intel s5000pal power supply - grounding problem / requirements ??

[idata]

The server board is an s5000pal, the power supply is an Antec 650 watt tri rail supply. The symptom is:

A. the system boots repeatedtly and predicably 'on the bench' or in the case if the board is not touching the case

B. Add just one screw between the board and the case to mount the board - the boot process hangs just after fan turn on. There are no POST leds at all. (see S5000pal http://communities.intel.com/click.jspa?searchID=5900440&objectType=2&objectID=139637 http://communities.intel.com/click.jspa?searchID=5900440&objectType=2&objectID=139637 thread)

What are the s5000pal PS ground requirements?

1. Are the mounting pads connected to the DC returns

2. Does the board assume no connection between earth and DC return

3. Does it assume the PS connects earth and DC ground externally?

Thanks in advance - Dan

[DSilv11 · Accepted solution]

dansawyer wrote:

The server board is an s5000pal, the power supply is an Antec 650 watt tri rail supply. The symptom is:

A. the system boots repeatedtly and predicably 'on the bench' or in the case if the board is not touching the case

B. Add just one screw between the board and the case to mount the board - the boot process hangs just after fan turn on. There are no POST leds at all. (see S5000pal /click.jspa?searchID=5900440&objectType=2&objectID=139637 http://communities.intel.com/click.jspa?searchID=5900440&objectType=2&objectID=139637 thread)

What are the s5000pal PS ground requirements?

1. Are the mounting pads connected to the DC returns

2. Does the board assume no connection between earth and DC return

3. Does it assume the PS connects earth and DC ground externally?

Thanks in advance - Dan

1) yes

2) no

3) Does not care unless the PS is faulty and has a ground loop through the mother board. Board still does not care, but the power supply would crowbar.

Most common causes of working on bench , but not in box are:

1) extra standoffs in the box short circuiting on the back of the mother baord.

2) power supply fault where a votlage potential exists between case ground and DC ground.

3) power rail sequencing. (see ATA specification)

[idata]

Thank you. There were no extra standoff, however the loose spring from empty cpu slot 2 may have been touching the cabinet. An extra set of standoff were added to the right three mounting holes. The board is now installed and operates in the cabinet. Dan