The AM Forum

The schematic shows the CT of the modulator and final filament transformers being switched by a DPDT relay.  In the transmit position they are connected to ground, in standby they are connected to ground through a 50k resistor.

I am not familiar with this scheme, how does it work?

http://www.amwindow.org/tech/htm/813/813.htm


Rich

Quite well!
I use the same setup on a 4-400 and a single 813.

Now, the theory? I think the cathode is now far enough above ground to stop the flow of RF. (Biased "off")

Someone chime in please and correct me if I am wrong.

Bill KA8WTK

I was wondering about the pros/cons of this method too. Wouldn't you need to use an HV relay in this application?

You're right,  I've seen that setup that JJ uses,  the tubes are biased off.  

The only thing that may become a problem is the no load condition on the HV plate supply.  The high voltage may rise above the voltage ratings of the filter caps.  Just something to keep in mind.

Fred

Yep, this is a very common method used in many commercial and amateur linear amplifiers. I've used it exclusively on all my rigs, class C plate modulated or linear for 40 years now. (Brrrrrrrrrrrrrrrrr)

When unkeyed, the 50K resistor drops a voltage across it and this produces a positive bias on the cathode. This acts as a negative bias relative to the grid and cuts the tube off.

You can use virtually any 1A or more relay there. There is not much voltage there when keyed or unkeyed. I sometimes put a 0.1 ufd non-polarized capacitor across the resistor to help absorb the spark on the relay contacts and also reduce keying transients.  It will not be an issue with a small rig, but on big mauls everything gets magnified.

I like this method for its simplicity. Plus I see it used everywhere in the handbooks for big linears, so that makes me feel good.

BTW, the 50K resistor is not critical at all. I've used anywhere from 30K to 100K as long as the tube gets cut off. The only problem I've seen is when I built a TPTG Hartley transmitter. It required about 200K to cut it off completely. But most rigs will be FB at 50K.

 T

Bill,
   Biased off, that makes sense.  I think a normal relay is ok as it is one the ground side.
Rich

'68 or '69 was the first time I used this method. You want the antenna relay closed before you bias the tubes on and you want the antenna relay to open after the power control turns off. Reason you want a load on the rig before you key the power and you want the antenna relay lag to open so all energy is dissipated before the antenna goes to RX. This avoids big sparks.

Yep, sequencing a BIG rig is important as we both have found, Frank... ;)

Jay/ W1VD makes a very nice solid state sequencer that will handle four different relays. I used one on my class e-rig and it worked FB.     The 4-1000A plate modulated rig took some more work to tame the arcs across the mod iron gaps. As you said, Frank, it was all a matter of dumping the power into the antenna during unkey and then having an antenna load present during key up.  Adding some C across the cathode keying resistor helped the problem too. Getting a big plate modulated rig to key and unkey cleanly without mod xfmr or Heising reactor sparks takes some time to achieve, but is worth it. There is big risk having unquenched sparks across the iron.

T

Oh man, I was right for a change!

Reminds me of the concern I often see expressed regarding cathode keying of a final for CW. Some are fearful of the "full plate voltage" across the key contacts in key-up position. But when the cathode circuit is broken, the voltage rises only to the cut-off bias voltage of the tube.  When you place your hand across the key, your skin serves as the cathode resistor. The voltage you feel is exactly  the same as the grid bias voltage on the tube, which limits the plate current.  You might get a little sting off the key, but you are feeling a minuscule current and nothing lethal, except perhaps in the unlikely event that the tube develops a plate-to-cathode short.

Then licking the key is ok?  ;-)

Rich

Only if you're into that sort of thing. 

Yesterday I mailed Tom, K1JJ a question about his PTT scheme and the 50K resistor. Tom answered swiftly and explained the resistor is nothing more than a cathode bias, cutting the 813 off.

Today I red through a lot of technical postings here (New on the forum... ;)) and noticed a topic allready covering the question I had.

But another one came up, and instead of mailing Tom again, I thought it would be better to continue the topic, so everyone can benefit from a possible answer...

I'm in the planning phase of a single class C plate/screen modulated 813. Since I was looking for a safe way to switch (PTT) the transmitter, I looked at different schemes on the internet, and that's how I came to mail K1JJ.

I'm planning to use a dropping resistor to supply the partly modulated screen.
In my opinion that would mean, that if the cathode bias is switched on and the tube is cut off, the voltage on the screen will rise to B+ level since there is no screen current to provide a voltage drop.....

Is this OK for the 813, or does it mean the mentioned PTT scheme can only be used with a fixed screen supply?

Martin, PA4WM

Nice to meet you again, Martin.

Your question is very observant and one I've not really considered. All of my plate modulated rigs have used a separate low voltage screen supply because when using a high voltage supply (like 2-3KV) with a dropping resistor,  the wasted screen resistor power can get into the hundreds of watts, depending on the rig.  For example, with a 4-1000A, it is ridiculous waste of power.

That said, you are concerned that putting say, 2KV on the screen (unkeyed) may cause the screen to flash over to the grid, filament, etc.  That is a possible problem and I will leave it up to the guys here to discuss it for an 813 tube. My guess is it will be OK and has probably been done already - I've never heard  any warnings so far.  But lets leave it open to debate.

The other alternative is to  key both the cathode and also the HV supply.  I have done that, mostly to save on HV bleeder power waste.

The cathode PTT keying circuit is used in many modern linears and all of my HB AM rigs, so is a proven method, otherwise.

T

One related question for anyone out there who has used the cathode-resistor cutoff with 811s or 812s:

what value did you use ?

tnx, Peter

I once saw a guy who built a 4CX5000 with 6 kv on the plate, a little over 1k on the screen and I don't remember the control grid potential.

It used the lift the center tap method of keying and has been in service for 10 years.

FWIW.

In that respect it's worth mentioning that I plan to run the 813 on only 1500v. That means the screen voltage will rise to the same when the cathode is biased....
The question is still open.. ;)

Martin,

My guess is 1500V will be fine on the screen when unkeyed.

I wonder if there is data published for the breakdown voltage from the screen to grid/fil of an 813?  They sometimes list it for fil to cathode on indirectly heated tubes.


Peter,

The value of cathode resistor is not critical, as long as it cuts off the tube currents during unkey. (Plate, grid and screen currents)   I have used anywhere from 30K to 100K in various rigs. Makes little difference. I've not tried the 812/811 with it, but should be just fine.


Dave: If I'm reading you right, that's a 4X5 with 6KV on the plate and 1KV FIXED on the screen?  Was this a linear with fixed 1KV on the screen or class C rig with a screen dropping resistor?    Martin was concerned about having his class C final screen voltage soar to full HV potential when the cathode was unkeyed when the screen had no I*R voltage drop across the screen resistor.
 

T

You could add a string of power Zeners to limit the screen to a peak voltage......or just shut down the screen supply in RX.

Duh. Never mind it was a regulated supply.

Or........................

If you are running any fixed biass on the tube. (you have a biass supply) Simply open up the ground connection on the biass supply bleeded / voltage divider resistor
and let the biass rize up and cut the tube off.

Or if you are using a leaky grid for biass, just make a simple biass supply and use a couple of steering diodes. Switch in the biass to cut the tube off in standby mode.

Ur basic grid-block circuit. Grid blocking saves a lot of beating up on ur plate iron, rectumfryers and filter capz by not constantly firing on and off the HV.

Just my $.02 worth.....................

I suspect that letting the screen float at many times its spec is asking for a screen to grid arc. Using the series screen resistor as half of a voltage divider with another to ground at the socket is one way to limit the soaring as well as eliminating negative screen current during tuning. This can happen especially with the small ceramic tetrodes.

For big tetrodes I use a seperate screen supply and open either the CT or the ground of the FWB and use another set of contacts on the same relay to apply at least 2X cut-off bias to the tubes using either a 47K resistor or with a bias supply scheme.

Excellent suggestion, Carl!

There ya go, Martin. Add another power resistor to ground from the screen and the voltage will not soar.  The downside is it will add some power waste even when the rig is unkeyed, unless you add a HV screen relay in there.

Personally, I would go with a separate small screen supply and modulating series screen choke - with the supply controlled with a variac.  You will find that when you load the final at different levels, the screen current will change. Having the ability to change the screen voltage via a Variac wil enable you to set the screen current to spec no matter how the final is loaded, heavily or lightly.  I use this method on all of my class C plate modulated rigs.  It saves power in the end, too.

T

FWIW, on my 4X1 transmitter, I run a very heavy fixed grid biass. (-150v IIRC) and drive the hell out of it. For the PTT scheem, I switch the screen voltage on and off.
With the heavy class C biass, when you shut the screen voltage off and kill the RF drive to the grid, the tube cuts off dead, no measurable Ip at all.

Been running that way for over 10 years now with no problems at all.

Another method would use thyratron switching.

Has anyone tried something like this:

http://n4trb.com/AmateurRadio/GE_HamNews/issues/GE%20Ham%20News%20Vol%2001%20No%202.pdf

Carl:
I think you are right. 1500v (or even higher when there's no load on the HV supply) is asking for troubles...
Dropping the screen voltage with a voltage divider is a good option, but my plan was to modulate the screen of the 813 partly by using two dropping resistors. One from the modulated B+ and the other from the unmodelated side of the mod transformer.. (partly selfmodulating at a ratio of 60/40) I probably could use voltage dividers on both side of the xfrmr... But that doesn't make it easier.

Tom:
I still like the idea of switching the 50k bias resistor in and out the cathode lead to PTT the transmitter.
Letting the screen voltage soar is maybe not such a good idea then ;), so I have to look at a different way to supply the screen. I might have to search for a 20h or so inductor.  

Thanks everyone, for all the reactions so far.

Martin

Martin,

Thinking more about it, if you use a 1500V HV supply tied to the screen thru the resistor, there could be spikes of over 3KV on the screen when the rig is unkeyed - with a spark  produced in the modulation plate circuit. The modulation transformer adds to the 1500V at the plate as normal and it is common in the early stages to see keying/unkeying transcients when the rig is first shaken down.  The floating screen could get arced over, though the current would be limited by the resistor.

However, I know many guys use the screen resistor off the plate as a modulation technique. Maybe someone can tell me why there are few reports of screen damage? What if the HV keying or added grid bias was sequenced poorly and produced damaging voltages for an instant?  Maybe the risk is not as bad as we think.  I mean, even if grid bias is applied, the screen voltage still soars if the HV is left on. Safety may depend upon opening the HV circuit quickly with a relay..


That said, small 10-20 H chokes are quite common in the states at the flea mkts. I know there are few parts in Europe from what they tell me. Too bad the shipping cost is ridiculous to Europe.

If you find a choke, I would go with the small screen supply and a Variac. Being able to control screen current on the fly when antennas changes are made, power levels are changed, etc., is an advantage - as well as eliminating the potential HV risk to the screen as we discussed.

T

I never did but I have a couple of those FG17 tubes.