Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

How to get the most out of a linear transformer

Status
Not open for further replies.

Plecto

Full Member level 5
Joined
Jan 4, 2012
Messages
315
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,298
Activity points
4,979
Hi. I've been interested in audio amplifiers for a while and I've began thinking of how much I can actually get out of a linear transformer? Giving that music is extremely dynamic and that the maximum power peaks are often several times greater than average power, couldn't it be an idea to install some sort of automatic fuse just to make sure the long-term average isn't too high? This particular transformer I have in front of me now is +/-18V 50VA with insulation class B which means that It can handle a temperature up to 130C (if I've understood it correctly). What if I attached a thermistor to it and made it so that the output cuts out when the transformer reaches 110-120C? I would still have a regular slow-blow fuse rated slightly above the transformer rating, but would attaching such a thermistor be ok engineering practice? Another thought is to monitor the supply voltage with an ADC, store the highest and lowest measured values to calculate the total supply ripple. By knowing the capacitance of the supply, total current consumption is directly related to supply ripple so the MCU can figure out how much current is drawing. It would then have to do some long term averaging to figure out if it needs to cut the output or not. I just imagine that a lot higher perceived output wattage can be reached by still having a rather small transformer with techniques like these, but I'm not sure if they are considered good engineering practice? I assume that a melting transformer is not a pretty sight?
 
Last edited:

Hi,

you can do the mathematical solution:
What´s the heating in a transformer? Mainly loss in windings and loss in core.
Loss in core increas with frequency and voltage. It becomes extremely unlinear when there is saturation.
Loss in windings are mostely caused by current. P = I^2 x R. Heating is (about) linear to power, as long as the temperature is low enough so that radiated thermal power is low.
Some overcurrent protection devices use the integral of I^2 x t to switch off.

But is all this useful when you don´t know environment temperature, air flow and so on?

Another solution is to feed a small DC voltage into the transformer primary. This causes a current depending on the primary windings resistance.
(you can also use a DC current and get a DC voltage)
Now you can use the thermal drift in copper resistance. The higher the temperature of the primary windings the lower the current.
Use a microcontroller to calculate R out of U and I. and then calculate the temperture out of R. (or delta_R/R).

Klaus
 

A lot of domestic "hifi" products are made in this way, so they are quoted as 400W music power, which really means 20W RMS. I feel it is good engineering practice to maximise the bang for your buck. The real problem is that the transformer could be underated but the reservoir capacitor must be still larger as it has to store the energy to carry music peaks which the transformer is not capable of. This is also true for SMPSs where the smaller ferrite transformers, could be made smaller, but the reservoir capacitor must then be made bigger.
I do not know if there is a specification for peak music power, such as able to carry more then ten cycles of 1 KHZ at maximum level before dropping down in power.
Frank
 

So attaching a thermistor to a transformer is not an uncommon thing to do? The datasheet for this particular transformer states TYP 40C temperature rise, I assume that is at max power output so cutting it at 40C+Ta would probably be well within the 130C maximum limit.

I tried to do the second solution that I mentioned, monitor the ripple voltage to calculate total current draw. I did some testing, monitored the ripple on my scope and I managed to get the mcu to accurately measure it. One question is how quickly it should cut the power once the output current is too high. I set it to cut after 1 minute with 150% output current, 30s with 200% output current etc. One issue I see is how the output will cut at lower and lower output currents as the capacitance of the rectifying caps sinks with time.
 

I assume that a melting transformer is not a pretty sight?

Very high temperatures start to deteriorate insulation on the wires. If aggravated by rubbing, etc., then it may lead to neighboring wires making contact, creating a shorted winding. It ruins the transformer unless you can break the contact somehow.

Very high temperatures may start internal metal expanding, even to develop noticeable bulging. This is likely to ruin a transformer.
 

What you need is a PTC resettable fuse.

https://en.wikipedia.org/wiki/Resettable_fuse

These are very commonly used for motor and loudspeaker protection, and if sized very carefully should work fine for protecting a transformer.

Yeah I've used some of these in the past. A couple of issues I have with them though. There's often quite a large difference between the hold current and trip current so it's hard to say exactly when it will break, besides, it will trip rather quickly when above a certain point. Another issue is that it's resistance rises after it has tripped which makes it easier for it to trip again, I think it can take hours even days for it to settle completely.
 

so it's hard to say exactly when it will break, besides, it will trip rather quickly when above a certain point. Another issue is that it's resistance rises after it has tripped which makes it easier for it to trip again, I think it can take hours even days for it to settle completely.
Does that really matter ?
Its not as though your transformer can run in perfect safety forever at 150C but suffer total permanent destruction at 160C.

If your transformer has become seriously overheated, its not going to cool right back down to ambient in only one minute.
It may be terribly annoying when the music stops and will not come back for fifteen to twenty minutes, then trip again very soon.
But to thermally protect the transformer, that is exactly what it must do.
 

Does that really matter ?
Its not as though your transformer can run in perfect safety forever at 150C but suffer total permanent destruction at 160C.

If your transformer has become seriously overheated, its not going to cool right back down to ambient in only one minute.
It may be terribly annoying when the music stops and will not come back for fifteen to twenty minutes, then trip again very soon.
But to thermally protect the transformer, that is exactly what it must do.

I can make it trip at 90C, after that raise the trip level to 100C. That way the user will know to turn the volume down without interrupting the music and the transformer should be way beneath any critical temperature.
 

I suppose you could use a thermistor buried in the transformer to control the system gain.
That would respond gradually, as things heated up and be much less annoying.

It would need some extra circuitry to do properly, a thermistor would not work well enough just by itself.
But an op amp, and something like a FET opto isolator could be arranged to have a suitable gain slope that decreased power output gradually with rising temperature.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top