• 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.

Those Magnificent Television Tubes

Perhaps this has been covered before but i can't find it. What happens if we tie g1+g2 together and treat them as control grid? Do we simply end up with a beam triode; i.e high rp high mu triode like 6HV5, 6HS5?

Do not ask me why :) but this operation mode is also called the "right handed triode". Google will bring you some hits, there are some curves for LS50/GU50 in this mode.
 
Have a search for »twin« or »crazy« drive.

My experiments called this "dual drive." I made an amplifier using dual drive several years ago, but haven't explored it much further recently. The output impedance is definitely as high or maybe higher than pure pentode operation. Tube to tube variation is much worse than driving G1 or G2 alone. Some means to adjust the idle current, and possibly the Gm is needed for each output tube (4 tubes or more in PPP per channel).

The biggest difference between dual drive and a beam triode is that you don't need high plate voltages to get any current flowing in the tube. The 6HV5's and others require positive grid voltage or 600+ plate volts.
 
what i mean is apparently called right handed triode

I think that grid1 tied to grid2 operation would only apply to tubes with very low gm1, like some old transmitting tubes. That would maybe give a more effective total gm than grid 1 drive alone.

Modern high gm1 tubes would likely burn up if g2 voltages were put onto grid 1.

The "dual drive" idea was to proportion the drives to the two grids according to how g2 only drive versus grid 1 only drive would operate. (roughly the internal Mu ratio) George experimented with this. Will give another 25% to 33% effective gm over grid 1 drive alone for typical internal Mu 3 to 4 Sweep tubes.

The "Twin/Crazy drive" thing came from an old RCA Ham Tips article. (no idea if there are earlier references) The J. H. Owens article (May-June, 1947 RCA Ham Tips), called it "special triode". See .pdf below.

It had a resistor between grid2 and grid1 of an 807 modulator tube. It shows graphs of the increased gm achieved but no mention of increased linearity. Tony found a more recent amplifier, initially, using a 4D32 with this setup, and further searching found an amplifier with a 12G-B7 in this mode. None of these mention anything about increased linearity. Only after the configuration was put onto a curve tracer did it become apparent that the R connection could be optimized for increased linearity besides increased gm. Further testing with the curve tracer made it apparent that a bottom 2nd resistor, grid1 to cathode, was helpful, and also that the initial drive signal needs to go negative to cathode somewhat, to overcome plate voltage effect for full turn-off. Vneg = 2 x (B+) x 1/(gm1 x Rp)
The auto-biasing Rk in the two cited amplifiers helps with this for class A SE operation.

http://n4trb.com/AmateurRadio/RCA_Ham_Tips/issues/rcahamtips0702.pdf

.
 

Attachments

  • 4d32_2.jpg
    4d32_2.jpg
    141.8 KB · Views: 633
  • 12GB7SE-screendrive.gif
    12GB7SE-screendrive.gif
    19 KB · Views: 672
Last edited:
this operation mode is also called the "right handed triode

Oh, that's coming back to me now. This was from some discussions between OldEurope and Kenpeter, referring to operating a tube like a bipolar transistor, with current drive to g1 and/or g2 (in the positive V region). Attempting to get constant current gain instead of constant Mu V gain. It's another interesting region of tube operation that could be optimized.

It is strongly related to (mostly unknown) vacuum tube current mirrors, using a diode for 2/3 power I to V conversion, followed by a tube grid for 3/2 power V to I conversion. Resulting in linear I_input to Ip output gain, like a bipolar transistor with current gain Beta. (after receiving some complaints about refering to "bipolar" tubes, their discussion was "coded" as left handed, or right handed mode of operation, leaving everyone else behind! Hehe..)

Seems like "left handed triode mode" would fit better? (the left side of triode plate curves (positive g1), versus right side (neg. g1), is where this notation comes from)

In the more normal realm, Voltage drive to g1 and g2 tied together would have to limit itself to a safe +V region for grid 1, and so grid 2 would only add 1/Mu (internal) additional gm2. So this would essentially look like a beam triode (with no grid 2). Not too useful, unless internal Mu is really low, like 1.1 to 2 range.

Twin/Crazy drive is actually using both modes of operation together. V drive for grid 2 and current drive for grid 1. The dropping off of grid 1 Beta (current gain) is what linearises the normally rising grid 2 transconductance.

Haven't heard from OldEurope or Kenpeter or MichaelKoster in a long time.
 
Last edited:
One might also mention the "Aikido" 1st gain stage as another example of bipolar, or "left handed", or current mirror operation. Just reverse order from the tube current mirror scheme.

The bottom triode (or could be a pentode) performs 3/2 power V to I operation, and the top triode load performs 2/3 power I to V loading. The result being linear V to Vout gain. I doubt if many actually understand the un-usual mode of operation going on there. Have never seen any user extensions of that, using a pentode on the bottom for example. Or using different tube types for top and bottom for higher gain.
 
Perhaps this has been covered before but i can't find it. What happens if we tie g1+g2 together and treat them as control grid? Do we simply end up with a beam triode; i.e high rp high mu triode like 6HV5, 6HS5?

This connection gives you a high-u power triode like an RF "zero bias" triode like the 811A. You don't want to do this for audio power, but it has often been done for RF power finals. For audio finals or AM plate modulated RF finals (this is an audio application) you want triodes with low-u factors and low rp, and that's what you get when you connect the screen to the plate.
 
Oh, I forgot to mention, that the Dual drive scheme (with approx. Mu proportioned drives to g2 and g1) reduced the maximum peak grid 2 voltage by 1/2, versus grid 2 only drive, required for peak output. This is a significant safety feature for the tube's grid 2, as well as reduced requirements for the preceding driver tube.

Twin/Crazy drive seems to typically reduce the peak grid 2 voltage by around 25% to 33% from grid 2 only drive, so is helpful there too. (measured for Sweep tubes at least) One has to be careful however about grid 1 dissipation, since it is being driven positive for Twin/Crazy drive. (Dual drive usually has grid 1 negative still) But it only takes maybe 0.01 to 0.1 Watt of diss. for grid 1 for Twin/Crazy drive.
A frame grid 1 might be of concern possibly. I have tried the frame grid 12HL7 with Crazy/Twin drive on the curve tracer successfully. (for a few minutes at least) The saving grace there is that the consequent high gm1 only needs a tiny positive excursion to do its job. (and resulting total tube gm linearity was excellent)

a) 12HL7 Crazy/Twin
b) 12HL7 g1 drive
.
 

Attachments

  • rsz_12hl7_g2_g1_80v_7vs.jpg
    rsz_12hl7_g2_g1_80v_7vs.jpg
    63.6 KB · Views: 501
  • rsz_12hl7_100v_4vs.jpg
    rsz_12hl7_100v_4vs.jpg
    65.4 KB · Views: 498
Last edited:
significant safety feature for the tube's grid 2

This is the limiting factor for pure screen drive. If G2 gets warm enough to start emitting electrons, those electrons will be immediately attracted to the plate which is on the average, far more positive. This will start a current flowing from G2 to the plate which can not be controlled and within a few milliseconds, there will be a flash, followed by the aroma of fried parts, or worse, in a high powered amp, the shrapnel of exploding mosfets, capacitors and resistors. This happened to me a few times before I captured the event on two storage scopes, and figured out what was happening.

Driving G2 positive while the plate gets pulled to zero makes G2 very hot since it eats most of the cathode current. On the next half cycle that hot grid is at zero volts while the plate is somewhere north of 1KV....tube arc, parts explode, game over!

Two scope pictures. 50 volts per division, 0 volts is the second horizontal line from the bottom. This is a screen driven 6CB5 push pull amp with 550 volts B+, and a 3300 ohm OPT.

The first picture shows the amp being driven to a clean 80 watts out. The plate voltage (rounded trace) being pulled down to about 50 volts. The screen voltage is the flat bottomed trace. It swings from zero volts to about 180 volts. The amp runs fine at this level.

The second picture shows the amp being driven to a clipped 130 watts out. The plate voltage (distorted trace) being pulled down to 0 volts, settling to about 30 volts before heading toward 1200 volts. The screen voltage is the flat bottomed trace. It swings from zero volts to about 230 volts. The amp exploded but this trace was still in the scope's memory.

The third picture shows the test amp before blowing up, running at some mild mannered power level.

The fourth picture shows a mosfet with most of its case gone.

Dual drive began as a method to eliminate these kinds of events. This may never happen in "normal" use, but I have never been "normal" and any amp I build must be able to survive me plugging my guitar into it, and "setting the controls for the heart of the sun."
 

Attachments

  • BlownMosfet.jpg
    BlownMosfet.jpg
    278.6 KB · Views: 180
  • TestAmp6CB5A_2_a.jpg
    TestAmp6CB5A_2_a.jpg
    552.6 KB · Views: 185
  • PlateAndScreen_130W.jpg
    PlateAndScreen_130W.jpg
    288.5 KB · Views: 106
  • PlateAndScreen_80W.jpg
    PlateAndScreen_80W.jpg
    278 KB · Views: 433
Last edited:
Putting a HV rectifier/diode in series with the grid2 might prevent that kind of run-away. Haven't tried it. One might wonder if it would pick up some capacitive + voltage on grid 2 then with the diode being off (from the plate swinging to 2 x B+) until conduction starts again. May not make any difference. Could always put a high value resistor across the diode I guess.
 
Last edited:
Hello Folks,
I just like to share my recent amplifier.
It uses television tubes throughout.
The circuit is inspired by the Synola 509 design and uses screen-driven PL519 (40KG6A) with fixed bias on g1.
I changed the preamp and driver tubes to PCL82 (16A8) and employed EY88 rectifiers. The output transformers are the ones originally designed for the Synola amplifier by Reinhöfer Electronic (1.7k primary).

The amp sounds very good, delivers about 10 Wpc.

I might want to try converting it to crazy drive just to see the difference.

Philipp
 

Attachments

  • PL519SE.png
    PL519SE.png
    17.5 KB · Views: 713
There's really no need to DC heat the PCL82. About 15 years ago I've built a similar stereo amplifier, featuring PL519's and PCL82's (as your's), PY82's as two way rectifiers and EAM86's as VU indicators. I've connected the heaters of one PCL82, one PY82 and one EAM86 in series, to feed them form the same 40 Vac windings that supply the PL's heaters. No hum at all.

Best regards!
 
re: Jazbo8:
Ahh, the Seybold article, figure 6, as you show.
I guess I goofed trying to follow the page continuations there.

---------------

Don, what if instead of single ended, this tube amp is made into a push pull amp instead?

Not sure which tube amp you are referring to. Crazy/Twin drive should work fine for P-P class AB too. The idle current is likely lower using Crazy/Twin than for conventional grid 1 drive for smoothest crossover, but should be similar to grid 2 drive examples. You only want to (class A) overlap the lowest portion of the plate curves (for class AB) where effective gm is dropping off, otherwise it will produce gm doubling in the overlap region.

Class A P-P would work fine too. gm doubling everywhere, so constant gm.

Mickeystan built a P-P Crazy/Twin Drive amplifier here:
Those Magnificent Television Tubes
 
Last edited:
Oh, the 4D32 Amp.
I don't see why it couldn't be converted to P-P. Only thing is the CFB in that design is really making for excessive drive voltage requirements at grid 2. Not sure how much the CFB helps if one already has the stage gm linearized. (maybe it is mainly for lowering the output Z, since Twin/Crazy drive alone is high output impedance)

The CFB may be a small percentage there, since it is being played off against g2 drive voltage, so not having a lot of N Fdbk effect. Just making the drive V situation worse. The driver distortion may go up faster than the CFB brings down the output stage distortion.

One -WILL- have to address the output impedance in some way, since the Twin/Crazy Drive is high output Z by itself. Getting sufficient CFB to do much Z lowering though will incinerate the driver stage when combined with grid 2 drive requirements. Like a total cathode follower output stage running on grid 2!!! Suggest some other (global) approach. Or "Schade" back to the driver cathode.
 
Last edited:
Oh, the 4D32 Amp.
I don't see why it couldn't be converted to P-P. Only thing is the CFB in that design is really making for excessive drive voltage requirements. Not sure how much the CFB helps if one already has the stage gm linearized. The CFB may be a small percentage there, since it is being played off against g2 drive voltage, so not having a lot of N Fdbk effect. Just making the drive V situation worse.

i plan to use the newer IXYS depletion mosfets, 500V 6 ampere parts, 100mA should be easy with these newer devices...they will bias the cathodes of the 4d32...
 
I guess just having some links to the other thread in this thread are sufficient. Whatever.

---

Any luck finding the missing page in the Susumu MJ 1999/2 article?

While browsing through the MJ archive, I came across the "HY65 CF Direct Coupled G1 G2 Simultaneous Drive Ampliifier" by Soya Susumu, which looks like the "Crazy Drive" scheme from smoking-amp, unfortunately, page 121 is missing from the article, thus leaving out an important part - the Ep-Ip characteristic of the HY65 with both G1 G2 driven. Would anyone know how to get a copy of the missing page?

Here is the link to the article: MJ^9902_(04)HY65-single-power-amp_SoyaS.pdf - Google Drive

Looks like all the schematics using the Twin/Crazy drive scheme in the article end up with a big capacitor between grid 2 and grid 1, which won't work for linearization. (and would be dangerous for grid1 !!)
 

Attachments

  • Susumu_MJ_(04)HY65.png
    Susumu_MJ_(04)HY65.png
    143.9 KB · Views: 530
  • Susumu_MJ_schematic.jpg
    Susumu_MJ_schematic.jpg
    99.7 KB · Views: 498
Last edited by a moderator: