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

Re: Magnetizing current in SSTCs, my previous posting



Original poster: "Stephen Conner by way of Terry Fritz <teslalist-at-qwest-dot-net>" <steve-at-scopeboy-dot-com>

At 22:12 15/05/03 -0600, you wrote:
>Original poster: "jimmy hynes by way of Terry Fritz <teslalist-at-qwest-dot-net>" 
><chunkyboy86-at-yahoo-dot-com>
>
>Hi Steve,
>
>I am actually not driving it at either frequency. I am driving it near the 
>"zero" between the poles.  Using the secondary as feedback will give you 
>this frequency, however primary current feedback would give you a 
>pole    :-( . This means that the power factor is not perfect, as I had 
>thought, but it is close, especially at low coupling. This brings up the 
>problem of switching loss, now that I know it isn't perfect ZCS. The times 
>that the phase is furthest off is at the beginning and end, where the 
>current is lowest. Lower coupling just gained another advantage in my 
>system, but we'll have to see how the sparks react to low coupling  ;-). 
>Why don't you and Richie do your "thinking" up on the list?

We're just not sure how the darn thing works. As far as we can tell, a coil 
with a series tuned primary has to have a slower rate of energy transfer 
than the same setup with an untuned primary. That is because when you turn 
on the inverter, you have to ring up both the primary and the secondary. If 
you like, the energy from the inverter goes into the primary circuit and 
then couples out into the secondary. With an untuned coil it all goes 
straight into the secondary.

Therefore we think untuned coils ought to perform _better_ in short pulse 
applications than the DRSSTC would. The improved efficiency due to the 
shorter rise time (hence achieving the same energy transfer in a shorter 
burst) might more than offset the losses due to magnetizing current. That 
is, assuming that the untuned coil _Does_ have a shorter rise time. You 
claimed that the DRSSTC is faster.

However the suspiciously low top voltages and massive primary currents 
reported by K.C. Herrick from simulations suggest that all might not be 
well. Also in the ancient tome of Craggs & Meek ("High Voltage Laboratory 
Technique", 1954, pp. 101-2) it says that untuned coils are more efficient. 
That was for CW applications, though, where the output was rectified to 
drive x-ray tubes etc.

In order to clear the confusion I think we should start talking in terms of 
M=(Energy input to secondary/(primary I^2*t)) taken over the length of one 
burst. I^2*t (i squared t) is found by using the RMS primary current 
(calculated over one burst), squaring it, and multiplying by the burst 
length. It is a rough measure of the heating energy in the inverter 
devices. Energy input to secondary is found by 0.5*Ctop*Vtop^2 at the first 
peak after the burst ends.

This is all easily done in MicroSim Probe or whatever. The higher "M" a 
design has, the bigger a burst it will be able to produce before fusing its 
transistors, thus the better it will be. To get the ball rolling: The OLTC 
that I built has M=23 roughly (burst length 25us, RMS primary current 380A, 
energy to secondary 0.2J)

Richie has built SSTCs with both tuned and untuned primaries, and found 
that the tuned coil sparked further with a lower coupling. He doesn't 
subscribe to the Pupman list any more.

Steve C.