Watts Up – Episode 3: PSU-Control Reverse-Engineering

The NV1 replacement PSU has a soft-off/soft-on capability, so its’ time to investigate exactly how it works. I have the “-N” model which has (quote) “TTL-compatible” PSU control, unlike the “-N3” model which has (quote) “ATX-compatible” PSU control. There is only a single control (input) signal.

Well, “ATX compatible” is pretty unambiguous, but what exactly does “TTL compatible” *really* mean – is the PSON signal active-high or active-low, is there an internal pull-up/down resistor in there somewhere, does it require any extra external current-limiting, and is there an internal latch (should the signal be pulsed, or continuous)?

Hopefully Simple
Due to the total lack of published schematics or *detailed* specs for the NV1, the only way to find out is to do some judicious (and cautious!) probing. In the following video, the purple wire is +5VDC Standby, grey is Standby Return, black is Power Output Ground, and white is the PSON line.

I knew that ancient noisy old 1Gb IBM disk drive would come in useful, someday!

It turns out that the NV1 PSON signal is non-latching, 5V active-low and incorporates an internal current-limiting pull-down resistor (ie: if left floating, the PSU defaults to “on”). Strangely, however, to deactivate (turn off the PSU), PSON must be connected directly to the “5V standby power” line: connecting it to any other 5V line (even the one powering the disk drive!!) or to any ground or return, via a 1K resistor or direct, simply has no effect.

Odd, but at least it should be usable.

What about the Darned Sun?
Now for the fun: the SS10 mainboard provides *two* separate PSU control signals, PON and POFF. That’s a bit brain-dead, what happens if both signals are activated at the same time (or at least within each others’ hold time)? The original APS-39 PSU contains an internal latch, the PON and POFF signals are pulsed by the SS10. Needless to say, there are no available specs for the setup-and-hold timings of those pulsed signals.

The actual soft-power-off activation in the SS10 is (ultimately) invoked by a call into the BIOS, which does god-knows-what at the hardware level – documentation is lacking, even the specs of the NCR system-reset-and-EBus controller don’t shed any light on this particular matter. It turns out that the POFF signal is 5.5V active-low. The SS10 simply shorts it directly to ground to turn the APS-39 off. No external resistor is involved – using a wire or wire-with-various-resistors to manually connect ground to POFF at the motherboard power-connector proves that – any resistor more than a few Ohms stops the power-off.

Soft power-on is activated by pressing the magic “power” key on a Sun (serial RS-423) keyboard, which suggests that there is a standby 5V power-line in the system somewhere. However, when in soft-off state the APS-39 5V output lines are all at 0V. What gives? What voltage do we see on the PON line, in both “power-on” and “soft-off” state?APS39_PON2APS39_PON1

Aha! the PON line *is* the standby power supply! Whilst the system is running or in soft-off state, PON is at 5.6V. In soft-off state, momentarily shorting it to any of the main power output ground lines pulls it low (below 4.5V seems to do the trick) and turns the APS-39 *on*, so PON is an active-low signal (latched by the APS-39). Pressing the “magic” key on the keyboard really does (ultimately) short the PON signal to directly ground; again, no external resistors involved:

Decisions, Decisions
Obviously, I’m going to need at least one external latch-circuit (powered from the 5VDC 2A standby power line) to adapt the NV1. There are several ways to skin that particular cat: a custom TTL/RTL circuit (asymmetrically-biased cross-coupled transistor SR-latch); an electromechanical low-signal relay (but what sensitivity?) with an external debounce cap; an asymmetrically-biased cross-coupled-logic-gate circuit; or something else…

My heads is starting to hurt a bit, so perhaps that decision is best left for later on.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s