Power supplies for modern MCU Geiger counters

For quite a while I have been using a good quality 12V 4A switch mode power supply then a 7805 regulator and ancillary components to run my small modern electronic MCU type counters. NetIO/CA JOE/ESPGeiger HW etc. I had to remove that PSU as I need it for something else, but noticed that other SMP's have a hell of a lot of noise or something. I currently have an STS-5 on my breadboard counter and with the original PSU nothing happened if I touched the metal wall of the tube or picked it up. The first PSU I tried as a replacement just made the counter go nuts. Constantly counting, so that went in the bin. The second and third PSUs seem to work OK, but when I touch the tube wall or pick it up, the counter goes nuts, again counting like crazy. Putting the old PSU back on stopped this happening and I could quite happily touch the tube. I have had to use the old PSU for something else, so now when I touch the tube it goes nuts. This is also being regulated by a 7805 so the ripple or interference must be pretty bad.

I highly suspect it is noise from the PSU, and my touching the tube is somehow grounding it or the opposite, but the electrical path is enough to interfere with the counter. I think the way forward is to build an old school linear PSU with some decent capacitors on it. Has anyone else experienced this kind of thing? And if so, what did you do to fix the issue?
 

Thanks for your replies. The PSUs I was trying are all used that have come from working environments. Its pretty unknown as to how long they have been in use, but were working when I acquired them as I removed them from equipment I was upgrading. I'm just going to assume they have either had it, or are low quality, high ripple etc. One thing that did come to mind, and I can test this, is one of the polarities at mains, or close to mains potential. I have seen it before where (even on some half decent) either the positive or negative is at mains potential. If that was the case and I am earthed, or even not earthed, and I touch the tube, it could be getting some interference from mains itself. Big Clive has identified this kind of thing more than once.

With regards to the filter capacitor, I wasn't aware that it had to be a certain value, tuned to the PSU. My understanding, (and I doubly checked with chatGPT. It has been known to be wrong, so have I), is that obviously the voltage has to be above that which it is fed, but the value simply has to be big enough to filter the ripple. Increasing the current draw will increase the ripple, so a larger value would be needed in that case. As far as I am aware the value of the capacitor is based on the current that is drawn and the ripple that is created. Some manufacturers may think that 30mv ripple is OK at 2 amps, others may go with 10mv. Depends on what the manufacturer I guess and the intended application for the PSU. I do have (somewhere) a very nice Meanwell, I think, medical grade PSU. The output is completely isolated from mains, completely. It's designed for use with medical devices that may have electrodes or things that attach to the human body like tens machines or something like. It would be interesting to put that on the scope and see what kind of ripple that is giving, and then check some of the cheaper PSUs I tried.

I think the best option for me though is to build a simple linear PSU, just for Geiger counters. I wish you could still buy the old linear PSUs but seems you can't except in certain cases such as high end HiFi, and they are not cheap. I do have a very basic 16v AC 1A transformer wall-wart, so I could easily build a simple rectifier/smoothing circuit and feed my 7805 with that. That would be completely isolated from mains, linear, so no switching ripple.

Right before I post this I just checked the PSU that is running two counters with my scope. It was quick and easy to at least check the 5v rail. This is post 7805. So I'm checking the ripple after the 7805, and it's bad. 30mVPP. It's messy also, as you can see in the picture. I didn't check the 12v side, the actual PSU output, but 30mv ripple after a 7805 is completely unacceptable. I think a linear PSU is really the best way for me. 

 

I was referring to the 5v PSU that supplies the HV PSU, not the HV PSU itself. The 5v PSU I have is causing some noise in certain circumstances that covers all the electronics on the counter. Two counters in fact as I use that PSU to power more than one Geiger counter.