Easy ways of measuring output impedance of tube stages

[stigla]

Hi all!

Going to start a tread to shed some light about measuring the dynamic output impedance of tube stages.

The general idea is to get you guys to teach me this, since I have no clue of how to do it, just vague ideas. (AC measurments?)

There are of course ways to calculate this, but sometimes it's nice to have a more precise (and real) number to work with (with RIAA's for example.)

So how's it done? What gear is needed? Is there any nice "tricks" to easily do this? How accurate can I get it?

[SY]

It's exceptionally easy. Make a wild estimate of the output Z. Let's say it's 10K ohms. Choose a resistor of 4-5 times greater, say 47K ohms in this example, as a load resistor. Inject a signal into the input and measure the output voltage without the load attached. Now attach the load resistor and measure the new output voltage. Use the voltage divider equation to solve for the output Z.

You can also do this by injecting a signal into the output of the stage, but that can be a bit trickier.

[Tom Bavis]

And if you don't like doin' math... measure output voltage, connect a variable resistor to output and adjust resistance until out drops in half. Disconnect the pot and measure it - there's you're output impedance.

[stigla]

Thanks for the input!

Hmmm, that was really more easy than I had imagined!

One question more: Will the output Z be the same for all frequencies? Does it matter what frequency I make the measurments on? My guess is that this will of course vary from one type of tube-stage to another, but It would be nice to do measurments under 1K since I can use only a multimeter then...

stigla

[EC8010]

No, it won't necessarily be the same at all frequencies. The most usual reason is that the coupling capacitor in series with the output causes the impedance to rise at low frequencies. If the amplifier is transformer coupled, then there's a whole selection of reasons for changing impedance with frequency. Amplifiers with feedback are also likely to have a non-constant output impedance.

I use the variable resistor method to measure output impedance, and temporarily fit a much larger coupling capacitor (>3uF), but you need to be careful about level. Stages like mu-followers have low output resistance, but it doesn't mean they can drive any significant voltage into a load of equal value. I monitor distortion at the same time. If distortion <1% into the load, then the output impedance measurement is likely to be valid.

Unless your DVM is very posh, it will be even worse accuracy at 1kHz than it is at 50Hz or 60Hz (and they're none too hot there). An oscilloscope or dedicated AC test set is a better bet. Using an oscilloscope, if you can't see distortion on the waveform, then the measurement is probably valid.

Old electronic saying, "If you don't know what the answer from the measurement should be, you shouldn't be making it."