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

Re: Discharge impedance questions. (RG-8, matching)

Hi Jim,

At 08:02 AM 03/13/2000 +0000, you wrote:

>> However, I was always under the impression that if you have a 50 ohm
>source
>> and 50 ohm cable and the forward power is say 400 watts and there is no
>> reflected power, that you must be running into a 50 ohm real load.  In
>that
>> case, the cable length makes no difference other than phase delay.
>
>True.. but is the source truly 50 ohms?  Most high power sources have some
>sort of matching network inside the generator to transform the impedance it
>sees into the cable into whatever the power amp puts out.  At that point
>the reflected power (at the generator end of the cable) will be zero, even
>if the impedance isn't 50 ohms.  [ The whole match vs reflected power vs
>impedance issue is a source of many misconceptions in the ham radio
>business where someone thinks that an antenna tuner at the transmitter end
>fixes the mismatch loss in the cable. ]
>
>In your case, the cable is SOOOO short (compared to a wavelength) that the
>effect is negligible, though.  If you were looking for gnat's eyelash
>precision, I'd worry more about the connector mismatch.
>

The generator does have fancy load match capability but I am not using it.
I was trying to run it into some other loads and it seemed very sensitive
to not having a nice 50 ohms on the output.  Running into a wirewound 50
ohm resistor was a disaster!  I was running some MicroSim models of the
thing and there are indeeded nice close to 50 ohm points in the tuning
range which are no doubt what I am hitting.  Problem is there are a lot of
varibles in the model so it is had to get a good feel for what is going on
but I am starting to see a few finer points.  I think 50 turns on the
primary is now too much and is hurting the matching.  I will try and
experiment more with this.  The generator is "supposed" to look like a true
50 ohm source but I can hook up my fiber probes to see the exact current
voltage and phase at the load end of the cable.


>
>> 
>> MicroSim says I should be using 50 turns on the primary instead of 10 to
>> get about 50% more power to the load (resistive part).  Perhaps that
>would
>> match the resonant frequency to the impedance match points better...
>> Really, all we are doing is matching the 50 ohm real source impedance of
>> the generator to the impedance of the arc.
>
>Exactly.. and since the arc isn't a nice impedance, it will be tough to
>keep a stable match (as your designers of automatic matching networks for
>plasma etchers have found, no doubt).

Actually the generator does seem rock stable when running.  If it is not in
tune there is some oscillation.  I think this is due the the response of
the Tesla coil and the response of the generator's control loops not
matching well.  However when in tune it is not a problem.  I think the coil
has a bit differnet response than what the generator is designed to run but
no big problems.

>
>> 
>> I am not a genius at all this load match stuff but I'll see what I can
>come
>> up with.  I'll dig up a manual to the power supply too.  I think it spent
>a
>> lot of time talking of such things.
>
>Impedance matching is not something that is trivial or initially
>intuitive.. Well it is in a general sense.. you turn the knob until the
>reflected power is zero.  However, if you want to know what power is really
>getting transferred, and so forth, it is a bit tougher.  You can have a
>great match, but have surprising losses in the cable. A 20 dB resistive pad
>makes a fine matching device... the source sees 50 ohms almost regardless
>of what's on the the other end. Not much power gets to the load though...
>
>The other real problem is that most signal generators that put out any sort
>of power (more than a few mW) are often not very close to a 50 ohm source. 
>I think my 100 watt ham transmitter is 50 ohm resistive output at only 2 or
>3 frequencies.  That 1:1.5 VSWR spec allows for a lot of variation!
>

The manual just had bare basic stuff in it.  I may be dissipating
considerable power in the secondary coil's resistance which may be a
problem at higher power.  The #24 wire may have to carry and amp are two of
350kHz RF which may be too much.  Just have to try it.  Fiber probes can
confirm the impedances and loads easily so I'll look into that.  These
generators are supposed to be very well behaved and controlling 350kHz is
far easier than 30MHz.

Thanks for the great ideas!

Cheers,

	Terry


>
>
>> 
>> Cheers,
>> 
>