[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Tube coil related questions II - Level Shift Power Supplies
Carl, Terry, Tesla List
----------------------------
This message was mailed earlier and didn't make the list. There is a method
presented for calculating LS capacitor sizing with given input power and
voltage
for a LSPS VTTC. This methodology appears to work and the values produced
matches within +/-10% of values used on commercial Microwave ovens.
Regards
DS
----------------------------
The bypass capacitor is actually across the HV winding of the MOT. See
ASCII art below
(go to full screen view)
Clevel shift 50 ohm/200W
---- -------------+------] [------+-----[===]-------------- + HV
level shift out
| || | |
Pri | || HV sec === C byps ---
| || | ^ HV diode
(cathode UP)
---- -------------+---------------+------------------------- - to ground
Cbyps (bypass) should be 0.1 to 1.0uF. The smaller Cbypass, the less loading
presented to the MOT. It prevents RF swings back into AC line, and protects
the HV sec of MO transformer. Level shifting capacitor size can be
approximately
determined by a little intermediate physics:
1. Calculate Peak Input Voltage to charge level shift capacitor
(pk) = Vsec(rms) *1.414 (sqrt(2))
2. Pavg Output desired (assume 1000W during positive alternation)
3. PCcharge = Pdischarge; Poutput to VTTC = 50% duty cycle, 1000W (transformer
contributes 500W of pulse [50% of 1000W] while diode is off and capacitor
and transformer are discharging additively)
3. Capacitor Discharge Energy per pulse (at 60 discharges /sec) ~ 4.16J
Ec = Pdischarge(W) * Tdischarge (secs) (500W*0.00833 sec)
4. Clevel shift ~ 2E/Vpk^2; this equation is derived from E=1/2*C*V^2
for this case assuming 1800Vrms (~2.5kV typical)
C= (2*4.16)/(2545^2) = 1.28uF
This seems very reasonable since most 700W to 1kW microwaves that I've
gutted for
parts have as small at 0.8uf to as large as 1.1uF (at 2.5kVAC rated).
Diode sizing
is determined by Vdiode = 2.5 X Vsec pk (~ 5.0kV in this example). In
this case
Vdiode ~5.0kV*2.5 ~ 12.5kV (most MOT diodes are 12kV).
As for 200W resistor wasting power, consider that if an 833A is a pulling
constant 1.5A
plate current, Rp = 1.5^2*50/2 (50% duty cycle) = 62.5W (force air cool
with a fan).
Consider also that tube peak input power is Pbb=Ep pk * Ib pk
With above numbers Pbb = 5.0kV * 1.5A = 7.5kW (pk). Resistor power loss is
therefore approximately 1.6% of total peak tube input power. Imax with
tube flashover
with 50 ohm resistor is 5000/50 (100A). However, if circuit loss is
assumed as only
1 ohm (transformer and series capacitor ESR -at- 60 Hz) peak currents will
approach
5000A without a ballasting resistor.
Tell me, which condition is more likely to NUKE a tube??? Ballasting load
resistor(s)
are routinely used in high power VT broadcast and scientific equipment to
limit fault
currents to safe values in event of a circuit flashover due to lightning,
load swing or
tube flash over. If a tube is run near its breakdown voltage rating (and
we never do
that with VTTC's do we ;^) ), a gamma ray (or other energetic radiation) can
spontaneously initiate a tube breakdown with stable circuit operation.
I'm presently working on a staccato circuit for a 1kW PP Hartley oscillator
VTTC. Other
folks on list (David Trimmell, John Freau) have used derivatives of a
circuit I devised.
SCR's seem to work better then TRIAC's or transistors (far more robust,
much higher
peak power ratings, and surge i^2t ratings then IGBT, transistors or FET's).
For diodes, I have constructed series strings with 1kV, 3A diodes in series
on perf boards. I do suggest 1M, 1W resistors across diodes for voltage
equalization. (1/2W is 500V rated across body, 1W is 0.75kV across body).
Number of diodes should be commensurate to hit both of the following
restrictions:
1. Sum of diode Vrrm => 2.5X Vpk sec, AND
2. Vbreakdown across resistor bodies => Vpk from doubler.
And finally, W A R N I N G!!! This circuit is very dangerous and should be
considered
LETHAL. Persons have been killed by working on MO without taking proper HV
precautions. Make sure capacitors have bleeder resistors across them to
remove HV
automatically. A 1M resistor of 10 -100K 1W units wired in series on a
perfboard
will discharge residual energy in a 1uF level shift cap in less then 10seconds.
HV MOT should be energized through a variable autotransformer to gradually
raise
voltage. Instant ON here will guarantee the magic smoke release from the
diodes
(and maybe the capacitors too).
Hope this information will be useful.
Regards
Dave Sharpe, TCBOR
Chesterfield, VA. USA
cwillis-at-guilford.edu wrote:
Hi Dave,
I think what you are recommending is that the condenser be placed in
parallel with the
rectifier. Right now I have
two
microwave-oven caps in series there, for about 0.5 microfarad (with my
MOT modified for
25% more voltage, this is
clearly not a good enough voltage rating even in series, but as long as
they don't
explode things are cool!) Do you
know if there are "standard" values for the capacitor in the
level-shifting part of the
circuit? I use as many
oven-caps
as I have on hand for this part, in parallel, and I have seen that the
more, the
better. But certainly there is a
point
where more capacity becomes irrelevant? and what if the capacitor is
resonant with the
MOT secondary?
I have considered the idea of a limiting resistor in the B+ supply. It
will waste good
power, as your 200-watt value
indicates. How did you decide on the 50 Ohm value? Would making it of,
say, 10 Ohms
at 40 watts still be effective
in any way?
And finally, have you ever had problems with blown devices for staccato
controllers?
(like 555-timers or the
switching triac of transistor?) I have been postponing construction of a
staccato
system since I felt that if I could
not
keep my rectifier safe, I had no chances with much more delicate silicon.
Best Regards,
-Carl