[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Toob coil questions



Original poster: FutureT-at-aol-dot-com 

Terry,

I'm glad to hear that you're doing some tube coil work.
Most folks let the 811A plates get red-orange or so, although the
specification is for it to be barely red for intermittent use.  The thinking
of a tube coiler is usually that it's OK if the tube lasts just 500 hours
instead of 5000 hours.  Since the builder used a voltage doubler (it appears),
the tube is being run way over its ratings, but that's the tradition.



>2.  Grid resistance.  There is a resistance to ground in series with the
>feedback coil and the grid of the tubes.  Like R1 in this very similar
>circuit (no SCR stuff) at the bottom of John's page (note sure if these
>things have "official" names?):
>
>http://members.aol-dot-com/futuret/page3.html
>
>Right now that resistance gets really hot and I note a number of people are
>using big variable power resistors like Digikey sells.  That allows for a
>lot more heat dissipation and I guess grid current adjustment.  So I was
>just going to replace it with the big variable power type.  I was just
>wondering if the inherent inductance in the big power resistors is a concern?



Inductive resistors work fine.  I too prefer a variable resister to make
it easy to make adjustments.  A range from 1500 ohms to 12k ohms
or even 20k ohms is a good range.  Usually 2500 ohms to 5k ohms is
what's needed for most coils.


>I would like to add an ammeter and make this adjustable.  I will get
>Svetlana 572Bs if (when ;-)) I blow these up...  The 811A is out of
>production now and the 572B is supposed to be even better...  But the
>811A's are cheaper ;-))
>
>
>3.  How long do the tubes have to heat up before plate power and all is
>turned on?
>
>I think I blew the one tube up by not letting them heat up enough before
>turning on power.  I also note that the heaters are very bright but I think
>that is normal for this type of tube...  I can check the heater voltage I
>guess ;-))



I like to be sure the heater voltage is either exactly correct, or perhaps
a little high, but not more than 5% high as measured at the tube socket.
Some of these tubes don't need much heat-up time, but I like to let
them heat for  30 seconds or a minute.  When selecting a tube for a tube
coil, I generally select a tube which is rated for a high voltage.  811A's
and 572B's are rated for 1200VDC and 2500VDC respectively I think.
833A's are rated for 4000VDC.



>4.  Do I need the plate chokes and resistors like the schematic above has
>in series with the plate connectors?
>
>"Plate chokes:  10 turns 14awg magnet wire wound on 30 ohm 20W
>resistor."  Right now, it does not have these.



These chokes prevent parasitic oscillations and protect the tube
from over-current spikes in case there's an internal tube arc- over.
I like to have them in place, just in case....



>5.  I note the "primary" has a big heavy inductor and big heavy tank
>cap.  Are these "tuned"?
>
>It does not seem like they are tuned to the right frequency now (not
>sure...).  Maybe they just have to be "sort of close" since these coils
>tend to self tune?



A tube coil will function with a wide range of tuning relationships, but
it will work best with a particular tune point.  Perfect tuning will not make
the sparks much longer after some point, but it will make the tubes
run cooler.  So it that sense a perfect tuning will result in longer
sparks from the point of view of what the tubes can withstand.  In
other words the coil will be much more efficient if it's tuned correctly.
Correct tuning does not necessarily mean the primary and secondary
are tuned to the same frequency.  Rather it means that whatever tune
point should be used that gives best results.  A means of variably
tuning the unit can be very useful.  I use a large air variable capacitor in
oil for this purpose.

In some careful tuning work I did, I found that there was a range of
tune points which gave the same spark length.  At one end of the range
the spark was long, but drew minimal current.  At the center of the
range the spark is the same length but is a little brighter and draws
more current and the tube plate may get red.  At the other end of
the range, the spark will still be long, but dimmer again and will
draw less current.  I generally aim to tune at a range-end point
for slightly better efficiency overall.



>6.  I imagine the feed back coil produces enough voltage on the grid to
>make the thing turn on and off (oscillate) within the tubes operating
>curves and all?  Any insight into that?



If the amount of feedback is not correct, all sorts of problems can
occur.   BTW, the output of the coil should sound somewhat smooth,
not real raspy.  Some coils give a sizzling sound to the spark which
is OK also.  It should basically sound like an SSTC.

Tube coils seem to have some self-tuning aspects that work in
conjunction with the tank tuning, and can cause phase shifts, etc.
The modified Armstrong type oscillator typically used is somewhat
unstable by nature, and becomes even more unstable with the
variable voltage and load of an AC fed tube TC.  Some folks use
Hartley push-pull oscillators which are known for their better
stability and efficiency, but I haven't seen any evidence that their
performance is actually better.  I actually think the instability of
the Armstrong may be an advantage in our application.



>7.  Sometimes it seems like the thing has a little trouble getting
>oscillation started.  This may just be because I don't hear it breaking out
>yet or something too...  But if it has trouble getting the oscillation
>started, is there something I can play with to help that?



The breakout point or spike on top of the coil should be  1) physically
sharp and  2) high enough above the toroid.  If the spike is too short,
the toroid shields it, and prevents breakout.  If the spike is rounded,
it can't emit an ion stream to initiate breakout.  It should be sharpened
with a grinding wheel so it's very sharp.  I'd try this before raising it.
As the coil runs, the spike will get dull again and have to be resharpened.
Some folks use a tungsten tip so it tends to stay sharp longer.
A fluorescent tube held nearby will show that the coil is oscillating.
How the coil breaks out also depends on how the coil is tuned.  It can
be tuned to break out more easily, but then it will give shorter sparks.
Basically it's being tuned to account for streamer capacitance for
longest sparks.



>When the tube blew up, it tripped out my $4000 tek scope, but the scope
>protected itself (thank goodness!!).  But I have beefed up the input stuff
>with some MOVs and all to help protect other stuff on the AC line.  I think
>these coils could use a little study into the fault protection and
>optimization areas.  If anyone has any thoughts as to how tube coils could
>be improved or studied, maybe I can help?  I am definitely no expert on
>tube stuff, but I do have the equipment to checkout just about anything 
>here...



A big factor with tube coils is impedance.  They generally run in a
much higher impedance mode than SSTC's.  Many of the coils are not
optimized.  The high impedance of the tube coils means that they
cannot break-out well without a breakout point.  Breakout can sometimes
be forced, but can cause tube arcs.  Breakout can be forced by retuning
which effectively changes the impedance, but the sparks will be much
shorter.  This all varies slightly from coil to coil though.  An interesting
experiment would be to use very high pulse rated tubes which can
handle up to 20kV or so, and see if this gives a little better efficiency, etc.



>As you can see, "I know nothing"  0:o))  so any advice, tips, insights,
>links, and answers are welcome!!



I use the formula

   spark length (inches) = 0.5*sqrt input power (watts)

for non-staccato tube coils.

In any tube coil, tank impedance is the key to efficiency.  The
tank impedance must be within a certain range, based on the
operating parameters of the tube and other factors.  Generally
speaking if the tank impedance is too low, the coil will draw
heavy current, the tube will get red, but the sparks will be short.
If the tank impedance is too high, the coil will not draw enough
power, the tube plate will stay cool, and the sparks will be short.
Unlike a spark gap coil, it's not enough to simply have a coil that's
in tune.  The tank impedance must also be matched to the overall
system.  Tank impedance is one of the factors that controls the
current draw in a tube coil.  There's a range of tank impedances
which generally work OK for a given coil.  Grid-leak network
parameters also control the current draw.

Tubes generally require a negative grid bias to operate properly and
especially to operate efficiently in Class C.  This bias is created by
the grid leak network as the coil runs.

In general, for highest efficiency, use the highest possible plate
voltage, along with a high tank impedance, and high (negative)
bias, and high drive levels-- which will reduce both switching time
and conduction time.

Tuning also affects the appearance of the sparks.  I prefer to see
long straightish sparks, but some folks tune for shorter more
branched sparks.

John



>Cheers,
>
>      Terry
>
>
>
>
>