UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 8th Nov 2016, 12:39 am   #1
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Design-and-building a Valve Tesla Coil

I am in the process of building my fourth ever Tesla coil, a large class C Armstrong oscillator, and I’d like to share the journey with forum members.

DISCLAIMER: this project involves pulsed DC EHT of over 4kV from a very low-impedance mains source, which is likely one-touch lethal. It also involves very high RF currents and voltages. Extreme caution is needed in every aspect of design and operation.


The mods have asked me to put this disclaimer in place. However, further discussions around the safety aspect of this circuit are probably not appropriate for this thread. Nobody would dream of grabbing the antenna of a transmitter, or of touching the power supply. Same goes here, and there are far more interesting things to say about it.

OFF-TOPIC for this thread (mods, am I right?) is any discussion of free energy or of Tesla’s other work. I respect Tesla as the genius inventor of the polyphase motor and of the first competent system for transmitting AC power, but I have no interest in any ‘free energy’ or other Tesla cult theories whatever.


Background

I began building electronically-excited Tesla coils around 15 years ago. I have previously built a design based on a high-powered transmitting tube, as well as two solid-state designs - a crystal-controlled IGBT class E design, and one based on driving a tank-circuit with an IGBT bridge. I currently have time on my hands in recovery from a health-scare and have begun to construct my largest-ever model, this time, a valve version again. I have been scavenging parts for around 3 years, biding my time.

The purpose of this build is to make a table-top design that is electronically and mechanically robust, which produces substantial plasma, can operate in CW mode for at least five minutes at a time and which looks good enough to live on display permanently.

The meta-purpose of this all is to produce Kirlian photographs by discharging the plasma through various substances that are normally insulators, but which break-down at the very high RF voltages produced by a Tesla coil.

The challenge

There are very few reference designs for valve Tesla coils. Most are from the US, where it is considerably easier to source some of the components. I do not know of a single British reference design. Reading through the experience of designers over many years, it is clear that in the quest for the largest possible plasma at the antinode at the top of the coil, builders frequently experience catastrophic failure of the valve through excessive grid or anode dissipation, or internal flashover.

Many designs on the internet are not really Tesla coils at all, but simple flyback drivers, usually based on the Royer oscillator. They are low-powered and neither interesting nor relevant to me.

There are no affordable reference books on the subject of valve Tesla coils, although one does exist with a price tag of around £1,500. I have not seen it. I have found ‘Principles of RF Amplifiers, RF power amplifiers’ (RCA Institutes, inc.) helpful reading in the past, although it does not, obviously, reference Tesla coils specifically.

The arrival on Ebay of ex-Soviet, military transmitting valves, has produced a couple of designs of central European origin in the last few years. However, the design using such a valve that caught my attention this summer was only assembled by the designer as something of a sprawling mass of wires, bread-board style, and not contained in a cabinet of any kind.

This tradition of building on a large breadboard also exists in the US, although I have seen exceptionally beautiful models that are in entirely the opposite direction of effort and crafstmanship. The lash-up assembly is appealing to first-timers, as it vastly much easier to change components, rebuild certain inductors and get some kind of response in the least possible time.

However, I am not tempted to repeat this kind of assembly as I feel it is something I tried out a very long time ago, greatly increases the inherent dangers, and produces very poor RF circuit design and response.

So the first major challenge is to contain everything in a way that leads to the best results and is electronically resilient.

To add a bit of spice, I don’t actually have a workshop. All cutting, brazing and working with solvents is done outside on a small picnic table on my balcony. See picture.

Soldering and other light assembly is done on a carefully-protected prized table in my living room and put away each session. I’m not complaining – I have two DMMs, an AVO, and a dual-trace oscilloscope, so plenty of tech to back me up.

The circuit

The circuit is just a variety of tuned-anode, tuned-grid oscillator, the Armstrong, in class C mode. It is described as a Tesla coil since Tesla first theorised a way of inducing very large voltages in an inductor using a tank circuit and a form of very high current capacity switch, originally a spark-gap. Mechanically interrupted Tesla coils are OT for this forum, however, the valve version may well be of interest to some.

On paper, it looks remarkably simple. However, in practice, it takes a lot of very careful planning, design and tuning to make a circuit that is robust enough to operate for more than a few seconds at a time without breaking down.

My experience is that the turns ratio between primary and secondary has remarkably little to do with the length of plasma at the antinode, although this is apparently a rather more complex subject than it appears. But it’s true to say that large secondary coils wound on fine wire have a high innate series resistance which lowers their Q. And it's all about Q with these things.

And it's also about raw power in, although it is of course entirely the control of that power that makes the difference between success and disaster. Hence extremely careful tuning and adjustment is needed along the way.

My first ever secondary coil used the Svetlana 572B, high-mu transmitting tube, with just 440 turns of wire on a 28cm long, 3 inch diameter former. This produced a plasma discharge of around 25cm with 400 watts of power, although this destroyed the tube in very short order through excessive anode dissipation, because of poor grid leak/biasing arrangements. That is a learning experience not to be repeated, of course.

One of my IGBT designs used 880 turns of wire on a 28cm former, which produced a brush-shaped plasma discharge around 4cm long with 150 watts of power. Nice, pretty and stable, but not that spectacular.

My current coil will use the GU81M transmitting pentode, and will operate at around 1.3 MHz.

You’ll note that I use the word ‘plasma’. It feels to me definitely not appropriate to describe this as a ‘spark’ as it is a discharge into free air, without any conductor present. It is extremely hot, and will melt a brass wire if this is used as a break-out point on the toroid.

Progress to date

I have built the basis of the chassis from two, six-bottle wine crates, bonded together. Only - since I wanted the proportions in the golden ratio - I cut one so that the combined length-to-width ratio of both is roughly 1.6:1

Over three years, I salvaged 4 capacitors for the EHT power supply from four separate microwave ovens that were scrapped and left on the street. I also salvaged two microwave oven transformers (MOTs), as well as some high-voltage cable.

I tested the GU81-M pentode before starting work. It needs a 12V, 10A filament supply, which I will be feeding with a toroidal transformer.

The biggest RFI component is a pull from avionics equipment.

I bought a vaccum variable capacitor for the tank circuit from Ebay. This is unorthodox, although of course common practice in transmitters. I don’t know if the break-down voltage will be enough. Much more common place is to use doorknob type capacitors, although only the much larger ones will cope, or to home-brew one from series-parallelled caps designed for pulse applications.

It will be a challenge to mount this physically large, aluminium component – half the power-house of the tank circuit- but one that I’m looking forward to.

I have previously used postage tubes for the main inductors (varnished and baked multiple times – works brilliantly) but am using Lexan this time. I made a trial primary coil on the larger of the two formers to do a bit of inductance modelling, and it got a bit scratched, so I have now painted it with several coats of red stove enamel before winding. I think this will look great with green magnet wire overwound.

The main power supply RF bypass capacitor needs to withstand approx. 20kV and so this is home-made. I have built one of approx 18nF by putting 9 x 0.15uF in series. It’s physically way larger than you might imagine, considering its capacitance, but is an important component.

The filament RF bypass caps are ex-Soviet types. I have hard-wired them to the ceramic base of the valve, which is mounted in a purpose-cut aperture in the chassis, in an old cigar case from a junk shop.

I have built a simple digital controller that will eventually interrupt and make the circuit between the filament centre-tap and ground (either using an IGBT or an SCR as the power device) thus lowering the duty-cycle of the valve. I have also started building the simple relay delay circuit to shunt, after 0.5 seconds, two inrush-limiting resistors. This is to avoid blowing the fuse when the microwave oven transformers are energised, since they initially pass a huge current, even off-load. (All microwave ovens have this in place, but I didn’t salvage the circuit or relay.)

I have also covered the top of the chassis with copper sheet, both to form an RF shield and to simplify ground arrangements.

The EHT power supply is also nearly complete. This is built from two MOT’s in parallel, with a 4uf capacitor and an EHT diode in the voltage doubler circuit. The anode of the diode is grounded.

The top-load I am using began life as a so-called ‘gazing ball’ from a garden centre. This is a fake-copper-covered object designed to reflect light and look all pretty. As I knew it wasn’t really copper coated, I heated it to blue heat on the stove to burn off the paint and to reveal the stainless steel underneath. This component forms a capacitor of some low pF, with reference to the ground. It is important as it allows a larger charge to build up between cycles.

In a previous thread, I considered using a mercury vapour rectifier in the power supply, and I even completed the build of a suitable power supply in this way. However, because it required big inductors for phase-shift reasons, it was extremely heavy and cumbersome and in the end, impractical. I still have the valve, of course.

There are two meters, one to monitor grid current, and the other, to monitor anode current.

I have assembled and partly built the components for the grid-leak circuit. This will probably be very fiddly to get spot-on. Several coilers whose work I have followed experienced catastrophic destruction of their valves by over-running grid current and exceeding anode dissipation, in the interests of producing the longest plasma possible. I repeated this error in my first build, all those years ago. I don’t want to do this again.

Next steps

The next stage is to wind the main inductors. One is for the tank circuit, and will be wound with 2mm wire. The other is for the secondary, and will be wound with 1mm wire.

I was lucky to discover that the two top caps of the huge tube are not both anode caps! One is the anode, and the other is connected to grid 3. Since I have grounded grid 3, a short circuit would have resulted if I hadn’t checked this out.

I will also map-out safety earth arrangements.

I hope this is of interest and I’ll keep you posted. More photos to follow soon!

Attached Thumbnails
Click image for larger version

Name:	IMG_0800.jpg
Views:	530
Size:	94.3 KB
ID:	132432  
__________________
Al

Last edited by Al (astral highway); 8th Nov 2016 at 12:48 am. Reason: small omissions
Al (astral highway) is offline  
Old 8th Nov 2016, 7:54 am   #2
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Very interesting Al, I can see you've put a lot of thought into it, both from the functional and aesthetic aspect. Not easy to do with no workshop, I take my hat off to you.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 8th Nov 2016, 8:48 am   #3
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Smile Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
Originally Posted by Dr Wobble View Post
I can see you've put a lot of thought into it, both from the functional and aesthetic aspect ...
Thank you, Andy, I appreciate your support.

On the aesthetics side, I realise that the finished product will likely have a few signs of the journey it takes to get there. There's already an extra drill hole in the main RF ground plane that will sit under the two big inductors. But I'm at ease with this, having discovered last year a Japanese tradition called Wabi Sabi, in which the history of a object's imperfections is not destroyed or apologised for but lightly contained and preserved (even beautifully foregrounded). I won't go further to take this OT for the thread but it was an important factor for me before deciding what to do if I made trial-and-error changes or mistakes with expensive materials (like 3mm sheet copper, in the photo up there). I will simply patch the hole with a tiny piece of copper tape.

Looking forward to posting progress photos later today.

Cheers!!
__________________
Al
Al (astral highway) is offline  
Old 8th Nov 2016, 10:16 am   #4
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

I thought the Armstrong oscillator was either tuned grid or tuned anode not both?

Lawrence.
ms660 is offline  
Old 8th Nov 2016, 11:10 am   #5
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Hi Lawrence, do call it something else if it bothers you, but the name works for me. Here's why...

We have a conventional parallel LC tank in the anode circuit, then we have a grid leak network comprising its own L (a small feedback winding coupled to the tank circuit L) and its own parallel RC. This parallel RC forms a time constant, obviously, and determines how long the valve is switched off during a cycle. The R determines bias, with self-evident results. The two are interrelated since we are obliged to change bias if we adjust the time constant by changing the R.

The critical nature of the grid leak circuit means it needs careful design and testing and incremental adjustment during building and testing. It is also ideal to be able to move the grid feedback coil up or down to change the coupling. All of those things makes it 'tuned' in my view. Even if we want to, we can't just grab any old RC from another circuit and valve and expect it to work with a particular grid feedback winding...very busy calculator during this project, btw!!

Cheers,
__________________
Al
Al (astral highway) is offline  
Old 8th Nov 2016, 11:18 am   #6
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

The confusion might have been me wrongly understanding your original description, my apologies for that, what you have just described above is an Armstrong type oscillator (one tuned circuit) so no problem for me now.

Lawrence.
ms660 is offline  
Old 9th Nov 2016, 2:52 pm   #7
dominicbeesley
Octode
 
dominicbeesley's Avatar
 
Join Date: Nov 2004
Location: Hebden Bridge, West Yorkshire, UK.
Posts: 1,885
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

More pictures please Al! A rough sketch of the circuit (or a block diagram) would be helpful too. Sounds very interesting, I've always wanted to build one of these but never had the courage or space. I am a true breadboarder too so would probably zap myself!
dominicbeesley is online now  
Old 9th Nov 2016, 3:04 pm   #8
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
Originally Posted by dominicbeesley View Post
More pictures please Al! A rough sketch of the circuit (or a block diagram) would be helpful too...
Sure, Dominic !! ��I'll be posting some pictures and a schematic diagram probably tomorrow... finished the EHT power supply yesterday.
__________________
Al

Last edited by Al (astral highway); 9th Nov 2016 at 3:31 pm.
Al (astral highway) is offline  
Old 9th Nov 2016, 4:57 pm   #9
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Here are some photos , folks!

Pictures are

...an approx 20nF capacitor made from a series of 0.15uf polypropylene caps rated 2,000VDC. This is the main RF bypass cap for the power supply. A doorknob type could be used instead. This is now mounted inside the cabinet on a paxolin base , a polythene back and using silicone and cable ties to fix it securely and provide some corona loss protection.

This is expensive and time consuming but a definite essential component.

...top view of valve base in situ and cabinet covered in copper foil as part of the ground and also RFI protection for what's going underneath. To the bottom right you can see a thicker copper plate which is the base for the primary and secondary coils and is mounted on brass stanchions that I braved on underneath. It will be bolted on with 25mm bolts. I am happy to have achieved a method of joining the copper foil cladding with zero resistance between all sections and ground. I used thinner copper tape over the gap between adjacent foil sections and soldered the joins one by one.

...front meter panel. This no longer looks like this. There was a flashover from the copper foil to the RHS ammeter , when I earthed the copper foil -- I didn't realise the mounting screws of the meter would arc inside -- which also violently bent the needle in half. I am not sure if the meter can be saved. I hope so, will take a look tomorrow , but that copper has to go, clearly.

Work will now start on the main inductors.
Attached Thumbnails
Click image for larger version

Name:	IMG_0803.jpg
Views:	384
Size:	77.4 KB
ID:	132498   Click image for larger version

Name:	IMG_0801.jpg
Views:	397
Size:	97.6 KB
ID:	132499   Click image for larger version

Name:	IMG_0802.jpg
Views:	372
Size:	91.6 KB
ID:	132500  
__________________
Al

Last edited by Al (astral highway); 9th Nov 2016 at 5:09 pm. Reason: Small typos
Al (astral highway) is offline  
Old 9th Nov 2016, 5:40 pm   #10
Guest
Guest
 
Posts: n/a
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

You are utterly mad!* Do carry on and lets see the resultant sparks etc..

*Love it.
 
Old 9th Nov 2016, 7:27 pm   #11
dseymo1
Retired Dormant Member
 
Join Date: Dec 2007
Location: Shropshire, UK.
Posts: 3,051
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

I'm sure you've read Tesla's original work on discharge coils.
Did you find it of practical use in designing your coils?
While full of practical details, it all seemed rather obscure to me, due partly to the unfamiliar use of the technical terms of the time, partly to the incomplete (or alternative) understanding of the phenomena.
dseymo1 is offline  
Old 9th Nov 2016, 8:05 pm   #12
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
Originally Posted by dseymo1 View Post
I'm sure you've read Tesla's original work on discharge coils.
Did you find it of practical use in designing your coils?
Ahah, I did. Theory on mechanically excited could has no relevance to electronically-excited coils, however.
__________________
Al
Al (astral highway) is offline  
Old 11th Nov 2016, 1:24 pm   #13
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

The EHT power supply is now complete , only I need to finish repairing the anode current ammeter (possible and looking good so far, I'm relieved to say. Thank you to Ed Dinning for suggesting that a bent /broken needle can be fixed with s sewing needle bonded behind the broken meter needle with superglue ������)

Pictured is a previously wound secondary coil. I will not be using this it has a high series R and hence a lower Q than I need. I am winding a new secondary coil on a Lexan tube using 1mm wire -the largest diameter I have ever used, to keep R down. . I have also started designing and building the tank circuit , critical to the success of the project.

Photo three shows a cu tape cut to fit the bottom of a vacuum variable capacitor. This has an aluminium housing that is at a very high potential and with a high pulsed current - so I will contain the unit inside a Lexan tube and mounted on a ceramic standoff.
Attached Thumbnails
Click image for larger version

Name:	IMG_0732.jpg
Views:	347
Size:	77.7 KB
ID:	132588   Click image for larger version

Name:	IMG_0811.jpg
Views:	305
Size:	52.9 KB
ID:	132589   Click image for larger version

Name:	IMG_0813.jpg
Views:	286
Size:	68.2 KB
ID:	132590  
__________________
Al
Al (astral highway) is offline  
Old 11th Nov 2016, 4:43 pm   #14
Guest
Guest
 
Posts: n/a
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
needle can be fixed with s sewing needle bonded behind the broken meter needle with superglue
Model shops sell some hair thin carbon fibre, less mass.
 
Old 11th Nov 2016, 6:18 pm   #15
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Good idea! Thank you
__________________
Al
Al (astral highway) is offline  
Old 11th Nov 2016, 7:37 pm   #16
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

London... Fife.... I think I'm at a safe distance Should be fun!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 11th Nov 2016, 9:24 pm   #17
Ed_Dinning
Dekatron
 
Join Date: Jan 2004
Location: Newcastle upon Tyne, Tyne & Wear, UK.
Posts: 8,171
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Hi Al, have you thought of using Litz wire for the coil?
It is possible to make your own with many thin strands of say 0.1mm enamelled wire. Not so good as the bought stuff, but a lot cheaper.

From the meter saga I see you are learning that these HF currents go where they want to unless there are huge amounts of insulation, preferably of low permittivity in the way!!

Some model shops also sell capillary tubing which can be used for pointer arms.

Ed
Ed_Dinning is offline  
Old 11th Nov 2016, 11:24 pm   #18
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Smile Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
Originally Posted by Radio Wrangler View Post
London... Fife.... I think I'm at a safe distance Should be fun!
Ahaha!! Yes!!
__________________
Al
Al (astral highway) is offline  
Old 11th Nov 2016, 11:28 pm   #19
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Quote:
Originally Posted by Ed_Dinning View Post
It is possible to make your own [Litz wire]with many thin strands of say 0.1mm enamelled wire. Not so good as the bought stuff, but a lot cheaper.
Good idea, Ed. Thank you ...I know this is your terrain precisely . I did notice Litz wire being used in a pulse transformer when the operating frequency was only 30KHz, so sure it's a potentially great improvement to be used on the primary of this project.
__________________
Al
Al (astral highway) is offline  
Old 12th Nov 2016, 10:13 am   #20
Ed_Dinning
Dekatron
 
Join Date: Jan 2004
Location: Newcastle upon Tyne, Tyne & Wear, UK.
Posts: 8,171
Default Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Hi Al, real Litz is pretty expensive as the layup of the wires is such that each strand occupies all possible positions in the bunch over a given length.
The SMPS people use a cheaper, modern version that is simply a twisted wire, but probably 95% as good.
PM me with your needs and I'll see what I've got.

Ed
Ed_Dinning is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 4:55 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.