0

I'm looking at a design for a IEC class II device, and can see that there is a Gas Discharge Tube being used for suppressing high voltage transients/surges. Since the device doesn't have an earth return, my gut reaction is that this is not a good idea (or wonders exactly what benefit it brings), and that if this is done then everything sharing the ground the GDT shorts to needs to have appropriate isolation from any conducting part or chassis.

Is this the proper use of a GDT when there is no earth return? What benefit does it bring in this case?

schematic

simulate this circuit – Schematic created using CircuitLab

milambergt
  • 3
  • 2

1 Answers1

0

Surges (indirect lightning etc.) will be earth referenced and so the GDT connection you have shown in your diagram needs to be tied to earth. Of course, if your actual circuit doesn't use an earth then there's no real need to use a GDT as shown; you can allow a surge to raise live and neutral to 2 kV (or whatever surge voltage spec you are working to) and, protection will come from isolation (the right hand component marked XFMR1).

And, if you can dispense with the GDT, you can also dispense with varistor 2 and varistor 3.

Of course XFMR1 needs to have adequate isolation capabilities and may need to be insulation rated as high as 6 kV.

The dotted block immediately to the left of XFMR1 will also be ineffective because it needs the two CM capacitors (C2 and C3) to be connected to earth to reduce CM emissions or susceptibility. I also have concerns about inductor L1 and how it will naturally imbalance CM noise but, it depends what is regarded as the noise source. Personally I would place it to the right of C2 and C3 (should they be connected to earth).

If you leave the GDT in place, surge currents that it conduct will be passed through to the ground on the secondary and, there is no hint in your question that this is proper protective earth. Hence, it's very likely a bad idea.

Since the device doesn't have an earth return, my gut reaction is that this is not a good idea

I agree.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • I'm glad to hear that this is nagging me for a reason! Good point with L1; would expect to see a second inductor for differential mode (and probably a cap across the lines between the common mode choke and XFMR1), or for L1 to not be present. And I'm not sure why I labelled the choke as XFMR2, but anyways... XFMR1 is actually a component power supply independently qualified as a Class II device; I get the feeling that someone copied an existing design where earth was available and put it in front of the component power supply. – milambergt Dec 09 '20 at 19:17
  • It looks that way. I did think it was a power supply proper but didn’t want to assume etc.. – Andy aka Dec 10 '20 at 00:14