Ahh, I see what you guys are getting at. Yes, once the conventional RC cathode bias has been removed, and the grid current limited, you can then build a negative supply, regulate it, then attach the bottom of the grid leaks to it. Or, for five bucks and a free light show, you can get the same instantaneous bias voltage recovery and no more impedance penalty than the usual idle current sensing resistor.
Yup, I wasn't thinking. LM 317 is a source not a sink.
I have both output tubes biased up with the LED stacks now. I trimmed the screen supply resistor to get it to 200V and all is up and running nicely.
I'll work on a couple screen supplies next, then I have something in mind to try in place of the diodes, but will refrain from sticking my foot in my mouth again till I do some more research.
Come to think of it, you could make it work if you really wanted to by loading the LM317 voltage regulator with a resistor that draws more than the maximum cathode current.
I was thinking of using a string of LEDs to drive the base of a NPN with it's emitter tied to ground and a resistor from base to ground to set the bias current of the LED string, collector tied to top of LED string.
Would the transistor amplify the noise of the LEDs and would result in more noise than the LEDs alone.
Bias the LEDs for 10mA.
Now I need 40mA from the collector so base current will be 40ma/Beta.
Most of the LED current and noise would be in the bias leg. The only portion in the collector of the transistor would be the amplified base drive. Both the noise and dc current should be beta times that injected in the BE junction.
This should not end up much noisier than the LED strings together should it?
Or do the values of the parallel strings of LEDs form a Square root of the sum of the squares function?
Would the transistor amplify the noise of the LEDs and would result in more noise than the LEDs alone.
Bias the LEDs for 10mA.
Now I need 40mA from the collector so base current will be 40ma/Beta.
Most of the LED current and noise would be in the bias leg. The only portion in the collector of the transistor would be the amplified base drive. Both the noise and dc current should be beta times that injected in the BE junction.
This should not end up much noisier than the LED strings together should it?
Or do the values of the parallel strings of LEDs form a Square root of the sum of the squares function?
I've routinely used that configuration as shunt regulator, wich is nothing else than a voltage source or a "Zener Booster"
And with Varistor as voltage reference:
Should work with LED string also but I can't tell about noise . . .
Yves.
And with Varistor as voltage reference:
An externally hosted image should be here but it was not working when we last tested it.
Should work with LED string also but I can't tell about noise . . .
Yves.
I've used suppression diodes like the 1.5KE350 as a HV ref diode for regulation and had thought about MOVs (Metal Oxide Veristors) that are used across power lines for noise suppression, but thought they would have too loose of tolerance. I'll look at the L14K60 series.
Does anyone have a link to Bas's article? the link
Compromised URL removed by Moderation
from post 86 is 86'd.
I'm trying to find out how Sy measured the z of the LEDs adn can't find it in this thread so I presume it is outlined in the article.
Does anyone have a link to Bas's article? the link
Compromised URL removed by Moderationfrom post 86 is 86'd.
I'm trying to find out how Sy measured the z of the LEDs adn can't find it in this thread so I presume it is outlined in the article.
Last edited by a moderator:
Last night I tried an RLD LED bias circuit. It sounded good but the bias was off from expected.
Tonight I tried multiple LEDs in series driving the base of a TIP41 with a 68 ohm resistor to ground and a 2.5ohm collector resistor to the LED anode. Emitter to ground. I ended up with two LEDs (4.7V gnd to Cathode) to get .241 volts across the 2.5 ohm resistor to set the bias current for the two 6P1P cathodes at 96.4 mA (.241/2.5ohm = 96.4mA) Split between the two tubes that is about 48mA each.
To get this bias current I ended up with the cathodes at 4.7V, not the 10.5 I expect from the data sheet. Screen to cathode is 199V vs 200V in the data sheet. Anode to cathode voltage is 256 vs 250 in the data sheet. Not enough to account for the cathode bias difference.
What gives with these 6P1P-EVs? I can't seem to get close to 250V Plate, 200V screen and 10.5 V cathode bias for 50mA.
I guess the next step is (1) start an engineering notebook, and (2) build two screen bias regulator circuits.
Don de Dieu me.....
Tonight I tried multiple LEDs in series driving the base of a TIP41 with a 68 ohm resistor to ground and a 2.5ohm collector resistor to the LED anode. Emitter to ground. I ended up with two LEDs (4.7V gnd to Cathode) to get .241 volts across the 2.5 ohm resistor to set the bias current for the two 6P1P cathodes at 96.4 mA (.241/2.5ohm = 96.4mA) Split between the two tubes that is about 48mA each.
To get this bias current I ended up with the cathodes at 4.7V, not the 10.5 I expect from the data sheet. Screen to cathode is 199V vs 200V in the data sheet. Anode to cathode voltage is 256 vs 250 in the data sheet. Not enough to account for the cathode bias difference.
What gives with these 6P1P-EVs? I can't seem to get close to 250V Plate, 200V screen and 10.5 V cathode bias for 50mA.
I guess the next step is (1) start an engineering notebook, and (2) build two screen bias regulator circuits.
Don de Dieu me.....
I Have found the same thing happening when I have used LED's for bias. I believe it is because the lack of resistance effects DC degenerative feedback.
Shoog
Shoog
The whole POINT of using LEDs is to have no degenerative feedback. That's what a "voltage source" means.
Gimp, the biasing is done via the screen regulator- all grids are created equal. If it doesn't bias up right, that's where to start looking.
Gimp, the biasing is done via the screen regulator- all grids are created equal. If it doesn't bias up right, that's where to start looking.
I've not found data for 200V screen.
But . . .
Since plate voltage has almost no effect on plate current (as long as it remains higher than screen voltage), I've used triode data from there:
http://frank.pocnet.net/sheets/113/6/6P1PEV.pdf
At 200V and 50mA anode + screen current, you need, say, +7.5 Volts cathode to grid bias.
Well, still not 4.7 . . . nor 10.5 ! !
Even God doesn't know why 😕
If you remove those 4.7 volts from the screen voltage, target for 48mA anode + screen current, you're not so far from the ballpark.
And . . . does the current in the LEDs flow thru the 2R5 sensing resistor ?
Yves.
Darned if I know where I got the 200V Screen voltage from. I went back and looked at all the 6P1P(-EV,-EB) and 6AQ5 specs and all say 250 Screen.
I'll up the screen voltage tonight.
The 2R5 does not sense the LED current, but the two screens are pulling near 4ma each so the actual plate current is not off by but 1.5mA per plate.
I'll up the screen voltage tonight.
The 2R5 does not sense the LED current, but the two screens are pulling near 4ma each so the actual plate current is not off by but 1.5mA per plate.
Last edited: