Power transformer questions...

OK, so I'm trying out the old step-down-and-then-step-up trick that you all seem to be using to get high voltages from cheap transformers...

I've heard people say that sometimes their transformers overheat, melt, etc., and I'm wondering what makes or breaks this setup?

So I got two 25VA Amvecos from Digi-Key (a 115+115:12+12 and a 115+115:15+15) and hooked them up like this: AC from the wall > 115:12 > 15:230 > 9.05K load (four 9K1/1W resistors I had lying around, in series/parallel)

I measured 120V/37mA coming out of the wall, 14.7V/263mA coming out of the first transformer and into the second, and 184V/20.2mA coming out of the second transformer and into the 9.5 KOhm load, which agrees with Ohm's Law.

Now, my questions are: Is this OK to do? Will conditions change much when I hook up a rectifier and filter caps instead of the purely resistive load? Neither transformer showed any signs of heat (at least not anything I could notice by touching them), but I only left them plugged in for about 5 minutes or so. I've read posts here from people who say the transformers overheat after 4 or 5 hours of being on! Does that make sense? What determines how much the transformers will overheat and how long it will take for them to get there?

So many questions :?...

Thanks, guys!

Peace,
Al.

> Is this OK to do?

Yes.

No winding should see an AC voltage higher than its rating. Maybe 125V or 130V on a "120V" winding, but any more will saturate the core, inductance drops low, a lot of reactive energy sloshes through the copper and heats it.

184VAC is less than 230V, so this should be fine. 250V in a "230V" winding could be problem.

The immediate "drawback" is that, in this case, you are only using 230/184= 80% of the voltage rating, the current rating is unchanged, you paid a 25VA price to get 20VA of performance.

Another issue is: if each transformer is 20% regulation, full-load voltage about 0.8 of no-load voltage, two such transformers is cascade gives 0.8*0.8= 0.64= 36% regulation. The first transformer is carrying about 20% more current than a rough estimate says, to cover the losses in the second transformer.

SInce 36% regulation is usually terrible, you might well want to derate the transformers to half their VA rating. This on top of the approximately 2:1 difference between AC VA rating and DC Watts due to ugly waveforms in cap-input rectifiers. So your 25VA cores might only be good for 6 Watts DC.

Something I had not thought about before. Because transformer ratios are fudged to give the right full-load output, the no-load output is higher. So 120V:12V>12V:120V would give, for small trannies at no load, around 173VAC on a 120V winding. So your trick of 12V>15V was wise.

You're the man, PRR... Thanks! :thumb:

If I end up going with this setup, I'm going to need to also pull .3A (regulated DC) from the 12V winding for a 6.3V tube heater. That's why I grabbed a 25VA core. But how would this affect the performance of the circuit? Would it still be "safe"?

And something I still don't understand, how can a transformer overheat only after being on for five hours? How can it take that long?

Thanks again!

Peace,
Al.

And something I still don't understand, how can a transformer overheat only after being on for five hours? How can it take that long?

Click to expand...

Think of an RC circuit driven by a voltage source. The time taken for the capacitance to ramp up to the driving voltage is literally infinite - it will never actually get there, it only approaches it asmptotically. However, the thing will be at 63% of the voltage at T=R*C, and 63% of the remaining voltage at 2*R*C and so on. Five time constants are typically taken as fully charged.

Why did I run you through that rigamarole? A transformer heating can be modeled as a simple RC circuit. The heat flux inside created by core losses and Isquared-R losses are the driver, the mass of copper and iron are the capacitance and the interwinding insulation and impregnation is the resistance.

For high mass, high resistance between windings - like for a well-isolated small transformer where the losses are largely in the secondary or primary and have to cross a thick insulating barrier to the other winding - the time constant is often hours. Especially where most of the heat actually escapes through the outside of the transformer and only the residual "excess" heat remains in to drive the temperature up.

Yes, it can easily be hours if the transformer is dissipating only a little more than it can.

One thing I like to do when I'm nervous is measure the winding resistances before operating the loaded supply for a while. Then during a test run periodically disconnect one side of one or more windings after powering down and remeasure the R's. Plug in the tempco of copper and get an idea of the average temp change.

It is limited in accuracy since the middle layers will typically have a bigger contribution, but you get an idea both of the magnitude of the change and of the thermal time constants.

Some have considered doing this in real time, i.e., while the supply is operating and using the information to throttle back the load where that is practical, for example in a power amp. It would seem like the trouble and circuit complexity would be way more than it is worth, compared to just resizing the trafo, but who knows.


Brad

Thanks, guys!

Peace,
Al.