Any way to extend OPT low-frequency response?

[bst]

I found an interesting pair of audio transformers at my local surplus store. They are UTC KR-652 potted transformers, very nicely made, and bear the following legend:

UTC KR-652
1,2,3 30 Ohms center-tapped
2A PK BAL D.C.
4,5 3000 Ohms 160 MADC
300 - 20,000 cps
30W MAX WV 535PK
5 MIN ON 10 MIN OFF

I thought I'd found some neat output transformers for a headphone amp, until I saw the frequency response. Ughhh!

Is there any way to extend the low-frequency response of a transformer, perhaps by externally adding to the primary inductance?

Or should I stuff these under the bench and quit haunting the industrial/military surplus warehouses?

[dhaen]

Well, the lower frequency is set by the primary inductance (which you could measure) and the source impedance ie the stage driving it, so the limit is not set in stone, but just "as measured" in the test or application circuit. If you feed it with a low Z source - it'll be better.

[Eli Duttman]

A small amount of loop NFB will help. Don't go "hog wild". Too much NFB will saturate the trafo cores.

A combination of low impedance drive, limited power handling, and loop NFB will improve L/F extension.

Parafeed topology might help too, as there is zero DC in the primary. Hmmm, CCS loaded ECC99 cathode followers driving the cap. coupled trafos is a thought.

[bst]

dhaen, thanks for your reply.

Please bear with a neophyte's basic questions. I'm just getting started in this; if someone would have mentioned Kirchoff's Law to me a few months ago, I would've thought they were talking about a new television series.

Will the frequency response be wider with less power through the transformer? If used for a headphone amp, these might see 250 volts and 350 mA, far less than their maximum ratings.

Also, I don't know if the rated response is at 0 db. It's possible that it might extend quite a bit further at 2 or 3 db down. Is there a simple way to plot this with only basic test equipment? I have an ancient RCA audio oscillator, a 20Mhz scope, and a couple of DMMs.

[dhaen]

bst,

We're all learning

Will the frequency response be wider with less power through the transformer?

It might be, but don't depend on it.

Will the frequency response be wider with less power through the transformer? If used for a headphone amp, these might see 250 volts and 350 mA, far less than their maximum ratings.

350mA? A low impedance driver will tend to be low voltage / high current but nothing of this order. Think 1/10th of this and half the voltage.

Is there a simple way to plot this with only basic test equipment? I have an ancient RCA audio oscillator, a 20Mhz scope, and a couple of DMMs.

Yes. Firstly make a calibration chart by feeding the osc to the dvm at various frequencies. Use this chart to correct for equipment inaccuracies. Note that some DMM AC frequency responses are better than others.
Determine the output impedance of the oscillator. If you are lucky it will be low: 600 ohm or so.
Use this to feed the transformer primary and measure the secondary AC - there may be a problem getting enough signal level to measure though, as the voltage will be reduced by the root of the impedance ratio. If you have an amplifier you could remedy this.
Another area where measurement inaccuracy can creep in is the wave-shape at low frequency which your meter might mis-read. Not much can be done about this.
This will just give you a rough idea...

Listen to what Eli says. Feedback is your friend here.

[bst]

Thanks to all for your responses.

350mA? A low impedance driver will tend to be low voltage / high current but nothing of this order. Think 1/10th of this and half the voltage.

I knew I shouldn't have posted until after the second cup of coffee - this should've read 350 mW, not 350 mA.

I'll see if I can generate a rough curve of these transformers using your suggestions. It might be that I'm getting ahead of myself, but
it's about time to string together a design of my own, rather than just copying someone else's.

I've been building from other peoples' schematics, using them to better understand biasing and load lines, coupling techniques, effects of bypass caps, and so on.

If I can lash up a simple amp that actually makes music instead of blue smoke, it'll be a satisfying step forward.

Thanks again for your help.

[Jeb-D.]

The less DC current through it, the better the low end. Also you can try putting a circuit in your amp design that boost the LF's higher.

[tubelab.com]

Most transformers have better low frequency response at lower powers. Since you want to run this at way below its rated power, you might get lucky. Frequency response is also limited by the inductance.

It works the other way too. A transformer that is rated for 50 Watts at 20 Hz can pass over 100 Watts in a guitar amp where the lowest frequency is 88Hz.