I found the following statement in Norman Crowhurst's High fidelity sound engineering (Newnes, 1961, p. 124):
"Many valves, operated as pentodes, require usage of a screen voltage lower than the anode supply voltage. This is because the screen grid gets much more than its share of current in the upswing (anode voltage downswing), so that its voltage must be reduced. Working as triodes, the same valves can use a higher screen voltage, the same as the anode, because the high current upswing is avoided by signal voltage drop. The same thing, happening to a lesser extent, permits the full anode supply voltage to be used also on the screen in ultra-linear working." [my emphasis]
Presumably this means that as long as we don't violate the maximum anode voltage, we can stop worrying about the maximum screen voltages quoted on spec sheets when we use pentodes as triodes. What do our gurus think?
"Many valves, operated as pentodes, require usage of a screen voltage lower than the anode supply voltage. This is because the screen grid gets much more than its share of current in the upswing (anode voltage downswing), so that its voltage must be reduced. Working as triodes, the same valves can use a higher screen voltage, the same as the anode, because the high current upswing is avoided by signal voltage drop. The same thing, happening to a lesser extent, permits the full anode supply voltage to be used also on the screen in ultra-linear working." [my emphasis]
Presumably this means that as long as we don't violate the maximum anode voltage, we can stop worrying about the maximum screen voltages quoted on spec sheets when we use pentodes as triodes. What do our gurus think?
It is not the only way to get triode characteristics by connecting screen grid to anode. Another way is to supply AC from anode to the control grid, this way is more flexible and more linear than better known one, "screen to anode".
However, input resistance is lower such a way, but considering Miller capacitance anyway similar driving current is needed for better than average performances.
Wavebourn, I realise that you have some different ideas about how to emulate triode performance with a pentode but that is not what I am asking for. I want comments on Crowhurst's statement that screen grids can be supplied with the same high voltage as anodes when the valve is triode-connected. If you cannot address this issue, please do not comment.
When you first build an amplifier this way, be sure to include a resistor between anode and screen - then you can be certain that the anode voltage is (at least a little) higher than screen.
The KT66 data sheet INSISTS you do this anyway, to prevent oscillation, using 100 to 300 ohms fitted right on the socket.
The more you want to violate the spec, the bigger the resistor you should use!
Test the amp in the dark when you first put the power on it. Look out for glowing screen, if you can see in there. Some valves can do this better than others, I believe.
The KT66 data sheet INSISTS you do this anyway, to prevent oscillation, using 100 to 300 ohms fitted right on the socket.
The more you want to violate the spec, the bigger the resistor you should use!
Test the amp in the dark when you first put the power on it. Look out for glowing screen, if you can see in there. Some valves can do this better than others, I believe.
Screen current is high when the anode voltage dips below it in pentode mode, This doesn't happen in triode (and in ultralinear, since screen voltage dips with plate). As long as screen stays within its dissipation rating...
Fritz Langford Smith (who was reputed to know a bit about valves...) used 807s in triode well above their screen rating in his version of the Williamson amplifier.
Fritz Langford Smith (who was reputed to know a bit about valves...) used 807s in triode well above their screen rating in his version of the Williamson amplifier.
I would propose a 'yes', but a qualified one:
A friend of mine did some tests on some ordinary small-signal pentodes and found that in triode mode, cathode current was often shared almost equally between anode and screen, or that screen current could even exceed anode current! (Perhaps because the anode is big, but the screen is closer to the cathode). This would vary somewhat with applied voltage. He managed to destroy at least one screen grid in the process of testing this!
Connecting the screen to anode through the usual 100 ohm stopper prevented this, so that screen current dropped to a much smaller fraction of the anode voltage, and I have been religious about using such stopper resistors ever since.
I haven't yet done my own tests to confirm his findings, but I don't have any reason to suspect he was mistaken.
By the way Baretter, where are you based?
I think we're gradually converging on a "treat with caution" verdict on this statement. I started this thread because of contradictions in 6V6 spec sheets pointed out in another thread on this forum.
I live in Todmorden, on the Yorkshire/Lancashire border, but I don't have any divided loyalties as I'm originally from North Lincolnshire!
I live in Todmorden, on the Yorkshire/Lancashire border, but I don't have any divided loyalties as I'm originally from North Lincolnshire!
I have also managed to blast the screen grid out of a 6EJ7 with the screen and plate connected directly together. I think oscillation might have been the cause. A resistor will fix this in either case.
Yes, some caution shound be advised when exploring the limits of a tube that has not been explored before. As pointed out, a resistor of at least 100 ohms should be placed in series with the screen grid as a current limiter and an oscillation stopper. 100 ohms is a good place to start with most tubes. Some tubes need a higher value, with 1000 ohms being the likely upper limit. The EL34 is one tube that wants more than 100 ohms, with 1000 ohms being mentioned in the original Mullard literature. If distortion measuring equipment is available test the distortion with several resistor values and pick the highest value that does not raise the distortion.
I have successfully operated several tubes well above their published screen ratings without issues in triode mode. In many cases you can not approach the plate voltage ratings. This is true with all sweep tubes and some RF tubes like the 807.
It has been stated that the usual 1/2 watt "stopper" resistor is sufficient since the dissipation in the resistor is low. I have found this not to be the case if the amp will ever see clipping. An EL34 can draw some fairly high screen grid current peaks when overdriven. I have fried a few 1/2 watt resistors, and now use a 1 or 2 watt resistor in all of my amps.
This is true especially for beam tetrode tubes, becaues of the tight grid aligment. I used 0.5W resistor for my 6550s without any problem. This is true for PL509s as well
Sajti
Beat me to it. Inordinately high values can generate a large number of high distortion harmonics.
My understanding was the screen limitation is power, not simply voltage.
Some tubes (especially sweep tubes) tend toward runaway with high screen voltages. This is obvious in pentode mode because the screen voltage is constant. A similar thing happens in triode mode with no signal present, since the plate and screen are at nearly the same voltage.
I was playing with some triode wired sweep tubes at something close to twice the normal screen voltage rating and found an "avalanche" effect. The tube will make gobs of power when driven, but will go into runaway with no signal present. As the B+ (plate and screen) is increased the current will naturally rise as with most triodes, but will reach a point where it just takes off regardless of control grid voltage. I blasted a perfectly good 6LW6 into oblivion at 450 volts. I was getting over 100 watts out of a pair with no objection until I turned down the audio oscillator. I know that a 6CD6 will do the same thing, but a 6AV5 seems happy at over 300 volts in triode despite the 175 volt screen rating. Some 6AV5's will work at 350 volts.
When you go exploring outside the normal ratings you must realize that every tube is different and even two from identical batches may behave differently.
barretter,
You may be on to something, I hope so! Much of the tube date we have to work with supports pentodes as radio transmitter tubes operating as class c amplifiers not class a audio. It would be a time trip to sit down Norman Crowhurst, buy him a cup of coffee and talk it over. My impression is that in pentode operation the diffrential voltages between the anode & screen can spike causeing lighthing in the bottle and an early death of the tube. I think that it was Crowhurst that called that flashover. If the anode and screen are tied together we leave the tube data behind. The anode and screen are at the same voltage potiential.
I am going to install the 100 ohm resisitor turn off the lights and crank up the plate/screen voltage untill the smoke comes out. Then turn it down.
You may be on to something, I hope so! Much of the tube date we have to work with supports pentodes as radio transmitter tubes operating as class c amplifiers not class a audio. It would be a time trip to sit down Norman Crowhurst, buy him a cup of coffee and talk it over. My impression is that in pentode operation the diffrential voltages between the anode & screen can spike causeing lighthing in the bottle and an early death of the tube. I think that it was Crowhurst that called that flashover. If the anode and screen are tied together we leave the tube data behind. The anode and screen are at the same voltage potiential.
I am going to install the 100 ohm resisitor turn off the lights and crank up the plate/screen voltage untill the smoke comes out. Then turn it down.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.