DIY tube testing Curve Tracer

[whitelabrat]

From what I've been reading, most tube testers don't accurately determine the condition of a tube. As well a proper test would involve tracing a tube's curve and likely require higher voltages than what most testers are able to provide.

Let's say that an old 350-0-350 tube amp transformer where handy, along with a variac, and a multimeter. Could these be used to determine a curve? Of course a DC supply would be needed. In the interest of being simple and inexpensive, what are the bare necessities for some sort of crude but accurate testing?

I would figure that some sort of signal would need to be passed through the tube.

[soulmerchant]

Hmmm. you have just about described the set-up I am using to get some values on a few TeeeVee toobs...

Its tough going though...

I measure voltage gain, and current flow for a given signal and operating point like 150V...

But then if you change the signal from something like 1khz to 10khz do you get the same result? What about at 20khz or lower values like 100hz...

Then you repeat all of that for 1mv, 2mv, 5mv, etc signal strength.....

Then try a different operating point like 170V, 200V, 250V...

Then you try a different tube, because they are all a *little bit* different... And try not to fry them or your power supply or your DMM...

All of the sudden it is a lot of work.

Oh yeah... what do you use for Rk when you have nearly no values to go on?

[Dave Cigna]

Seems to me that if you want to determine the condition of a tube then you need to answer two questions:

1. What changes as the tube wears out?

2. How do I measure that?

I'm not sure what tracing curves has to do with either of those questions....


Anyway, you can display curves on your scope, *one* grid voltage at a time, with the circuit shown here. The AC supply could be your 350V transformer on a variac. The plate voltage will be negative half of the time, but so what? The DC grid supply needs to be especially clean or it will spoil the curves

Set your scope to X-Y mode. Connect the X channel to the plate and set it to AC coupling. Connect the Y channel to the top of the 1 ohm current sense resistor. 1 ohm means 1 volt per amp, so if you have your Y axis set to 10mV per division then you can read it as 10mA per division. You might have to set that axis to invert.

With both supplies off you'll get just a dot. Position it somewhere convenient in the lower left corner of the display, then turn up the supplies and amaze your friends.

-- Dave

[soulmerchant]

it helps a lot to use a 'scope. I picked an old one up that was still very serviceable for about $80 from my local electronics supply...

It would be nice to have one of those new fancy tektronix jobs that you can hook up to a pc...

[resident]

Dave,
thanks for this simple curve tracer! I'll try it!

[wa2ise]

Originally posted by Dave Cigna

1. What changes as the tube wears out?

2. How do I measure that?

Anyway, you can display curves on your scope, *one* grid voltage at a time, with the circuit shown here. The AC supply could be your 350V transformer on a variac. The plate voltage will be negative half of the time, but so what? The DC grid supply needs to be especially clean or it will spoil the curves

Set your scope to X-Y mode. Connect the X channel to the plate and set it to AC coupling. Connect the Y channel to the top of the 1 ohm current sense resistor. 1 ohm means 1 volt per amp, so if you have your Y axis set to 10mV per division then you can read it as 10mA per division. You might have to set that axis to invert.

With both supplies off you'll get just a dot. Position it somewhere convenient in the lower left corner of the display, then turn up the supplies and amaze your friends.

-- Dave

The main thing that wears out in a tube is cathode emission of electrons. Thus most tube testers only measure cathode emission, and you compare this measurement to a table in the tube tester's manual for that particular tube type. Most everything else in a tube is dictated by the physical spacing of grids, plates and so on, and does not change as the tube ages. Or change so slightly as to be ignored. But an emission tester doesn't tell you much about an unknown tube (say one where the markings were washed away). Or, back in the day, characteristics of a new tube in a vacuum tube development lab. Then you'd need the curve tracers and gm measurers. There were special curve tracer oscilloscopes built for this kind of work. You'd characterize several tubes to get the curve diagrams that you are going to place in the data sheet of the new tube. I used to do stuff similar to this with analog semiconductor chips.

----------

The grid voltage supply should be returned to the cathode, so the voltage due to the current thru the cathode resistor doesn't confuse things. Also, you could use a bridge rectifier (or a full wave rectifier if the high voltage secondary has a centertap connected to ground. This would eliminate the negative voltage swings on the plate and make them positive so you get a brighter display on the scope. This because otherwise the plate curve traces are at zero half the time because of the time spent with the negative swings.

Even better if you can build a stairstep waveform generator that creates 0V, -1V, -2V, and so on to feed to the grid. And use a camera to capture the oscilloscope display to get the data you see in tube manuals.