Hi,

I am trying to build a tube power oscillator running at 25MHz from a
Russian tube (GI6B), and although I have it running, I can't seem to
ramp up the power to anywhere near the level I want out of it. I am
aiming for 200 to 400 watts, but I seem to be running at just a
handful of watts. Everything I have tried that I assumed would control
this parameter seems to have failed to work...

The RF is needed not to transmit but to drive a plasma, but for now I
just want to get the tank coil running the right voltage level. But
with a B+ of about 1000V, I am only getting about 50 volts in the tank
coil.

Circuit diagram is available here http://geocities.com/peterpion/oscillator.jpg

Parameters a
Voltage = 1000V, full wave rectified and smoothed
Tank coil, 5mm copper tube, 7 turns, 5cm diameter, 7cm long, 1.3uH
Tank Cap, air dielectric, 5mm spacing, 13cm square, 31pF, 1.6mm thick
copper plates
Tank circuit is capable of taking at least 100A, very heavily soldered
Feedback coil, 5 turns 2mm copper wire, same diameter as tank coil, 5
cm length, placed on same former with gap of 3cm between tank and
feedback coil.
B+ choke 120 turns on 32mm diameter former
Bias, external 12V supply, fed via a diode to the ground side of the
grid coil, with a 33k resistor going to ground from the same point,
bypassed by a 1nF cap.

Although I have it running, both increasing or decreasing the feedback
turns reduces the tank voltage.

The tube spec is 350W anode dissipation, mu of 18-26, 2.5KV max, but
with heavy grid current requirements it seems, although I am not quite
clear on this.

http://www.nd2x.net/gi6b.html

Ive never built anything like this before, and I am lost as to where I
am going wrong. Is it that the tube has too little gain to drive
itself up to those power levels? I assumed that the low gain would
just mean that I had to couple more power back to the grid to make it
work. But if I increase the turns on the feedback winding, it actually
reduces the power - as does reducing the turns. 5 seems to be
optimum.

I have taken care to keep all leads short - very short (1 to 3 inches
mostly). Rats nest construction. Have tried several types of caps, all
apparently rated for heavy RF use. Even rolled my own caps from copper
sheet and polyethelene which work the same as the other caps.

On the scope, I can see that I never drive the grid into positive.
This seems wrong to me, but nothing I do makes it go positive. On the
plate, all I see on the scope is a tiny 50v drop (at 25MHz) when the
tube is supposedly hard on, ie the tube at the most on condition
through the cycle, still has about 950V across it. So I assume that I
am not turning the grid on enough.

I am wondering if the problem is gain, IE is there too little gain to
make it work properly? Do I need to add another tube in a darlington
type of arrangement to add enough gain to make it work better?

I want this type of free running oscillator because it should stay in
tune as I add small amounts of reactive load (the plasma during warm
up etc) - or at least thats what I believe, not being an RF guru I am
not certain of this.

If anyone can help me I would appreciate it very much. Having spent
the last 2 weeks trying, and read everything I can and searched the
web extensivley, I am at a bit of a loss as to what to do next.

I do apologise for probably stupid mistakes / questions (you cant do
that! you need a driver stage...? who knows!) but I am sure someone
must have done this before so I appeal to them to share their
experience!

Thanks, Pete

I think the fact that you do not have a load on the output tank, is what is
limiting the power output. Power tubes require a particular load Z to
deliver 'rated' power. Right now all I see is a parallel LC tank which will
be high Z. You quote a voltage of 50V "in the tank coil". How are you
measuring this?
Can you monitor plate current? If so, that times the applied plate voltage
will give you your DC input power. Power out will always be less than
this amount.
Also, are you sure your HV supply has the current capacity for the 200-400W
you want to get out? (This may mean something like 300-600W, or more, DC
input power).
Are you taking steps to prevent radiation of the generated RF into the wider
environment? The level of power you are talking can reek havoc with other
licensed users. The FCC could pay you a visit.
Finally, BE CAREFUL! These DC voltages can kill, and you have not lived
until you have gotten an RF burn from 100W, much less 300-400W.

Bob WB0POQ
wrote in message
oups.com...
Hi,

I am trying to build a tube power oscillator running at 25MHz from a
Russian tube (GI6B), and although I have it running, I can't seem to
ramp up the power to anywhere near the level I want out of it.

Hi Bob thanks for the attention,

Re the load, I am using a soft iron 'poker' which I bring close to /
into the tank coil to provide a load which I can easily tell when
power is delivered into. So I fire the circuit up with the scope
connected to a certain point, then increase the load by bringing the
poker closer to or into the coil. The iron looks like a very close to
resistance load, the Z varies with the proximity. So I cat test all
loads from high Z to a few ohms. If the iron gets hot, I am delivering
a decent amount of power - got some decent power with a circuit like
this running at 7 MHz with the iron getting very hot.

While doing this I monitor the voltage on the plate or tank coil with
a scope. So the 50V is the best peak to peak voltage across the tank
coil.

Loading the circuit more reduces the pp voltage in the tank. And since
the tank Z is about 200R, p=e(squared)/r so p=1.6 watts in the tank
circuit.

Plate current is around 100mA but this is mostly dumped into the anode
as heat.

PSU is half amp MOT plus variac, bridge rectifier and 10uF cap. Shunts
are removed from MOT. Plenty of power available, about 1kW.

As for shielding, when running at full power the whole thing will be
screened including the load. Till then though, the tops off although I
dont think that much will radiate from the exposed leads even when the
power gets up to decent levels.

As for RF burns, yes I am rather nervous about that too... I am used
to working around up to about 20kv but this RF angle is a new thing,
so I am being cautious. Painful experience is the best way to learn
your lesson and I have had it :-(

Anyway thanks for the suggestions.



On Jul 25, 10:09 pm, "Bob Liesenfeld" wrote:
I think the fact that you do not have a load on the output tank, is what is
limiting the power output. Power tubes require a particular load Z to
deliver 'rated' power. Right now all I see is a parallel LC tank which will
be high Z. You quote a voltage of 50V "in the tank coil". How are you
measuring this?
Can you monitor plate current? If so, that times the applied plate voltage
will give you your DC input power. Power out will always be less than
this amount.
Also, are you sure your HV supply has the current capacity for the 200-400W
you want to get out? (This may mean something like 300-600W, or more, DC
input power).
Are you taking steps to prevent radiation of the generated RF into the wider
environment? The level of power you are talking can reek havoc with other
licensed users. The FCC could pay you a visit.
Finally, BE CAREFUL! These DC voltages can kill, and you have not lived
until you have gotten an RF burn from 100W, much less 300-400W.

Loading the circuit more reduces the pp voltage in the tank. And since
the tank Z is about 200R, p=e(squared)/r so p=1.6 watts in the tank
circuit.

Where are you getting the "200R" value from? (And what is 'R'?)
Another thought is to change the L/C ratio which should give a different
loaded Q with a given poker position. This may provide more power out.

On Wed, 25 Jul 2007 21:22:56 -0500, Bob Liesenfeld wrote:
Loading the circuit more reduces the pp voltage in the tank. And since
the tank Z is about 200R, p=e(squared)/r so p=1.6 watts in the tank
circuit.

Where are you getting the "200R" value from? (And what is 'R'?)

200R is British for 200 ohms.

(which does seem a bit low for a proper match, for 1kv plate voltage and
~300w RF out shouldn't the tank Z be closer to 3,000 ohms?)

(or is my sleep deprivation showing?)

Well im not completley certain - heres some of my maths.

Far as I have read, for Class C operations,

Load Z should = Plate voltage/2 * Plate current

And since I am aiming for plate current of 300mA or so (for power of
300 watts at 1kV) then thats

1000/2*0.3 = about 1500 ohms (R).

Then tank Z should = load Z / Q

For a Q of 7.5, which is slightly below average I think, thats
1500/7.5 which is 200R tank Z.

Then via Xl = 2piFL, Xl is 2 * 3.14 * 25M * 1.3u which = 204R
and
via Xc = 1/2piFC, Xc is 1 / ( 2 * 3.14 * 25M * 30p) which = 212R

So the tuned circuit is going to buzz at a slightly higher F than
25MHz but only slightly, at a Z of about 208R. I think stray values
are larger than the difference here so this is not worth worrying
about.

Thats my understanding of the maths having studied it over the last
few weeks and I am pretty sure its in the right ballpark, although the
Z could be a bit high or a bit low for the tube conditions. But since
the thing buzzes at 25MHz the F is right, and if the Z is a bit off
this just means the Q is a bit off and I understand this is not a
critical parameter for power.

Also to back this up, the tank Z and tank coil and tank cap values are
about the same as you see in published designs for this tube (GI6B),
this power level, and this frequency (eg http://www.nd2x.net/yu1aw/1xgi7-6m-schem.GIF
- its a GI7B but apparently they are almost identical). Ive compared
with a few circuits for these tubes and they are all similar.

Only thing of course is that they all use grounded grid, which is not
anything like the grounded cathode I am using.

Of course if I have my tank Z wrong, this would surely cause power
problems. But as I say, I cant see anything wrong with these values,
but maybe thats my lack of knowledge...

Thanks for the help so far...

On Jul 26, 4:35 am, Doug Smith W9WI wrote:
On Wed, 25 Jul 2007 21:22:56 -0500, Bob Liesenfeld wrote:
Loading the circuit more reduces the pp voltage in the tank. And since
the tank Z is about 200R, p=e(squared)/r so p=1.6 watts in the tank
circuit.

Where are you getting the "200R" value from? (And what is 'R'?)

200R is British for 200 ohms.

(which does seem a bit low for a proper match, for 1kv plate voltage and
~300w RF out shouldn't the tank Z be closer to 3,000 ohms?)

(or is my sleep deprivation showing?)

wrote:
Hi,

I am trying to build a tube power oscillator running at 25MHz from a
Russian tube (GI6B), and although I have it running, I can't seem to
ramp up the power to anywhere near the level I want out of it. I am
aiming for 200 to 400 watts, but I seem to be running at just a
handful of watts. Everything I have tried that I assumed would control
this parameter seems to have failed to work...

The RF is needed not to transmit but to drive a plasma, but for now I
just want to get the tank coil running the right voltage level. But
with a B+ of about 1000V, I am only getting about 50 volts in the tank
coil.

Circuit diagram is available here http://geocities.com/peterpion/oscillator.jpg

Parameters a
Voltage = 1000V, full wave rectified and smoothed
Tank coil, 5mm copper tube, 7 turns, 5cm diameter, 7cm long, 1.3uH
Tank Cap, air dielectric, 5mm spacing, 13cm square, 31pF, 1.6mm thick
copper plates
Tank circuit is capable of taking at least 100A, very heavily soldered
Feedback coil, 5 turns 2mm copper wire, same diameter as tank coil, 5
cm length, placed on same former with gap of 3cm between tank and
feedback coil.
B+ choke 120 turns on 32mm diameter former
Bias, external 12V supply, fed via a diode to the ground side of the
grid coil, with a 33k resistor going to ground from the same point,
bypassed by a 1nF cap.

Although I have it running, both increasing or decreasing the feedback
turns reduces the tank voltage.

The tube spec is 350W anode dissipation, mu of 18-26, 2.5KV max, but
with heavy grid current requirements it seems, although I am not quite
clear on this.

http://www.nd2x.net/gi6b.html

Ive never built anything like this before, and I am lost as to where I
am going wrong. Is it that the tube has too little gain to drive
itself up to those power levels? I assumed that the low gain would
just mean that I had to couple more power back to the grid to make it
work. But if I increase the turns on the feedback winding, it actually
reduces the power - as does reducing the turns. 5 seems to be
optimum.

I have taken care to keep all leads short - very short (1 to 3 inches
mostly). Rats nest construction. Have tried several types of caps, all
apparently rated for heavy RF use. Even rolled my own caps from copper
sheet and polyethylene which work the same as the other caps.

On the scope, I can see that I never drive the grid into positive.
This seems wrong to me, but nothing I do makes it go positive. On the
plate, all I see on the scope is a tiny 50v drop (at 25MHz) when the
tube is supposedly hard on, ie the tube at the most on condition
through the cycle, still has about 950V across it. So I assume that I
am not turning the grid on enough.

I am wondering if the problem is gain, IE is there too little gain to
make it work properly? Do I need to add another tube in a Darlington
type of arrangement to add enough gain to make it work better?

I want this type of free running oscillator because it should stay in
tune as I add small amounts of reactive load (the plasma during warm
up etc) - or at least thats what I believe, not being an RF guru I am
not certain of this.

If anyone can help me I would appreciate it very much. Having spent
the last 2 weeks trying, and read everything I can and searched the
web extensively, I am at a bit of a loss as to what to do next.

I do apologize for probably stupid mistakes / questions (you cant do
that! you need a driver stage...? who knows!) but I am sure someone
must have done this before so I appeal to them to share their
experience!

Thanks, Pete

With only 1KV on the anode you are unlikely to get more than 40W out.
You need to be nearer to 2.5KV with a lot of cooling to get anywhere
near 400W. Good layout is vital with this tube, they are good to 3GHz
and will take off without some sort of stopper in the anode cct.

Steve H

Also... borrow a GHz Spectrum Analyser and chech that the unit is
running at the expected freq. What bandwidth is your scope ?

An anode APC is also a very good idea. A few turn of 16AWG around a 2W
4R7 maybe a good start.

Any chance of cranking Va up to 1500V just to see how much power
increases ?

Don't forget a fresh change of underwear and a fire extiguisher

Regards,

Mark


On Jul 26, 8:53 pm, Steve H wrote:
wrote:
Hi,

I am trying to build a tube power oscillator running at 25MHz from a
Russian tube (GI6B), and although I have it running, I can't seem to
ramp up the power to anywhere near the level I want out of it. I am
aiming for 200 to 400 watts, but I seem to be running at just a
handful of watts. Everything I have tried that I assumed would control
this parameter seems to have failed to work...

The RF is needed not to transmit but to drive a plasma, but for now I
just want to get the tank coil running the right voltage level. But
with a B+ of about 1000V, I am only getting about 50 volts in the tank
coil.

Circuit diagram is available herehttp://geocities.com/peterpion/oscillator.jpg

Parameters a
Voltage = 1000V, full wave rectified and smoothed
Tank coil, 5mm copper tube, 7 turns, 5cm diameter, 7cm long, 1.3uH
Tank Cap, air dielectric, 5mm spacing, 13cm square, 31pF, 1.6mm thick
copper plates
Tank circuit is capable of taking at least 100A, very heavily soldered
Feedback coil, 5 turns 2mm copper wire, same diameter as tank coil, 5
cm length, placed on same former with gap of 3cm between tank and
feedback coil.
B+ choke 120 turns on 32mm diameter former
Bias, external 12V supply, fed via a diode to the ground side of the
grid coil, with a 33k resistor going to ground from the same point,
bypassed by a 1nF cap.

Although I have it running, both increasing or decreasing the feedback
turns reduces the tank voltage.

The tube spec is 350W anode dissipation, mu of 18-26, 2.5KV max, but
with heavy grid current requirements it seems, although I am not quite
clear on this.

http://www.nd2x.net/gi6b.html

Ive never built anything like this before, and I am lost as to where I
am going wrong. Is it that the tube has too little gain to drive
itself up to those power levels? I assumed that the low gain would
just mean that I had to couple more power back to the grid to make it
work. But if I increase the turns on the feedback winding, it actually
reduces the power - as does reducing the turns. 5 seems to be
optimum.

I have taken care to keep all leads short - very short (1 to 3 inches
mostly). Rats nest construction. Have tried several types of caps, all
apparently rated for heavy RF use. Even rolled my own caps from copper
sheet and polyethylene which work the same as the other caps.

On the scope, I can see that I never drive the grid into positive.
This seems wrong to me, but nothing I do makes it go positive. On the
plate, all I see on the scope is a tiny 50v drop (at 25MHz) when the
tube is supposedly hard on, ie the tube at the most on condition
through the cycle, still has about 950V across it. So I assume that I
am not turning the grid on enough.

I am wondering if the problem is gain, IE is there too little gain to
make it work properly? Do I need to add another tube in a Darlington
type of arrangement to add enough gain to make it work better?

I want this type of free running oscillator because it should stay in
tune as I add small amounts of reactive load (the plasma during warm
up etc) - or at least thats what I believe, not being an RF guru I am
not certain of this.

If anyone can help me I would appreciate it very much. Having spent
the last 2 weeks trying, and read everything I can and searched the
web extensively, I am at a bit of a loss as to what to do next.

I do apologize for probably stupid mistakes / questions (you cant do
that! you need a driver stage...? who knows!) but I am sure someone
must have done this before so I appeal to them to share their
experience!

Thanks, Pete

With only 1KV on the anode you are unlikely to get more than 40W out.
You need to be nearer to 2.5KV with a lot of cooling to get anywhere
near 400W. Good layout is vital with this tube, they are good to 3GHz
and will take off without some sort of stopper in the anode cct.

Steve H- Hide quoted text -

- Show quoted text -

GHZ spectrum analyser would be a nice bit of kit to borrow, but I dont
know anyone with one - which is a pity. The scope is 40MHz so 25MHz is
pretty close to its max and the resolution is not the best. Waveform
is fairly symmetrical so I think the fundemental is right at 25MHz.

Ill do the APC tonight, what difference will it make?

As for 1.5kV - yup, I have had it cranked to full, 2.2kV, and no extra
power really (maybe a slight amount, but I am looking for 1kv swing
where I am only seeing 50v so its not even close). Underwear got
messy... makes the hair on the back of my neck stand up from fear to
be that close to a 10u cap charged to 2.2kv...!


Also... borrow a GHz Spectrum Analyser and chech that the unit is
running at the expected freq. What bandwidth is your scope ?

An anode APC is also a very good idea. A few turn of 16AWG around a 2W
4R7 maybe a good start.

Any chance of cranking Va up to 1500V just to see how much power
increases ?

Don't forget a fresh change of underwear and a fire extiguisher

The APC is intended to be resistive (damps Q) at high freq and very
low reactance at operation freq. I should have mentioned that the 4R7
should be a carbon resistor (not wire wound).

I hope that the anode isolation inductor is mounted away, so can't
form part of the tuned circuit... What value is the inductor you are
using ?

Just for getting the anode tuning sorted out, have you considered
driving the grid with a crystal oscillator source and a suitable
buffer - this splits the problem in two. A few volts of drive should
be a good start, maybe a CMOS oscillator driving some paralleled
74AC04 buffers, that will give you 5V P-P drive into a 1:2 torridal
transformer, 10V P-P may be useful dive ?

Just out of interest, why haven't you gone for Grounded Grid
configuration ?

-Mark

On Jul 27, 1:15 am, wrote:
GHZ spectrum analyser would be a nice bit of kit to borrow, but I dont
know anyone with one - which is a pity. The scope is 40MHz so 25MHz is
pretty close to its max and the resolution is not the best. Waveform
is fairly symmetrical so I think the fundemental is right at 25MHz.

Ill do the APC tonight, what difference will it make?

As for 1.5kV - yup, I have had it cranked to full, 2.2kV, and no extra
power really (maybe a slight amount, but I am looking for 1kv swing
where I am only seeing 50v so its not even close). Underwear got
messy... makes the hair on the back of my neck stand up from fear to
be that close to a 10u cap charged to 2.2kv...!



Also... borrow a GHz Spectrum Analyser and chech that the unit is
running at the expected freq. What bandwidth is your scope ?

An anode APC is also a very good idea. A few turn of 16AWG around a 2W
4R7 maybe a good start.

Any chance of cranking Va up to 1500V just to see how much power
increases ?

Don't forget a fresh change of underwear and a fire extiguisher - Hide quoted text -

- Show quoted text -