Live data from Hacker News

Intel Introduces New ATX PSU Specifications

intel.com

21–22 of 22 posts

Re: Intel Introduces New ATX PSU Specifications

#21
post #19
post #2

What caught my eye was the new spec includes side channels for the PSU to communicate with add in hardware. Specifically to report the PSUs capabilities. This looks like a red flag to me. This could be used as a way to vendor lock certain features of hardware to certain brands or capacities of PSU. Get ready for DRM that checks your system all the way down to the PSU level. I prefer a dumb PSU that just gives up the…

I went looking for the specs because I was specifically interested in how that power negotiation works. Turns out it's only two wires with plain static logic signals, no communication or anything. https://cdrdv2.intel.com/v1/dl/getContent/613768 page 25. Sense0 Sense1 Initial Max Gnd Gnd 375W 600W Open Gnd 225W 450W Gnd Open 150W 300W Open Open 100W 150W

Very interesting. So it says what each rail can handle.

I wonder if there is any wiggle room in the spec to piggyback a proprietary signal. For example, do you think a PSU that toggled the output of one rail between two specific values between spec polling intervals might be able to indicate to proprietary software the specific brand or model of PSU?

Re: Intel Introduces New ATX PSU Specifications

#22

Includes a revision on the 12V only one that I'd dearly love to take off. Running a PC on a misc 12V supply instead of also having to cobble together 5/3.3/etc at much lower wattage seems much better.

Seems like 20v might be a better choice. Then you have 20v available for USB-C monitor ports.

You could give every device its own modular swappable mini PSU that takes 20v and gives whatever the card requests, so if the onboad converters go bad on a GPU you can swap them.

At least it seems like negative voltages are mostly on the way out!

Post reply on HN