 12:00:59 | Temperature #0x20 | Upper Non-critical going high | Asserted
 300 | 03/18/2015 | 12:01:34 | Temperature #0x20 | Upper Critical going high | Asserted
 318 | 07/06/2009 | 00:00:22 | Fan #0x42 | Upper Critical going high | Asserted
 330 | 11/13/2016 | 13:25:47 | Fan #0x41 | Upper Critical going high | Asserted
 348 | 11/13/2016 | 13:33:00 | Fan #0x41 | Upper Critical going high | Asserted
 360 | 11/13/2016 | 13:33:47 | Fan #0x41 | Upper Critical going high | Asserted
 378 | 11/13/2016 | 13:44:58 | Fan #0x41 | Upper Critical going high | Asserted
 390 | 11/13/2016 | 13:45:48 | Fan #0x41 | Upper Critical going high | Asserted
 3a8 | 11/13/2016 | 13:47:45 | Fan #0x41 | Upper Critical going high | Asserted
 3c0 | 12/01/2016 | 17:00:29 | Fan #0x41 | Upper Critical going high | Asserted
 3d8 | 12/01/2016 | 17:01:53 | Fan #0x41 | Upper Critical going high | Asserted
 3f0 | 12/01/2016 | 17:04:02 | Fan #0x41 | Upper Critical going high | Asserted
 408 | 12/01/2016 | 17:31:34 | Fan #0x41 | Upper Critical going high | Asserted
 420 | 12/01/2016 | 17:43:42 | Fan #0x41 | Upper Critical going high | Asserted

11/13/2016 it happened to me first time, I thought it could be hardware watchdog, so we disabled it in BIOS.

The server has 2x1TB disks, 2x3TB without optical drive. 365 watts non-hot plug, non-redundant power supply.

Now, we recommended to replace the box, but as far as I am concerned I cannot explain why this is happening (I assume it's some sort of mechanical mainboard failure). I wonder if You have any other ideas.

** Update, mr Chopper3 asked what I meant with but CPU one is not standard. So, original hatsink has been damaged like this:

Time and bad material choice, plastic was not meant to be durable under constant pressures. I've never seen plastic mount since that setup anymore in any other boxes...

Server has been kept in fair conditions, never overheated, not in the direct influence of the Sun, nobody touched it during work.

It was about 1,5 year ago. We couldn't find original HP part anymore on the market. We replaced it with 3 times larger one, because AM2 sockets were not so popular at given time. I cannot remember now if it has had 2 signaling wires plus VCC and GND (4) like stock posted above. It could have just three. VCC + GND and rotation signaling (3). From that point of time we had multiple power outages and situation like this never happened.