Can you determine impedance ratio of completely unknown transformer?

[b_force]

Please be aware that unless a special model, a regular Variac is an auto-transformer,so secondary is directly connected to live mains, so anything from dangerous to deadly.
Be very careful,or better yet, use a standard (isolating) 12V transformer.

An 1 to 1 isolating transformer before or after the VARIAC will also do.

That being said, as long as you just don't touch anything/stay (far) away while testing and use a quick switch (plus fuse) to turn everything off, there is not a lot that can go wrong.

Otherwise anything from very unconformable to potentially lethal yes.

 

[bananasplit_00]

An 1 to 1 isolating transformer before or after the VARIAC will also do.

That being said, as long as you just don't touch anything/stay (far) away while testing and use a quick switch (plus fuse) to turn everything off, there is not a lot that can go wrong.

Otherwise anything from very unconformable to potentially lethal yes.

I really don't think I'll be doing the variac method as I have a 0-25V AC source I could use instead. I'll probably just use a signal generator to injct a 1KHz sine into the secondary and seeing what comes out on the primary with either a scope or multimeter(just got to check that I grab a multimeter that's fine with 1kHz)

Impedance matching criteria, it is not marketing. For an output stage it guarantees maximum power transfer to the load.

However, you don’t always want to match impedance, for example, in a line out application you want to transfer the most voltage (signal information) to a load (the next stage) without voltage drops. This is called bridging impedance, you want the apparent output impedance to be much lower than the input impedance of the next stage.

So your example, 10k:600 would be a bridging impedance. You amplifier has output impedance of 10k which you want to lower to 600 so that the stage it connects to is transferred most signal. These line level transformers also provide isolation which negates ground loops.

Edit; forgot to answer the final part.

I think what you are saying is 10k:600 has the same ratio as 100k:6000 so what’s the point of the actual numbers?

The actual windings have a parasitic resistance and the certain magnetising inductance (low frequency) and leakage inductance (high frequency). The winding resistances cause insertion loss, the more turns the higher the inductances, which will start to shift your frequency response.

A lot of effort is put in to optimising all the various values to get the best frequency response possible for an audio transformer. As a part of this process you need to know what impedance will be connected on the primary and secondary. That’s why the actual values are specified.

Oh that's pretty interesting actually but seems like nothing is care much about seeing as I am already getting my amp components from salvage and trash 😂

 

[bananasplit_00]

So, measured up all of the suspects.

Found the primary and secondaries, connected a signal generator to the primary and a scope to the secondary and applied a 1Vpp 1KHz sine to the primary and read the amplitude on the scope and calculated it all out.

None of them would be any use what so ever for 8 ohm or 16 ohm applications but for 50 ohm and higher I found that two of them should be decent. Tried some different frequencies and the amplitude on the scope seemed fairly constant so they should be pretty alright! 😄

The ones I found were 21.7k:50 and 17k:50, don't know off the top of my head what those would be a good match for but I am pretty happy with finding those. Runner up to that would be 1.25k:50 and the rest worse ratios than that.

These transformers are also all quite small so a headphone amp seems like the perfect application to me! Might do some tests to see if they can handle a few W load and report back here but for now that's a pretty satisfying end to a thread 😀

 

[kodabmx]

They just transform.
So in theory, 21k7:50R == ~3k4:8R
Indeed maybe useful but small as you said.

 

[exlabs]

GREAT! Glad to hear you could make progress. With your setup, you could easily measure the bandwidth by varying the frequency and finding the 3dB drop on the low and high frequency side.

But the way, how did you figure out the secondary was aimed at 50r specifically? What is the winding resistance of the secondary?

 

[bananasplit_00]

GREAT! Glad to hear you could make progress. With your setup, you could easily measure the bandwidth by varying the frequency and finding the 3dB drop on the low and high frequency side.

But the way, how did you figure out the secondary was aimed at 50r specifically? What is the winding resistance of the secondary?

Well they could be anything I suppose, just calculated for 50 ohms as that's what I have for headphones and the numbers seemed pretty alright. Could definitely be 32 ohm too but I don't think it would be lower than that. Could be wrong of course, I have barely any experience with output transformers

As the transformers are only about 25mm(1") cubed I assume they would be for a headphone application or some smaller active speakers.

Could definitely do a more comprehensive frequency sweep next time, only tried 100Hz, 1kHz and 5kHz and saw no amplitude change and was happy with that seeing as these are scavenged from audio equipment but it would definitely be nice to actually know what they can do! Might as well set up a simple load test for the two good ones I found too to see what kind of power I could push

 

[martyh]

The other thing you want to consider is inductance of the primary. Using an output transformer designed for a 3K4 primary at 21K7 is not likely to work well.

 

[bananasplit_00]

The other thing you want to consider is inductance of the primary. Using an output transformer designed for a 3K4 primary at 21K7 is not likely to work well.

What sort of inductance am I looking for on the primary then? I thought the ratio was the most important part and that if you were chasing the best then inductance was more important? That's what I gathered from this thread at least obviously would like to use the one that matches the best for my case

 

[martyh]

In general, a hi-fi transformer designed for a higher impedance will have a higher inductance. It is hard to say what you might need because it will depend on how you want to use the transformer. Sorry, not trying to be vague. I just wanted to point out that more goes into it than the ratio.

 

[exlabs]

Ya. Start with an Ohm meter to map the windings. Then you need an AC stimulus, best into the high-Z windings so that you don't ~short your source and/or generate any high voltages. And best to use a middle frequency (~1KHz), avoiding inductance loading at the low end and ringing at the high end. Remember the impedance ration is the square of the voltage ratio. Note the inductance of the winding will be about the rated impedance at the bottom of the frequency range, ie L=2πFZ, or Z=L/2πF, (F~=20Hz) Better transformers have more inductance, ie the low frequency limit is lower.

Here is why the inductance is important. It determines the low frequency cutoff. (The high frequency is determined my the leakage inductance.)

 

[bananasplit_00]

Here is why the inductance is important. It determines the low frequency cutoff. (The high frequency is determined my the leakage inductance.)

In general, a hi-fi transformer designed for a higher impedance will have a higher inductance. It is hard to say what you might need because it will depend on how you want to use the transformer. Sorry, not trying to be vague. I just wanted to point out that more goes into it than the ratio.

Alright, the inductance of the primary of the one I labelled 21k7:50 is 3850mH, I am not really making sense of the formulas honestly so I don't really know what that would match well with.

The measured 3dB cutoffs are 40Hz and 45kHz at least, not great honestly but should still be usable?

 

[exlabs]

I think that’s pretty good! But actually, might be time to consider the impedance of the generator and load to get accurate values. I know I didn’t mention that before - I’ll actually want to understand better myself.

For example, have generator impedance around 21k and load the secondary with 50r resistor and repeat the measurements.

QUESTION: It would be great for someone with more experience to comment - would you measure a different frequency response with the two methods?

1) Signal generator (unknown impedance)-> primary (open circuit)

2) Signal generator with 21k impedance -> primary loaded with 50r

 

[Gnobuddy]

...why certain transformers are marked stuff like 10k:600 or similar if all that matters is the actual turns ratio...

The turns ratio is not the only thing that matters.

The actual self-inductance of the windings matters, too. So does the amount of "iron".

The inductive reactance of the windings appears in parallel with the windings, and limits the available frequency response. A transformer with 10 turns on the primary and 1 turn on the secondary won't work the same way as one with 1000 turns on the primary and 100 turns on the secondary.

The amount of "iron" also matters, because it determines how much power the transformer can handle without the core saturating. More iron in the windings also tends to increase the self-inductance, improving the low frequency response as well as the power handling.

-Gnobuddy

 

[bananasplit_00]

I think that’s pretty good! But actually, might be time to consider the impedance of the generator and load to get accurate values. I know I didn’t mention that before - I’ll actually want to understand better myself.

For example, have generator impedance around 21k and load the secondary with 50r resistor and repeat the measurements.

QUESTION: It would be great for someone with more experience to comment - would you measure a different frequency response with the two methods?

1) Signal generator (unknown impedance)-> primary (open circuit)

2) Signal generator with 21k impedance -> primary loaded with 50r

The signal generator is a Rigol DG812 model which specifies 50 ohm output impedance, the scope has a 10M input impedance if that helps, can't do any more meassureing for a few days as I'm going on a trip tomorrow

 

[exlabs]

Oh that’s quite fortunate. In that case, if you are interested, why not connect the signal generator to the SECONDARY, but a 21k load on the primary and measure the frequency response.

With these impedances being closer to the design impedance I wonder if you get different results? This is a bit of a grey area for me, would be very interested to see what happens.

Have a good trip!

 

[Gnobuddy]

...would you measure a different frequency response with the two methods?

1) Signal generator (unknown impedance)-> primary (open circuit)

2) Signal generator with 21k impedance -> primary loaded with 50r

Transformers are one area where I'm not 100% confident either. But as far as I know, yes, you would (get different frequency responses)!

If you drive one winding from a low source impedance, then you will get extended bass response (at least for small enough signals).

Imagine a transformer with a 10000 ohm load on the primary, and a primary inductance of 20 Henries. Let's say it's driven by a constant current source, which means the driving signal comes from a very high impedance.

(The anode of a pentode has very high impedance, so for our purposes, it acts more like a current source than like a voltage source. So the above assumption is not too different from driving the transformer from the anode ("plate") of a pentode.)

Now the bass response will drop by 3 dB at the frequency where the inductive reactance of the 20 H coil equals the 10k load, because the combined impedance of the 10k and 20H in parallel drops, and so does the voltage across it. This frequency works out to about 80 Hz. Good enough for an electric-guitar amplifier.

Now let's suppose the same transformer is driven from a nice Rigol function generator with a 50 ohm output impedance. Because of its low output impedance, the Rigol can keep the voltage across the transformer primary nearly constant until its reactance drops all the way down to 50 ohms (not 10000 ohms). It's only then that the voltage from the Rigol starts to drop.

Because we're now looking at 50 ohms rather than 10,000 ohms, the -3 dB frequency w ill be (10000/50) or 200 times lower than before! This works out to an astounding 0.4 Hz!

(This will only be true if we drive very, very little signal into the transformer for the test, because the iron will want to saturate at these very low frequencies).

Still, the main point is clear: if we drive the transformer from a low source impedance, we'll measure an extended (and unrealistic) bass response.

To get a realistic frequency response, the transformer must be driven from the same source impedance that it will see in the final circuit. The same goes for the load impedance (on the secondary side).

-Gnobuddy

 

[exlabs]

Thanks Gnobuddy,

Now the bass response will drop by 3 dB at the frequency where the inductive reactance of the 20 H coil equals the 10k load, because the combined impedance of the 10k and 20H in parallel drops, and so does the voltage across it. This frequency works out to about 80 Hz. Good enough for an electric-guitar amplifier.

This was really helpful way to think about it, thanks. 😄

So as the secondary is approx. 50r and the signal generator is 50r, if you put the signal into the secondary and place a 21k load in the primary, this would give quiet a realistic measurement of frequency response.

I know that the signal is going “backwards” but it’s doesn’t matter.

 

[bananasplit_00]

Thanks Gnobuddy,


This was really helpful way to think about it, thanks. 😄

So as the secondary is approx. 50r and the signal generator is 50r, if you put the signal into the secondary and place a 21k load in the primary, this would give quiet a realistic measurement of frequency response.

I know that the signal is going “backwards” but it’s doesn’t matter.

Back from my trip now, frequency response is a lot worse with the load. About 250-300Hz on the low end up too about 85kHz on the high end and starts looking pretty distorted at anything under 100Hz(though this gets quite a lot better at very low signal levels)... Also measured the impedance ratio to be more like "20k:50" with the load attached and calculated from the voltage ratio.

 

[exlabs]

Back from my trip now, frequency response is a lot worse with the load. About 250-300Hz on the low end up too about 85kHz on the high end and starts looking pretty distorted at anything under 100Hz(though this gets quite a lot better at very low signal levels)... Also measured the impedance ratio to be more like "20k:50" with the load attached and calculated from the voltage ratio

So, you put the signal into the secondary? What voltage? The low end sound like core saturation, give everything you said, about it improving with reducing voltage.

 

[bananasplit_00]

Another thing that points to this transformer I have been measuring being for headphone use is that it can't take even 0.13W for one second before the core starts making melting sounds 😅 I first set up the test at 1/2W and some wax(???) started boiling out of it the instant the power was switched on...

So, you put the signal into the secondary? What voltage? The low end sound like core saturation, give everything you said, about it improving with reducing voltage.

I did most of the testing with 1Vpp into the secondary so 20V out of the primary, I also agree that this sounds like core saturation which is not really that surprising seeing as it is about one inch in all dimensions I suppose 😄

At 40Hz the signlal input needs to be about 40mV for it to not be a triangle wave, at 100Hz it's more like 180-200mV. Either way that's practically 0W into 50ohm headphones