roline: please tell me more about your amp. schematics? measured performance? output transformers? more?
Thanks!
Thanks!
All; trying to find thread 94 per tubelab's suggestion but have not been able to figure our how to do that! can't see a thread number anywhere on a given thread line and a search using 94 as an argument yields no results.
Help!
Help!
I experimented with using a VR tube to drop screen voltage. It kinda works, but the VR tube extinguishes when the plate voltage swings down below the dropout voltage of the VR tube. The zener would have the same problem.
You could do this for situations where you didn't need to lose a lot of volts.
6CD6 testing
It's been a while since I lit up a 6CB5 at high power. When I did the experiments in the AES thread I had an old Fluke 407D power supply that goes to 555 volts at 300 mA, which I repeatadely overloaded to 500+ mA without issue. I now have a 650 volt 1.7 amp HP, so some of these experiments will be repeated.
If the tubes are up to it, you should get about 125 to 150 watts per pair at 3300 ohms on 650 volts, a bit more on 680 volts. The 6CB5 is an old design. 680 volts is a bit above where I would run them because of the the idle dissipation. If you biased the tubes at 30 mA the idle dissipation is over 20 watts per tube. I have one of Petes big red boards running 6HJ5's at 650 volts and 35 mA without issue, and the tubes look the same internally, so it should work.
I have connected dozens of sweep tubes to Petes red boards. All worked well at 150 volts, but a few worked better with a bit more voltage, up to 175.
I would guess that they would, but I have some really BIG 1250 ohm OPT's that I want to put to good use, so I might get a chance to try it soon. With 4 tubes and 680 volts you will see well over your 200 watt limit. I ran 4 35LR6's into those 1250 ohm OPT's on 650 volts and got 525 watts!
Based on extrapolating on data seen in post #66 of Petes driver board thread, you would lilkely do better on 450 to 500 volts. I got 91 watts per pair on 450 volts, and 106 watts per pair on 500 volts using the $1 13GB5 into 3300 ohms. If you used 4 tubes into 1500 ohms each pair sees 3000 ohms.
Click on this link. Then go through several pages until you get to post #94. The post # is in the upper right corner of each post. If you are using the default 10 posts per page it should be on page 10.
http://www.diyaudio.com/forums/tubes-valves/128533-tube-sale-aes.html?highlight=tube+sale+AES
You can safely have more screen grid voltage if running Ultraliner or evem nore if running triode.
The reason for a limit on screen grid voltge is to limit screen grid dissipation.
Notice that screen grid current increases as the plate voltage drops.
Screen grid current increases somewhat once the plate voltage is less than the screen.
Electrons do not attach to the screen if there is something larget and at a higher voltage like a plate.
If using 50% UL taps, the screen grid voltage drops in half at the worst cast power dissipation point if running pemtode operation.
If running as a triode, the plate voltage is never less than the screen voltage. Not much screen current is drawn.
One example is I am running an 865 as a triode. The screen grid voltage limit is 175 volts.
I see no problems running the plate and screen at 500 volts. Screen power dissipation is below limits.
The reason for a limit on screen grid voltge is to limit screen grid dissipation.
Notice that screen grid current increases as the plate voltage drops.
Screen grid current increases somewhat once the plate voltage is less than the screen.
Electrons do not attach to the screen if there is something larget and at a higher voltage like a plate.
If using 50% UL taps, the screen grid voltage drops in half at the worst cast power dissipation point if running pemtode operation.
If running as a triode, the plate voltage is never less than the screen voltage. Not much screen current is drawn.
One example is I am running an 865 as a triode. The screen grid voltage limit is 175 volts.
I see no problems running the plate and screen at 500 volts. Screen power dissipation is below limits.
Assuming a triode connection, would you recommend connecting the screen grid to the plate through a 100-300R resistor or straight?
You could run triode without a screen resistor it will be OK for typical operation.
If you use something with plate caps you may want a resistor.
A small screen resistor could save a tube if you have a plate cap that get loose though.
Having screen grid voltage and none on the plate will result in excessive screen dissipation.
The catch is you need to pick something low enough that it does not impair performance.
And it needs to be high enough to prevent burning the screen grid in you loose plate voltage.
Tubelab: thanks for the help on finding the post.
Understood fully about the screen voltage vs instantaneous plate voltage, of course I would watch that and make sure the zener drop is not violated during expected operations. Your experience in testing this gives me hope that this approach, properly implemented, would work.
_carter: Got it, thanks.
Screen voltage limits are not just due to power dissipation. The high G2 gm exhibited by sweep tubes suggests that the physical placement of the G2 is much closer to the G1 and K than in other types of beam tubes. Same, to a lesser degree, is true of RF tubes such as the 807. I personally would be leary of strapping the G2 to the plate and running a sweep tube to voltages along the lines of the plate rating. I think that's inviting G2 to G1 arcing. Anyway, my goal is UL operation, though now I'm also educating myself about G2 drive 🙂
Thanks all!
Understood fully about the screen voltage vs instantaneous plate voltage, of course I would watch that and make sure the zener drop is not violated during expected operations. Your experience in testing this gives me hope that this approach, properly implemented, would work.
_carter: Got it, thanks.
Screen voltage limits are not just due to power dissipation. The high G2 gm exhibited by sweep tubes suggests that the physical placement of the G2 is much closer to the G1 and K than in other types of beam tubes. Same, to a lesser degree, is true of RF tubes such as the 807. I personally would be leary of strapping the G2 to the plate and running a sweep tube to voltages along the lines of the plate rating. I think that's inviting G2 to G1 arcing. Anyway, my goal is UL operation, though now I'm also educating myself about G2 drive 🙂
Thanks all!
That's the point.
Also explains the incredibly high current capacity exhibited by sweep tubes ... which is meant to be used only in relatively narrow pulses.
Just read the datasheet but mainly have a close look at the Plate curves.
The compromise paid is low screen voltage capability.
Personally I'd feed rated voltage (150V) to those screens and , if any, go wild with plates, (safe because as you say, tubes work properly even when biased quite cold.
In the old wild days of mega open air festivals (around 1970), our then star PA man, Teddy Goldman, powered them with lots of 250W class B 6x6DQ6 amplifiers, with silly high +B (over 600/650V).
As a side note, he surrounded the people with tons of 4x12" columns (then the standard PA cabinet of choice) .
If you were inside that semicircle, your brain would melt.
Since this required long runs of cable, up to 60 meters each side of the stage, the homebuilt 6DQ6 power amps had only 32 ohms output, and the 4x12" 8 ohms columns were connected in series, in clusters of 4.
This minimized cable losses big way.
Of course, sound was "all mids", no real bass ... but that was the standard, even in England (WEM).
Even the mixers were home made:
An externally hosted image should be here but it was not working when we last tested it.
Sorry I didn't find a power amp picture, even if fuzzy.
i have several sweep tube amps lined up as wip, and based on my experience with he 6LU8 operation in pentode mode with the screens zener regulated at 120 volts dc yielded a very nice sounding amp, at 12 watts output, those who heard it liked it better than a pp 6L6 operating ultralinear...
This is not always true, especially with sweep tubes. I have blown up enough tubes to know. In my experience a triode wired sweep tube will not fail at full crank. It will go into runaway at IDLE after it has been cranked for a while. A tube at idle doesn't know if it is triode, UL or pentode wired. It just has too much voltage on the screen.
Some sweep tubes like the 6AV5 can be triode wired and run at 350 volts or more. Some like the 6CD6 will blow up at 250 volts in triode. I built a triode wired amp using my favorite the 6LW6. I sorted through about 20 tubes and picked 6 or 8 that worked at 450 volts. It played fine for several months then a tube ran away. This pattern repeated even after lowering the plate voltage twice, so the amp was dismantled. Want a tube that you can triode wire up to 750 volts, find some 7403 / 3D21WB.
True, I dissect any tube I kill. The screens are usually closer, and wound tighter using finer wire. The European E130L uses a frame grid for both G1 and G2. Don't try triode wiring these. Yes, I found out the hard way.
The 807, the 1625, and the 6BG6GA ARE 6L6GB's with different bases and the plate on a top cap. Again, I have dissected a bunch of these. There are some late production 807's with 6L6GC guts, and a very few straight glass 6BG6GA's with 7027 guts inside.
The 6LU8 is a vertical sweep tube. They are a different animal and were designed to run as a linear amp. I have run its twin, the 6LR8 at 300 screen, 400 plate without issue. Some will live at 400/400 and they do work well in triode or UL. Some say these are a 12AT7 triode, and a 6W6 Pentode, but a 6W6 will melt at the conditions I have run 6LR8's at.
Horizontal (line) output tubes are a different beast, that requires a little different technique to tame, but big power and stunning dynamics await those who master them.
I've done designs with these TV HD pents as audio power finals: Le Renard
They're not at all like the other types you mention here, in that they don't seem to make nearly so much higher order garbage. At least with the 6BQ6GTB/GA, it mainly makes H3, and sounds a bit "edgy" or overly "aggressive". Local NFB isn't required at all, just enough gNFB to take the edge off. I wouldn't bother, and especially not with UL, as the HD types have very limited screen voltage ratings. To use UL, you'd require a special custom (BIG $BUX) OPT with tertiary windings to take off the screen feedback voltage (and you'd best hope that whoever did the OPT did it right, and that the screen coupling windings have low leakage inductance -- not guaranteed by any means).
If you feel you just have to have lNFB, look into cathode feedback lNFB. It accomplishes the same thing as UL, but doesn't require any connections to the screens, which should be supplied from a stiff power source, preferable an active regulator.
yes, that is also my experience, you d not have to go ultralinear, pentode works fine as long as you take good care of the screens....
hey George, i get you, i guess what i really wanted to say is that TV sweep tubes can work wonders in an audio amp...
hey George, i get you, i guess what i really wanted to say is that TV sweep tubes can work wonders in an audio amp...
So,
With all the sweep tube talk, I'm interested, but wonder how a pair of 20W plate dissipation sweeps makes 100W clean whereas a 42W plate dissipation pair of KT88s in UL only makes 60-70W?
Is it the screen treatment? I'd like to try the sweep tubes for an amp, but I'm a bit lost about how a PP pair of 6CB5s makes 125W.
Can someone explain this in some form of layman?
Thanks!
Blair
With all the sweep tube talk, I'm interested, but wonder how a pair of 20W plate dissipation sweeps makes 100W clean whereas a 42W plate dissipation pair of KT88s in UL only makes 60-70W?
Is it the screen treatment? I'd like to try the sweep tubes for an amp, but I'm a bit lost about how a PP pair of 6CB5s makes 125W.
Can someone explain this in some form of layman?
Thanks!
Blair
tubelab has tortured a lot of these horizontal sweep tubes, and despite the low plate watt ratings on datasheets, many of these tubes can run at higher plate wattage without red plating...
higher power is realized by lower plate loads and higher plate currents, these two goes hand in hand...horizontal sweep tubes have bigger cathodes and can supply more currents than the usual audio finals tubes,
higher power is realized by lower plate loads and higher plate currents, these two goes hand in hand...horizontal sweep tubes have bigger cathodes and can supply more currents than the usual audio finals tubes,
Thanks. I'm really intrigued by this. I now have a pair of transformers that can be used with choke input for ~520-540VDC @ 500+mA along with a pair of 200W 1.5K output transformers. I'd like to try a quad of the 6CB5A tubes with all of this. They are so much cheaper than KT88 type tubes. How does the distortion compare to "audio" tubes?
Blair
Blair
There are two big factors at work here, and one has already been mentioned.
First, a sweep tube (horizontal and vertical) doesn't get a break. There is no volume control on the sweep circuit, it must run at full power all the time. In the early days of TV the average familly had one TV, and it was on for several hour a day, nearly every day. The sweep tube "sweeps" the electron beam across the face of the TV screen and if it doesn't make full power, the picture shrinks. The Horizontal sweep tube also generates the 10 to 30 KV needed to accelerate this electron beam toward the face of the screen. When it gets weak, the picture gets dim due to low high voltage.
The highest failure item in a tube TV set was always the horizontal output tube. A tube manufacturer needed to make sweep tubes last at least as long as the TV sets warrantee, which was usually 1 year. For this reason most sweep tubes were conservatively rated. This is obvious when looking at tubes that were rated for audio use and rated for sweep tube use. The venerable 6V6GT is one of those tubes. Its plate dissipation rating is 12 watts for audio, but 9 watts for sweep use.
As TV sets went color, and got bigger, they demanded more from the horizontal sweep tube. My old favorite, and the physically largest sweep tube that I have ever seen is the 6LW6. There are several versions of these, and the biggest ones have multiple heat sink fins welded to the plate. It is rated for 40 watts of plate dissipation, but I have ran one at 90 watts dissipation for hours and there was no signs of distress. So, yes, many sweep tubes carry conservative ratings compared to a similar sized audio tube. Compare the plate structure of that 40 watt KT88 to the 26 watt 6CB5A. They are about the same size, in fact some 6CB5's are bigger than some KT88's. I have run a 6CB5A at 40 watts and saw no signs of distress.....
Note: not all sweep tubes have this much margin. Some may only handle 125% overdissipation. Conversely, the early Chinese KT88's glowed red at 25 to 30 watts. The new EH KT88's have no problem with 45 watts though, and I have extracted over 100 watts from a pair in pentode mode.
Compare the size of the cathode in the 6CB5A to the cathode in the KT88, or the really wimpy 6L6GC. Which one can reach right through your OPT and yank those 15 inch woofer cones hard enough to follow a big kick drum hit???? That big cathode that contributes to great dynamics also contributes the second factor, efficiency. A 6L6GC conducting full on can pull its plate down to 75 volts, maybe a little less with a light load. A G1 driven sweep tube, 25 volts or so. I have seen and posted the scope trace of a screen driven 6CB5 pulling its plate down all the way to zero volts working into a rather hard load (3300 ohms).
A sweep tube can usually exhibit adequate linearity at about half the idle current of a similar sized audio tube. This allows for a higher plate voltage for a given idle dissipation. The big fat cathode allows for higher peak currents, so a lower load impedance can be used. I have seen 250 watts from a pair of 35LR6's (a 40 watt sweep tube) running on 650 volts into a 2500 ohm load, and 525 watts from 4 of them (1250 ohms). They just started to glow red at full power in a darkened room. They were dissipating about 60 watts each.
A horizontal sweep tube sees peak currents near 1 amp in a 25 inch color TV. It does this all day long, so it can support much lower load impedances than a similar sized audio tube.
The typical KT88 making 70 watts into a 5 K ohm load might see 55% plate efficiency. This means you need to put 127.5 watts of DC into the tubes to get 70.125 watts of audio out. This leaves 57.375 watts to be burned up between the two tubes, or 28.7875 watts per tube.
I have measured 70+ % plate efficiency on G1 driven sweep tubes. I have measured near 80% plate efficiency on screen driven sweep tubes. In an experiment with 6AV5GA's (a small sweep tube rated for 15 watts of dissipation) I was getting 80 watts of audio with 105 watts of DC input. This burns only 12.5 watts in each tube.
Details and schematic at the bottom of this page:
6AV5GA testing
- Status
- Not open for further replies.