0
\$\begingroup\$

On the primary side, I have a low-impedance AC supply that can be adjusted to a minimum of 100V, with a frequency variable from 5 to 800Hz (and I need to operate over this whole range). I need to somehow transform this down to 2.5V or lower, naturally as efficiently as possible, with the secondary side drawing up to 12A.

I considered just buying a transformer of roughly the right ratio and maybe a bit higher power rating than I need, but there's rarely any detailed information with regards to resistance, loop counts, wire gauge and/or inductance that I'd need to properly calculate the losses caused by running a 50Hz transformer at 5Hz. I have no idea if it would just get a little hot, or melt through the casing; and even if the losses were acceptable I wouldn't know if it would even transfer enough power to the secondary.

I have looked at audio transformers, as they are designed to operate at a wide frequency range, but even they seem to only go down to 20Hz.

What would a more experienced engineer than I do in this situation? Are there certain specs I can look for in pre-made components, or will I need to look at having one custom-made?

\$\endgroup\$
4
  • \$\begingroup\$ Read all you can about high fidelity valve (vacuum tube) audio amplifier design, then design (and if necessary) wind) your own. This is discussed in detail in Wireless World issues from 1947 - search for "Williamson amplifier" should find them - he was designing for 15 or 20Hz but it gives you the information you need to scale down. \$\endgroup\$ – user_1818839 Jul 7 '17 at 15:04
  • \$\begingroup\$ XY problem. Common transformers are tuned for one frequency, they even lose efficiency with a 50/60Hz change. What you really need is a digital power supply that makes your choice of frequency at your choice of near-2.5V voltage. It can run on your 100V supply, but it'll just rectify it into DC first. If you need the 2.5 to be phaselocked with the 100V, maybe you could do a buck converter of some kind. \$\endgroup\$ – Harper - Reinstate Monica Jul 7 '17 at 16:36
  • 1
    \$\begingroup\$ If you add a bit of context and some detail about the bigger picture of what you are trying to solve, and not only how you tried, would surely land you a good solution. Your question seems to "stand too close to the problem to actually see it"... It is no ordinary situation to require 12A at 5Hz, on ~2.5V... \$\endgroup\$ – ppeterka Jul 7 '17 at 19:54
  • \$\begingroup\$ If you want something off-the-shelf (and supported, calibrated, packaged nicely), I'd look at amps for vibration test stands. Common frequency ranges are DC to few kHz. \$\endgroup\$ – Nick T Jul 7 '17 at 21:41
3
\$\begingroup\$

Using a transformer to convert your existing AC supply to a lower voltage does not make sense due to the very low frequency. You're not going to find such a transformer off the shelf. Even if you did, it would be huge, heavy, and expensive.

It sounds like what you really want is something that can put out up to 2.5 V AC at up to 12 A. Start from there. Your existing supply is not a solution. Forget about it.

Your spec is something a class D amplifier can do. Off the shelf audio amps don't quite work here due to the 5 Hz low end requirement. You will therefore have to design your own. There are a variety of class D chips out there. It's been a while since I looked at the offerings, but if I remember right, most of them actually work down to DC and expect the external circuit to add the high pass filter for audio use. You put the parts around one normally, except put the high pass filter below 5 Hz instead of below 20 Hz.

Another possibility is to do this yourself using a microcontroller with a PWM generator meant to drive a H bridge. You haven't specified harmonic content, so doing this open loop might even be good enough. Or, you can close the loop with error feedback like a real class D amplifier. One advantage of this is that it is then very easy to synthesize the waveform you want directly in the micro.

\$\endgroup\$
5
\$\begingroup\$

To avoid saturation at 100 volts and 5 Hz, a 50 Hz transformer would need to be rated at 1000 volts or higher. To avoid saturation with reduced frequency, the ratio of voltage to frequency must not exceed the ratio for which the transformer was designed. If that ratio is exceeded, the transformer will draw a high current with no load on the secondary and get very hot very quickly.

An old microwave oven transformer (MOT) might work if you use the original output winding as your input winding. However the ratio will probably not be high enough to make the secondary voltage as low as you want. You may need to connect the output of the MOT to another transformer to reduce the voltage some more.

Another alternative would to connect the primary windings of more than one transformer in series. Three 400 volt transformers would work. If you use three identical transformers, the secondary windings could be connected in parallel. Otherwise, connect them in series.

The above alternatives are not going to be very efficient.

\$\endgroup\$
4
\$\begingroup\$

What would a more experienced engineer than I do in this situation?

This one would ask why you are solving the problem that way. That voltage and current could be supplied by a DC coupled audio power-amplifier driven by a suitable signal generator.

enter image description here

  • 2.5 VRMS = 7 VP-P approx and this could be supplied by a 12 V audio amplifier running in bridge mode on a single-ended 12 V supply.
  • 2.5 V x 12 A = 30 W.
  • Load resistance appears to be = 2.5 / 12 = 0.2 Ω.

Some of the high-powered in-car audio sub-woofer speakers may get close to handling this sort of load. You could find out more about this late on a Friday or Saturday night if you go to the right part of town and look for brightly painted cars with spoilers and what looks like collapsed suspension.

Fill in more details in your question ...

\$\endgroup\$
3
\$\begingroup\$

What would a more experienced engineer than I do in this situation?

The first thing to decide is if the problem is XY i.e. are you thinking a transformer is the right approach to deliver what you want at the output when in fact there may be a superior solution that you haven't considered.

To investigate this you need to put in words the range of input voltages and frequencies - stating the minium input voltage is not good enough because, at a higher voltage, a transformer (if indeed that is the best solution) might start to saturate its core. So an input spec is needed.

Then an output spec is needed.

Then someone can start a design or make recommendations. It's quite feasible that using a SMPS and a PWM driven MOSFET power amplifier can deliver what you want (rather than try and find a transformer that maybe only achieve half of what you require).

I'm not ruling out a transformer; I'm just saying that what you are asking for isn't necessarily justified by your requirements.

\$\endgroup\$
3
\$\begingroup\$

Your output VA is 2.5V * 12A = 30VA at 5Hz.

The minimum size 50Hz transformer you need to meet this requirement is 30VA * (50/5) = 300VA, which though large, is available.

Your 2.5v winding at 5Hz would need to be a minimum of 25v at 50Hz. I expect you could find find toroidal mains power transformers for audio amplifiers in this sort of VA and voltage.

5Hz is the difficult end for core saturation.

800Hz is the difficult end for core losses. A standard mains transformer has only been designed with laminations good for 50/60Hz, so will have high losses at 800Hz. You might be able to find power transformers rated for 50 to 400Hz, but they will be more expensive due to thinner laminations.

\$\endgroup\$
2
\$\begingroup\$

Get thee to a classic audio antique store, and borrow a transformer from the "Williamson" output stage. You'll get some losses at 5Hz, but you've expressed no compelling need for flatness.

\$\endgroup\$
1
  • \$\begingroup\$ hah! snap! .... \$\endgroup\$ – user_1818839 Jul 7 '17 at 15:06
1
\$\begingroup\$

5Hz to 800Hz, transformer? Stuff that!

Rectify, smooth and such, caps will be largish, but you only need 30VA so well doable, this gets you ~145V (RMS vs Peak), use that to drive a switch mode supply down to maybe 12V or so, an off the shelf module if you just buy an isolated DC/DC converter, but not hard to design if you want to do it cheap in quantity with a flyback.

If you need the output to be AC like the input then having gotten your isolated 12V rail, a class D power amp (TI have suitable chips aimed at car audio) will get you your 5Hz to 800Hz back at a few volts, feed the input with a sample of the original waveform.

I figure this fits on a eurocard easily, and weighs maybe a few hundred grams or so, as opposed to the transformer approach that is about 5kg of iron and at least an order of magnitude more expensive.

This assumes that the output has to be AC and basically a smaller copy of the input, if you just want DC at a few volts, then you can leave the power amp out, and can probably find a 'quarter brick' converter that will get you 100-160V straight down to 2.5V or so, I expect MuRata or the like have something off the shelf. You will need the rectifier and big cap, but that is hardly a major issue.

\$\endgroup\$

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.