• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Compactron Output Tubes

Hey guys,
Have any of you wandered across any compactrons suitable (or equivalent to another) for use as an output tube? I see a few here and there, but there aren't any curves for them. There are also some with fairly decent looking pentodes in them and an associated triode/diode. I wouldn't mind using the pentode only section I suppose.

I would like to stay with 12 or 6 volt heaters, and also the 6T9 is not beefy enough.
 
I picked up a pair of Novar (9pin) compactron sockets at Hamcon.
I have three 6V beam power tubes for that pinout, I hope at least
two are still good. Two of the boxes are marked "low", but
there is no telling for sure but to try em out.

My SICO emission tester has no Novar socket, I may have to rig
up some kind of an adaptor... I think the other compactron had
12 pins... I don't have any of those.
 
Most any of the TV horizontal output compactron tubes are suitable for audio outputs as long as you are not using the typical Ultralinear mode from xfmr taps (since the screen voltage ratings are usually lower than the plate V ratings). The 6 and 12V heater ones have been pretty well picked over by now though. Lots of odd fil. voltage ones yet though. Sockets may not be too easy to find. Best to find some sockets first before ordering the tubes.

Don
 
6LU8 or 6LR8 are dual compactrons. Small signal triode makes a good driver, large signal pentode makes a good output stage. Search the forums for that type number to find some schematics. They make about 15 to 20 watts in P-P and 2 watts in SE triode 5 watts in SE UL. 6JZ8 is their little brother.

As mentioned before there are dozens of single pentodes that were designed for use as the horizontal output tube in TV sets. The 6LF6 is the granddaddy of them all a pair can crank out 150 watts in P-P pentode mode. 6JS6, 6JE6, 6LQ6, 6KD6, 6KG6, 6KN6, 6LB6, 6LR6, 6ME6, 6MJ6 are all slightly smaller (30 watts dissipation) versions of the same idea. 100 watts from a pair is possible. 6JG6, 6JM6, 6JN6, 6JR6, 6JT6, 6JU6, 6JB6, 6GY5, 6GV5, 6GT5, 6GJ5, 6GE5, 6GB5/EL500 are all horizontal output tubes in the 15 to 20 watt range.

Most of these tubes have a relatively low screen grid voltage rating. Sometimes this rating can be blatantly violated, sometimes it can't be pushed at all. Carefull observation of the tube at hane is needed to find out. I have tested the 6LF6 and its octal brother the 6LW6 to 500 volts in triode mode without issue.
 
  • Like
Reactions: 1 user
I have tested the 6LF6 and its octal brother the 6LW6 to 500 volts in triode mode without issue.
Tony's recent "like" on my 14 year old post reminded me that some info posted long ago is not valid. I realized long after this thread went silent that my 6LW6 in triode mode amp did not exactly work out as planned. The amp worked great, but it ate 4 tubes in less than two years. I wound up putting the EH KT88's back into the amp and ran them at a similar 80 to 100 mA bias point with just over 400 volts across the tube. The EH tubes are still alive and well but have lost their neat blue glow that danced around on the glass.

It seems that some older design sweep tubes can live with excess screen voltage, but the last of the big tubes don't like it. All died due to a runaway event while idling. The SSE amp was used to power my computer speakers for several years, so it saw a lot of idle time. After two tube failures I lowered the overall B+ voltage to about 350 volts and still had two failures. At idle, the output tube does not know if it is in triode, UL or pentode mode since all of the element voltages are constant and roughly the same.

Sometime after these failures I had a 6W6GT die by red plate runaway, again at idle. It was running well inside the published specs for TV vertical sweep use, yet it died. It was a crusty old tube, so I got five brand new 6W6GT's and wired them up to a power supply and let them cook. After about two months of steady state idle one began to run away. I stopped the test before it died and experimented on it, and it was exhibiting signs of excess control grid current probably due to "gas."

In a TV set the sweep circuits never idle. They run at full power all of the time. This lowers the average plate and screen dissipation compared to idle in a class A design. Some users have used big sweep tubes in triode wired push pull amps biased well into class AB1 without failure.

The 6W6GT is exactly the same tube as the 50L6GT used in countless AA5 radios except for the heater voltage. Those things can sit at idle or low output for years, but the screen grid sees about 120 volts.
 
I used to have the odd sweep tube do exactly as you described. I sorted it by using lower value for grid leak (100k) and automatic grid bias.
I haven't has a tube run away in any of my amps in several years now.
I only use 320V B+ though. I don't know what they would do with 400V B+.
Oh, and they are all push pull, biased between 35mA and 120mA.
 
The screen grid sees about 120 volts, and the plate not much higher, if at all in those designs. Stable ideal bias appears to require keeping the screen voltage AND plate dissipation (bulk heat) in check. If you only need a fraction of the maximum possible cathode current, then dropping the screen is possible, at least in regular pentode mode. All the sweeps I’ve experimented with so far seem to be happy at 120 volt screens - but the bigger they are the less linear they (or a pair in push pull) are with them running that close to “pinch off”.

The smaller ones seem to be fine idling at or near rated plate dissipation - many of them nice and linear at those operating points. *JA5, *LU8, and *JN6 deserved enough attention for me to have them on the breadboard for a while. But those make 30 to 50 watt amps. That LU8 pentode is sort of a 7591 equal, and the JN6 can almost sub for an EL34 (at sane Pdiss and 120V screen).

I started playing with the old 6CB5, and those were pretty. But I can’t seem to get enough of them to matter, so the 6(26)DQ5 became my next victim. They seem to act an awful lot like a 6550 when you run the screen at 120V. Happy enough at 25 mA. I tried running at 50, and it does sound better at high frequency but I don’t have enough running hours on it that way to KNOW that it won’t run away. I‘ve got some “big plans” to run the bigger ones at kW-ish levels, but I’ll probably run the screens down at 100 and use lots of global NFB to deal with the higher intrinsic distortion. If the DQ5’s stay stable at 25+ watts idle dissipation they will end up in the “tweeter amp”. And if I end up retooling my 6550 monoblocks with them because the 6550’s become unobtainium they are the first choice. The 21LG6 looks like another good candidate, and I’ve already secured a good supply of them.

Koda’s circuits (above) have mostly used the Russian (old stock) TV sweeps. They may very well be more rugged. And an automatic bias circuit could re-adjust Vg1 continuously as the tubes age. Turning it “down” as required to prevent runaway. Lower grid leak will help in ANY event - even some of your basic audio tubes have lower Rg1 requirements (Which were often violated, and the new production 7591 EH’s didn’t like that very much and would run away).
 
I started playing with the old 6CB5, and those were pretty. But I can’t seem to get enough of them to matter, so the 6(26)DQ5 became my next victim. They seem to act an awful lot like a 6550 when you run the screen at 120V. Happy enough at 25 mA. I tried running at 50, and it does sound better at high frequency but I don’t have enough running hours on it that way to KNOW that it won’t run away. I‘ve got some “big plans” to run the bigger ones at kW-ish levels, but I’ll probably run the screens down at 100 and use lots of global NFB to deal with the higher intrinsic distortion.
I have a box full of 6CB5's but only one of them looks like a KT88 with a plate cap. Too bad they do "look" powerful.
I have a good size collection of used 6DQ5's collected from hamfest bargain bins for $1 or less. I also have a few new 26DQ5's that are International brand "Made in Japan." Their melting point is considerably lower than the used USA made tubes. I do not trust them at high power.

Virtually all my horizontal (line output) sweep tube experiments that used tubes designed in the late 50's or after use 150 volts on the screen.

I also have "big plans" for a 1KW (500 WPC) tube amp. I have no need for such an amp, but I have all the parts to make it except for a chassis that's big enough and light enough. I will use 26LW6, 36LW6 or 26HU5 tubes because they work and I have seen over 300 watts flow from one pair! The chassis size will determine whether each channel gets 2 pair or 3 pair of tubes. Post # 55 and 56 here have the details:

https://www.diyaudio.com/community/threads/unset-is-coming.340856/page-3

The cheat sheets that show the contents of some Compactrons are included. Somewhere I remember reading that the pentode in a 6LU8 / 6LR8 is a 6EZ5 and the triode is a 12AT7. Both can be used in the SSE, so that led to the experiment shown in the picture, which led to the board shown in the next 2 pictures.
 

Attachments

  • GE Compactron List June 1962.pdf
    583.2 KB · Views: 59
  • GE Compactron List Spring 1964.pdf
    116.8 KB · Views: 60
  • Furball_3.jpg
    Furball_3.jpg
    125.6 KB · Views: 76
  • PCboard.jpg
    PCboard.jpg
    126.3 KB · Views: 77
  • Spud_Amp_2.jpg
    Spud_Amp_2.jpg
    109.5 KB · Views: 128
Last edited:
  • Like
Reactions: 1 user
All my DQ5s are the IC “Made in Japan”. They seem to run fine with no increase in bias when running at 475V/25 mA. They made the basis for a 130W bass guitar amp using a quad. I’m thinking with an increase in bias and a little more B+ I could push that to ~150 ish for 2” compression drivers. If that fails, there are the 21LG6’s which I have about 3 dozen of.

The LU8/LR8 looked interesting - especially with a 12AT7 triode in there. But unfortunately, I found out just how poorly matched those triodes were between any 2 given tubes. Not an issue with running one as a gain stage and the other as a concertina, but trying to make an LTP phase splitter out of two just wasn’t happening. That design got reworked around the 10JA5, which is the same pentode. And 6BQ7’s which I could match. Not as nice as that 12AT7 triode, but two 40-mu‘s gives as much gain as a 60 and a 20. It’s sitting in a semi-working condition waiting on my new work facility, however long it takes.
 
"New" 30 to 35 Watt COMPACTRON kids on the block for output tubes, 6JD5, 6HZ5, 6JH5, 6HV5, 6HS5, 6JK5 TV Beam Triodes with special "Flat Plates".

Most of the Beam Triodes around have traditionally been limited to Multi-KV use for Electro Static speaker Amps. But a select group of these tubes have a different plate structure that allows use at as low as 600V B+ (When using up to 6V positive grid drive). They have the bog standard old fashioned flat plate near to the beam plate.

While the multi-KV types have a long "boxlet" or "cavity" at the sheet metal joint facing the electron beam, next to the plate supports. Pics below. These "boxlets" or "cavities" normally have a couple of Barkhausen fins inside them to eliminate secondary electrons by multiple reflections within. Many Horiz. Sweep tubes have them. with resulting smaller kinks in their plate curves. Unfortunately, the Beam Triode tubes are missing the Barkhausen fins inside, leaving just the empty boxlet, placing the exposed plate (thru the beam plate slot) more than 2 or 3x as far away from the grid/cathode assembly, Leading to the multi-KV plate voltage requirement.

The "Flat Plate" Beam Triodes seem to be largely confined to RCA types, but NOT ALL RCA types are that way. A few Sylvania and Zenith tubes have the flat plates too. Oddly, all of them have round "ventilation" holes (ha-ha) in the skinny plate edges. (helpful for IDing them)
I have one odd case of a Zenith tube that has a "thin" boxlet/cavity (room for just -one- Barkhausen fin inside, but no fin present) that tests quite -near- to the flat plate tube curves. It has the more common square "ventilation" holes along the skinny plate edge.

Oh, another point to keep in mind with these is a typical 7500 Ohm Ri for the plate. (the 4600 Ohm Ri on the datasheets is at a pulsed 300 mA) I just got an 8K OT in to try these in a 2+2 parallel P-P class AB 100 Watt amp. Since the frame grid 1 is being used in positive territory, +6V, 200 mA, with 1.2 peak grid Watts, I plan to drop the heater power down by a Watt to compensate.

A) flat plate pic
B) usual Boxlet/Cavity plate
C) Flat Plate curves 20 mA/div Vert., 100V/div Horizontal, +0.5V grid steps, 4th curve up is 0V on grid 1
D) usual Boxlet/Cavity type plate curves, same 20 mA/div Vert., 100V/div Horizontal, +0.5V grid steps, 4th curve up is 0V on grid 1
E) Flat Plate curves with a HV Mosfet driver combining grid current with plate current for lower "knee" voltages. Excellent curves!

6JH5_flat_plate.jpg


6JH5_box_plate.jpg


6JD5_ZEN_Half_norm.JPG


6JK5_norm.JPG


6JH5_Mosfet.JPG
 
Last edited:
I did a version of Crowhurst's Twin Coupled Amp with a pair of 6LU8s.
~35W on 400V plates, 300V on G2s. So Unity Coupled just like a McIntosh,
but this one also partial UL. Very good measurement results on the bench.
 

Attachments

  • Twin Coupled Amp Front for RHS Veiwing.jpg
    Twin Coupled Amp Front for RHS Veiwing.jpg
    195.6 KB · Views: 148
  • Twincoupled Amp B.jpg
    Twincoupled Amp B.jpg
    136.7 KB · Views: 142
  • Like
Reactions: 1 user
Somewhere I remember reading that the pentode in a 6LU8 / 6LR8 is a 6EZ5 ...
Not so, the pentode in 6LU8 has gm 9.3 mA/V while 6EZ5 is only half of that 4.1 mA/V ...
its the 6MF8 which contains a 6EZ5, though;
also, some more nice compactron tables ...
 

Attachments

  • Compactron_Survey_v1-06.pdf
    3.1 MB · Views: 66
  • Compactron_Table_1.png
    Compactron_Table_1.png
    120.4 KB · Views: 84
  • Compactron_Table_2.png
    Compactron_Table_2.png
    118.9 KB · Views: 86
  • Compactron_Table_3.png
    Compactron_Table_3.png
    159 KB · Views: 85
6JC5/6JB5/6HE5 vertical deflector tubes. The JC5 is good for 19W, the JB5 15W, and the HE5 12W. The plate is rated for 350V, and the screen for 300V. I've exceeded the voltage ratings with Sylvania 6JC5 with big beefy plates with over 400V on the plates and about 390V on the screens in a push-pull guitar amp. It sounded really clean, but the overdrive tone wasn't great. I have a 6JC5 SE amp with a 5879 driver, and it's wonderful.