Vacuum tube VCO with expo control.

Rene Schmitz uzs159 at
Mon Nov 29 11:35:34 CET 1999

At 19:50 28.11.99 -0500, media at wrote:
>At 11:20 AM +0100 11/25/99, Rene Schmitz wrote:
>>I have no clue how to make an expo convertor with tubes, especcialy for one
>>with a current output.
>I don't either.  Paul Perry said that it has been done in the past.  There
>are some tube "op-amps" at  that
>show where their various resistor values would plug into their solid-state
>counterparts.  However, I don't know if replacing the resistor in the
>negative feedback path of a tube amplifier with a diode would work, or what
>sort of compensation would be necessary if it did.

We certainly would have to use a device with an exponential current vs.
voltage behaviour. Like a diode or BJT. But there are automatically
problems with temperature then. The tube opamps at that site are either AC
coupled or they use semiconductors, and difficult to get tubes like the
ECC86 "low-voltage" triode.
I'm sticking with the semiconductors here. I don't see the advantage that 
avoiding opamps and transistors here has. I'd rather want to see something
Beam power modulators, heptode modulators, magic eye tubes, unusual filters
and the like. (I have to admit that the output of that VCO is just a high
voltage sawtooth. Nothing really exiting.)

>>Why should we refuse to use the often discussed transistor expo convertor?
>>Its a well known circuit, and there are many possibilities to make it
>>stable. Following your idea: Why not use a oven stabilized 3046 here?
>>Generates his own heat too.
>I'm not saying we shouldn't!!  Until I heard about your circuit I thought
>it was my idea :) I hadn't thought of converting it to a current.  However,
>avoiding the use of op-amps would simplify the power supply and possibly
>avoid etching PCB's.  It would also be more in the spirit of the "Tube

The current control was the key to the interfaceing, with retaining the
To avoid opamps use a PNP/NPN expo convertor like the one used in the
minimoog filter. It doesn't need an opamp. One could generate +/-6V from
the filament supply to power these.

>>>Do you think a thyratron can be made to function as a current controlled
>>>oscillator??  I was under the impression that their response was
>>>non-linear, and hence the need to switch between hand selected resistor
>>>values in the Phattytron.
>>The *grid* of the thyratron has a nonlinear response. Like a triode grid
>While a triode grid is non-linear over its entire range, in a triode
>amplifier it is biased to operate within a linear region.  Although I don't
>think that would apply to a thyratron oscillator which needs to turn
>completely off.

Yes, a triode is linear for small signal swings, but in DC-applications
with large swings there is distortion. 
As long as the thyratron is not fired it would be out of the circuit,
leaving a linear ramp over the cap.

>>(I first had just VT1 in the circuit, the cathode grounded thru a resistor,
>>and steered it from the grid, you get nonlinear response as well.) The
>>magic of this circuit is in the current steering thru the cathode. (Grid
>>grounded amplifier situation is linear for currents: Ic=Ip.)
>That makes sense.  The neon oscillator, the neon tube and capacitor in
>parallel,  is a current controlled two-terminal oscillator.

A fun thing that I'm going to try is to use two such CCOs in series. 
Should make a complex interaction. Coupled pendulums?!

>I just took a look at Eric Barbour's thyratron VCO.  It's not your typical
>oscillator circuit.  I thought that thyratrons were gas-filled triodes, but
>this tube, a "2D21 or 5727," seems to be a tetrode.  I'm not exactly sure
>how it works.
>A sync input is AC coupled by a .001uF cap and connected to the other grid
>by a 220K resistor.  (It appears as though both grids are tied together.)
>I think I understand this part:  A pulse at the grid ionizes the gas
>allowing the current to conduct from anode to cathode, discharging the
>parallel .22uF cap, the plate current drops as the cap re-charges and
>conduction through the tube stops, until it receives another pulse.  Does
>that sound right??

I'm not sure, here is my explanation:
The thyratron is sort of a neon lamp with a heated cathode. Plus one or
more grids which can be used to prevent fireing, by applying a negative
voltage. This prevents that the electrons can reach the anode. Or rather
that they can be accelerated towards the anode. When fired they give back
their energy to the neon gas in interactions. Thus ionizing the gas. The
negative field arround the grid prevents that. The negative grid forms a
potential barrier which most electrons can't pass. The less negative the
grid becomes, the more likely to find electrons with high enough energy to
get past the grid, and accelerate in the plate field. So with the grid
voltage we can control how much plate voltage is needed to "soak" enough
electrons into the plate field. Once enough neon atoms are ionized, there
is sufficient energy for most electrons to overcome the grid barrier, the
thyratron is fired.

I'm not sure if that is what happens. Anyone else?!

>The CV in goes to the grid of a triode (1/2 a 12AT7) whose cathode is
>connected to one of the thyratron grids by a 10K resistor.  The cathode is
>also connected to 150VDC by a 120K resistor.  Any idea what's going on

The cathode of the input 12AT7 is connected to -150V, this is a level
shifting amplifier, which translates the positive input to the negative
voltage used for the thyratron.

>The output is taken from the plate of the thyratron through a .22uF cap, a
>pot set up as a voltage divider ("clipping level"), to the grid of the
>other half 12AT7 which appears to be set up as a common cathode amp.
>So if I understand your idea correctly, you want to bias the grid and then
>control the oscillation by controlling the plate current??

Rather the cathode current, but that would be the same. 

>>I think keeping the grid of the thyratron at a fixed potential and forcing a
>>current thru the cathode would be analogous to a Korg style thyristor VCO,
>>with its "overhead" triggering.
>You lost me there!!  What's a Korg thyristor VCO??

I mean that three transistor discharge circuit like in some korg synths, or
in the TB303, this acts as a thyristor or programmable unijunction
transistor. Which is similar to the thyratron.

>>Let the current charge the parallel cap, until the voltage across the
>>thyratron is high enough to fire.
>That seems quite similar to your neon oscillator.

Indeed, I'd use the thyratron in a similar fashion as the neon lamp. The
only difference would be the heater, and the grid biasing. The heater would
probably help with the temperature drift of the neon circuit.


opinionsexpressedher   | uzs159 at
einarethatofmyemploy   |
erandnotofmyself....   |  


More information about the Synth-diy mailing list