I was just wondering if anyone here has tryed screen drive with an el84 tube. Any ideas on how to design for it?
It's a tube whose characteristics scream, "AGGGGH! I AM NOT MADE TO BE SCREEN-DRIVEN!"
What you want is a very high rate of change of the plate current with respect to screen voltage. Sweep tubes are far, far better candidates. Excellent 17.5W tubes can be gotten for under $8, sometimes far under.
What you want is a very high rate of change of the plate current with respect to screen voltage. Sweep tubes are far, far better candidates. Excellent 17.5W tubes can be gotten for under $8, sometimes far under.
Many times for way, way under that $8 mark.
Though they can be used to good effect in SE applications (I have a nice, but humble-looking 6CD6GA screen driven amp to prove it), you can get the most out of screen-driven sweep tubes in push-pull where those big cathodes with high peak current ratings can be used to best advantage. As far as I've seen, 200V p-p screen drive is enough to squeeze just about all the juice out of any sweep tube I've encountered. This is keeping in mind that you're kicking one tube of the push-pull pair into the stratosphere, while biasing the other way into cutoff. The followers needed to drive the screen current will need to come out of cutoff gracefully to avoid misbehavior at the crossover point. Mosfets work well, though if you use a more ambitious scheme like an active current source at the bottom, It'll also need to recover gracefully from cutoff. I arbitrarily put point of maximum juice squeezing at ~50V at the plate, though George (Tubelab), is capable of ruinning plates into the ground with his test setup. This practice can melt the screens, though. A more modest drive level will still yield lots of power.
Though they can be used to good effect in SE applications (I have a nice, but humble-looking 6CD6GA screen driven amp to prove it), you can get the most out of screen-driven sweep tubes in push-pull where those big cathodes with high peak current ratings can be used to best advantage. As far as I've seen, 200V p-p screen drive is enough to squeeze just about all the juice out of any sweep tube I've encountered. This is keeping in mind that you're kicking one tube of the push-pull pair into the stratosphere, while biasing the other way into cutoff. The followers needed to drive the screen current will need to come out of cutoff gracefully to avoid misbehavior at the crossover point. Mosfets work well, though if you use a more ambitious scheme like an active current source at the bottom, It'll also need to recover gracefully from cutoff. I arbitrarily put point of maximum juice squeezing at ~50V at the plate, though George (Tubelab), is capable of ruinning plates into the ground with his test setup. This practice can melt the screens, though. A more modest drive level will still yield lots of power.
Yeah, I tried it and it doesn't work. Neither do any of the usual audio tubes. You can get some sound, but the output is low and there really is no advantage to using screen drive.
I have come to the conclusion that sweep tubes work best, along with a few other oddball TV tubes. You want to look for the lowest maximum screen grid voltage rating that you can find. This implies good screen sensitivity. The common audio tubes that can live with 400 volts on the screen generally require too much drive voltage to work in screen drive mode. Some sweep tubes work better than others, and the only way to find out is to try them.
As stated screen drive is better suited to push pull applications. An SE amp must run class A, so the output tube must be biased with a high quiescent current. The advantage of screen drive with the right tube is good linearity at low currents which is maintained up into the glowing tube zone. This allows for some efficient amplifiers with high power outputs, running on the edge of class B.
Yes it is possible to drive the screen grids hard enough to make them glow. This is a problem if you want to extract 100 watts of pure sine wave power continuously from a pair of tiny 6BQ6's. It is OK to run them with music that may see a 100 watt peak once avery few seconds. Carver sold a zillion solid state amplifiers like this (no heat sinks), I still have one. Or you could just call it an 80 watt amp.
Top secret scoop....... The 6GV5 is on sale at ESRC for $1 each. It is a 6AV5GA in a compactron base. I have extracted 80 watts from a pair for hours with no ill effects.
I have cranked stupid power out of 6BQ6's. They are probably still affordable, but we will likely never see them for 98 cents again after last summers feeding frenzy on them.
I have a bunch more tubes to try, but no time right now.
I bought 10 to play with. They are GE's with the large plate structure found in the 6FW5, which means that the screen grid will be the weakest link in a screen drive amplifier.
One more thing to consider, the 6GV5 is a 12 pin compactron with a plate cap. This means that you will need a 12 pin socket and a plate cap. The price of the 6GV5 + the socket + the plate cap may exceed the cost of a 6AV5GA and an octal socket, especially if you already have the octal socket. The 6AV5GA does come in many flavors of varying plate capability, but I have seen the wimpiest ones crank out 80 watts in screen driven push pull without violating any specs.
I have collected a few 6AV5's over the years, so that is what I would use to build any amplifier that I intended to keep, but all of my experiments will be done on the 6GV5's. I feel better about torturing a $1 tube.
If you want to try screen drive with 9-pin tubes, 6DW5, 6DT5, 6EM5, 6DS5 are some you might try... screen curves for 6DW5 look promising - more linear than G1 curves.
Several vendors seem to be almost giving the 6GV5s away. I'd be tempted, except for that teeny plate cap. It's a pain to find a connector that will fit.
I've been looking ay a fair number of sweep tube data sheets, and most have data for intermittent operation at a low plate voltage (40-70V) with zero g1 bias and 100-150V on the screen. The point is chosen so that most of the current is still flowing in the plate, with a screen current of ~10% of plate current. It's a reasonable indicator as to what kind of screen drive is needed at the max power point, and how hefty your screen drive follower needs to be. It's also a good indicator as to the sensitivity of the screen.
Some caveats on screen drive mode. The plate resistance of the sweep tubes I've seen is pretty high using screen drive, so expect to need about as much negative feedback as one would use with a pen-toad. I've used the usual feedback from the output to the input tube with good results. A partial feedback scheme like used in the simple SE or the RH-type amps would probably also work.
I've been snapping up various high transconductance pentodes for use in the input stages of both screen driven sweep tube amps and some of the usual audio pentode types employing partial feedback. There are lots of pentodes with gm ~10-20k umhos at reasonable plate current, and most are still inexpensive. I won't name any names just yet, but a diligent search through a tube data book will be rewarded. Some are on the $1 lists. Others are pricier, but nothing unreasonable.
I've been looking ay a fair number of sweep tube data sheets, and most have data for intermittent operation at a low plate voltage (40-70V) with zero g1 bias and 100-150V on the screen. The point is chosen so that most of the current is still flowing in the plate, with a screen current of ~10% of plate current. It's a reasonable indicator as to what kind of screen drive is needed at the max power point, and how hefty your screen drive follower needs to be. It's also a good indicator as to the sensitivity of the screen.
Some caveats on screen drive mode. The plate resistance of the sweep tubes I've seen is pretty high using screen drive, so expect to need about as much negative feedback as one would use with a pen-toad. I've used the usual feedback from the output to the input tube with good results. A partial feedback scheme like used in the simple SE or the RH-type amps would probably also work.
I've been snapping up various high transconductance pentodes for use in the input stages of both screen driven sweep tube amps and some of the usual audio pentode types employing partial feedback. There are lots of pentodes with gm ~10-20k umhos at reasonable plate current, and most are still inexpensive. I won't name any names just yet, but a diligent search through a tube data book will be rewarded. Some are on the $1 lists. Others are pricier, but nothing unreasonable.
I purchased 10 each of several likely candidates, but I haven't had the time to test them all yet. A few look like winners on paper, but the proof is in the sound. I too am keeping quiet until I can do some testing. I think that I spoke too soon on the 6CB5, so that they were gone before I could get a bunch. During the tube sale at AES feeding frenzy I called them and inquired about the 98 cent 6BQ6's. They said that they had plenty and were reducing their inventory. We talked them up, and now they no longer have ANY!
Same plate cap as the 6BQ6. 1/4 inch, Stan has the ceramic Chinese ones.
I shouldn't have read this thread. Wallet is lighter now. I've got a bunch of 2E26's that may be entertaining too.
tubelab.com said:
6GE5 is VERY similar, but has no plate-cap....(Looking at data-sheets, could even be the same guts in different glass....)
Nice screen-drive amp possibilities. Maybe even quite high power,
--considering its tiny size....
I made a 'junk-box' screen-drive with these, (using Tubelab Power-drive into screens) and was quite pleased with the result. I'm limited on power due to the iron I had available, but it still sounds nice, --One good thing, It doesnt heat the whole room up after a while, and was able t4o make it on a plastic 'project-box'!....
😀
I haven't had a lot of luck with the small Chinese-made plate connectors - some fit (albeit indifferently), and some don't even come close. I've taken to replacing the Chinese stuff with NOS plate caps wherever I can. For my EL509 screen drive amp, I'll probably resort to making my own plate caps with some phosphor bronze sheet and silicone/fiberglass composite rod stock.
Hey, George - mum's the word on those pentodes, though there are a few cherries in the bunch with 550V max plate voltage, making them especially useful, though not necessarily more expensive... Running a JFET cascoded by a triode also gives one a lot of gain - it's the scheme used in the "Shrine" amp, which is back in the living room and playing nicely as I write this. It's a real kick watching a 6" cube pump out some watts (tubes on top). The next amp in the series should be better.
Hey, George - mum's the word on those pentodes, though there are a few cherries in the bunch with 550V max plate voltage, making them especially useful, though not necessarily more expensive... Running a JFET cascoded by a triode also gives one a lot of gain - it's the scheme used in the "Shrine" amp, which is back in the living room and playing nicely as I write this. It's a real kick watching a 6" cube pump out some watts (tubes on top). The next amp in the series should be better.
- Status
- Not open for further replies.