Power rail switching

[Torbjřrn Heltne]

Hello all,

I am looking into a small project where I (for various reasons) want to
use a single ATX power supply for 2 motherboards. I know reasonably well
how the power is working and how it uses external circuitry for switching
on/off. I have come up with a simple circuit which will ensure that the
ATX power is switched on whenever at least one MB requires so.

My main concern now is what to use for switching all the various power
rails. I have checked a lot of datasheets on optorelays, power
transistors and a few relays but I'm kind of stuck. My main concerns are
the +3.3V +5V and +12V line which all delivers a good number of Amperes.

Optorelays and power transistors are fine wrt contact bouncing, but all
the ones I have found will introduce an unacceptable voltage drop. With
relays the opposite is the case with virtually no voltage drop, but what
will happen due to contact bouncing...?

There is always the possibility of using both kinds in parallel or
maybe even several optorelays to minimize voltage drop (or what?).


Anyone here with experiences from something similar?

Thanks for reading.

--
Torbjřrn Heltne

 

[Torbjřrn Heltne]

Walter Harley wrote:

The ATX power system includes a signal line for "power good". The idea is
that the supply does not raise this signal until all the power supply
voltages have stabilized.

Thanks!

I had a slight misunderstanding of the "power good" signal, but looking
at it again I see that you are, of course, right.

Relays, then.

--
Torbjřrn Heltne

 

[Roger Hamlett]

"Torbjřrn Heltne" <torbjorn.heltne@amelektronikk.no> wrote in message
news:411c92ac$1@news.broadpark.no...

Walter Harley wrote:

The ATX power system includes a signal line for "power good". The
idea is
that the supply does not raise this signal until all the power supply
voltages have stabilized.

Thanks!

I had a slight misunderstanding of the "power good" signal, but looking
at it again I see that you are, of course, right.

Relays, then.
Seriously, you should look again at the specifications of FETs. Devices

like the IRFP3703, have 'on state' resistances, that you will be hard
pressed to match with your wiring, and this sort of device can be
paralleled to give even lower resistances (at the cost of increasing the
gate loads that have to be driven).
It is worth saying, that such supplies are available. Some 'server' cases,
have dual supply/dual board systems, where either board can run off either
supply, with the full control logic to handle this.

Best Wishes

 

[Watson A.Name - \"Watt Su]

"Torbjřrn Heltne" <torbjorn.heltne@amelektronikk.no> wrote in message
newsan.2004.08.12.20.36.41.545560@amelektronikk.no...

Hello all,

I am looking into a small project where I (for various reasons) want
to
use a single ATX power supply for 2 motherboards. I know reasonably
well
how the power is working and how it uses external circuitry for
switching
on/off. I have come up with a simple circuit which will ensure that
the
ATX power is switched on whenever at least one MB requires so.

My main concern now is what to use for switching all the various power
rails. I have checked a lot of datasheets on optorelays, power
transistors and a few relays but I'm kind of stuck. My main concerns
are
the +3.3V +5V and +12V line which all delivers a good number of
Amperes.

Optorelays and power transistors are fine wrt contact bouncing, but
all
the ones I have found will introduce an unacceptable voltage drop.
With
relays the opposite is the case with virtually no voltage drop, but
what
will happen due to contact bouncing...?

Nothing, since the power busses all have filter capacitors on them. But
the extremely high inrush current, probably hundreds of amps, might weld
the relay contacts closed if they aren't heavy duty.

There is always the possibility of using both kinds in parallel or
maybe even several optorelays to minimize voltage drop (or what?).

Big old power FETs have very low Rds(on), in the milliohms, so the V
drop would be very low. But they require 10 volts higher than the
supply V on the gate to fully turn on.

Anyone here with experiences from something similar?

If you turn the current completely off while the ATx PS is still
running, it's likely that it will go into overvoltage protection and
shut down. So you will have to make sure that that doesn't happen. I'd
say that you have got a lot more difficult project than you originally
thought.

Thanks for reading.

--
Torbjřrn Heltne