What would be the simplest way to check Gm of various tubes without a tube tester? I am thinking of just setting up a quick and dirty jig. Don't really want to get fancy with opamps, A/D, and meter displays.
I have three multimeters, three oscilloscopes, audio xfmrs for isolation, and function generator. Have no problem setting up my conditions to match actual operating conditions in my amps.
Maybe measure mu, too, since I'm at it --
Thanks !
I have three multimeters, three oscilloscopes, audio xfmrs for isolation, and function generator. Have no problem setting up my conditions to match actual operating conditions in my amps.
Maybe measure mu, too, since I'm at it --
Thanks !
You can do this in-circuit. Put one meter, set to AC volts, on the grid (referenced to ground). Insert another meter, set to AC milliamps, into the plate circuit right 'above' the plate. Supply the grid with a sine signal that is well within its capabilities. For a front end tube, say 0.1Vac; for an output tube, 1Vac. Measure the AC milliamps on the plate meter and divide out. gm is specified in mA / V.
Poinz
AudioTropic
Poinz
AudioTropic
To properly measure valves, it is paramount to regulate all voltages.... Without regulation the data is eroneous.....
95% of tube testers old and new are useless as far as I am concerned...
WIth tetrodes and and pentodes the screen voltage needs to be locked down tight with regulator... The plate should be regulated but can get away with no reg on the plate under the conditions of small signal testing and that you are in the satuaration region of curves.... Then you will use a AC mA meter in series with the plate to read gm...
Another problem I see is many testers don't regulate the heaters... This can be disaster when pluging in various number of tubes in the tester, thus a load change and input line variation....
A tube's heater is also not consistent from valve to valve of the same type... You can choose to voltage reg the heaters and you will see variation of the heater current....or you can choose to use a current regulator and then you will see variation of the heater voltage developed...
Keep in mind that gm will move about based on the operating conditions the valve sees...
So if you are operating the output valves in Class AB1, then you will have a dillema in testing... If you use small signal testing as is typicall, then the test will be at same conditions at amplifier idle voltages and bias, thus a Class A test is conducted with sine wave injection at input grid.... But the valve in the amplifier will swing into cut-off durring large signal excursion... This is where averaging the gm comes in.... Your best off testing the valve in a Class A condition which is approx 1/2 the operating conditions of the amp...this small signal gm WILL be the average gm you would get at full operting conditions...
cerrem
95% of tube testers old and new are useless as far as I am concerned...
WIth tetrodes and and pentodes the screen voltage needs to be locked down tight with regulator... The plate should be regulated but can get away with no reg on the plate under the conditions of small signal testing and that you are in the satuaration region of curves.... Then you will use a AC mA meter in series with the plate to read gm...
Another problem I see is many testers don't regulate the heaters... This can be disaster when pluging in various number of tubes in the tester, thus a load change and input line variation....
A tube's heater is also not consistent from valve to valve of the same type... You can choose to voltage reg the heaters and you will see variation of the heater current....or you can choose to use a current regulator and then you will see variation of the heater voltage developed...
Keep in mind that gm will move about based on the operating conditions the valve sees...
So if you are operating the output valves in Class AB1, then you will have a dillema in testing... If you use small signal testing as is typicall, then the test will be at same conditions at amplifier idle voltages and bias, thus a Class A test is conducted with sine wave injection at input grid.... But the valve in the amplifier will swing into cut-off durring large signal excursion... This is where averaging the gm comes in.... Your best off testing the valve in a Class A condition which is approx 1/2 the operating conditions of the amp...this small signal gm WILL be the average gm you would get at full operting conditions...
cerrem
How about using a 1 or 10 ohm resistor, and measuring voltage drop? Measurement of AC voltage is easier than AC current, and typically more accurate.
Am I correct in assuming Gm is (within reason) independent of load ? That is, plate voltage and current determines Gm, not the particular load ?
Am I correct in assuming Gm is (within reason) independent of load ? That is, plate voltage and current determines Gm, not the particular load ?
zigzagflux said:
Yes Zigzag, you can measure AC voltage across a 10 resistor. For a Gm of 1000, 1 ma of AC current must flow. (referenced to 1v AC on grid) A 10 ohm resistor would provide 10 millivolts across it. You will need a good sensitive AC voltmeter to accurately measure that. That's what I do on my rack setup for transmitting tubes. The resistor is on the bottom (negative end) of the B+ supply just before ground. I use an HP 3400A TRMS meter in case the sine is distorted. (seldom is) An average responding meter would also serve ok.
And yes, Gm is a function of current through the tube. Not plate voltage or load. My load is the PS itself, nothing else. Of course, grid, screen and plate voltages will affect tube current, and so the Gm.
Measuring AC voltage (at 15kHz) across a 10 Ohm resistor is exactly how the AVO VCM163 valve tester determines gm.
The first will work. The second needs grid bias or a capacitor across that cathode resistor. Also, you're going to need a transformer coupled via a capacitor to extract your (very small) audio signal without blowing up your measurement electronics.
I might be a little confused. If my source is DC coupled, I would not need the grid leak, correct ? If AC coupled, it is needed. If it is transformer coupled (probably what I will do) no grid leak required.
If I use a CCS on top, RL is infinite, and the capacitor bypass across the cathode resistor should not be required ? Can't hurt, of course, just thought it shouldn't be needed.
Very good point about the capacitor for AC coupling the output signal. If using a Fluke multimeter, it shouldn't be necessary, but if tying into the scope, this is a very good idea.
If I use a CCS on top, RL is infinite, and the capacitor bypass across the cathode resistor should not be required ? Can't hurt, of course, just thought it shouldn't be needed.
Very good point about the capacitor for AC coupling the output signal. If using a Fluke multimeter, it shouldn't be necessary, but if tying into the scope, this is a very good idea.
For whatever reason, I've seen anomalously high distortion when I use a CCS with an unbypassed cathode resistor.
One of these days when I have time, I'll dig into it. It may just be the higher effective rp of the tube fighting the CCS, but some experiments will tell all. It did light the "Aha!" light in my head as to why I was not happy with CCS loading when I first tried it (late '70s in a PAS3 mod) but found it to be an excellent technique when coupled with LED bias..
I don't remember if I saw higher distortion with an unbypassed cathode resistor and CCS load, but you certainly see more power supply noise getting through due to the increased ra.
Completed
Project completed.
All spare parts except for the choke and B+ capacitor. I knew those variacs had a calling in life 🙂 Left variac is for heating up the tube (0-6.3V, 0-4A), right variac is for setting B+, which is good for about 600V at 100mA. Can test 300B, 46, EL34, 6N6P, and 12A*7 variants. So it covers all my existing equipment so far, at the operating points used.
Seems to work pretty well, though the accuracy is possibly dubious. In the end, I'm more concerned about verifying a good/bad tube, and finding matched characteristics, so I'm not overly concerned about accuracy. Maybe on a rainy day I'll pursue some real investigations as to accuracy.
Currently going through my stash (couple hundred I think) and numbering/documenting.
Thanks for the assistance !!
Project completed.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
All spare parts except for the choke and B+ capacitor. I knew those variacs had a calling in life 🙂 Left variac is for heating up the tube (0-6.3V, 0-4A), right variac is for setting B+, which is good for about 600V at 100mA. Can test 300B, 46, EL34, 6N6P, and 12A*7 variants. So it covers all my existing equipment so far, at the operating points used.
Seems to work pretty well, though the accuracy is possibly dubious. In the end, I'm more concerned about verifying a good/bad tube, and finding matched characteristics, so I'm not overly concerned about accuracy. Maybe on a rainy day I'll pursue some real investigations as to accuracy.
Currently going through my stash (couple hundred I think) and numbering/documenting.
Thanks for the assistance !!
Cool! But with reference to the original thread title, it sure looks like a purpose built tube tester to me.
Hi, Gentleman:
Does anyone can tell me how to construct a tube tester, ( simple one ) it should test GM, IP/EG, and how to connect the tube socket. I only want to test power tubes as 6v6, 6l6, 6aq5, 6bq5 etc. Thanks
Does anyone can tell me how to construct a tube tester, ( simple one ) it should test GM, IP/EG, and how to connect the tube socket. I only want to test power tubes as 6v6, 6l6, 6aq5, 6bq5 etc. Thanks
how about posting the schematic? I'd like to build one and compare the readings to my transconductance tube testers.
Can't recall how I got each number in here, but they were specific to my scope input and probes.
Power supply is a bunch of surplus parts I had hanging around, so you could throw just about anything together that has reasonably low ripple and provides the v/i you need. I had full wave bridge - 0.33uF - 20H - 50 uF for the power supply. I would not hesitate to overload parts, either, such as the
transformers and choke; they are used intermittently for the test.
It really works quite well for matching tubes, I have to say. My results, while maybe not extremely accurate, are fairly well repeatable, so on a relative basis I can be assured of good matching where matching is required.
Power supply is a bunch of surplus parts I had hanging around, so you could throw just about anything together that has reasonably low ripple and provides the v/i you need. I had full wave bridge - 0.33uF - 20H - 50 uF for the power supply. I would not hesitate to overload parts, either, such as the
transformers and choke; they are used intermittently for the test.
It really works quite well for matching tubes, I have to say. My results, while maybe not extremely accurate, are fairly well repeatable, so on a relative basis I can be assured of good matching where matching is required.
- Status
- Not open for further replies.