An unusual problem for a side project

An answer from the other forum:
The apparent problem is that a DMM that functions correctly when powered by 3 volts of batteries does not function correctly when powered from a 3 volt regulator supplied by a 5 volt switching power supply.
- You understood the problem perfectly ! It is exactly what I did.
It does function, but with an additional error that was not there while powered normally from 3V batteries.
The error is like this:
-1- while NOT measuring anything, when the probes are on the table, and importantly, my arms or hands are very close to them or touching them, the display is picking random frequency, around 300Hz. Why I say important, because when I lift away my arms or hands from a certain area around the DMM, it starting to behave normal, without any random fv reading.
-2- I only tested measuring Resistors. Its the only thing it is VERY good for. In my experience and I love it for that. It is searching automajically for the x R value, no need to switch the knob to a certain range like the old DMM. Some of you probably already know already this function. Anyway, AFTER measuring a 10k, it was continuously search for other R values but not stopping from the search as before. It is randomly jumping from M to null to k.
Notice how in the picture attached, I keep my hands in air, but it is displayed a value there. That value actually is switching randomly at an interval of 1/s.

These are the 2 major errors I observed. Maybe there are others. I didnt do all the DMM operations. I did measure voltage one time and AFTER measuring, it is jumping in values as with the R, but for the V values.
My opinion is it's autosearch function mest up somehow.
-3- All the measurements are FINE, are ok, with this error that happens AFTER the measuring. So the values measured are OK. It is doing its thing as a DMM. No problem there. But After, instead of stopping from its usual automatic search, it is jumping around.
-4- If im putting back the batteries, everything gets back to normal as before !!! No errors of any kind. This tells me loud and clear it is the power supply Hack that I made.

--5--Thats why I wanted you to watch my movie because I did ALL this explanations here in writing, there live in the video and more. Including changing the batteries. It is a very good and efficient way of explaining a problem through a video. It's like showing you the problem instead of explaining it in writing. It is how I see it. But I also understand your point of view as well, the music is too much perhaps, it is a german military march from 1945 btw, my beautiful carphatian voice, accent or speed might not be on the same resonance with some ears. heh. Or movies in general excepting me personally. Hmm....I make the best ones on the whole world, haha.

I just made the test with the value I didnt think to put. Before I put 100nF and a 10uF in parallel.
Now, I put 2 capacitors in parallel, one of 100nF and another of 220uF.
- It didnt work. No effect. The same error is present.
I told you is a very strange problem.
How do I put a 3 volt source that is completely floating, but from this SMPS that I already have? What circuit/component can simulate/emulate such floatingness?

Ive just added 5 capacitors in parallel. (its a gray 5th one behind the green one)
220uF - 10uF - 100nF - 150pF - 22pF
Still no effect. The same error.

I don't have time to watch the video but what you probably have is common-mode interference. I.e. an AC voltage with respect to ground, coming along both conductors of the PSU DC output. When other parts of the DMM are coupled to ground, e.g, by capacitance between your body and the probes, some of this voltage appears across parts of the DMM circuit that are not protected against interference (because it's designed to work on battery.)

Because it's common-mode, you can't get rid of the interference by capacitive shunt filtering or regulation in the DC power lead. You could try a longitudinal inductor in both conductors, or change to a linear PSU.

It is a common problem that SMPSU makers have to compromise on. If there's no capacitance between AC input and DC output, the PSU tends to give out a lot of its own switching frequency as common-mode. Makers include some Y-capacitors to tame that, and instead you get more 50Hz. Larger PSUs tend to have the DC output referenced to ground, which helps in many situations but not in yours where you need it floating.

Lucien Nunes said:

a longitudinal inductor in both conductors

Click to expand...

How this would look like ? -I did tried already and I present it in my video as well, a coaxial cable, but only for the + wire. The - wire of the coaxial was linked to the - wire coming from the SMPS. It didnt work ! The same error persisted. If you are refering to something other than coaxial, please tell, Im all ears and eyes.
Thank you for responding.

I mean that both the +ve and -ve conductors pass through two separate windings on the same inductor. It is a standard configuration for suppressing cable-borne RF interference. I have no idea whether it is a practical solution in this case as we don't know whether the problem is indeed RF or whether it is perhaps ordinary 50 or 100Hz leakage current.

Personally I would not attempt to power a battery DMM from a standard SMPSU because I would expect leakage current to be a problem. You can never consider the DC output truly isolated from the AC input at all frequencies and current levels of interest. For applications requiring ultra-low leakage, there are special 'medical grade' SMPSUs but even these would not necessarily be suitable. Consider the SMPSU as a source of DC voltage attached to a 0.1mA, 110V AC current source, and you will start to visualise what you are up against.

I would stick to batteries. I have high dynamic-range audio measuring equipment that is designed to be AC powered via a linear (not SM) PSU, but for the best accuracy at the highest gain it is recommended to unplug the AC and run on battery.

Lucien Nunes said:

I mean that both the +ve and -ve conductors pass through two separate windings on the same inductor. It is a standard configuration for suppressing cable-borne RF interference.

Click to expand...

Please show me a picture of it. Also a full official name of this inductor you speak. Its a new thing to me.

A popular name is common-mode choke.


If the interference is line-frequency leakage, it won't have enough inductance to stop it. There is a reason DMMs don't have a DC input jack!