I need to get some 6GW8/ECL86 tubes, but it seems all of the ones I am buying are out of spec, at least the pentode is.
I have a +600V supply, 47K plate resistor, and a -30V adjustable supply for the control-grid voltage.
According to the datasheets, I should have the control-grid voltage around -7V to get about +300V on the plate. But with almost all of the tubes I've bought I have to set it down around -30V and sometimes that's not even enough. I've bought 12 tubes all from major dealers, supposedly NOS, marked GE Made in Great Britain, Sylvania Made in Great Britain, and Amperex Made In Holland. 10 of the 12 tubes are out of spec. A friend has some recently purchased Westinghouse labeled and all of them were in spec, but the supply has dried up [came from same dealer as my Amperex].
Has anyone worked with these tubes and can offer any advice?
I have a +600V supply, 47K plate resistor, and a -30V adjustable supply for the control-grid voltage.
According to the datasheets, I should have the control-grid voltage around -7V to get about +300V on the plate. But with almost all of the tubes I've bought I have to set it down around -30V and sometimes that's not even enough. I've bought 12 tubes all from major dealers, supposedly NOS, marked GE Made in Great Britain, Sylvania Made in Great Britain, and Amperex Made In Holland. 10 of the 12 tubes are out of spec. A friend has some recently purchased Westinghouse labeled and all of them were in spec, but the supply has dried up [came from same dealer as my Amperex].
Has anyone worked with these tubes and can offer any advice?
> Where is G2 attached to?
Agree: this is more important than pentode plate voltage.
> I have a +600V supply, 47K plate resistor, and a -30V adjustable supply for the control-grid voltage. According to the datasheets, I should have the control-grid voltage around -7V to get about +300V on the plate. But with almost all of the tubes I've bought I have to set it down around -30V...
Now waitasec. You feed 600V through 47K and aim for 300V at the plate. (600V-300V)/47K is 13mA, not the 36mA specified for P=250V, G2=250V, G1=-7V. As Gm at 36mA is about 10,000uMho or 100Ω, to get from 36mA to 13mA you need more than (36mA-13mA)*100Ω=2.3V more bias, probably 3 or 4V more since Gm will fall as current falls. But that only gives -10V which is not what you see.
That G2 voltage IS critical. If I assume you leave the G2 at 600V, and G1-G2 Mu is 21, then you need to approach -29V to cut-off the tube, and 13mA is awful close to cut-off for this 100+mA tube. This is similar to your observation (though it is odd that some tubes are different). You are also far over the G2 rating, enough that the tube may be damaged, which would put it out-of-spec.
Let's hear where your G2 is.
When you get funny results from an unusual test rig, try something more like normal use. Feed plate and G2 from +300V through separate 1K resistors. Start with G1 at full negative voltage, let the tube heat, and sneak the grid bias up until plate or screen voltage drops. When the 1K plate resistor is dropping 36V (say from a sagged 290V to 254V) and the 1K G2 resistor drops a bit less (about 9V), then you are near the book condition. Note that required G1 voltage WILL scale roughly with change of G2 voltage, so you want to test with G2 voltage fairly close to either book-value (for verification) or proposed operating conditions (for actual use). Plate voltage will be much less critical (look at how flat the plate curve is for fixed G1 and G2 voltage as plate is swept from ~100V up to maximum).
Agree: this is more important than pentode plate voltage.
> I have a +600V supply, 47K plate resistor, and a -30V adjustable supply for the control-grid voltage. According to the datasheets, I should have the control-grid voltage around -7V to get about +300V on the plate. But with almost all of the tubes I've bought I have to set it down around -30V...
Now waitasec. You feed 600V through 47K and aim for 300V at the plate. (600V-300V)/47K is 13mA, not the 36mA specified for P=250V, G2=250V, G1=-7V. As Gm at 36mA is about 10,000uMho or 100Ω, to get from 36mA to 13mA you need more than (36mA-13mA)*100Ω=2.3V more bias, probably 3 or 4V more since Gm will fall as current falls. But that only gives -10V which is not what you see.
That G2 voltage IS critical. If I assume you leave the G2 at 600V, and G1-G2 Mu is 21, then you need to approach -29V to cut-off the tube, and 13mA is awful close to cut-off for this 100+mA tube. This is similar to your observation (though it is odd that some tubes are different). You are also far over the G2 rating, enough that the tube may be damaged, which would put it out-of-spec.
Let's hear where your G2 is.
When you get funny results from an unusual test rig, try something more like normal use. Feed plate and G2 from +300V through separate 1K resistors. Start with G1 at full negative voltage, let the tube heat, and sneak the grid bias up until plate or screen voltage drops. When the 1K plate resistor is dropping 36V (say from a sagged 290V to 254V) and the 1K G2 resistor drops a bit less (about 9V), then you are near the book condition. Note that required G1 voltage WILL scale roughly with change of G2 voltage, so you want to test with G2 voltage fairly close to either book-value (for verification) or proposed operating conditions (for actual use). Plate voltage will be much less critical (look at how flat the plate curve is for fixed G1 and G2 voltage as plate is swept from ~100V up to maximum).
G2 was attached through the 470 ohm resistor to the +600V supply...maybe this should have been 470K or go to the plate. I haven't tried connecting it to the plate yet.
I changed the 470 ohm to a 330K to the +600 supply, a value I had on hand. Now here are my voltages (to ground)
Plate = 300
G2 = 275
G1 = -9
Cathode = 0.8
These seem to be within spec and no problems. This set of tubes was G.E. labeled Made it Great Britain, so probably Mullard.
I have another set of tubes, Amperex made in Holland is the labeling, that I tried. I have different results with these tubes
Plate = 300
G2 = 75
G1 = -3.5
Cathode = 0.8
So the 330K g2 resistor value is too high with the Amperex tubes, maybe something like 47K would work better.
Is one solution to replace the g2 resistor with a pot or just go with a small value like the 470 ohms and connect it to the plate?
I changed the 470 ohm to a 330K to the +600 supply, a value I had on hand. Now here are my voltages (to ground)
Plate = 300
G2 = 275
G1 = -9
Cathode = 0.8
These seem to be within spec and no problems. This set of tubes was G.E. labeled Made it Great Britain, so probably Mullard.
I have another set of tubes, Amperex made in Holland is the labeling, that I tried. I have different results with these tubes
Plate = 300
G2 = 75
G1 = -3.5
Cathode = 0.8
So the 330K g2 resistor value is too high with the Amperex tubes, maybe something like 47K would work better.
Is one solution to replace the g2 resistor with a pot or just go with a small value like the 470 ohms and connect it to the plate?
> G2 was attached through the 470 ohm resistor to the +600V supply
Pardon my language, but.... why the %$#@! are you connecting and testing 250V-300V tubes on a 600V supply?????
Of COURSE you will need tremendous G1 bias to get reasonable plate current, when G2 is 2.5 times the spec-sheet value and twice the MAXimum rating.
Pentode plate current is roughly proportional to (Vg1)+(Vg2/20). Big G2 voltage makes big plate current, or needs big G1 bias to keep current down to reasonable levels.
Also, the metal in small tubes is not treated to to stand over 300 volts. Those 600V-propelled electrons SLAM into G2 and Plate with enough velocity to knock-loose multiple secondary electrons each. While G3 will drain some of these away, if you get a real electron-storm going in there, bad things happen. 600V power tubes have extra treatment of the plate metal.
Stuffing arbitrary resistors in there doesn't fix the overvoltage at cold-start or if you do get the current low. It will just be confusing.
Set the G2 voltage AT 250V, solid, just 470Ω-1K in series to protect against shorts. The plate voltage is far less critical. At book-value G2 bias, and G1 bias, you should get book-value plate current +/-30%.
Pardon my language, but.... why the %$#@! are you connecting and testing 250V-300V tubes on a 600V supply?????
Of COURSE you will need tremendous G1 bias to get reasonable plate current, when G2 is 2.5 times the spec-sheet value and twice the MAXimum rating.
Pentode plate current is roughly proportional to (Vg1)+(Vg2/20). Big G2 voltage makes big plate current, or needs big G1 bias to keep current down to reasonable levels.
Also, the metal in small tubes is not treated to to stand over 300 volts. Those 600V-propelled electrons SLAM into G2 and Plate with enough velocity to knock-loose multiple secondary electrons each. While G3 will drain some of these away, if you get a real electron-storm going in there, bad things happen. 600V power tubes have extra treatment of the plate metal.
Stuffing arbitrary resistors in there doesn't fix the overvoltage at cold-start or if you do get the current low. It will just be confusing.
Set the G2 voltage AT 250V, solid, just 470Ω-1K in series to protect against shorts. The plate voltage is far less critical. At book-value G2 bias, and G1 bias, you should get book-value plate current +/-30%.
- Status
- Not open for further replies.