NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/sci/electron/22224

Text File | 1993-01-07 | 6.9 KB | 124 lines

Newsgroups: sci.electronics
Path: sparky!uunet!psinntp!newsserver.pixel.kodak.com!kodak!ornitz
From: [e-mail] (Barry x24904/ER/167B-TED)
Subject: Re: Microwave Oven Plasma Etcher? (definitions)
Message-ID: <[e-mail]>
Followup-To: sci.electronics
Keywords: microwave, Reike and Hartree diagrams, Gerling Labs
Sender: Barry L. Ornitz
Organization: Eastman Chemical Company
References: <rich.726361718@pencil> <[e-mail]> <[e-mail]>
Distribution: net
Date: Fri, 8 Jan 93 01:15:50 GMT
Lines: 109

In article <[e-mail]> [e-mail]
(John Whitmore) writes:
>    In the context of the original question (lowering the power
>density in the microwave oven), this is irrelevant.  First, the
>load is the contents of the microwave oven (and this can, we all know,
>vary quite a lot without stalling the oven).  Second, we don't
>care about the exact frequency (nor the exact power); only large
>power-output changes are of interest.
>    The load would only change markedly if the feed horn
>from the magnetron were altered.

Microwaves are no different than high frequency RF sources in that they expect
low VSWR from their load.  High VSWR means that the magnetron must dissipate
reflected power as heat - decreasing the tube life.  Most home cooking
applications which present variable loads to the microwave occur for short
times.  Cooking a roast, however, gives you a fairly stable load.  Note that
many ovens warn you to place a glass of water in them (giving them a stable
load) when cooking other relatively non-absorbing items.  I would assume the
plasma etching application is a long duration operation.

>>A Hartree diagram ...
>    This is obviously of concern.  I would have thought that the
>magnetron was just a diode, and that changing the input current (by
>substituting a smaller capacitor in the charge-pump power supply)
>would alter the output power.  The implication here is that at too-low
>voltages the magnetron doesn't oscillate.

Quite true.  The efficiency also changes rapidly with voltage too.  Magnetrons
are far from simple diodes, however, in their characteristics.  Changing the
capacitor in the doubler circuit is NOT the way to do this.  Remember what I
said about the transformer having high leakage reactance.  This is there for
a reason, not because cheap transformers are often made this way.

To get a reasonably stable output from a microwave, it is necessary to provide
substantially constant anode current to the magnetron independent of the
supply voltage and load conditions.  For magnetrons with non-adjustable fields
(i.e. permanent magnet types like those found in home ovens), some means must
be found to adjust the voltage to regulate the current.  In home ovens, the
technique used is very similar to that in Sola constant-voltage transformers.
The power transformer uses its leakage reactance much like a saturable
reactor which is resonated with the capacitor at a frequency higher than the
line frequency.  As the anode current increases, the inductance drops which
raises the impedance of the (series) resonant circuit which in turn drops the
voltage and limits the current.  This is a low-cost system since it has no
active components but it has the disadvantage that power adjustment is not
possible.  This is also the reason that home microwave ovens are so sensitive
to supply frequency variations.  [It also explains why microwave ovens often
fail quickly with non-frequency stabilized camping generators and solar/hydro/
etc. inverter systems.]

>    Again, I think this is not prohibitive: magnetrons are more than
>85% efficient, and this implies that the input voltage (which is
>unfiltered) is in the operating range for a substantial fraction of the
>whole 60-Hz duty cycle.  That implies that the input voltage can vary
>by a factor of four or so before the magnetron stops oscillating.

Home magnetrons are rarely this efficient.  Most are in the 50 to 60% range.
Read the section above about the resonant power supply again and then measure
the actual waveform as seen by the magnetron.  It is hardly sinusoidal - it
will be more like a square wave (like the output on non-harmonic neutralized
constant voltage transformers).  The voltage over a cycle varies less than
you think; the regions where the voltage falls are short compared to the
regions where the voltage is relatively constant.

Changing the value of the capacitor is a fast way of destroying your oven.
The design of the oven power supply is fairly precise, but the results from
such a simple system are surprizingly good.  Many home ovens can maintain
less than a percent variation in anode current with more than a ten percent
change in line voltage.

>    It might be impossible to get a 100:1 output power range, but
>10:1 still looks reasonable.  As long as the magnetron doesn't overheat
>from the current during the 'wasted' time, the microwave oven will
>still operate even if the magnetron has output efficiency of zero
>for most of the cycle.  It just means that the output power drops
>to zero faster than proportionally to the input power.

I doubt if you can get a 10:1 range from home ovens without excessively short
tube life or major modifications of the power circuit.

However, if you can change the magnetic field of the magnetron, things are a
lot different.  Using an electromagnet rather than permanent magnets gives
you lots of control.  A 1% change in electromagnet current can often give
greater than a 10% change in output power. Commercial microwave systems often
have 30 to 40 dB of dynamic power range by using a feedback loop and a power
sensor to control the magnets.  Commercial systems also often use circulators
to protect the magnetron from reflected power from load changes.  The
purpose of the circulator is to divert reflected power to an absorbing load
rather than back to the magnetron.  You waste energy this way but protect the
expensive tubes (at least they are in multikilowatt systems).

I agree with John DeArmond, just use some water elsewhere in the oven to
absorb the excess power if this is a simple one-shot experiment.

As a final note to end this discussion, the plasma itself is a load with
negative resistance characteristics under certain operating conditions.
Adding a resistive load in parallel is a good idea anyway to stabilize the
generation of the plasma.  Commercial plasma etching systems cost big bucks
for a reason.
                        Barry
 -----------------
|  ___  ________  |    Dr. Barry L. Ornitz          WA4VZQ
| |  / /        | |    Eastman Chemical Company
| | / /         | |    ECC Research Laboratories, Engineering Research Div.
| |< < K O D A K| |    Process Instrumentation Research Laboratory
| | \ \         | |    P. O. Box 1972, Building 167B
| |__\ \________| |    Kingsport, TN  37662  (615/229-4904, FAX [phone])
|                 |    INTERNET:   [e-mail]
 -----------------