Vacuum tube amp RF plate choke design.

[ashok.das81]

Hi all,

I am building a vacuum tube RF power amplifier for SSB. Here is the link for the amplifier.

https://www.qsl.net/vu2awc/ham_radio_india_v/vu2atn_qro/vu2atn.htm

I need to clarify first that I am building for 100 watt RF power approx. I want to operate 14,21,24,28 MHz bands.
Almost all the forums I have seen, when talk about vacuum tube based amplifier, all are at the KW level.
All the calculations, assumptions, suggestions are for kilowatt level RF power amplifier. So I could not get anything
for my build. Ultimately I started this thread.

My question is how to make a plate choke, that goes between DC power supply and anode.
There is a reference here
https://www.w8ji.com/rf_plate_choke.htm

But I cannot understand it properly. Appears that lots of analysis required.

How many turns of what size copper wire needed to make it. Assuming 750V supply and max 250mA. For coil former I do not have ceramic tube or any exotic material like fiber glass tube or Teflon tube. I have 2 options as coil former:
1) A high quality thick walled glass test tube approx 150mm long and 25mm in diameter
2) Delrin rod 16 or 22 mm in dia 150mm long.

Please help me on this.

[T3sl4co1l]

Back-link for posterity: https://electronics.stackexchange.com/questions/718385/vacuum-tube-amp-rf-plate-choke-design

[Wallace Gasiewicz]

Since you are building a 100W tube amp...Perhaps look at the schematics for the Kenwood TS 520.   

https://kenwood-hybrids.groups.io/g/main/files/03%20-%20TS-520%28s%29%20Manuals/TS520SE%20Service%20Manual%20Addendum%20%208.5x11%20Contrast%20Adj.pdf

 They use two pentodes and about 800 Volts. The Plate Choke on this is  205 turns of #34 wire. It is 165 uH. I have rewound several and it is a bit tedious especially if you accidentally break the small gauge wire. However the exact value of the choke or the exact number of windings is not critical.   

 The ceramic core is about 14 mm diameter and 10 cm long, but there are other components in the Plate Choke System in the 520.-- Your Schematic calls for a much larger 2.5 mH choke (is that correct?). I think Delrin would be OK but I have not used it.    Perhaps someone else has.     
The larger value choke would be difficult to achieve with this sort of winding.

[G0HZU]

Since you are building a 100W tube amp...Perhaps look at the schematics for the Kenwood TS 520.   

https://kenwood-hybrids.groups.io/g/main/files/03%20-%20TS-520%28s%29%20Manuals/TS520SE%20Service%20Manual%20Addendum%20%208.5x11%20Contrast%20Adj.pdf

 They use two pentodes and about 800 Volts. The Plate Choke on this is  205 turns of #34 wire. It is 165 uH. I have rewound several and it is a bit tedious especially if you accidentally break the small gauge wire. However the exact value of the choke or the exact number of windings is not critical.   

 The ceramic core is about 14 mm diameter and 10 cm long, but there are other components in the Plate Choke System in the 520.--

Those choke winding details don't make a lot of sense to me? I'd expect that choke to have a series resonance somewhere in the HF band.

I had a go at simulating an alternative choke using some old software I wrote many years ago. To get the series resonance up near 50MHz I ended up with 144 turns with an 11mm diameter and a choke length of just 55mm. The first parallel resonance is at about 21MHz.

See the screenshot below:

The software tries to predict the inductance, the parallel resonance and the first series resonance. It can also compute the Q (as measured on a Q meter) although this is usually only valid up to the peak of the Q curve shown in the diagram below.

It looks like this choke would have a peak Q of about 160 at about 10MHz.

The neat thing about the software below is that it computes it all in real time so it's often easy to optimise the design for a given task.

The design below isn't meant to be optimal for ashok.das81's PA design but it probably represents a good starting point?

[Wallace Gasiewicz]

As I said before, there is more to the circuit than just one choke. There is another smaller choke and a HV cap. If you are interested, look at the schematic of a TS 520.   
The point I was trying to make is that the choke on ashok schematic (2.5 mH) is hard to make with a single winding on a form.   
The choke (and rest of circuit) is there for high impedance in the HF Band,

[T3sl4co1l]

I expect the original was a pie-wound type, though that might not have enough bandwidth either.

The easy answer is: just make a component that does the job.  But therein lies the challenge.

The hard answer is: use whatever value, and design and tune the whole thing, choke and all.  The choke's reactance is balanced by the nearby tuning cap, and pi network generally.  This may require more components to change when shifting bands, and in any case will need more alignment/tuning to work in the first place.

Tim

[G0HZU]

I looked at a TS520 and this is a radio that works down to 3.5MHz. The OP only wants to operate over a bandwidth of one octave starting at 14MHz. So a simple solenoid should be fine.

[David Hess]

How many turns of what size copper wire needed to make it. Assuming 750V supply and max 250mA. For coil former I do not have ceramic tube or any exotic material like fiber glass tube or Teflon tube. I have 2 options as coil former:
1) A high quality thick walled glass test tube approx 150mm long and 25mm in diameter
2) Delrin rod 16 or 22 mm in dia 150mm long.

The coil former can be tested for dielectric loss.  Place the candidate inside of a microwave oven, along with a calibrated cup of water.  Run the microwave for a constant amount of time, and then check the increase in temperature of the prospective coil former.  Higher increases in temperature indicate greater dielectric losses.

[ashok.das81]

@Gasiewicz The link given to kenwood TS520 opens up the home page, I cannot locate the circuit. can you post the PDF directly here?

Ok, Matt Eriksson KK5DR suggests the use of delrin as former in his blog. As he did a lot of amplifier building, delrin is ok to use. As I do not need lower end of HF spectrum, the choke should be relatively simple. Now on the inductance, Matt recommends 200uH is maximum inductance needed to work safely in HF band, but in the circuit that I have found having a inductance of 2.5mH. So is it that builder just blindly picked up 2.5mH or he did it on purpose. Achieving 2.5mH with air-core coils is not possible practically in small amplifier. So should I go for a 200uH choke or find some way to make 2.5mH choke ?

Ashok

[Wallace Gasiewicz]

https://kenwood-hybrids.groups.io/g/main/files/03%20-%20TS-520%28s%29%20Manuals/Kenwood%20TS%20520S%20Operating%20Manual%20OCR.pdf   

This seems to work for me. file is too big to attach, I think. You may have to join groups.io, which is easy and free.  The manuals are in the Kenwood Hybrids Group in the files section. There are also other groups with schematics for 100 Watt Transmitters, like Hallicrafters, Drake, Collins and more.  If you are going to build a tube transmitter there are many models to look at and perhaps to copy parts of.

[ashok.das81]

@Gasiewicz. Thanks. yes I need to join the group. Waiting for moderator approval

[ashok.das81]

I am seeing some similarities in amp circuits plate choke. So far checked 3 designes and all use 2.5mH plate choke and all are close to 100W designs. What i can see is that all use a special kind of ready made choke that have multiple windings on a single former separated by a distance. It is called Layer-pi wound choke. Below is a sample pic:



This choke was used in many amplifier and it is of 2.5mH. As I cannot obtain this now, so a suitable replacement of this choke needed.

[G0HZU]

Here's the result for Lp and Rp if I model the 144 turn choke I described back in reply #3.

This is based on a physical (distributed) model of the choke and the model usually does a reasonable job of modelling Rp up to about 70% of the frequency of the (50MHz) series resonance. So the model is probably good to just over 30MHz in terms of predicting Rp.

Therefore, the Rp for this choke is probably going to be >500k ohm across 14MHz to 29.7MHz if it was wound and measured for real. The capacitance of the choke above 21MHz is probably going to be less than 1.2pF according to the model.

With such a high Rp, there is going to be very little heating in this choke. However, this assumes your amplifier doesn't produce lots of energy at the second harmonic frequencies. As long as you load it up correctly with your Pi network on the output of the tube then this shouldn't be an issue.

[Wallace Gasiewicz]

Perhaps learn to make a Toroid choke yourself. You have to know the Mix of the Toroid and then can calculate the needed Henry Value.     
The Pi wound chokes are not the best because of their lack of power handling characteristics.     
I am not sure you need such a high value though. But if you think you need one, Here is a link to one that I include for the picture only.
Also Mouser India sells lots of Inductors

https://outerrim.galacticjunk.com/product/1000uh-1-mh-inductor-rf-choke-coil-toroid/

[T3sl4co1l]

I doubt a toroid would help here: the self-capacitance to the core, and the slow-wave resonance through the core, likely give a resonance near 10MHz, at a fairly low impedance (couple kohms?).  A core isn't useless though; a ferrite rod would be okay, and a high-mu* core will be effective even with the winding spaced away from it (i.e. use a bobbin with ID > core OD).  The effective mu will be small (2-4?) but that's still an advantage over an air core coil.

*"High mu" for a rod core, is at least some hundreds; basically any random grade of ferrite will do.

You can build bobbins out of cardboard if nothing else (but, preferably some actual engineering material like phenolic, fiberglass (stripped PCB if nothing else?), etc.), soaking it in varnish or melted wax for stability and to hold the windings in place. Many hardware store glues are also effective.  Bobbins can be stacked on the ferrite rod to make something like a pie-wound choke.

But again, this is in the realm of custom designs, and I'd have to try several arrangements and characterize them to land on the ideal design.

Tim

[G0HZU]

A toroid would be an odd choice because the input and output wire ends will end up close to each other. With the large dc + RF voltages involved, this would seem an odd choice as it increases the risk of breakdown across the hot output to the cold input end of the choke windings.

The amp only has to work over an octave of bandwidth so a basic air cored solenoid ought to be fine. At 100W power you could probably wind it on a hollow PTFE rod. It's generally best to go for a long solenoid with a relatively thin diameter as this will give a higher series resonance. It also provides a useful amount of separation (isolation) from the hot end of the choke to the cold end.

[Wallace Gasiewicz]

Air core solenoid will be very large to be 1.5 mH     
I know it might be a problem but you could use a large toroid to put in and out reasonably apart from each other. There are modern toroids that can go through the HF bands.

But your point is well taken. I have worked on many 100 Watt Ham Tube Radios and the only chokes I remember are about 3 inch long Solenoids which are maybe 100 to 200 uH.  Oh wait, there was one that used a 235 uH pi wound, I think, but no where near 1.5 mH for the Plate Choke. But the OP's schematic calls for one.

[G0HZU]

OK but I wasn't suggesting that the air cored solenoid should have 1.5mH inductance. To act as an effective plate choke across 14-29.7MHz the choke can be designed to have about 50uH inductance at low frequencies.

The 144 turn choke I described in reply #3 only has 50uH inductance at LF. Due to its distributed capacitance, by 14MHz this will look more like 130uH and by 29.7MHz the choke will probably look like a 1.3pF shunt capacitor.

The model I produced indicates that the parallel resistance of the choke will be over 500k ohm across 14 to 29.7MHz. So even with 500 Vrms of RF voltage at the anode the power dissipation in the choke will be quite small.

[BrokenYugo]

My understanding is 2.5mH is a sort of HF band amateur radio "one size fits all" value going back to the 1920s. A big enough "choke" you didn't really need to properly design a filter, just throw a 2.5mH at it and it would work.

[Bud]

Any ARRL Handbook ftom 70's...90's would have plenty of information on practical plate choke design.

[Wallace Gasiewicz]

I think Yugo is correct.   If the Plate choke is such a high value maybe you do not need to include it in the calculations of the Pi matching network.     I think things like this were used in homebrew RF amps that by today's standards were pretty low power mostly.   
But the commercial 100 Watt Radios used a smaller value choke then a cap to ground and a smaller choke to B+.  See my previous attached screenshot. Open it in a new tab and enlarge.     
The high value Pi wound chokes are not great for lots of power since they are wound with very thin wire. but maybe good enough for 100 Watts?

[Andy Chee]

I am seeing some similarities in amp circuits plate choke. So far checked 3 designes and all use 2.5mH plate choke and all are close to 100W designs.

This choke was used in many amplifier and it is of 2.5mH. As I cannot obtain this now, so a suitable replacement of this choke needed.

If you are operating only 14,21,24,28 bands, then I don't think you need 2.5mH

Big inductance is needed for 1.8, 3.5 bands.