6
\$\begingroup\$

As the title suggests; when connecting the PS_On signal directly to ground in order to switch on an ATX PSU, is there a required amount of resistance?

-

To try to clarify my particular problem; I have a fully modular ATX PSU (a Seasonic Platinum Series) but am not using it to connect a motherboard, so have no need of the bulky motherboard cable. I decided to order a custom-made cable compatible with the modular connectors on the PSU itself, with a wire connecting PS_On to ground (via a 2-pin header so I can connect a suitable switch or LED if I want). However, when I connect the pin directly, although the PSU switches on, it almost immediately switches off again.

The PSU is not at fault, as connecting the regular motherboard cable using the paper-clip trick works just fine and the PSU starts providing power normally. The only difference I can see between the working and non-working cables is that the non-working cable uses much shorter wires and of a finer gauge (AWG22 at about 5-6cm as opposed to AWG18 at around 60cm). So I'm wondering; is my cable failing to work because of a lack of resistance?

If so, I'd appreciate anyone that can also clarify how I would work out what type of resistor I would need to add to correctly complete the circuit.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Is there any load connected to the 5v line? some PSU don't start unless there is min load in the 5v line sinking a couple of amps \$\endgroup\$
    – alexan_e
    Jan 22, 2014 at 16:49
  • \$\begingroup\$ I believe my problem is a lack of signal connections as DoxyLover pointed out, though I wouldn't be surprised if I need resistance on the 5v line too, I'll have to get myself plenty of wire, terminals and some resistors and just try it all I guess! \$\endgroup\$
    – Haravikk
    Jan 23, 2014 at 14:31
  • \$\begingroup\$ It is a logic “0” level to gnd so since the impedance input is high, The source resistance is only affected by the charge capacitance of the input gate and the pullup R value >100k. You can always increase the load with a330 Ohm series and LED to pullup but then a lower switch resistance will improve the logic “0” Vol. \$\endgroup\$ Nov 19, 2017 at 0:02

6 Answers 6

6
\$\begingroup\$

I suspect your problem is lack of connection for the sense lines. If you take a normal power supply, with just separate insulated wires in a bundle, and look at the motherboard connector, you will notice that several pins have two wires crimped together. The larger gauge wire is the actual power wire and the smaller gauge wire is the sense wire.

The purpose of the sense wire is to overcome voltage loss due to resistance in the wire harness: the power supply increases the 3.3V (for example) so that at the other end of the wires, you get true 3.3V. These may also be a sense wire on ground.

If the sense wire is not connected, the power supply detects a fault and shuts down. If you cannot tell which wires are sense in your original harness, you will need to connect together ALL +5V wires, connect together ALL +3.3V wires, connect together ALL +12V wires, and connect together all ground wires.

Edit: I just checked the ATX specification and it appears that there is probably only a sense line on the +3.3V. Try just connecting together all of the +3.3V (orange) wires.

Additional edit: the answer to the OP's specific question is no, no resistor is required, as shown by the fact that the power supply works with its original cable harness.

\$\endgroup\$
9
  • \$\begingroup\$ Ah! I think you're right, it looks like there is a signal connection on one each of the +12v, +5v, +3.3v and a ground line, and I've managed to completely miss all of them on the pin-out diagram when I had the cable made. Now, this is the answer to my problem, but I'm not sure if I should mark it as the correct answer or not due to how I've asked the question? In any even I'll to order up more terminals and some wire… \$\endgroup\$
    – Haravikk
    Jan 23, 2014 at 14:29
  • 1
    \$\begingroup\$ I'd suggest that if this solves your problem, it should be accepted. \$\endgroup\$
    – DoxyLover
    Jan 23, 2014 at 18:18
  • \$\begingroup\$ Only the 3.3v Sense is part of the ATX spec, and even then, resistance in a foot of wire is consistent, foreseeable, and completely negligible, especially considering that the main rails have ±5% tolerances. \$\endgroup\$
    – Passerby
    Jan 24, 2014 at 2:05
  • \$\begingroup\$ I'll try the +3.3v sense wire only first, as I'll need to re-crimp some of the other wires to do the other voltages, but it does look like my PSU has sense wires for +12v, +5v and ground as well, or rather one of each (i.e - one +12v pin has it, one +5v etc.). I'm going to mark this as an answer, but you might want to add a brief note that resistance shouldn't be the problem, as that's kind of how I phrased my question so "no" is actually the most correct answer ;) \$\endgroup\$
    – Haravikk
    Jan 24, 2014 at 18:41
  • 1
    \$\begingroup\$ @IncnisMrsi are you saying that the motherboard cable is intact but unconnected (except for the PS_ON# jumper)? That's fine. The sense wire(s) are connected through the cable. The OP's issue was the lack of the MB cable itself. \$\endgroup\$
    – DoxyLover
    Aug 1, 2016 at 5:37
6
\$\begingroup\$

There is no need for a resistor. Any voltage <0.8 (or ground connection) will enable the psu

Refer to ATX12V Power Supply Design Guide section 3.3.2

PS_ON# is an active-low, TTL-compatible signal that allows a motherboard to remotely control the power supply in conjunction with features such as soft on/off, Wake on LAN*, or wake-on-modem. When PS_ON# is pulled to TTL low, the power supply should turn on the four main DC output rails: +12VDC, +5VDC, +3.3VDC and -12VDC. When PS_ON# is pulled to TTL high or open-circuited, the DC output rails should not deliver current and should be held at zero potential with respect to ground. PS_ON# has no effect on the +5VSB output, which is always enabled whenever the AC power is present. Table 14 lists PS_ON# signal characteristics. The power supply shall provide an internal pull-up to TTL high. The power supply shall also provide de-bounce circuitry on PS_ON# to prevent it from oscillating on/off at startup when activated by a mechanical switch. The DC output enable circuitry must be SELV- compliant. The power supply shall not latch into a shutdown state when PS_ON# is driven active by pulses between 10ms to 100ms during the decay of the power rails.

enter image description here

\$\endgroup\$
2
\$\begingroup\$

The PS_ON is a signal line. Does not carry/need much current to work as intended. Technically, no, it might not need a resistor, but small resistors are cheaper than dirt, good practice would dictate one be used, in case you short it to the wrong thing.

But it's not working because ATX supplies are complicated, potentially dangerous devices meant for non-technical consumer use. They have protection circuits intended to safeguard you, the pc, and itself, while maintaining it voltage output within spec. The most common version of this is a open line detection. No load on the supply, it shuts down. Remember, these are designed to only run when connected to a matching AT speced motherboard, not by themselves.

Connect a low value, high wattage power resistor to one or multiple rails. 5v, 12v, 3.3v, try experimenting. Keep in mind it will get hot, so use Ohm's law to make sure the resistor will survive. I = V/R, P = I * V. Attach it to the PSU's cage as a heatsink.

\$\endgroup\$
1
\$\begingroup\$

Despite all the ATX rules that say it isn't necessary, a few ATX PSUs do require a load on the 5V line.

enter image description here

From Dangerous Prototypes

I found a Dell ATX supply where my multimeter showed roughly 12V, 5V at the outputs even without the PS_ON line grounded, however a LED on the the PWR_OK line was flickering/blinking very intermittently.

Using a load resistor persuaded it to work more correctly. The voltages on the 12V and 5V lines were then closer to the spec.

\$\endgroup\$
1
  • \$\begingroup\$ How is it an answer for the question given (about managing the PS_ON# line)? \$\endgroup\$ Jul 31, 2016 at 18:00
1
\$\begingroup\$

I am working with a PSU and I had to put a load on it to make it work. I added a 10 ohms 25W to the 12V output of the ATX connector.

\$\endgroup\$
1
\$\begingroup\$

I did the same except designed an ATX breakout board with a 10 watt for 5 volt or a 15 watt resistor for the 12 volt rail. Some power supplies are different requiring the load resistor on the 5 volt rail and some on the 12 volt rail. I used a on off on switch attaching to both rails then to the load resistor that way I can use any such power supply regardless of rail load. each power rail voltage has it's own layer so I can ramp up the amperage to 20. Still looking for the 3D binding posts that handle 30 amps. I added the 4x4 CPU also because it has 4 12v lines and 4 ground lines as the 24 pin only has two 12 volt lines and I did not feel safe with just two at 20 amps so the plug was added. ATX breakout board with load resistor

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.