tube op-amps

[Eric Barbour]

Paul Perry wrote:

> The level shifting is normally done by resistor networks - there is always a
> high voltage negative rail in a valve analog computer system - sometimes
> aided by neon constant voltage tubes (from memory 0B2 - the 0 probably means
> zero heater votage)

Classic tube opamps, such as the Philbrick K2 series, used a combination
of
resistors and NE-2 neon lamps. Simple and effective.
 
> it is possible to make a basic differential input op amp from a twin triode
> long tail pair input stage and a pentode output, with carefully adjusted
> positive feedback to get the gain tweaked up to something like 20,000 - but
> it is nothing as good as the first 709, let alone the 741!

Quite true....the K2R used two 12AX7s and had open-loop gain of about
2000.
The later K2XA changed to a 12AX7 and a 6BR8, the pentode serving as the
gain stage--then the gain is more like 5000. Some positive feedback is
used. Don't forget the cathode follower in the output. 
It may not have fantastic performance, but it WORKS.

> I have seen a 100,000 gain dc valve op amp that worked by modulating a valve
> chopper and then consisted of a string of rf amplifiers.....this was the
> size of a normal valve cro and had a bandwidth of 100KHz

Believe me...the readers of this mailing list do NOT want to experiment
with choppers! It would be vastly easier to use solid-state op amps.
There is a company in Arizona, Apex Microsystems, who makes high-voltage
op amps. I think they are hybrid thinfilm ICs. Some can operate on
+-150v,
a perfect match to the tubes.
 
> >       >more if tubes are used as voltage-variable transconductances.
> >       >This
> >       >is workable
> 
> a lucky guess on my part!
> 
> >I have no idea how to do this.
> neither do I!, but I think  pairs of variable mu pentodes as in a ladder
> filter, or
> for a non-resonant filter, I would modulate an rf signal and change the
> carrier frequency to lose more or less of the sideband to a rectangular
> tuned circuit....there are various reactance modulation schemes as were used
> in the old rf sweep test generators
> 
> for anyone wanting to look at this seriously, a good book is Analogue
> computing at Ultra-high Speed D.M.MacKay and M.E.Fisher
> 
> some general analog tips but no valves in Korn & Korn electronic Analog and
> Hybrid Computers

It is difficult to find these old analog-computer books....see my GLASS
AUDIO
articles for more references.
 
> BTW, sometimes the valve heaters are run on rectified dc with a 6v car
> battery floating across it.....apparently this gives a more stable heater
> supply (valve synth tip of the day)

Not necessary! Unless you need to have an uninterruptable power supply.
Conventional, well-filtered DC power is adequate. Regulation is not
needed
for heaters.

-- 
Eric Barbour
Svetlana Electron Devices
Portola Valley CA USA