Hi, my PP amplifier is under construction, and I want to drive the g1 of the output tubes from a CF. The output tubes work in class AB1. How can I determine the necessary (minimum) current through the CF?
Would the following idea work? Tube: E88CC (6DJ8), B+ = 375V, Ug= 240V, Ia = 5mA. I would use an Rk = 47k. The coupling capacitor would be 2.2uF, the grid leak resistor of the output tube would be 100k.
Would the following idea work? Tube: E88CC (6DJ8), B+ = 375V, Ug= 240V, Ia = 5mA. I would use an Rk = 47k. The coupling capacitor would be 2.2uF, the grid leak resistor of the output tube would be 100k.
My rule of thumb of any driver -under A1-: at least three times larger steady current as power stage's grid "current" (due to the grid leak resistor//Z of all capacitance at 100kHz).
Mine is the more the merrier, even though AB1 is not supposed to draw grid current; just stay below Pd/2 for safety.
The prob I see is you may need to elevate the heater supply, which may require a separate winding just for the CF.
Not sure if the E88CC is the right guy for the job. Maybe a 12BH7, a 6CG7, or a E80CC.
The prob I see is you may need to elevate the heater supply, which may require a separate winding just for the CF.
Not sure if the E88CC is the right guy for the job. Maybe a 12BH7, a 6CG7, or a E80CC.
What Euro said. Lots more than the load current, if possible.
Zung points out that grids may be "no current", conceptually. However there is capacitance everywhere and it all sucks. Figure grid capacitance and its impedance at the highest frequency which might happen.
But more to the point: WHY use cathode followers into no-current grids? Yes, there may be reasons. Extended treble response (video modulators). A high-strung gain-stage (>500k loading) into a fat hungry grid (6550 fix-bias wants <50k in the grid return).
Zung points out that grids may be "no current", conceptually. However there is capacitance everywhere and it all sucks. Figure grid capacitance and its impedance at the highest frequency which might happen.
But more to the point: WHY use cathode followers into no-current grids? Yes, there may be reasons. Extended treble response (video modulators). A high-strung gain-stage (>500k loading) into a fat hungry grid (6550 fix-bias wants <50k in the grid return).
“Elevating” a heater to a negative voltage - hmmmmm….. Something just doesn‘t sound right about that. But it does need to be between the bias voltage and whatever negative rail you tie the tail resistor to. My 6550 monstrosity used the 12BH7s as direct drivers so I could easily drive three pair and keep the Rg1 low. Running them pretty hot though, 10 mA each. All 3 heaters floated down to the -68V bias supply.
Recently I’ve run across the 6GU7, which looks an awful lot like a BH7. Curves are a bit different, but same 17 mu and 3100 gm. Probably not much different as a k-follower, and at the price I’m not going to complain. I’m also looking into video output pentodes wired as followers for both direct drive on finals and driving 200-600 ohm balanced lines. Gms are high so the Z out is low, and they’re quite happy running as high as 20 or 25 mA. Good for 5 to 7 watts, and many of them come packaged with a general purpose triode. And cheaper than buying EL84’s or 6V6’s which ive used in the past for this sort of thing.
Recently I’ve run across the 6GU7, which looks an awful lot like a BH7. Curves are a bit different, but same 17 mu and 3100 gm. Probably not much different as a k-follower, and at the price I’m not going to complain. I’m also looking into video output pentodes wired as followers for both direct drive on finals and driving 200-600 ohm balanced lines. Gms are high so the Z out is low, and they’re quite happy running as high as 20 or 25 mA. Good for 5 to 7 watts, and many of them come packaged with a general purpose triode. And cheaper than buying EL84’s or 6V6’s which ive used in the past for this sort of thing.
6mn8. 6md8 if you don't mind paralleling plates and grids, they are dirt cheap...http://www.nj7p.org/Tubes/PDFs/Frank/123-GE/6MN8.pdf
Last edited:
Nothing much to add really only the beefier your LTP or driver is the better. Any hefty triode 12BH7's, ECC99's/6H6N 6S4A, 12BL7 12B4 run at high Va 350v+ and at about 4-6mA in a CCS LTP will drive pretty much anything. Pat Turner (see - http://www.turneraudio.com.au/ ) favoured using parallel strapped double triodes in a LTP like the 12BH7 or EL84's as wg_ski also mentions. He didn't like KF's for AB1, it's another stage so more phase shift.
Andy.
Andy.
As I understand the heater should be at a more positive voltage than the cathode. This is in order to prevent a "tube within a tube" situation, where the heater acts as cathode, and the cathode acts as anode. Also one should respect the maximum heater-to-cathode voltage limit. But I wonder why? Is there any risk of internal breakdown due to proximity of heater to cathode? Has enybody ever experienced such thing?
You normally think of elevating a k follower heater because the cathode may sit 100 or more volts above ground, possibly exceeding Vh-k, or at least “harmlessly” causing the damn thing to hum. Direct driving output tubes (presumably with fixed bias), the cathode sits negative. Might be only -25 or -30 volts but at start up it might be -200. If you do anything with the heater, you run it below ground so Vh-k stays reasonable under normal quiescent or start up with no current.
I've done several CF Drivers with no problems. 6J5s in a Circlotron, 4BQ7 driving PP 6V6s in Class AB2, 6LU8 triodes in a Crowhurst
Twin Coupled Amp Clone.
Twin Coupled Amp Clone.