The AM Forum

A friend of mine dropped a "project" on my workbench today - a Viking II in quite nice physical condition that doesn't work.  Anyway, I fixed enough stuff to get a good carrier, although there is some hum on it - deal with that later.

Now on to the modulator.  The owner of the rig agrees that the driver transformer has to go, so I am building a direct coupled solid state driver for the 807s, and I want to put them in a triode connected configuration.

It has been MANY MANY years since I've messed with 807s triode connected.  Any suggestions?  I have a 300 V audio swing available at the grids (+/- 150V).  I seem to remember in my deepest memory that when I simply connected the two grids together and drove 807s as triodes that it didn't sound very good.   I wonder if, with these tubes, it is better to only drive the screen, and connect the control grid to a fixed source. 

I remember something in the RCA transmitting tube handbook, where there was a resistor in series with the control grid... but that doesn't seem very sound from an engineering standpoint.  The grid is a non-linear load, and putting a resistance in series with such a load doesn't seem very smart...

Anyone ever done this?  I know 6DQ5s would be much better, but I'm not sure I could convince the rig's owner that taking out the 807s would ever be a good idea  !!

Regards,

Steve

Most of the audiophools strap the screen and plate for triode connection.

Just use 809s if you want triodes.

Stevo, that is for low pwer class A operation. IIRC the ARRL handbook sez to feed the audio directly to the screens, and tie the grids to the screens with a resistor for class B operation.


"Just use 809s if you want triodes."
Now yer talkin!!

yieee ...  Bacon is smokin' ... according to Orr and the RSGB handbooks a 20k resistor between screen and control grid gets it done ....sounds interesting, Steve ....be curious to see how it turns out ...73...John

Steve,
I have never used 807's other in Tetrode operation.  I did find this on the web showing the screen tied to plate through a 100 ohm resistor.  
Here is the link:

http://www.dvq.com/hifi/images/acro.pdf

Regards,
Joe, W3GMS

Steve I tried it  once by driving the screen with the G1 tied through a resistor to G2. Look at old handbooks for the value of the resistor. I didn't have enough drive power at the time so went back to tetrode. I bet you could also ground G1 and drive the screen.  I think the screen is going to need to hit a good +250 volts to saturate the 807 based on the stuff I did in my last V2.

Ok, here's another monkey in the wrench, Steve.
Why not just leave well enuf alone and replace the driver transformer with a phase inverter? The Window has a few or you can design a solid state one of your choosing.

IMHO it wud be a whole lot easier!

 This'll work  (http://www.amwindow.org/tech/htm/valiant/audio.htm)

My first plate modulated transmitter used the triode 807 configuration.  I used the 20k resistors.  It would develop good modulation peaks but the audio was crappy because of the modulation transformer (from an ARC-5) and poor driver stage (single 6AQ5).  The driver tube had too much voltage on it and distortion got worse after about 60 seconds, so I fixed it by inverting the tube into a glass of water.  I used that thing for several months before building a better version using a Stancor modulation transformer and 6L6 driver stage with feedback.  Probably ended up with about the same audio quality as a stock Valiant or Viking II.

From when I was a penniless university student in 1960.

Well, I'll be! A liquid-cooled 6AQ5!

I'll give it about two years before the audiophools pick up on that idea as a way to reduce distortion, and about two weeks after that before they start burning their houses down.

You bring the marshmallows, I'll bring the weenies!  ;D

Excellent!

I have a rubberized medium for sale, that when added to the cooling water, completely eliminates Eddy currents.

Sorry for the hijack Steve I couldn’t help it.

Allright, allright, back to the original thread  ;D

Stuck with the original modulation transformer, tetrode connected tubes won't be able to develop enough swing to give high modulation in the positive direction....  You can get a lot closer to 0V (tube fully on) with a triode than with a tetrode, and that's what I'm trying to do.

I'm going to have to do a bunch of experiments.

On the plus side, the result will be a practical direct coupled driver for class B grids  8)

So far, it looks good - uses only one "real" power supply (I'll use the +300 volt supply in the viking II) and a negative supply just a little more negative than the negative bias needed (it can be lower - more negative, but no higher)  Everything in the circuit is at 0V (ground-chassis) reference and there's no unusual circuitry.  This thing is dirt simple !!

I'm going to put a loop of feedback around the mod transformer back to an earlier stage.  We'll see !!!  Also will  add a 3 diode negative peak limiter as a final touch.

Regards,

Steve

The only way you can add feedback around the mod transformer is to control the phase shift of the driver to around 20 degrees over the audio range.
Tom and I had some interesting tests to get there and it was cool to have the simulation match the test results.
Best to simulate the circuit. Sounds like a good idea to just put a small negative bias pulldown rather than a high negative voltage. 20K ohms seems a good number for resistor value if I remember. But you will need to get the screens to about 200 to 250 volts to saturate the tube. I dropped my screen down to get more plate voltage swing. I was driving the G1 to about positive 15 volts.

That's the problem.  You can't pull the plate voltage down less positive than the screen or control grid at the crest of the peak. I believe that is the purpose of the 20k resistor, to adjust the control grid voltage when driven positive to give the maximum plate current at the lowest grid voltage. Maybe you  could experiment to see how much grid swing is required to saturate the plate current.

The 809s is a good suggestion, if they will fit into the space.  They have the same plate dissipation rating as an 807, and the filament is 6.3v, but IIRC, the envelope is physically the same as that of an 811A. The filament current is almost 3 times that of the 807, so that could be a problem unless an auxiliary filament transformer is added.

http://www.shinjo.info/frank/sheets/049/8/809.pdf

I'm running Zener regulated screens in my V2 CDC and the screen dissipation is getting a bit high when you drop the voltage too much. My G1 voltage is limited by the level pot of all things. I have considered a solid state driver after the nice one Tom and I did for his rig. I would think the 20 K resistor allows the grid dissipation to track so nobody gets cooked. fc 

Steve,

 The attached file shows the 6L6GC plate curves where G1 is grounded, and G2 is the control element. The 807 and 6L6 have similar if not exact curves, although other ratings are different mainly due to base, plate cap, plate material, etc.

  At first glance, the curve shows that G2 at about +50v quiescent gives ~ 30ma plate current (600v X .03A = 18W PD). Swing G2 up to 300v peak, and you got 1/4 ampere plate current flowing when the plate voltage is << 100v.  Looks like the G2 drive needs to be upwards of 250v peak, or 500v P-P!!

   Yes a 6DQ5 would be much better, and done the same way will behave a lot like an 809.

Jim
WD5JKO   

Interesting curve. My old V2 PDM ran 6DQ5s. I got the saturation voltage down to about 60 volts with G1 grounded and about 90 on the screens.
The 6DQ5 doesn't need a lot of screen voltage so you can run it lower and get more clean plate swing but Steve is stuck with the 807.
These curves match my guess that you need to swing 200 to 250 screen volts.

How do you pull the plate voltage below 100 volts while G2 is driven to +250 volts?  The instantaneous plate voltage cannot be pulled down below the instantaneous screen voltage.

  Don, Good question! I think if the screen current curve had been included, once the screen rises much above the plate, the screen current rises very rapidly. This is an area where beam power tubes have an advantage over pentodes. Maybe I'm wrong, but one reason the 6L6 took off over pentodes of the day (6F6, 42, 6K6) was that they had higher efficiency. To get the higher efficiency for the same B+, the plate swing had to be larger, hence over a portion of the curve, the plate voltage went further below the screen grid than did the old pentodes..

  Back to those QIX P-P 807's, a SS driver capable of 400-500v P-P drive with a low impedance would drive the 807's to saturation where we get an Eb min of << 100v at EG2 max of ~ 250-300v. I would not be surprised if the peak screen current was 100ma or more at the peak level of drive. Even if it was the average drive power would be a few watts at most, but might peak at 10X that for just a small portion of the audio cycle.

Edit, I found another plate curve that shows screen current at VG1=0v, and VG2 fixed at +250v. Looks like G2 current peaks at ~ 20ma when plate voltage minimum goes < 100v.

 

Jim
WD5JKO

Just a thought to ponder:

Has anyone ever thought about using 6080s as modders? ? They would prolly need a lot of grid drive, but they have the lowest plate resistance of any tube I know of.
And they are designed to handle high plate currents.

having such a low plate resistance, the plate / grid curve should be very linear.
I have pondered using them for cathode modulation, but they might work very well as class B plate modders if you could come up with enough audio to drive them.

They are pretty wide but a cool bottle.

The 6C33 is another REALLY REALLY low impedance tube.  I think it's the 6C33C-B that's the best, but it has been a while since I've closely examined the ratings.  I do have a pair of them here - somewhere in my endess piles of ----- stuff !

Frank,
I seem to remember the 6080 used as shunt regulators in some vintage regulated Lamda power supplies.  They would make sense in that application.
Regards,
Joe, W3GMS

Joe,
      most of the stuff that I have seen with them used them as pass regulators. As shunt regulators, they would prolly pull enough current to fry something  ???  ???

I seem to remember seeing them in HP stuff. I would think pass tubes in the negative lead. They would also work as shunt regulators with a current limit resistor in the plate lead. Actually is was fairly easy to saturate the 807 on the V2. I would think it is more modulation transformer limited.

that always seemed to be the trouble with Valiants.

The old Tektronics scopes which weighed a ton used 6080s.

ah yes the 545

There are plenty of schematics for using 6080's as audio amplifiers out there on the tube audio sites.

They supposedly work very well.

About 8 or 9 watts output from push pull circuits in class a. 2K output impedance on the output transformer.

That would be more than plenty for 807s and perhaps you could parallel two 6080s and drive tubes like 805s. Sure would be cheaper than 2A3s.

Jim, I'd say your beam deflection theory is right. Maybe the deflection plates aim electrons directly to the plate and away from the screen, allowing the plate voltage to drop to a lower value.  I have some articles on the theory of how beam deflection tubes work that I'll have to review.

I don't have any 6080s, but I do have a bunch of 6AS7Gs, a similar tube.  The problem with using the 6AS7s is the maximum plate voltage rating.  IIRC, it's something like 150 volts. I have used 6AS7s as bias regulators, but never tried them as class B drivers.  They ought to work well in a lower power modulator that requires less grid voltage swing.

Fortunately for me, I had already accumulated an adequate supply of 2A3 spares long before the audiophools discovered them.

Here is a schematic for the 6as7 as a push pull audio amplifier dating all the way back to 1949.

It uses a tube rectifier with a 800v ct power transformer and two stage choke input filter. So it must be running on something in the 250 - 300v range.

Perhaps the tube can handle a bit more voltage in this type of circuit.

http://www.bonavolta.ch/hobby/en/audio/6as7_1.htm

Dave

  Dave, It is always cool to see these old circuits. I wonder though, the 6N7 audio driver plate load resistor, R6 0.1 meg-ohms, seems like a typo. I bet it was supposed to be closer to 5-10K WW power resistor.. It just doesn't look right with such a high resistance as a plate load on a high level audio driver tube.

Jim
WD5JKO

Use a 6336 instead. Good for 400 VDC.