Hi Martin,

Thank you for some great input!

- Rotary encoders will be added to the list, i have had them (or ten-turn pots) in mind, but wasn't sure how i would implement the control using them. Good input.

- self resetting fuses; I planned to use them on the fixed voltage outputs, but not on the CC/CV variable voltage part, but maybe i'm thinking wrong (it happens a lot). I got this idea from the Dangerous Prototypes ATX breakout board which uses "a 1.25 amp resettable polyfuse on each supply help protect your project from the full current of ATX supply." i did plan to have a switch or bridge to select different fuses/amp levels or bypass them, but i might need to rethink that then.

- on the timer function; very valid point, i figured that it would be good to have because i am forgetful and at times wont visit the lab for weeks, so distrustful of my own capabilities and quality of my work, i thought i would add the timer and the residual-current circuit breaker. I can always use power plug versions of these at the outlet until i build up the confidence. :)

- Single point mains connector; I planned to use the ATX for this as it has built-in fuse and switch, and it is the natural way to power a case. It does mean i will need to open it up, and caution - as mentioned in the above point - made me think. Then again, its not the first time i rewire a 220v device, and if i connect the second PSU right at the socket on the inside, it shouldn't push the ATX PSU ratings in any way, right?

I might still consider optional timer and/or remote on/off functionality for some of the outputs, controlled by MCU, as a way of controlling remote monitored/logged experiments (my lab is in the basement and I am lazy :) ). Mains input to the power supply i might just put on a zigbee/wifi switch with enough capacity. 

Excellent input guys, i really appreciate it!

All the best,

bjorn

We have a RCD in the apartment installation and together with a solid PE connection this should be enough to protect you from faults in your PSU. Against faults in your DUT/experiment it can not protect, but this is LVDC anyway. Yes, that was my suggestion, to connect the wires onto the input socket. Use some heat shrink tubing, think about possibilities and dangers of wires coming loose and where they could touch - and where to put a cable tie if you want to prevent this. Don't rely on mounting screws which hold together the device for PE connection. This would be considered bad practice, probably because there is a higher danger on forgetting that dangling ground wire when reassembling. Better use a dedicated screw and dedicated PE wires for grounding multiple case parts or PSU modules. But probably a ready made ATX case should be considered safe.

Use a big red (or blue) "POWER" LED to minimize the chance of leaving it on without intention. :-)

Hi again Martin,

We live in an old house from the sixties and only "protection" is the 16A ceramic fuses in the fuse box.

Faults in my experiments are a certainty and happens daily, no device can ever protect me from those, I was more thinking of protection if my 5-year-old decides to pour her slushie on it (or my beer falls over before i finish it). No higher risk here than with my workstation in the apartment, and that doesn't have any PE even (just learned something: PE = Protective Earth... or Poly Ethylene, but i think you mean the first here :-) ), the mains outlet does have PE socket("EU style") so i will make sure to have a dedicated fastener if there isn't one already in the ATX PSU. I assume that - since it is already mounted metal-to-metal in the case, and the secondary 180w PSU will also be so, it's all the same  if i connect a PE wire between the two? or perhaps better to do it anyway in case i do something silly in the future (like putting it in a drivebay case), and the PE connection through the PC case is broken.

Adding big, bright, obvious power indicator LED now!

Big thanks once again!

If you have a 5 year old child you could consider having an RCD installed for your house. Childs do not only pour slushies around, sometimes they decide it could be interesting to stick something into the holes of a mains outlet.

Against slushies and beer in the PSU I would recommend a closed top of the housing. Necessary vent holes can be on the underside and at the rear. This also prevents small pieces of wire (clippings from THT components) from getting into it.

Some inputs:

Use (at least additionally to your touch screen) real rotary encoders. You can use the button to switch coarse/fine. Touch screen alone can be really annoying if you want  to e.g. slowly change your voltage to test how your DUT behaves under varying voltage.

Make the design so that any fans are (at least nearly) silent under no/low load conditions.

When you have electronic current/limiting or constant current operation external fuses or self resetting fuses are not necessary and can be annoying because they have quite high resistance.

Individual output switches are a good thing.

A timer is not so important, at least it has to be possible to disable it. It makes sense on a battery operated device (DMM) but on a PSU it is much worse to have a timer interrupting your test run than to waste a small amount of electricity if you forget to turn it off.

Single point mains connection. I would open one PSU and connect some wires to the other one internally. If you really do not want this, at least built a power distribution strip into your ATX case and lead the cables back in. USE RUBBER GROMMETS (or strain relieve clamps) where cables go through sheet metal.

A RCD is useless in a (galvanically isolated) DC power supply. Only if you want to build an extra protected mains outlet into the ATX case it has some use. But I would not do this. If you want to work on mains powered equipment it is better to use an isolation transformer.

Agreed. RCD is not really required as is the timer function, but maybe and RTC would be a good addition so capacity measurements could be made?

RTC shouldnt be a problem unless it needs to be milliseconds-precise, i have some battery backed RTC modules and suitable LIR cells, and the ESP8266 can sync with NTP. Good one!

Edit: Just remembered that the Nextion Display i have has an onboard RTC that i should be able to use.