I have been using this old audio amplifier for some time now and recently it stopped working - music was playing and it suddenly went out, you could only hear the 50 Hz buzz, magic smoke coming out and the thing was warm.

After changing a bad fuse and powering it up, I found out that the transformer was heating up so much its wax insulation melted. I tested the output with a multimeter and it roughly checked out, but when I disconnected all secondary wingdings and powered it back up it still got hot, so I figured that it was faulty.

Since it was pretty specific (2x20V with center tap plus 16V) there were no replacements available online, so I decided to replace its wingdings manually. I measured the primary winding to be 70Ω, 0.2mm copper wire and calculated that it has to be roughly 130m long. After taking the core and old coils apart, I winded 960 turns of the same wire type onto it, then 10 turns secondary so I could measure its voltage per turn, and carefully powered it up with a fire extinguisher nearby. It didn't trip any fuses, but it buzzed pretty loudly and heated up to 80 degrees Celsius before I disconnected it.

My question is: what did I do wrong? Although I never built a transformer, I have been playing with electronics for over a decade now, so I thought I could handle this.

For reference, the sockets here are 230V 50Hz.

  • \$\begingroup\$ Copper wire was insulated? \$\endgroup\$
    – Andy aka
    Apr 4, 2020 at 11:54
  • \$\begingroup\$ Yes, it was the one made for coil winding. The primary was 60 or 65 ohms, so no short circuit. \$\endgroup\$ Apr 4, 2020 at 11:56
  • 1
    \$\begingroup\$ Maybe your laminations are shorting out. \$\endgroup\$
    – Andy aka
    Apr 4, 2020 at 11:59
  • \$\begingroup\$ Did you count primary turns when unwinding it? NOTE that resistance measurements on an already damaged primary may not be accurate : if there were already shorted turns, that may have fooled you into reducing your turn count estimate, giving too low a primary inductance (too high primary current, too much heating) \$\endgroup\$
    – user16324
    Apr 4, 2020 at 12:00
  • \$\begingroup\$ Yeah, it was probably dumb to try and recreate a faulty transformer... I'll just double the turn count and see what happens. Thanks. \$\endgroup\$ Apr 4, 2020 at 12:08

1 Answer 1


The right way to get the needed turn numbers when rebuilding a transformer is to calculate from a known ORIGINAL winding the right turns per volt -factor. I guess your mains AC side winding has too few turns.

Beware: If you test a transformer with no load you can accidentally break it by switching the mains AC input OFF. The inductive peak can break the insulations if you happen to be able to separate the transformer from the mains input fast enough (=no arc in the switch, but one inside the transformer) A load or another still connected device in parallel with the primary would absorb the peak. See this: How does the inductor ''really'' induce voltage?

  • \$\begingroup\$ I thought I would double the turns, but I didn't want to waste another 6 hours just to find out the problem was somewhere else. Thanks. \$\endgroup\$ Apr 4, 2020 at 12:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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