Re: Grounded grid VHF front-end

john larkin <JL@gct.com> wrote:

On Sun, 10 Nov 2024 08:00:17 +0000, liz@poppyrecords.invalid.invalid
(Liz Tuddenham) wrote:
 

john larkin <JL@gct.com> wrote:
>

On Sat, 9 Nov 2024 20:26:18 +0000, liz@poppyrecords.invalid.invalid
(Liz Tuddenham) wrote:
 

Cursitor Doom <cd@notformail.com> wrote:
>

On Sat, 9 Nov 2024 19:27:13 +0000, liz@poppyrecords.invalid.invalid
(Liz Tuddenham) wrote:
 

john larkin <JL@gct.com> wrote:
>

On Sat, 9 Nov 2024 16:35:45 +0000, liz@poppyrecords.invalid.invalid
(Liz Tuddenham) wrote:
 

My current receiving aerial system is very inefficient at 2 metres
(144 Mc/s) and I have thought about making a sleeve dipole for that
band.  My VHF receiver is an Eddystone 770R, which covers the band
but only in a small portion of the whole scale.  While I am
improvomg the aerial system, I could also make a crystal-controlled
down-converter, that would allow me to use an HF communications
receiver or the lower ranges of the 770R, so that the  band 2 Mc/s
wide would cover a much greater scale length.
>
It's been a few years since I designed anything with valves, so I
thought I might have a go at making a down-converter using valves -
but not necessarily the expensive 'cult' ones which everyone seems
to regard as having magical powers.  The EF91 is plentiful and
cheap as New Old Stock, so that seems like a good valve to start
playing about with.
>
The EF91 was used as an RF amplifier in the input stages of
television sets working at about 45 Mc/s, so it can't have too bad
a noise figure (although Mullard don't quote one in their data
sheet).  If I triode-strapped it and ran it in grounded grid mode,
that would reduce the noise and increase the maximum frequency it
could usefully amplify. From the data sheet, with 200v on anode and
grid 2 and an anode current of 6mA, the gm is about 6mA/V, which
gives an input impedance at the cathode of  160 ohms.  A 75-ohm
feeder could be matched to this with a Pi tank or by tapping the L
or the C of an input tumed circuit.
>
The voltage gain may not be as high in this configuration as in
grounded cathode mode, but it allows the valve to be triode
strapped for low noise without instability problems or the
dependence on neutralising that a cascode stage would have
(especially the need for correct neutralising to obtain the best
noise figure).  If I also use an EF91 as a mixer, I might need one
more stage of RF gain to get the signal up to a level where the
mixer noise is negligible - but this isn't such a bad thing because
it would allow extra tuned circuits to give better image rejection
and allow a lower output frquency if I wanted one.
>
Anyone with experience of doing something like this with valves?

 
 
How about a tube/valve XO and a diode mixer to start?
 
A good HF receiver may have a low enough noise figure that
atmospheric noise still dominates.

>
Good thinking but there are several snags with that system:
>
If the down-converter is at the aerial end of the feeder, the HF
receiver is almost certain to suffer from strong HF signals picked up
on the downlead.  If the down-converter is adjacent to the HF
receiver, there will be significant losses at VHF in the downlead, as
the aerial needs to be mounted as high as possible.
>
If there is no amplifier ahead of the mixing diode, the local
oscillator signal could be radiated by the aerial - especially if it
happens to lie at a frequency where the dipole has another resonance
or the dipole and downlead form a resonant system.
>
I was thinking in terms of the converter being right next to the
aerial (the sleeve dipole has a 'cold' bottom end and could be joined
directly onto the converter box).  The HT and LT could be supplied
either by a separate multi-core cable or by superimposing 40v A.C.  at
50c/s on the co-ax and feeding it into the 200-220-240v tappings.of a
mains transformer primary.  The full primary winding would act as an
auto-transformer to give 250v H.T. and the secondary could give 6.3v
or 12.6v to run the heaters.

 
This is really ham territory so I don't think JL - with all due respect
- will be able to assist you very much in this endeavour. However,
there should be tons of info on this in one of the old ARRL handbooks.
If you have any from the early 60s lying around it should be well worth
a look through.

>
I have read most of that sort of literature in the past and still have
copies of most of it but don't remember this particular approach being
used before - that was why I though it might make a good fun project.
There are some grounded-grid circuits but they use triodes intended for
the purpose.  There are cascode circuits with double (and sometimes two
single) triodes  but, again, the triodes are intended for that purpose.
The idea of using a bog-standard descendent of the ubiquitous EF50 for
frequencies it wasn't supposed to cover - and making it do that
adequately - appealed to me.
>
The only place I have come across anything like this is in Geoff
Woodburn's design for the Eddystone Panoramic Display Unit, where a
triode-strapped E180F is used as a grounded-grid untuned wideband
front-end amplifier.  I did copy that successfully with a ZTX450 as the
wideband front end of a noise-measuring set that I designed; it gave very
satisfactory results.

 
Seems to me that the lowest noise voltage gain - no noise in fact -
comes from a high-Q LC resonator. And that best drives a small
capacitance like a grid.
 
Driving a cathode can be wideband, but a cathode looks like a low
value resistor, a Q killer.

>
Series-tuned input circuit.

 
Or a parallel tank with taps. The Q killers are the radiation
resistance of the antenna and, a little bit, the ohmic component of
the grid impedance from electrons being ornery.

The point I was making about grounded-grid operation is that the input
impedance of the valve is very nearly the characteristic impedance of
the co-ax (voltage ratio 3:2 for a triode-strapped EF91 drawing 6mA from
a 200V HT line).   A Pi network or a 3:2 winding on a ferrite core could
be used to match them

 With some VHF triodes, the gm is around 13 mA/V so the ratio is 1:1
without any additional matching componments and you can join the cathode
through an isolation capacitor directly to the aerial socket, if you
don't mind the lack of selectivity.. 

If you want selectivity, insert a series-tuned circuit which will be
high impedance to unwanted frequencies and low impedance to the wanted
ones.  The input impedance of the valve and the characteristic impedance
of the feeder and dipole appear in series as the resistive component of
the series tuned circuit, so the lower they are, the higher will be the
'Q'.

Another possibility is to use two EF91s in parallel to bring the
impedance down to 80 ohms.  The phasing errors between different valves
is unlikely to be significant at 144 Mc/s.


--
~ Liz Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk