Date: Fri, 26 Jun 1998 17:57:48 +1200 (NZST) From: Andrew McNaughton <andrew@squiz.co.nz> To: Atipa <freebsd@atipa.com> Cc: isp@FreeBSD.ORG Subject: Re: Recommended SCSI Drives Message-ID: <Pine.BSF.3.96.980626173357.313A-100000@aniwa.sky> In-Reply-To: <Pine.BSF.3.96.980625102155.22172A-100000@altrox.atipa.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 25 Jun 1998, Atipa wrote:
>
> > > Insist on ATX form factor, and a case that has (2) fans in addition to the
> > > fan inside the PS. Our ATX Full Tower fits the bill, and has gobs of drive
> > > space. The extra hieght will also aid in airflow and dissipation.
> >
> > I've run into a problem with an ATX server's power supply where in case of
> > power failure, the machine would not reboot without someone manually
> > pushing the button on the front. The motherboard had no bios setting to
> > avoid this, and local service departments were not able to help other than
> > to suggest a new mother board.
>
> Yes, that is a problem. We sold some ATX machines to the local gas company
> to do real-time flow control w/ QNX, until they realized they could not do
> a remote reboot. We had to take back the machines.
>
> > I ended up making up my own circuit to provide a signal pulse signal to
> > the power supply triggered and powered off the standby power line. Works
> > fine.
>
> Like a watchdog or what? I don't understand your trigger mechanism, unless
> you NAND the power supply and MB power...
I hooked this circuit up off the wires between the power supply and the
motherboard
----------------------------------- Standby Power Line (+5 V)
|
|
R1
|
|
|------------S1----------------- Reset (+5 V signal)
| |
| |
R2 C1
| |
| |
----------------------------------- Common (0 V)
R1, R2 = resistors
C1 = Capacitor
S1 = Schmidt trigger (inverter with hysteresis).
When power goes on, S1's input is low, so it's output is high.
R1 charges C1. When the voltage threshold of S1 is reached, output
drops.
when power goes off, R2 discharges C1. R2 should be several times larger
than R1 so that input to S1 goes close to 5V. It should be small enough
that C1 discharges more quickly than the minimum length power interruption
you expect.
My circuit gave a pulse of about 0.1 second duration, which seems to be
long enough. I forget the exact values of the components used.
> > I'm curious to know how common this problem is and whether there's a more
> > conventional solution to it?
>
> Well, a reboot causes the machine to shutdown completely, so a BIOS fix
> would not work. There are some MBs that use a toggle switch instead of
> intermittant, but the cases are based on intermittant (reset-type) powers.
A motherboard could implement the same mechanism I did. I'm not sure if
the bios gets to use the standby power, but if not, then a jumper could be
used to enable the mechanism.
> Kevin
>
Andrew McNaughton
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.980626173357.313A-100000>