My observations:
- The right side triode section of the 12AX7 has both its cathode and anode connected to ground, so it doesn't do anything.
- The current through the left side triode section of the 12AX7 flows from ground through R8 (10K) to the anode, and than from the cathode through the string of zeners back to the negative terminal of the 170 V power supply. I think the amount of current is adjustable with the 47K pot but I don't see how a change of current through the left side triode section changes the grid voltage of the tube under test. As far as I see the grid voltage can only be adjusted with the 22K pot, but for that the string of zeners and the 12AX7 with associated circuitry is not needed.
- The 12B4A's are passing about 50 mA each at Va = 50 V and a bias voltage of about -1.7 V. I would feel very uncomfortable with such a low bias voltage (- 1.7 V is very near the starting point of grid current; grid current would disturb the funtioning of the 6267). Furthermore the datasheets for the 12B4A state a maximum average cathode current of only 30 mA.
- The voltage drop over the 470K anode resistor of the 6267 is only 1.7 V, so the anode current of the 6267 is very small. I'm not 100 % sure if it is too small but I would think it is. Adjusting the screen grid voltage of the 6267 will probaly be a very tricky thing to do. If I'm not mistaken, a tiny change of the screen grid voltage will give a relatively large change of the voltage drop over the 470K anode resistor.
Last edited:
As you rightly noted, the grid circuit had something too much. In this last configuration it should give 1 to 55v negative, enough to regulate most tubes. As for the anode and screen power supply, I had replaced the voltage regulator because I only have 9-pin sockets available and because I wished to do everything with tubes that I already have, but obviously the 12B4As are not sufficient for their function. I need to figure out what else to replace the 6Y6s with. The circuit is taken, with very small variations, from Lambda 25. The zener snake wants to simulate the work of a 5651 which is not present in the Microcap options. Screen voltage can be adjusted from about 85 to 150v.
Attachments
I've redrawn the supply for g1 to try to understand how it works. But to be honest, I still don't understand the workings and the purpose of the circuit. What puzzles me the most is that the control grid of the right side triode is positive with respect to its cathode since for current to be able to flow through the 220K resistor and the left side triode, its anode has to be positive with respect to its cathode and the lower side of the 220K resistor.
I wonder if the supply for g1 needs to be regulated (if that is the intended purpose of the circuit). The current draw of g1 of the tube under test is practically zero. So I would think that the job could be done with a voltage source, a multi-turn potentiometer, and some extra resistors. One resistor for shunting the pot so to make the current through the pot small, one (switchable?) resistor between the positive terminal of the voltage source and the pot to prevent the bias becoming 0 V (which you probably don't want when measuring mutual conductance of power tubes), and one large valued resistor from the wiper of the pot to the negative side of the pot to insure that there still is negative bias in case the slider of the pot makes bad or no contact inside the pot.
I wonder if the supply for g1 needs to be regulated (if that is the intended purpose of the circuit). The current draw of g1 of the tube under test is practically zero. So I would think that the job could be done with a voltage source, a multi-turn potentiometer, and some extra resistors. One resistor for shunting the pot so to make the current through the pot small, one (switchable?) resistor between the positive terminal of the voltage source and the pot to prevent the bias becoming 0 V (which you probably don't want when measuring mutual conductance of power tubes), and one large valued resistor from the wiper of the pot to the negative side of the pot to insure that there still is negative bias in case the slider of the pot makes bad or no contact inside the pot.
Last edited:
Through the years I gathered a lot of data/schematics/pictures on tube testers. Some months ago I uploaded all to Google Drive. It can be found here:
Tube testers
Tube testers
The general scheme is this (I can no longer find the other drawing with the sizes and types of components). If you have a better idea, maybe even a simpler one, feel free to post it. The initial idea was to have something simple, a bench with which to test the tubes and compatible in order to power a simple SE with a maximum of 100mA if necessary. I have a metal platform taken from an old jukebox perfect for this job because has a lot of sockets and the 3 most used sockets with which to test 90% of the main tubes are available too.
Your schematic shows the principle of a common type of regulated power supply. The power supply in the attached schematic of an active crossover I built is basically the same. The only real difference is that the voltage reference (the 51 V zener diode) is being fed from a point before the pass tube (the triode connected beam power part of the ECL82) while in your schematic the voltage reference (V3) is being fed from a point after the pass tube (V1). But this difference doesn't change the way the circuits work. Note that the filament supply in my schematic is 'lifted' to 71 Vdc to stay well within the limits for cathode to heater voltages of the tubes. Note also that my schematic only shows one of the two channels (the channels share the power supply).
What other tubes beside the 12AX7 do you have at hand? A triode section of the 12AX7 is fine for the role of the control amplifier (V2 in your schematic) but not such a good choice for the role as the pass tube (better are tubes with highish mutual conductance and lowish internal resistance).
What other tubes beside the 12AX7 do you have at hand? A triode section of the 12AX7 is fine for the role of the control amplifier (V2 in your schematic) but not such a good choice for the role as the pass tube (better are tubes with highish mutual conductance and lowish internal resistance).
I'm starting to understand a little more about power supply ... I see you take it in a different place ... what do you provide with that cross-over?
The blanket is always short when almost exclusively recovery components are used ... for example I don't have a power supply that is powerful enough to deliver more than 290v DC with 100mA of load to test a pretentious EL34 ... plus, I only have available 9-pins sockets to use and, among the power tubes I have of that kind, I only have some EL84s and, more massive, only 1 6GM5 which alone is not enough. I would also have 6080s as well as 6Y6s that can handle those loads but they are octal.
The problem of the second triode does not arise since, as you pointed out to me, it is practically uninfluential.
To simplify even further, even the screen circuit may be unnecessary since almost all tubes have measurements in datasheet with the same voltage on plate and screen.
Attachments
You could keep an eye out for a (broken) big tube oscilloscope (Tektronix, Telequipment, Hewlett-Packard, etc.). Many of them have (eleborate) stabilized power supplies so their power transformers are perfect for making your own stabilized power supply. Some 10 years ago I was surprised about how cheap they were (especially in non working order) but I don't know their current prices.
Finding or making a stable HV power supply is always the trickier part, however it is helpful to remember that if you are testing pentodes, only G2 and G1 supplies need to regulated.
For testing triodes a regulated plate supply is needed.
When I test, I use a dual 0-30V lab supply for heater and G1, and a pair of separate adjustable HV supplies for G2 and plate.
To measure VG/Ia characteristic, have 10 ohm precision resistor in series with plate supply, measure voltage across it, and derive plate current from Ohm's law.
To measure Gm, inject exactly 1V RMS at say 400Hz to G1 via coupling capacitor, and measure AC voltage across that 10 ohm resistor in plate circuit.
Example measurement of EL84/6P14P: Va 300V VG2 250V VG1 -7.3V (as per data manual), get 46.2mA Ia and Gm 11.2 mA/V.
If making up test circuit for this, remember to add 1kohm grid stoppers on G1 and G2, monitor G1 with a spare DMM so bias stays correct, and AC signal exactly exactly 1V (makes the math to derive Gm easy). Testing triodes is same, just no G2 supply needed.
PS Thank you PFL200 for tube data - very interesting
For testing triodes a regulated plate supply is needed.
When I test, I use a dual 0-30V lab supply for heater and G1, and a pair of separate adjustable HV supplies for G2 and plate.
To measure VG/Ia characteristic, have 10 ohm precision resistor in series with plate supply, measure voltage across it, and derive plate current from Ohm's law.
To measure Gm, inject exactly 1V RMS at say 400Hz to G1 via coupling capacitor, and measure AC voltage across that 10 ohm resistor in plate circuit.
Example measurement of EL84/6P14P: Va 300V VG2 250V VG1 -7.3V (as per data manual), get 46.2mA Ia and Gm 11.2 mA/V.
If making up test circuit for this, remember to add 1kohm grid stoppers on G1 and G2, monitor G1 with a spare DMM so bias stays correct, and AC signal exactly exactly 1V (makes the math to derive Gm easy). Testing triodes is same, just no G2 supply needed.
PS Thank you PFL200 for tube data - very interesting
I had found on the net a circuit very similar to the one you described which is certainly much more professional, but I discarded it because I don't have a dual power supply and I was hoping to be able to do without it like this ... do you think mine does not work well?
I have read the thread again, and it appear that you have limitations of parts etc that can use, I agree with making best of what you have.
Perhaps my comment was a little off topic, more about technique of testing.
Since the control grid does not draw current, maybe keep it simple with potentionometer across negative voltage reference? Good luck with your project
Perhaps my comment was a little off topic, more about technique of testing.
Since the control grid does not draw current, maybe keep it simple with potentionometer across negative voltage reference? Good luck with your project
About the revised circuit for the control grid supply in post #10: The triode section of the 12AX7 now functions only like a diode (which has some internal resistance) since the control grid is connected to the plate (through R8 of 10K), making the control grid very positive with respect to its cathode. It doesn't regulate in any way, and the diode function is unnecessary since the voltage source already supplies a DC voltage.
TKS! It was not off topic, on the contrary you gave a good contribution to the project and you gave some excellent suggestions for those who wanted to build one. If you have drawn it, it might be even more useful.
Maybe I could just use one octal socket for testing and use the other for a 6080 which would solve the power supply problem, wouldn't push the pot to the limit to get 250v and would give me a few more mA if I wanted to use it to power another circuit.
Attachments
The screen grid of the 6267 in your schematic is connected to its cathode. The screen grid should be positive with respect to the cathode (like it was in your first schematic).
You are welcome, meanwhile I made a quick sketch (and got my scanner working), may be of use
Warning: Do not test an EL84 at Va = 300 V, Vg2 = 250 V and Vg1 = -7.3 V. Under these conditions the anode dissipation would be about 15 Watt (P = I x V = 0.05 x 300 = 15) while the maximum is 12 Watt.