Transformer secondary voltage at different loads.

[Chris935]

Hello,

I recently built a small linear power supply and learned about how AC voltage specs and measurements are RMS rather than peak, diode voltage drop, calculating smoothing capacitor values, regulator dropout voltages etc. Lots of "traps for young players" as Dave would say.

I made a spreadsheet where I could plug in the rated transformer secondary voltage at nominal mains voltage, the voltage drop of the diodes, the regulator output and dropout voltages, and my current draw, and it would tell me the value of smoothing capacitor I needed to keep the regulator out of dropout at the minimum allowable mains voltage. Alternatively, I could change something else and if the capacitor value went negative I knew didn't have enough voltage upstream.

When I built the PSU I discovered that I had quite a bit more voltage margin than I'd expected, and at that point I learned that the rated voltage is at full load and will be higher at lower loads. I had another of the same transformer, so I put one multimeter across it's secondary to measure voltage, and another in series with the secondary and a variable resistor to measure and adjust current. I plotted the results on the attached spreadsheet. At the higher currents the resistor was getting extremely hot so I didn't have time to finely adjust the current to evenly spaced values, I just took what I could get and recorded it. It's rated at 10W but started to smoke at about 6W even with a pair of TO220 heatsinks attached to it. The 10W is probably assuming it's secured to a large metal panel.

This was a torroidal transformer rated at 7VA 12V at 230V. Mains voltage was 240V before and after this test, I didn't have a third meter to record it throughout.

I have three questions.

1/ Is this voltage/load curve a standard relationship that could be used to predict the behaviour or other transformers of other ratings at other loads, if I were to convert it into "percentage of rated voltage over percentage of rated load" or is it specific to this particular transformer?

2/ The ratio of input voltage to output voltage is dynamic, varying according to the load. Does this ratio operate independently of the mains input voltage, such that I could make a prediction based on 230V input that would hold at 207V or 253V (mains tolerance) if I just added or subtracted the 10% from the 230V number? IE. If it's rated for 12V at 582mA with 230V input will it produce 10.8V at 582mA with 207V input and 13.2V at 582mA with 253V input? I suppose it can't work like this because it would mean a 10% greater power output than it's rated for. Assuming the power output can't exceed 7VA, do I get a voltage increase per the windings ratio but less current to maintain the same power? It would be interesting to try this with a variac, but I don't have one.

3/ I can test this myself, but thought it was worth asking here too. As I increase current draw I'm "using up my allowance" of something, which is causing the voltage to drop, which must be either the iron core or the secondary winding. I did my testing with a dual secondary transformer with both secondaries in paralell, so I'd have approached both the "core" and "winding" limits simultaneously. If I was to test this with only one of the secondaries, so that I was using the winding to it's maximum but only half of the capacity of the core, would I get the voltage from full load in the previous test (because the winding is at full load), or the voltage from the half load (because the core is only at half load)? Why?

Thanks for any help,

Chris

[mariush]

Tiny transformers (let's say less than 20-25 VA) will have huge variations ,  by design they won't be great.  They output voltage will vary much more depending how big of a load you have attached to them, compared to bigger transformers.

Besides the issues caused by the small va rating of the transformer, you have other issues, like for example different forward voltages (depending on how much current you have flowing) for the diodes in bridge rectifiers (or individual diodes if you choose to go that route) , different kinds of linear regulators (which could have dropout voltages from as little as 10-30mV up to 2-3 volts) and so on..

The datasheet of the transformer usually specifies the percentage, how much it's expected to deviate from the advertised output value.

[orolo]

The primary side of a transformer has a winding series resistance. I took a small scrapped E-I transformer I have here and measured 740 Ohms at the primary. If you draw current from the transformer, that series resistance will steal voltage from the primary inductor, and it will see less than 240V at the primary, and therefore, less voltage at the secondary.

For example, your transformer has a voltage drop of 0.7 volts at 150mA, that is, 4.66 Ohms. Reflecting this back to the primary side (14.8:1 turns ratio) that means about 1K. Rather higher than mine, but not too unreasonable. I'm sure there are other nonidealities at play.

Edit: of course, the secondary also has a winding series resistance; in my transformer it amounts to 2 Ohms. That resistance has also to be taken into account. Considering both resistances (secondary + reflected primary,) probably they add to the 4.66 ohms in your transformer.