Silly question here.

I've wired up a full-wave rectifier using this dual-secondary transformer. If you want to check my wiring, I connected red-to-orange to "create" the center tap (I assume this is OK?), and use black and yellow to provide +/- 24VAC. The center-tap is my circuit ground, obviously. See the rough idea below.

rectifier circuit

This circuit works nicely. Using some filter caps and linear regulators, I get a nice DC voltage. Everything is as expected. I was running some relatively high current through a load (about 1 amp).

After a few hours of use, I noticed I was getting a "tingly" feeling when I accidentally touched the circuit ground. My multimeter says 77VAC between circuit ground and earth mains! Yikes!

I don't understand how it floated so high above earth ground. I understand that when you have two unconnected grounds, you have no right to expect them to not float with respect to each other.

But, in this case, it seems to me there should be no way for current to flow from myself (at earth ground) to the circuit ground--hence there could be no voltage developed between "us". I thought this was entirely the point of isolation transformers: voltage is developed with respect to each lead of the secondary, but not with respect to earth ground--so one is safe when working with high voltages. My transformer claims 4000VAC dielectric resistance between primary and secondary. So, what is going on here--how could there be such a voltage? And, should I connect earth mains to circuit ground here to fix this--or will I blow a fuse?


That's a big fat unscreened transformer, so there is plenty of capacitive coupling between the input and output. A high impedance meter could easily read a fraction of mains voltage between the center tap and earth. If you flip the input connections you'll likely get a somewhat different measurement.

The measure of health of the transformer is the leakage current at near maximum isolation voltage. It's not possible to measure that easily with a simple multimeter. The voltage measurement alone is not an indication of a problem, however the tingle could be. It would only take microamps to make the meter read something like 77VAC, but a perceptible tingle would reportedly take more like 1mA.

There is no problem in earthing the secondary center tap- if it blows a fuse, the transformer is faulty.

  • \$\begingroup\$ I see; can I model this as a capacitor between primary and secondary? If that is the case, wouldn't there be a large current always flowing to ground if I connected the center tap to main's ground? Well, "large" I suppose would depend on the amount of capacitance. This current, flowing constantly would cause 60hz voltage noise in my ground plane, no? (eventually this design is going on a PCB) \$\endgroup\$ – Paul L Jan 23 '15 at 6:54
  • 2
    \$\begingroup\$ More like two caps, one from each side of the mains. The current should not be large because the capacitance should not be very large. Say 50pF (WAG) at 120VAC/50Hz would be a couple uA, which should not be perceptible. \$\endgroup\$ – Spehro Pefhany Jan 23 '15 at 13:08
  • 2
    \$\begingroup\$ Or one cap to half the supply rail (i.e. halfway up the primary). I observed about 1000pf on a smallish (probably 20VA) toroidal transformer.After some surprise, a colleauge did a WAG calculation based on its dimensions and a guess at permittivity ... he came up with 1100pf! So it's real - and pretty harmless to humans, though probably not to microprocessors if you're not careful. \$\endgroup\$ – Brian Drummond Jan 23 '15 at 19:07
  • \$\begingroup\$ Thanks Sephro. You're right--I just measured the current and it's about 38uArms @ 60Hz, from circuit-ground to earth-ground, steady-state at 1A load, \$\endgroup\$ – Paul L Jan 23 '15 at 20:19
  • 1
    \$\begingroup\$ @PaulIL The best thing to do depends on what else you're connecting. If all your connections (in/out) are differential or floating then grounding it directly is probably best. If some are single ended, then it might be better to have it grounded in one spot. You could also consider grounding it through (say) a 100 ohm resistor so any ground loop current would be small, and the voltage would only be mV. \$\endgroup\$ – Spehro Pefhany Jan 23 '15 at 20:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.