this web page documents my latest whim, er , I mean project. A HF linear using thermionic valves
A HF linear using thermionic valves
"use a transistor, go to gaol, its the law !"
This project was prompted by the increasing RF noise floor  which renders DX  effectively out of reach to me. The solutions are more advanced modulation methods, ultra broadband spread spectrum or just plain old shouting louder.  My HF voice is only a modest 50 Watts  which no longer cuts the mustard.  May have seemed useable 30 years ago, but they did not have unfiltered SW power supplys, plasma screens LED light bulbs, un shielded dsl modems and ethernet over power. Goodbye HF, alas I know thee well and what it sounded like forty years ago.  The more advanced methods are denied to me, so that I shall just shout loouder instead in the hope that other hams that also have gone QRO  can also hear me over the cacophony.

Forty years ago I salvaged all these 6CM5/EL36 tubes from discarded TV sets  thinking that one day I would find a use for them. That day has come.  This project will attempt some difficult things and pioneer an old but neglected inductor design. I like to push the boundaries  so I shall attempt to harness  8  TV line output tubes in parralel and re-introduce  the conical inductor for both anode tank and anode RF choke.

At the recent Werribee hamfest, I bough an armfull of old 6CM5 bottles for only $2 each ! How to choose them ? Look at the getter. If its bright metallic and lustrous then this is a must buy bottle. If it it has become brown the tube has seen much use and the getter is loaded with tramp gas.  It may still be useable , test before use.  If the getter is white, this marks the true end of life for the bottle;  it means the tube is open to the atmosphere.


I am well aware of the traps of putting valves in parallel  and the resulting  strange untameable  VHF oscillator that can result.  I hope to work around this  with heavy grid loading, low Q  screen grid bypassses, heavy Kathode degeneration and generous use of lossy ferrites and the benefits of low output impedance that the parrallel design offers.  The use of the conical inductors that will exhibit fewer parasitic  resonances will also reinforce the tendancy towards stability.  



The 20 Watt plate dissipation of  the EL36/6CM5  of 8 of these old bottles  should allow me at least 200W  CW  and just under  400W PEP for ICAS   speech  if given plenty of forced airflow.  The power supply can be nicely obtained with a new very old but still serviceable  TV transformer  which I have hoarded for 40 years.

This is a "high power" project by Australian standards, the authorities do not allow us to even aspire to "kilowatt input".
I used these bottles because they have cost me nothing, cost $2  at hamfests and just work.  I am not comfortable about using  high power solid state  transistors simply because your first mistake is your last, and, they are costly.

This design and and more forgiving nature of thermionic valves permit opeartion into a wider variety of load impedances than an equivalent solid state design.  It is very hard to find an antenna that is "50 ohms"


The conical inductor for HF choke attempts to bypass a problem with cylinderical coils  and their strange resonances. At some frequncy, the winding length becomes like an equivalent 1/2 wave line, which turns the choke into a short circuit and the central part into a tesla coil , whoose corona will rapidly erode the choke.  The stepped pie wound chokes found on classical transmitters are an approximation of the cone inductor.  Conical chokes are making a comeback, a firm called CoilCraft is making them for microwave bias lines specifically for their broadband performance.  The cone has less common magnetic flux lines linking the turns,  this creates a model of many "perfect" individual inductors in series.  The cones were discovered in the 1920's but proved unpopular due to the mathematical difficulty in predicting  their inductance; and they required for special mandrels for winding them. My RFC cone has 137uH as measured on my GenRad bridge, which is a usefully large inductance for choke applications.  The choke will be placed at the "cold end" of the PI output coil. This will eliminate much of the troubles associated with broadband RF chokes.  This means that the tap selector switch and the coil and tuning capacitor will be at EHV potential.  This is a break with the standard transmitter design orthodoxy and also makes this design hazardous.  Be warned if you try to copy this idea.

The purpose of trying a conical inductor in the anode tank  is to improve the Q of the inductor. By tradition, the anode PI tank is tapped and switched.  The unwanted part of the coil is shorted out. Magnetic flux from the "active" part of the coil is still threading the shorted turns  causing a circulating  current. This current is subject to ohmic losses and serves only to reduce the amount of energy in the tank circuit, that same energy that you have commited so much time effort and money to produce !  By shorting out the tip of the cone, far less magnetic flux is coupled to the shorted out section. This has the overall effect of increasing the overall Q of the inductance.  There is also the benefit of less chance of exiting a stray resonance. Measurements of Q and Inductance on my bridge seems to confirm the soundness of this idea.  I should also add that winding the cone was extremely difficult with the heavy gauge wire .  The secret is to wind the small end first and not to stop untill the entire winding is completed. The spiral will unwide itself somewhat, never mind, the unwinding is small but does increase the diamter of the cone.  Take this into account if the available space is small.
XmtrRule1.jpg
img_0626-tx-case-valve-box.jpg thumb.img_0539-tx.jpg thumb.img_0540-tx.jpg
It is still amazing that you can still purchase bottles in original boxes at hamfests. The interior of the box lists hundreds of patent numbers
thumb.img_0541-tx.jpg thumb.img_0542-tx.jpg thumb.img_0543tx.jpg
now , will all the decks fit.  The chassis that loomed so large now reveals itself to be undersized. This chassis was once an ISEP frame that I found in the dump. Brass shim stock is attached to the valve deck plate to provide a solderable surface. Many bypass caps need a positive earth I do my metal work outdoors. no fumes, no mess, no dust !
thumb.img_0587-tx.jpg thumb.img_0588-tx.jpg thumb.img_0590-tx.jpg
thumb.img_0594-tx.jpg thumb.img_0596-tx.jpg thumb.img_0600-tx-tank-deck.jpg
silver plated teflon coax offcuts are used as flexible wide diameter RF conductors. They are heat resistant and flexible.
thumb.img_0601-tx-tank.jpg thumb.img_0602-tx-tank.jpg thumb.img_0603-tx-tank.jpg
Anode tuning deck is completed. Anode tuning capacitor shaft is posively grounded to take the dodgy wiper effectively out of circuit Vortex coil is mechanically stabiliese with these fibre glass rods and held together with expoxy  resin
thumb.img_0604-tx-tank-deck.jpg thumb.img_0605-tx-tank.jpg thumb.img_0606-tx-tank.jpg
Anode tuning deck is completed Anode DC isolation and RF coupling
thumb.img_0607-tx-tank.jpg thumb.img_0608-tx-tank.jpg thumb.img_0622-tx.jpg
Anode RF choke.  This is a trial of a conical inductor for choke service at HF. the anode tuning shafts are insulated from front panel valve deck construction permits verification of 6CM5 filaments. My stock of bottles light up. Filaments draw about 10Amperes in total
thumb.img_0623-tx-case.jpg thumb.img_0624-tx-case.jpg thumb.img_0625-tx-case.jpg
the 6CM5 fits the metal work
thumb.img_0627-tx.jpg thumb.img_0628-tx.jpg thumb.img_0629-tx.jpg
thumb.img_0630-tx.jpg thumb.img_0631-tx.jpg thumb.img_0655-tx.jpg
input TR control deck, grid 1 control rear panel fan pressurises the chassis. There is also a purpose fan to cool the valves.  The other port is a passive hot gas exit port.  A fan should never be compelled to SUCK  hot exhaust gas, it will have a very short life.

thumb.img_0657-tx.jpg thumb.img_0658-tx.jpg
cutting big holes for the meters.  They ever seem to line up ! this deck will control grid 1 input and bias with input TR relay, metering and sampling
thumb.img_0656-tx-cone-manderel.jpg
vortex coil mandrel vortex coil mandrel with coil ! I ran out of wire before running out of mandrel
thumb.img_0689-tx.jpg thumb.img_0690-tx.jpg thumb.img_0691-tx.jpg
front panel is taking shape. meters are for total anode current, and all indivdual  kathode currents and all screen grid voltages and a general multimeter. 12V psu utility and fan supply  and grid bias control deck pots, 24V relay power.
 I am still designing the grid 2 bias regulator & control supply.
Its harder than you think.
EHV rectifier with 3A diodes and  bleeder resistors. They are ceramic metal film and should have a long service life.
thumb.img_0692-tx.jpg thumb.img_0693-tx.jpg thumb.img_0700-tx.jpg
EHV rectifier and HV capacitor stack EHV rectifier and HV capacitor stack 12DC psu  and EHT transformer . the remaining gap will be filled in with the grid2 bias deck
thumb.img_0702-tx.jpg thumb.img_0703-tx.jpg thumb.img_0704-tx.jpg
gaze into  the Vortex input circuit and TR relay controls EHV anode mercury relay and EHV recifier Anode deck and valve deck
thumb.img_0725-hflinear.jpg thumb.img_0726-hflinear.jpg thumb.img_0727-hflinear.jpg
under the chassis, access to the valve sockets and valve deck components
thumb.img_0728-hflinear.jpg
showing here, EHT anode relay, input grid1 deck and filament inrush delay resistor
to be continued
Thu Dec 15 18:16:08 EST 2016
Tue Mar 14 18:40:25 EST 2017  added progress images.  It is not working yet.  So close but yet so far !

....they may laugh, but Ill show them....bwah ha ha ha ha