Winding a toroidal xformer for audio impedance matching

[dnyberg]

I'm adapting some headphones for use with a portable audio device, and have experienced some success: sound comes out and no smoke... but the sound is very faint. I think the reason is that portable devices assume efficient, low impedance earphones in the 15 ohm impedance range, while the ones I'm trying to drive are somewhere in the 150-300 ohm range.

The solution I'm thinking of trying is to wind a couple (one per channel) of toroidal transformers with a suitable turns ratio. I understand the n^2 impedance thing, and would choose a toroid of a size I can hand wind without use of a microscope, and since we're talking tens of milliwatts, I figure the wire wrap wire I have on hand should do okay.

But I have absolutely no idea what composition I should look for for use in the audio range. I'm guessing ferrites I've salvaged from PSU's were designed for much higher frequencies.

Can anyone offer useful advice regarding size/composition/sources of toroidal cores I should look to for this application?

Thanks!

[Gyro]

Most audio transformers (even the small signal ones) use laminated cores, high grade Silicon steel at minimum and Mumetal for the better quality ones. The laminations need to be thin for decent high frequency response. You need to achieve a decent inductance for low frequency response. I'm not sure that you will be able to achieve useable results in a compact form factor, you probably want powdered iron cores, certainly not salvaged PSU ones (as you've already guessed).

Without trying to dishearten you, you will probably achieve less frustrating results with off the shelf parts. You don't say what part of the world you're in but here are some examples:

https://uk.farnell.com/c/transformers/audio-transformers?ost=audio+transformer&rd=audio+transformer

As with all things, the higher the quality, the higher the price (mainly when it comes to low frequency roll-off), but you may be able to get acceptable quality from lower cost parts.


Edit: homing in a bit, maybe something similar to this one... https://uk.farnell.com/vigortronix/vtx-101-001/transformer-audio-6-3-6-3-1-1/dp/1674300 The DC resistance of the 3.75R winding might not be good for portable devices without output short circuit protection though.

[Kleinstein]

Iron powder cores don't make good transformers, espeically not at a small size. The permeability is too low and thus too much magnetizing current needed.
One more wants very high permeability. Nonocrystalline toroids would be a good option. They are found in some high performance (small size, high current) common mode chokes. Winding torroid by hand is quite time consuming.

Still the transformer adds additional losses.

Chances are it would be easier to buy suiteable headphones for the portable device.

[Zero999]

Why not simply buy one? You're not going to learn much from winding a transformer.

At a push, you could use a small mains transformer, run in reverse. I've tested them and they work up to around 20kHz, but I wouldn't guarantee it. 120V:24V would give a ratio of 1:5, which would give a decent match.

[dnyberg]

Well, these headphones are built into an aviation helmet, so replacing them would be expensive and present some mechanical problems. Likewise, transformers perfectly suitable for a desktop audio device won't fit well into my shirt pocket. It's disappointing to hear off the shelf ferrite cores likely wouldn't work well (are we assuming I need excellent fidelity? Anything noticeably better than "voice grade" would do). Can one buy toroids made of laminated silicon steel?

[dnyberg]

Oh yes, I forgot to mention one benefit I'd hoped for from hand winding (I need only 2) is I could rewind them until I found the right ratio.

[Gyro]

Does "aviation helmet" imply that you are near a source of power? If so, an active gain stage would be a superior option, certainly cheaper. Battery power might also be an option if output power levels are low.

[dnyberg]

Available source of power will be a 9V battery, but if I get 2 hours lifetime from one, that'd do, so it could be a simple amplifier and accept the inefficiencies would be a pragmatic solution. Something like this maybe (plus a vreg ofc)?

https://www.allelectronics.com/item/amp-8403/3w-stereo-amplifier-module/1.html

[T3sl4co1l]

Use a power transformer, say 120 to 36V.  Get the smallest one you can, it'll handle more than enough voltage, even at low frequencies.  Toroid is the correct choice over shell or split-bobbin type, the leakage inductance is lower meaning high frequency response will be acceptable (if maybe not ideal, YMMV).

Tim

[Gyro]

Available source of power will be a 9V battery, but if I get 2 hours lifetime from one, that'd do, so it could be a simple amplifier and accept the inefficiencies would be a pragmatic solution. Something like this maybe (plus a vreg ofc)?

https://www.allelectronics.com/item/amp-8403/3w-stereo-amplifier-module/1.html

Although the efficiency looks high, that board is optimised for putting out a fair bit of power into a low impedance load. A better solution would be a bridged output using cmos opamps. There are plenty that are spec'd to drive 32R loads these days.


Edit: A quick and lazy search on ebay (bed time), reveals a number of tiny TDA1308 based headphone amp PCBs. Typical quiescent current 3mA. Absolute max supply voltage is 7V though - I'm sure there are better boards out there.

[penfold]

Can one buy toroids made of laminated silicon steel?

Yes, called "tape wound cores", I've never seen them as a stocked catalogue item, probably made to order from most manufacturers, so small quantities will be pricy, same deal with small EI type cores also I guess. You could always strip the windings off a pre-wound transformer and re-wind... but I have a feeling you'd be somewhat disheartened on the idea of self-winding after that.

[dnyberg]

Hm, bridged op amps. Can you suggest likely (hopefully somewhat jellybean) parts off the top of your head? Or I guess I could dig into art/electronics and cobble up something that'd run directly from the 9V supply, though that's not the direction I was originally thinking.

[dnyberg]

Can one buy toroids made of laminated silicon steel?

but I have a feeling you'd be somewhat disheartened on the idea of self-winding after that.

Yeah, I'm kinda wilting here; thought I had it largely figured out. So now I'm just looking for a practical solution, probably wildly off from my original thoughts. :/

[bob91343]

How about a reality check?  If you correct the mismatch, how much more audio power will you get?  You might be surprised that the resultant volume may not be as much an increase as you hope.

[dnyberg]

How about a reality check?  If you correct the mismatch, how much more audio power will you get?  You might be surprised that the resultant volume may not be as much an increase as you hope.

Possible, but how would I know until after I've tried it?

[dnyberg]

Maybe something like this?

https://www.mouser.com/datasheet/2/294/njrc_s_a0009983567_1-2279521.pdf

[exe]

I think an amplifier is a good and simple idea. It's just a few parts: a battery, opamp, and a bunch of resistors and blocking capacitors. The only downside is you'd need to charge/replace batteries. Or find another way to supply it. There are might be suitable headphone amps on the market, I didn't check.

PS why do you think impedance matching is needed? Or do you mean you need bigger output voltage swing to drive high-impedance headphones? That's imho not an impedance matching.

[T3sl4co1l]

Or this: https://www.digikey.com/en/products/detail/hammond-manufacturing/146G/454914

Not particularly cheap though.  If your time isn't worth much, sure, can wind it yourself, in part or in full.  Iron cores aren't terribly easy to find (let alone other really nice signal cores like permalloy), but nanocrystalline is (used for EMI purposes), and has excellent permeability (in the right grades).  Downside, it's even more expensive!

The nearest toroid as I suggested earlier, would be something like this,
https://www.digikey.com/short/797hwjh0
though none of them are showing stock right now.  Mouser or others may have some.

Tim

[bob91343]

You can calculate the increase in power by use of a novel procedure called Ohm's law.  Calculate what you have, and then calculate what you'd have with a proper match.  If the difference is less than around 3 or 4 dB you won't gain enough to matter.

[The Electrician]

Yeah, I'm kinda wilting here; thought I had it largely figured out. So now I'm just looking for a practical solution, probably wildly off from my original thoughts. :/

This is the size transformer you need, and this one is near the impedance ratio you need.  You could get this and try it out on one channel to see if you get much improvement.

https://www.ebay.com/itm/334018573088?hash=item4dc50ef320:g:0yoAAOSwm2hgcKZd

[exe]

Btw, does the application needs a hi-fi transformer? I mean, it's an aviation helmet, I have doubts it was designed for listening to the music.

[dnyberg]

Btw, does the application needs a hi-fi transformer? I mean, it's an aviation helmet, I have doubts it was designed for listening to the music.

Quite right; anything better than voice grade would do. Nowdays we're spoiled in that we like to listen to decent stereo sources between bursts of radio traffic, but it's the radio that really matters, and that is most definitely not hi-fi.

Thanks for the help gang; you've given me some directions to explore.

[Terry Bites]

Its really unlikey that you power amp is going to complain about a higher load impedance. Unless its some tube madness.
You can just whack 15R resitors in parallel with your phones to make them look like 15R.
A long as the power amp has a reasonably low output impecance the volts across your phones wiil be nearly the same.

[Gyro]

Maybe something like this?

https://www.mouser.com/datasheet/2/294/njrc_s_a0009983567_1-2279521.pdf

Yes, that looks like a good choice, bridged output for maximum voltage swing and nice low quiescent current. The distortion is maybe a little high for HiFi, but probably perfectly acceptable for an aviation headset.

Just a note that bridged output requires individual access to both of the wires from each earpiece (no common pin). If that's a problem for a standard jack then 4 pin microphone connectors and 4 pin XLR connectors are commonly available.

The little TDA1308 based boards that I mentioned as an edit to my previous post is single ended output, so doesn't have any issues with a 3 pin headphone connection, but doesn't have as big an output voltage swing as your JRC part.

Just a though, are aviation headsets Mono anyway?

[Gyro]

Its really unlikey that you power amp is going to complain about a higher load impedance. Unless its some tube madness.
You can just whack 15R resitors in parallel with your phones to make them look like 15R.
A long as the power amp has a reasonably low output impecance the volts across your phones wiil be nearly the same.

The OP's problem is that the sound level is too low. Parallel resistors wont help that.