CPUs sometimes not posting

"*Vanguard*" <no-email.TakeThisOut@reply-to-newsgroup.invalid> wrote in message news:<qqqdnWyYtNvPajHd4p2dnA.TakeThisOut@comcast.com>...
 > Cliff Beveridge said in
 > news:f2939766.0405200842.1389e0a1@posting.google.com:
  > > I am working on a system that will include three passive backplanes
  > > with a single board computer (SBC) on each one. I am using a single
  > > power supply for the entire system. I am seeing that sometimes when I
  > > start the system with the soft power momentary switch, one of the
  > > three CPUs does not post. This problem seems to happen only right
  > > after I have cycled the A/C to the power supply. Whenever this
  > > happens, if I then shutdown the system and then bring it back up again
  > > using soft power, all three CPUs post just fine. From then on I can
  > > shutdown and restart the system using soft power with no problems.
  > > But as soon as I remove the A/C from the power supply and then bring
  > > it back up again, the problem recurs. Has anyone seen this before?
  > > Am I overtaxing the 5V standby circuit by driving three SBCs? By the
  > > way, I have seen this using just two SBCs as well. Thanks in advance
  > > for any insight into this problem...
 >
 > Although <a style='text-decoration: underline;' href="http://www.hardwarebook.net/connector/power/atxpower.html" target="_blank">http://www.hardwarebook.net/connector/power/atxpower.html</a> says
 > the max current for the 5VSB line is 10mA, my Fortron PSU says the 5VSB
 > will handle minimum = 0.0A, normal = 1.0A, maximum = 2.0A. My Abit
 > NF7-S motherboard's manual (page 3-10) say, "..., and 720mA +5VSB at
 > least for supporting some special features." So I suppose the mobo
 > might consume up to 720mA which means my PSU could handle twice that
 > load but not thrice. You'll have to check the specs for your PSU and
 > slot computers. Have you clipped on a multimeter to monitor the voltage
 > on the 5VSB line?

I have measured the voltage and it drops down to a range from 4.5 to
4.9 volts with the momentary switch closed, so that should be fine. I
will look at the specs for the PSU, but unfortunately the SBC
manufacturer tells me that they don't have specs for the soft power
circuitry on the board. Another interesting bit of information I
forgot to mention is that it is always the same SBC that fails to
post. I have tried moving the SBCs around to different backplanes,
but the same SBC always exhibits the problem. The weird part is that
I have played around with a sample of about 10 SBCs and in most of the
combinations I can get all of them to come up just fine every time,
but certain cards don't "play well" with others. These "pesky" cards
work just fine when they are the only one connected to the PSU, but
not when used in conjunction with other cards. Strange...<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: CPUs sometimes not posting 

On 20 May 2004 09:42:23 -0700, mybevman DeleteThis @yahoo.com (Cliff Beveridge) wrote:

 >I am working on a system that will include three passive backplanes
 >with a single board computer (SBC) on each one. I am using a single
 >power supply for the entire system. I am seeing that sometimes when I
 >start the system with the soft power momentary switch, one of the
 >three CPUs does not post. This problem seems to happen only right
 >after I have cycled the A/C to the power supply. Whenever this
 >happens, if I then shutdown the system and then bring it back up again
 >using soft power, all three CPUs post just fine. From then on I can
 >shutdown and restart the system using soft power with no problems.
 >But as soon as I remove the A/C from the power supply and then bring
 >it back up again, the problem recurs. Has anyone seen this before?
 >Am I overtaxing the 5V standby circuit by driving three SBCs? By the
 >way, I have seen this using just two SBCs as well. Thanks in advance
 >for any insight into this problem...

I could be wrong but think the bios has to be loaded that first time to
set chipset registers, that it's not a power related problem, the best fix
being to prevent loss of AC with an UPS. If you have reason to believe
otherwise please post more details.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: CPUs sometimes not posting 

Cliff Beveridge wrote:
 > I am working on a system that will include three passive backplanes
 > with a single board computer (SBC) on each one. I am using a single
 > power supply for the entire system. I am seeing that sometimes when I
 > start the system with the soft power momentary switch, one of the
 > three CPUs does not post. This problem seems to happen only right
 > after I have cycled the A/C to the power supply. Whenever this
 > happens, if I then shutdown the system and then bring it back up again
 > using soft power, all three CPUs post just fine. From then on I can
 > shutdown and restart the system using soft power with no problems.
 > But as soon as I remove the A/C from the power supply and then bring
 > it back up again, the problem recurs. Has anyone seen this before?
 > Am I overtaxing the 5V standby circuit by driving three SBCs? By the
 > way, I have seen this using just two SBCs as well. Thanks in advance
 > for any insight into this problem...

This from the ATX spec:

 > Page 19 ATX/ATX12V Power Supply Design Guide
 > Version 1.2
 > 3.3.3. +5 VSB
 > +5 VSB is a standby supply output that is active whenever the AC power is present. It
 > provides a power source for circuits that must remain operational when the five main DC
 > output rails are in a disabled state. Example uses include soft power control, Wake on
 > LAN, wake-on-modem, intrusion detection, or suspend state activities.
 > The +5 VSB output should be capable of delivering a minimum of 1.0 A at +5 V ± 5% to
 > external circuits. Because trends indicate a growing demand for standby power, it is
 > recommended that designs be scalable to 2.0 A to meet future needs. The power supply
 > must be able to provide the required power during a "wake up" event. If an external USB
 > device generates the event, there may be peak currents as high as 2.5A lasting no more than
 > 500mS.
 > Overcurrent protection is required on the +5 VSB output regardless of the output current
 > rating. This ensures the power supply will not be damaged if external circuits draw more
 > current than the supply can provide.

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: CPUs sometimes not posting