0
\$\begingroup\$

I have a transformer. It is a custom wound transformer for a old linear powersupply. As is common, it has two primary windings for supporting 115V and 230V operations (or, 120V and 240V as is now more common. the powersupply is old). So, the options are the following:

Transformer winding options

The owner of this powersupply bought the device in the USA and is now in Europe, so he wants to change the wireing such that this is possible. They desolered the transformer windings, but then got cold feet and stopped, worried they would damage the device (or themselves) if they rewired the transformer incorrectly. It was then handed to me and I now have to make sure it's done right.

However, the transformer doesn't have the wires labled. I wondered, what is the correct way to verify what pins are what coil, and more importantly, what orientation the coil has. It's obvoius in this case what the primary windings are since they were labled on the PCB, and those are the only ones unsolderd. I also know what two pins belong to the same windings with a simple continuity check.

What I'm now worried about is the "orientation". What I don't want to do is this:

Wrong orientation of second winding

I have only limited tools to my disposal for now since I cannot use univerity instruments. All I have is a simple multimeter and an old analog oscilloscope.

I was thinking about applying, with a sound card, a low-frequency (say, 1kHz) signal. Since it is so low voltage, and higher frequency as the transformer is rated for, I expect that the voltages on the secondary would be on the order of milivolts, not able to turn on any of the circuitry. I could then apply this signal to one of the windings. By then using two channels of the oscilloscope, one for each winding, I could see what way the windings are in respect to each other (if I have one connected the "wrong" way I expect a near 180 degree phase shift, otherwise almost perfectly in phase).

Is this a good method? how is this done "propperly"? I would rather not fully dissasemble the powersupply to unsolder the secondary windings as well since this requires me to pretty much take the entire thing appart.

\$\endgroup\$
  • 1
    \$\begingroup\$ You could also try it the other way round. Put, say, 6V AC from a suitable mains transformer across one of the secondary windings and fiddle with the primary connections until you get something that makes sense. \$\endgroup\$ – Leon Heller Jun 2 '17 at 14:13
2
\$\begingroup\$

Perhaps the simplest way is to wire the transformer up and test the whole unit, with a current limiting load in series with a primary. A mains filament light bulb would be suitable (make sure it's filament, not LED or CCFL). Or a mid power resistive load like curling tongs. Whatever load you choose, it should not draw more current than your transformer input is fused for. This does mean you'd have to connect things mains side. If you're not happy doing that safely, then you shouldn't be inside with the transformer mains connections anyway!

Wired correctly, the transformer will produce a significant output, and the bulb will be dim or off.

Wired incorrectly, the transformer will present a short circuit to the input. It will produce no output, and the bulb will light at more or less full brightness.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Ofcourse, this would be simplest. But it also requires me to remove the secondary load. I was hoping to be able to test it somehow with such a low input signal that the secondary load doesn't turn on (if the secondary voltage is less than a diode drop the rectifier will never conduct). However, if all else fails, wireing the device up is clearly the easiest. It's just a big system and to remove the secondary requires me to dive in and dissasemble the entire device \$\endgroup\$ – Joren Vaes Jun 2 '17 at 14:23
  • \$\begingroup\$ No, you don't have to remove the load, that's the whole point. Just plug it in, and see if it works, but with the current limit protection if it's wrong. It may not work 'properly' if the bulb drops too much voltage, but you should be able to tell the difference between that, and the bulb being full on. Check the rated power of the appliance first to see whether the bulb would be expected to run it. The great advantage of a filament bulb is its non-linear resistance, low when cold, high when hot. If the unit takes <50% of the full rated bulb current, it will work properly. \$\endgroup\$ – Neil_UK Jun 2 '17 at 14:40
  • \$\begingroup\$ Oh ofcourse, I see. Sounds like I have to see if I have an old lightbulb laying around... (they are not sold anymore, unfortunatly) \$\endgroup\$ – Joren Vaes Jun 2 '17 at 14:46
  • 1
    \$\begingroup\$ The 'ordinary' ones aren't, but there are still 'special' ones available, for a while. As the world switches to LED, these too will vanish from the shelves, but for the moment, they're being cleared out for cheap. Stock up now, as they're such a useful test component. I've just bought a dozen each 50W spot lights, 12v and 240v, for pennies. Should last me out! \$\endgroup\$ – Neil_UK Jun 3 '17 at 6:10
  • \$\begingroup\$ I shall hunt stores for these. Thanks for the tip! \$\endgroup\$ – Joren Vaes Jun 3 '17 at 6:10
0
\$\begingroup\$

Drive your sound card signal onto the secondary and look at the relative phases of each primary winding is my advice.

However if you did exactly this it would work: -

enter image description here

If you don't know how it was originally wired then you can't do that.

If there is any doubt about the transformer's state or condition get a new one. Mains AC voltages can kill and burn.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Yes, the problem is that I don't know the original wireing. Maybe I should state that more clearly in the original question. The problem is that I would rather not desolder the secondaries as this would require completely dissasembling the powersupply. Is it at all possible to correlate the phase shift from primary 1 to primary two to figure out which relative "polarity" is used? The transformer itself is in excelent condition, it's just that the windings aren't clearly marked. Given the device's age it is likely economicly unviable to get a new transformer. \$\endgroup\$ – Joren Vaes Jun 2 '17 at 14:18
  • 1
    \$\begingroup\$ Yes, you can correlate the phasing from P1 to P2. Mark the two wires that are in phase and imagine they are the positive terminals of a battery then, wire those batteries in series as you would to get double the voltage. \$\endgroup\$ – Andy aka Jun 2 '17 at 14:22
0
\$\begingroup\$

Apply 120 V to one of the primary windings, and connect ONE wire of the second winding to one wire of the first. Measure the voltage from the free end of the second winding to the other end of the first - if you have the polarity correct, you should measure virtually no voltage. If wrong, you will get 240 V.

schematic

simulate this circuit – Schematic created using CircuitLab

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ This assumes I have a 120V source available, which is not the case, unfortunatly. That said, I might be able to throw something together using a 240-12V transformer and a 240-24V transformer to get my hands on 120. \$\endgroup\$ – Joren Vaes Jun 2 '17 at 15:30
  • \$\begingroup\$ You could use a lower voltage instead of 120. The point is that you get virtually no voltage between the free ends if it is wired correctly, and double the supply voltage if wired incorrrectly. \$\endgroup\$ – Peter Bennett Jun 2 '17 at 15:34
0
\$\begingroup\$

I love when the quiet voices give the solution. @LeonHeller nailed it with his comment to OP.

Backfeed the transformer secondary with lower voltage than it's rated for, then check to make sure sensible voltages are coming out of the primary.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Sure, this is gennerally a very good method - but the thing is that I wanted to do it without touching the secondary since it requires more dissasembly. \$\endgroup\$ – Joren Vaes Jun 3 '17 at 6:03

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.