Homebrew 811A grounded-grid linear amp.

[G6Tanuki]

I acquired this - for £5 - at a radio-rally [Harwell? Newbury? can't remember] a few years back.

It's a nicely-constructed single-band grounded-grid linear using an 811A as the business-end. Looking at the tuned-circuits and capacitors I'm guessing it's for 14MHz.

Assuming that the 811A is still OK, from what I can work out it should be good for 200W PEP - the 1942 RCA "Guide for Transmitting Tubes" -

http://www.nj7p.org/Manuals/PDFs/Tub...-gfat-1942.pdf

says it is rated (ICAS) for 65 Watts plate dissipation.

But it needs a power-supply.

Sample circuit - from an ARRL book - here:

www.qro.it/amp/schemi/pdf/68hb187.pdf

I'm guessing I need at least 1Kv HT at 150mA or so - hence my thoughts are drifting towards a 250V isolating-transformer driving a classic Cockroft-Walton diode-capacitor quadrupler. These days, capacitors are cheap and semiconductor diodes have low forward-resistance so I can build a supply that will easily deal with high instantaneous current-demand.

I'll be adding a fan, too!

I have used CW multipliers for HT on many occasions, they work a treat (great for multi HT 1920's valve radios that need various HTs), for this I would go both up (positive) and down (negative) at the same time, may as well use both halves of the cycle! A modern transformer will easily cope with a kV or so of DC on the secondary WRT earth. Huge 55-0-55 toroids are easy to find, only three times each way.

[G6Tanuki]

Going "up and down" might present a few issues in this case with biasing of the output-bottle though. I need its negative to be at chassis-potential (both DC and RF).

I'm also pondering using a "Panel transformer" - https://uk.rs-online.com/web/p/din-r...rmers/0504139/

in reverse feeding a simple fullwave doubler. I've done similar things in the past to get 1.5Kv at an amp or so, and the result was good enough for MoD.

I need its negative to be at chassis-potential (both DC and RF).

Just connect the down to chassis.

[G6Tanuki]

That would work.

I'm currently figuring out the bias supply: It needs something up to -50V with a low source-impedance, so I'm thinking about a chain of Zeners and a 'backwards' LT-transformer fed from the heater circuit - that at least guarantees the availability of bias when the tube is 'lit'.

For TX/RX switching, some designs shove a couple of hundred volts of -ve on the grid to turn the valve hard off, but that means the B+ rail soars away to the peak-voltage the unloaded transformer can deliver - which stresses the smoothing components something horrid. I'm minded to fit a TX/RX relay that simply removes power from the HT transformer primary on receive - and accept that there will probably be a big core-thump/dimming-of-the-lights when power is called-for. I can over-spec the HT rectifier diodes so they can happily handle a few 50Hz cycles of 10A or so when the relay closes.

[Sean Williams]

Nice looking amp chassis.

Amazing how compact a monoband amp can be.

[G3PIJpeter]

Originally Posted by G6Tanuki

But it needs a power-supply. Sample circuit - from an ARRL book - here:
www.qro.it/amp/schemi/pdf/68hb187.pdf

The bias supply is needed only for standby. The 811 (and 811A) is a 'zero-bias' triode and so the control grid is at zero DC volts during transmitting (unless the applied EHT is inordinately high). Note the standby switch earthing the grid. Thus, the bias supply does not have to be 'stiff' - anyhow, that's my reading of the circuit diagram.

Peter (G3PIJ)

[G6Tanuki]

My understanding is that you need to supply a squib of adjustable-bias to set the standing current: while the 811A is generally used zero-bias, I intend to use this amp for both SSB and AM. When running in AM mode the static carrier shifts the working-point a bit and so additional bias can be used to throttle things back a bit.

How it will handle 'controlled carrier' AM (series-gate modulation) - where the resting carrier is pretty much cut-off in the gaps between syllables - will be interesting to see. I expect to have to treat that a lot more like SSB than conventional AM.

[G3PIJpeter]

Originally Posted by G6Tanuki

My understanding is that you need to supply a squib of adjustable-bias to set the standing current

Some '60s zero-bias GG amplifiers used a few dry cells to supply the bias to adjust the standing Ia. They have a low internal impedance and tend to keep their 'as new' charge state due to the grid current flow. The grid can also be made negative by inserting 1N5408 diodes in the cathode lead (or a zener), bypassed for RF - see diagram at https://www.w8ji.com/bias.htm
- Peter G3PIJ
P.S. See Figure 3 at http://www.813am.qsl.br/artigos/line..._Amplifier.pdf

[G6Tanuki]

Originally Posted by G3PIJpeter

The grid can also be made negative by inserting 1N5408 diodes in the cathode lead (or a zener), bypassed for RF - see diagram at https://www.w8ji.com/bias.htm
- Peter G3PIJ

Yes I'm familiar with the diodes in the cathode path; in times-past I've designed 'serious' multi-Kilowatt G-G pulse-amps where bias was provided by a parallel-pair of 2N3055 and Zeners from collector-to-base (it gets fun doing this when the peak cathode-current is around 10A).

And yes, batteries can be used too: I remember a 1960s RSGB-Handbook design which used a 22.5V hearing-aid battery and a PP3 to provide the bias. At least with batteries you know the bias will 'always' be there, which can't be said for power-supplies that use separate transformers/rectifiers for HT, LT and grid-bias.