Tetrode RF amps question

[Chris Wilson]

Tetrode RF amp, single valve. Motorised bandswitching via a powered rotary switch. Between contacts there is no HV to the anode. That means G2 needs to be killed as well as HV when "tuning", right?

Do they have to be removed in any particular sequence? Thanks

[vk6zgo]

If HT is removed from the anode,but G2 volts are left on,G2 will get "delusions of grandeur" & think it is an anode,draw too much current & may kill your tube.

You will need to remove the G2 supply as well as the anode supply.
Offhand,I would say,remove the G2 supply first,rather than simultaneously.
What do you do when initially turning the amp on,or when you rurn it off?

[ElectroIrradiator]

At least for some tubes, the solution is to connect a (very) high voltage diode between g2 and anode, diode cathode to tube anode (reverse biased). Should be done near the tube, but outside the RF compartments, of course. This way the screen supply also feeds the anode, if the anode supply is disconnected/not yet ready. The diode should be able to carry the full anode current, and be very generously specified. A failed diode would not be too much fun (but should be caught by the tetrode board).

You need to check whether this keeps the screen current sufficiently low to not cause any problems (running anode at g2 voltage), or if the total anode+screen current is too high for the screen PSU. In a good beam tetrode the total cathode current should stay the same if the screen voltage remains unchanged. So the anode 'fault' current is likely to stay around the same value as seen during normal operation, causing a significant g2 PSU current drain when the anode supply is absent.

Second option is an electronic interlock, which cuts the tube current to zero by reducing g1 bias voltage anytime the anode supply may be absent. Your PSU should already be able to do this. However the reason this option is second in line, is that the diode protection is much simpler and thus less likely to fail when you really need it. You may want to use both techniques. For everyday operation the g1 bias cuts the tube off as and when needed. Yet should this somehow fail, the diode routes the g2 current to the anode, possibly causing your Tetrode control board to trip on high g2 current, which still saves the day (and the tube).

[GK]

Generally, the solution is to not bandswitch with the key down. It doesn't matter if the anode potential drops out when the cathode circuit is open circuit (the screen won't draw any current under those conditions, regardless of the screen voltage present).

[ElectroIrradiator]

Generally, the solution is to not bandswitch with the key down. It doesn't matter if the anode potential drops out when the cathode circuit is open circuit.

Generally, the solution is to design appropriate interlocks, guarding against any fault condition, which could cause damage. This way it doesn't matter if an inexperienced/tired/impatient operator pushes the wrong button at an inappropriate time.

Some people even go as far as using 'outsize' tubes for a given power level, giving some headroom for errors should the interlocks somehow fail. This has been a strong tradition among radio amateur DIY amp constructors for decades, for this very reason.

[GK]

Well of course plate, screen and even grid bias and drive supplies should all have over current protection in a high power RF amplifier, to protect against fault conditions, potential internal tube arcing or operator stupidity. However that is all besides the point. None of the over current protection circuits will trigger if the amplifier is properly operated and band switched ONLY when the key is up. NO tube RF final PA that I have ever come across has been designed for "hot" (key-down) band switching, which is a definite no-no, for more than one reason.

[ElectroIrradiator]

Mr. Wilson intends to use motor control for the band switch. In this case the actual switching may take a non-trivial amount of time, potentially causing operator issues without appropriate interlocks. I have been involved in building at least one amp project, where the primary band tuning was by way of tweaking a huge roller inductor (for the band 'taps'), plus a pair of Johanson vacuum variable caps. The whole thing was/is digitally controlled, complete with push buttons for band up/down. Yet the actual band switch from 3.5 to 28 MHz took a fair amount of time.

Thus attempting to activate the amp while the band switching operation was in progress, is a real and present danger in an age of operators mostly used to instantaneous feedback in solid state transmitters. His question is perfectly valid, despite the inherent stupidity involved in attempting to operate the band switch while transmitting.

[GK]

His question is perfectly valid, despite the inherent stupidity involved in attempting to operate the band switch while transmitting.

This is getting tedious. I didn't say that his question was invalid, I simply pointed out that the screen with not draw (damaging) current due to the anode voltage dropping out if the amplifier is operated properly (not band switched while transmitting). Even if it ever was, the typical protection circuits employed will protect the tube (fingers crossed). In that regard, there is absolutely nothing unique about the OP's application, motor-driven band switching or not. Any protection circuitry beyond the usual or bare-bones minimum will only be required if some degree of redundancy is required or mandated for the application, and that of course, is entirely fine.

[JackOfVA]

G3SEK has a well thought out tetrode control board - http://www.ifwtech.co.uk/g3sek/boards/tetrode/tetrode-1.htm  -- the linked manual is 50 pages long and includes a great deal of useful information on the subject of protecting the screen grid.

[ElectroIrradiator]

This is getting tedious. I didn't say that (...)

Then how about contributing a new question, instead of pointing out the obvious. Something like this:

What does the amp schematic look like, if the anode current flows through the band switch tap(s) on the main tuning coil?

Using series DC feed is a wise move if you have the components to pull it off, as it hugely reduces the difficult problem of making a good, wideband RF choke. You would usually have the whole length of the coil wired into the circuit at all times, providing a solid DC path for the anode current, with the band tap shorting out the unused portions of the coil for band switching. Doing things this way reduces the likelihood of having parasitic resonances in the 'open' coil turns from potentially causing difficulties at the higher bands. However whether this is the way to do it depends on the details of the particular amp.

G3SEK has a well thought out tetrode control board - http://www.ifwtech.co.uk/g3sek/boards/tetrode/tetrode-1.htm  -- the linked manual is 50 pages long and includes a great deal of useful information on the subject of protecting the screen grid.

He is using that board AFAIK.

[Chris Wilson]

My issues are I am trying to combine three set ups. A Russian R-140 amp that comes without the original control boxes, a G3SEK Triode Board (they are actually two boards, but you get the idea... ), and the safety interlocks, display board and control board schematics of a French hams site who has done a lot with these amps. He doesn't want to get involved with me using a G3SEK board with his "other stuff". I had the Triode Boards and felt they were extremely well designed. I can appreciate the French ham not wanting to arbitrate between his schematics and someone elses. So I am slowly trying to get my head around the French stuff, G3SEK's board info, and the Russian schematics. I have the French "boite du commande" PCB etched, drilled and poulated, and the "safety board" and the LED display panel. I also have the G3SEK boards built. It's interconnecting them in a harmonious manner that worries me

http://f1frv.free.fr/main5b_R-140_Amp.html

http://www.ifwtech.co.uk/g3sek/boards/tetrode/tetrode-3.htm

This is by far the most challenging thing I have attempted so far. I also have the HV supply built, and the screen and grid supplies nearly done. I was happy to have the bandswitching not safety interlocked and rely on the Tetrode Boards in case I did something rash, but as the French ham has gone to considerable trouble to arrange interlocks via relays I might as well try and incorporate it. My original question was to try and get my head around the required timing sequences for valve voltage application and removal. I am not bothering with the keyer circuit as I can't use Morse and it doesn't interest me. Thanks for the advice so far. Feel free to study the schematics, especially if French is your native language, or one you are pretty familiar with

[GK]

This is getting tedious. I didn't say that (...)

Then how about contributing a new question, instead of pointing out the obvious. Something like this:

What does the amp schematic look like, if the anode current flows through the band switch tap(s) on the main tuning coil?

Brilliant, I can see the direct relevance to screen grid protection.

[vk6zgo]

In commercial communications,HF transmitters with motorised band switching have been around for over 60 years.

In all cases,the anode & screen supplies are removed using appropriate interlock circuits before the switch operates.

The switching control signal removes these supplies,either by linking into an existing interlock circuit,or a dedicated circuit,then after a short time operates the motorised switch.
When the switching is completed,a switch is operated which allows the anode & screen supplies to be restored.

[GK]

I am not bothering with the keyer circuit as I can't use Morse and it doesn't interest me.

Keeping the keying circuit (grid cut-off in this case) is still handy as it allows the final to be cut-off when not transmitting; otherwise the tube will be dissipating continuously if biased into class AB for linear SSB amplification.

[GK]

In all cases,the anode & screen supplies are removed using appropriate interlock circuits before the switch operates.

Well I know for a fact that that isn't true as I've seen vintage transmitters with primitive "preset" electromechanics that just lock out the keying circuit with a relay contact when auto setting. Switching the HT directly isn't trivial. I'm not sure what the case is with the OP's vintage R-140 transmitter.





 

[Chris Wilson]

The French site shows HT removed via a vacuum relay, as well as grid voltage, if I am reading things right. I am still unsure if the sequence of their removal is important, but I suspect it is.

[GK]

If you wish to relay-switch the HT it would be preferable to have a control circuit which removes grid drive from the tube briefly prior to opening the HT relay. This ensures that the tube is cut-off/not conducting when the HT relay opens - preventing potential "hot" switching and contact arcing.

BTW, the grid voltage (bias) should never be removed, under any circumstances. The tube can be deliberately cut-off by pulling the grid potential (sufficently further) negative. Removing the grid bias causes the tube to "switch on hard" , which, depending on the circumstances, can result in catastrophic failure.

[vk6zgo]

In all cases,the anode & screen supplies are removed using appropriate interlock circuits before the switch operates.

Well I know for a fact that that isn't true as I've seen vintage transmitters with primitive "preset" electromechanics that just lock out the keying circuit with a relay contact when auto setting. Switching the HT directly isn't trivial. I'm not sure what the case is with the OP's vintage R-140 transmitter.

Perhaps I overstated the case,but all the transmitters of this type that I've seen over many years do remove both these supplies when
bandchanging--------obviously if they do some form of "autotune" within a band,they need both supplies present.

"Not trivial"?-------The supplies are usually turned off with the same contactors that are used to apply anode HT & screen voltage in the first place,normally in the transformer primaries.
I agree that the control circuitry is quite complex,but then this never seemed to be a huge problem for the designers of these units.

The simplest answer to the OP's problem would be prevent the motorised switching from operating unless the HT & screen supplies are turned off first.

In other words,operate the "HT Off" switch,then the bandchange,& put the HT back on after it has finished changing.
This would simplify things considerably,at the cost of 2 more "button pushes" by the Operator.




 

[GK]

If you switch the HT in the mains transformer primaries your control circuitry has to be sophisticated enough to wait for the filter caps to bleed. If it's a big transmitter you then have to cope with the inrush charging current of the filter capacitors every time the HT is cycled off and on.
If you switch the HT directly you also need timer/interlocking circuitry to ensure the tube(s) is cut off when the relay opens and closes - unless you have the money for a very expensive relay capable for switching kV's at full plate current and don't care much about the occasional spark-gap spurious transmission.
So yes, IMO switching the HT in a reliable, failsafe and economical way is indeed non-trivial.   

BTW, your suggested solution isn't any different in principle from my initial suggestion of just not oprating the band change switch when holding down the key or PTT button 

Something that, amazingly, thousands of HAMs have been doing for decades with their conventional desk-top rigs.

[vk6zgo]

If you switch the HT in the mains transformer primaries your control circuitry has to be sophisticated enough to wait for the filter caps to bleed. If it's a big transmitter you then have to cope with the inrush charging current of the filter capacitors every time the HT is cycled off and on.

Maybe,but it was done that way all the time.
When another band was selected,you could see the HT fall as the caps discharged, before it started "motoring".
The inrush charging current is not a problem with well engineered power supplies.

If you switch the HT directly you also need timer/interlocking circuitry to ensure the tube(s) is cut off when the relay opens and closes - unless you have the money for a very expensive relay capable for switching kV's at full plate current and don't care much about the occasional spark-gap spurious transmission.

So yes, IMO switching the HT in a reliable, failsafe and economical way is indeed non-trivial.   
Primary switching is all of the above,& was used extensively in commercial comms gear.
Amateur amps may not have separate filament transformers,in which case,you will have to switch HT direct.

BTW, your suggested solution isn't any different in principle from my initial suggestion of just not oprating the band change switch when holding down the key or PTT button 

Something that, amazingly, thousands of HAMs have been doing for decades with their conventional desk-top rigs.

I did suggest that the motorised switch be locked out as long as HT is present.

[GK]

Jesus Christ! This could go on for ever! It is besides the point that such things may have been "done all the time" or that that certain things are not an issue with "problem with well engineered power supplies".

Every point I detailed should, by all rights, be thoroughly understood and considered by the OP if he is to successfully implement his own control and power supply circuitry for his transmitter along the lines described/suggested with the parts he has available and presumably within the constraints of a finite hobbyist’s budget. 

Pontificating in general/abstract terms over what grandpa did when building his companies broadcast transmitter two hundred years ago with who knows what kind of budget does not help anyone understand the exact engineering intricacies and possible traps and shortcomings involved in implementing some semblance of a replication of  these control, protection and power supply concepts in his own modified or homebrew desk-top rig.

I'm departing this room of the asylum now - have wasted enough time!

[Chris Wilson]

On a TETRODE I thought if G2 was supplied and the anode voltage was removed, nasty things could happen? Is just biasing G1 hard enough OK, or does G2 need removing completely? Please don't fall out, you are all being helpful Anecdotal stuff is also interesting to me Thanks everyone.

[vk6zgo]

Jesus Christ! This could go on for ever! It is besides the point that such things may have been "done all the time" or that that certain things are not an issue with "problem with well engineered power supplies".

Every point I detailed should, by all rights, be thoroughly understood and considered by the OP if he is to successfully implement his own control and power supply circuitry for his transmitter along the lines described/suggested with the parts he has available and presumably within the constraints of a finite hobbyist’s budget. 

Pontificating in general/abstract terms over what grandpa did when building his companies broadcast transmitter two hundred years ago with who knows what kind of budget does not help anyone understand the exact engineering intricacies and possible traps and shortcomings involved in implementing some semblance of a replication of  these control, protection and power supply concepts in his own modified or homebrew desk-top rig.

I'm departing this room of the asylum now - have wasted enough time!

Nice dummy spit!

Back then,grandpa needed a huge budget because he had to build racks of relay logic & get his delays from dashpots or thermal relays.
The OP can do many of these things with a PIC,or even a few 555s.

I don't see a problem with turning the HT on & off----why would it be any different to normal HT switching?
It doesn't need to be done any faster.
You can find "well engineered" power supplies in the ARRL Handbook,or RSGB Manual,so you don't have to re-invent the wheel------grandpa did it years ago!

He doesn't want a CW amp so the "keydown" situation should not apply.
If, however,there is a keying circuit present which opens the cathode,as suggested by GK in post#3.,there would be no necessity for a time delay between the anode & screen switching.

We don't know how much of the control stuff is still present in the OP's example of this amp.
Are there any relays or interlock switches which only need to be interfaced to the new control circuitry?