Grounding For Isolated SMPS Power Supply

Thread Starter


Joined Mar 28, 2011
I have project that uses a cheap DC SMPS (the low cost fixed-voltage metal enclosed units you see all over Ebay).

The circuit ultimately drives a couple of local relays whose output is interfaced outside the module, so I saw no reason to tie the ground of the circuit to anything, allowing it to simply float.

The supply itself has worked great, but I ran into an interesting problem: With some regularity, I found that I had a problem getting my programmer (an STLink V2) to program the MCU. It gave odd verification failure problems, and flash programming errors. Sometimes to the point I had to leave the module for a period of time (hour, next day) then try to reprogram. I even replaced some of the microcontrollers thinking this was an issue with the IC.

After a number of attempts to solve the problem, I finally found something that worked: If I used a jumper to hook the ground of the programmer to earth ground from the mains, and it alleviated the problem.

In my mind, the fact that the DC circuit was floating, and that I was programming through a non-isolated programmer, the programmer, upon being installed, would immediately tie the ground of the circuit the ground of my laptop. These programming issues seem to hint otherwise. My benchtop supply is floating, and I never have had a problem like this after years of powering all types of prototypes with no ground tether.

So I give this back story to ask the question: Ideas on why this has become a problem? Are these cheap SMPS supplies not really isolated? As a rule, is there some reason not to leave these circuits floating?


Joined Aug 7, 2020
Most switched-mode supplies have a class-Y capacitor between rectified-mains-positive and output negative, as an interference suppression filter.
It will charge up to about 120V DC, and will be grounded when you connect to the output of the power supply.
Connect the negative output of the supply to ground before you attach the probe.


Joined Jan 23, 2018
There are different levels of isolation, medical grade supplies have a much better isolation standard than others, and the cheap ones may meet some level of isolation. So certainly inadequate isolation could be a problem.

Thread Starter


Joined Mar 28, 2011
All useful information.

@Ian0 , I am only briefly familiar with safety caps, but was unaware they were configured in this way. How does this even provide any filtering if the post rectified +V and output ground are isolated via the transformer? This would seem to suggest that such a supply would expect the output ground to bonded to earth ground?

@MisterBill2 - I am wondering if this is the case, it still has enough tethering to simply cause a lot of noise on the ground of my device, particularily if I am hooking to another floating supply like a laptop.

@AlbertHall - Agreed, I was not concerned with being tied to earth ground, but more expected two floating supplies tied together would operate (program download) properly.

How would a typical consumer electronic device handle this? Certainly relying on an earth ground is not possible. Is this just a PSU quality issue?