Measurements on cheap "inline" mic transformer

[NewYorkDave]

Just for fun... During my lunch break today, I disassembled one of those cheap "inline" mic transformers--the kind you can buy for 10 or 12 bucks--and took some basic measurements. The following were measured at -50dBU input (except where noted) from a 200-ohm source impedance, with the transformer secondary loaded by 500K shunted by about 100pF to 120pF of test instrument/cable capacitance.

Advertised impedance ratio: 200:50K (1:15.8 turns ratio)
Measured voltage gain: 19dB
Freq. response (+0/-3dB): 27Hz - 36kHz
(+0/-1dB): 50Hz - 25kHz

THD
20Hz: 2.5%
27Hz: 2%
1kHz: 1.5%
20kHz: 1%
The THD figures include the residual noise of the millivoltmeter amplifier and the distortion analyzer. The 1% figure at 20kHz is almost entirely residual noise. A lot of gain was required of the millivoltmeter amplifier just to bring the signal level up to what the distortion analyzer would handle. The distortion analyzer is an ancient Heathkit, certainly contributing some noise of its own.

Since distortion is a problem at low levels and low frequencies with the cheap steel cores used in these transformers, I measured at -50dBU to get more of a "worst case" measurement. Low frequency distortion does improve as you increase level, up to a point (see below).

Maximum input:
(For this, I just cranked the input until the output waveform became noticably "bent", then took a level reading).
20Hz: -12dBU
27Hz: -8dBU
35Hz: -6dBU
50Hz: -2dBU

Square wave performance at 1kHz: Not bad at all, just a slight tilt and overshoot. The overshoot increases if you reduce the capacitance shunting the secondary, of course.

 

[chrissugar]

Who make this transformer, Dave?

chrissugar

 

[Michael_Joly]

Do you think these things could be used an intentional "dirty up" device as was discussed in another thread - DC biasing to saturate the core etc?

 

[NewYorkDave]

Chris, it's a Hosa transformer. But honestly, I suspect that most transformers of this type are cranked out of the same Chinese factory, regardless of whose name appears on the label.

 

[chrissugar]

Thank you.

chrissugar

 

[tk@halmi]

I have several of these, hacked out of different impedance matching units. Invariably, they sound nasty to me. How about using them backwards for DI?

 

[PRR]

> Advertised impedance ratio: 200:50K (1:15.8 turns ratio)
> Measured voltage gain: 19dB

The measured voltage ratio says 1:8.9 or 1:9 turns ratio.

The advertised impedances are logically mis-matched: use a 200Ω mike and a 50K input and you will get good results. You will get around 630Ω to the mike, which is good for 200Ω mikes.

You say yours is marked Hosa. I agree that they probably all come from the same factory. But the XLR-1/4" lo/hi-Z transformers I got from Radio Shack and Audio Technica some years back were clearly 1:7 ratio, not 1:9 or 1:10 as yours seem to be.

Your bent-bass numbers look like mine, which is logical if they are the same core and the same primary winding.

I'm not sure I can accept your THD numbers. Even with 1% residual, 1.5% at 1KHz seems high for any iron at -50dBm.

Are you really seeing NO peaking at the top of the band? That seems odd for a 1:10 or really 20K secondary winding without much loading. Might have high copper resistance or really lossy iron.

 

[NewYorkDave]

The measured voltage ratio says 1:8.9 or 1:9 turns ratio.

Yep, and the loading on the secondary was light enough that I don't think we can attribute it to copper loss. So, despite what it says on the label, it's really a 200:16K (or thereabouts) transformer.

I'm not sure I can accept your THD numbers. Even with 1% residual, 1.5% at 1KHz seems high for any iron at -50dBm.

Actually, you shouldn't, and I'm glad you brought it up because I now realize that I made a stupid oversight when going from the freq. response measurement to the THD measurement. I forgot to change my signal source. I had been using a function generator (THD: about 0.5%) for frequency response, because of its easy sweepability, and forgot to switch over to my low-distortion HP sine oscillator for THD measurements. So those THD figures are meaningless. I'll re-do that test if I have the time.

Are you really seeing NO peaking at the top of the band?

Really none to speak of, maybe 0.5dB at worst. But that's with the secondary shunted by 100+ pF of capacitance. If I reduce the capacitance (say, by switching my scope probe to X10) then yes, there is more peaking.