Since there hasn't been the announced firmware update to support AT style power switches I decided to take matters in my own hands. I've never done Arduino stuff before, but that turned out to be a fun exercise.
I decided to hardcode the AT power switch mode since it changes the logic quite much. Introducing it as a configurable option would have made things too complicate (at least for me). Unfortunately I have no means to professionally test if the overcurrent protection still works as intended. I did however do a rudimentary test by directly shorting 12V and 5V with paperclips and everything behaved as expected.
Here's what this firmware does:
- Displays "ATX2AT 1.13 AT Switch" as header instead of "ATX2AT CONVERTER 1.13"
- Turns the power ON and OFF according to the state of the switch (aka AT style)
- If the overcurrent protection triggers you have to turn off said switch to reset the protection
- Commands to trigger the power ON or OFF sent from the configuration tool will be ignored since it would make things super complicated
Fun facts I figured out while looking into this:
- The onboard POWER switch traces are combined with the EXTPWR header which means it now also works like an AT style switch. Good for a quick test but now useless beyond that. 🙁
- The configuration tool reads the firmware version from the EEPROM - which is not updated if you update the firmware. So don't believe that number.
- The extpwr variable has nothing to do with the EXTPWR header. That caused some confusion in the beginning.
You can deploy this firmware by:
- Installing Arduino IDE
- Installing Adafruit_GFX and Adafruit_SSD1306 libraries
- Set the board type to "Arduino Micro"
- Plugin an USB cable into the ATX2AT and your computer
- Uploading the sketch attached to this post
Be warned: I'm not an electronics engineer nor an embedded developer. Be aware of the risk.
If you want to have the old firmware back just get it from the original repository.
The attachment ATX2AT 1.13 AT Switch Style Firmware.7z is no longer available
Some time later:
I've mounted the converter in my case which gave me some serious headache. I understand that Samuel designed this from a different perspective, with open bench testing in mind but that causes some trouble if you want to use this in an actual retro gaming rig (regardless if AT or ATX case). Every voltage the computer needs is exposed on the board, especially at the connectors. I don't want this thing dangling around in a metal case nor let is touch my components. This thing is not build to mount in an AT(X) case, the short P8/P9 leads also restrict your options. 3D printing a case also is very difficult because of the board layout. I ended up using super thick (like 2mm) double sided adhesive for mounting but I still feel a little uncomfortable with the exposed connectors.
Don't get me wrong though: I still love that thing and I will try to acquire a second one.
More later:
Oh dear. So far I've tested this with my 486 testing rig, now I built the converter into my 233 MMX AT rig. And well, it triggers immediately on 5V when powering on the computer. I've removed all drives and cards, I've tried the original firmware, increased the response time to the maximum (500ms), increased the limit to the maximum (8A on 5V) - no dice. My original AT power supply can deliver 30A on 5V according to its spec. Is this the wrong device for my rig or is there something wrong with my motherboard? It has 2 tantalum capacitors and the lowest resistance between ground and 5V on the AT plug are 0.6 kOhms.