Elektor RPi ATX Interface

Was reading the latest issue of Elektor and an article on converting a PSU for use by a Pi - here is the schematic

Unfortunately the article is not really complete on the design, so I was enjoying filling in the blanks

  1. it seems odd to me that the schematic capture’s symbol library has the body diode on the SI2318CD but not on the 2N7002 - checking the data sheet for each, the manufacturer includes the body diode in their symbol. But im also struggling to find a good reason for using two different models here, rather than consolidate to a single part

  2. it’s not clear from the writeup, but SV14 is a programming header i do believe

  3. no idea what SV1 is for, but as a pin header it seems like shorting Vcc and Ground is a real possibility… i have actually read the code and this doesn’t seem to be used at all; PB3 is only mentioned as RX for debug UART (but TX is on PB5)

  4. My assumption is X4 is for power only, and X3/X2 would allow for externally powered USB devices (connect X2 to the pi, and a disk drive or something to X3)… wonder what would happen if you connected X3 to the pi where Vbus is powered

note: found the original schematic (even though it was in last months elektor, the design is from 2019!)

sorry, forgot to add an actual question!

given the Getting to Blinky Project, this doesn’t seem like the best, most reliable design - have I missed something or would anyone agree?

Seems awfully complex for what it does. I find it a bit odd to use a microcontroller to control an ATX PSU and a pretty old one to boot. The newer ATTinys are simpler to program via UPDI and could even be programmed from the PI (there appears to be a PI connection for that, btw) to make it easier to build/manufacture. IIRC, the older programming interface was proprietary. The newer one, UPDI, is well documented and simple. Yes, SV14 is the old style ATTiny programming header. SV1 was probably for an external input with power, maybe for a sensor. +5V in the middle means no reverse polarity if plugged in backwards.

SV1 appears to be support for plugging in some kind of RF serial receiver. The pull-down is good practice to prevent stray noise input. I’ve had the misfortune of building a device one time where the input floated when the device was not connected via USB, resulting in spurious bytes due to noise which then shut the unit down!

cheers @phil_from_seattle and @ToyBuilder - i agree it seems overly complex; but now I’ve looked at it a few times I wonder if they were looking at including possibility for future capability that never made it, was removed or was for debgging… but ok, then good, i’m making progress in reverse engineering :wink:

The current is different from the 2 FETs, that is why there is 2 different types

@kvk yes, you are absolutely right, I had misread the datasheets; the 2n7002 is handling milliamps where as the SI2318CDS will handle 5 Amps which makes sense as the SI2318CDS is controlling a fan whereas the 2n700s is just soft-controlling the ATX power-on/off signal from a GPIO on the ATTiny. Having said that I reckon it would be possible to replace the 2N7002 with a second SI2318CDS? clearly the SI2318CDS is able to be driven from a GPIO - but this would be at the cost of power consumption… so trading BOM simplicity for power consumption… in a hand assembled, one-off the reduced power consumption would be good…

but what I do find fascinating is why the 2N7002 schematic symbol is missing the body diode… a minor thing I guess?

image