microwave paint 2

[ssedlmayr]

I am trying to coat a glass vessel with a microwave paint that will contain the microwaves within the glass vessel. Can anyone tell me what thickness of paint I should use, as I assume the paint will be silver. I looked at specs for copper paint and it appears only to go to 1GHZ, and I need 2.45GHZ reflection. The field strength is around 50 w/cm squared.
If you have any other suggestions for the paint, I am open. I have tried building a stainless steel enclosure around the vessel, and it still emits quite a bit of radiation, and I also tried an aluminum vessel. Would coating the inside of the aluminum with silver paint help? the vessel is shaped like a cylinder, and so is the containment vessel, but I cannot get the levels down to consumer safe levels yet. Would several redundant cylinders help? Maybe attenuating a little bit each cylinder? Also, do each cylinder need to be grounded. From what I understand, the reflective layer should also be grounded.

thanks for your answer, and if you have a proprietary answer, then maybe we might have a employment situation or consulting arrangement. We are in need of a top notch microwave engineer with thinking out of the box ideas.

 

[biff44]

are you completely enclosing the vessel? If not there will be significant leakage. For exampe, if the vessel is a flask, you would need a metal stopper in the top of it too.

If youwant to coat a glass vessel, take it to a vapor deposition house and have them deposit pure metal on the glass surfaces. You can use any good conductor, like gold, silver, copper, tin. It would have to be at least 5 skin depths thick, which might require electroplating after the metal is sputtered on first.

You can use a metal loaded paint. Silver is a good one. But that is not going to be as good as a pure metal coating, as a paint has metal bits AND some sort of non-conductive binder. Depending on the fill, it will stop being very conductive at some high frequency, possibly in the 15 GHz region.

But if you havea crack anywhere where the vessel metal does not contact the metal stopper, there will be a lot of RF leakage.

 

[biff44]

BTW, if you had to have an open top of the vessel, you could make a vessel that had a long circular tube out of the top that was also conductively coated. If the diameter ofthe tube was very small compared to the RF wavelength, then the tube would be a "cuttoff waveguide" that will not let much of the RF leak out of the vessel. You would want the tube to be perhaps 2 to 4 wavelengths in lenght

 

[logbook]

Let’s talk numbers here. The skin depth in copper at 50Hz is 9mm. Scaling to 2.45GHz gives 1.3 microns. Hence a coating between 5 and 10 microns thick will be adequate.

Multiple layers separated by say 1mm will give a dramatic improvement in shielding. The layers do not need to be earthed and are probably best left isolated. The biggest problem is joining the shield to make it continuous. This is made worse when you enter the shield. If you enter the shield with say one power wire and you fail to use a filtered feethru, the shielding will be defeated (will not work). This would be the prime reason why your shielding, whilst looking complete, will not work.

 

[ssedlmayr]

thanks for the replies...They are great. I do have a question about copper though. Everytime I see stats on copper paint for shielding it usually states it is good up to 1GHZ, not much beyond. However, I am working at 2.65 GHZ. Am I missing something?

My first vessel I just wrapped with a thin copper metal and it worked very well.
That was before I looked up the specs for copper, and could not find where it was good for 2.45 GHZ.
Sometimes this microwave stuff is just magic to me. I am a physicist and electronics engineer, but I really salute you microwave guys. Sometimes things work, and the next time they don't.
I have tried coating it (painting it) with silver, then copper, then silver, and it worked pretty good the first time. The next time I tried it, it did not work very well. Maybe my technique and not paying as much detail to it as the first time. The two cylinders around the vessels were a failure. Maybe I need to put filters on the power lines to the transformer, as I am using the old transformer, capacitor, diode circuit to feed the magnetron.
I did notice the inverter supplied more and better power to the magnetron than the transformer circuit did.

thanks for the great posts.

 

[logbook]

Copper, silver, gold or any metal is going to be good to any frequency. I work at mmwave frequencies, amongst others, and beyond 100GHz. Metals contain and shield just fine at all frequencies of interest to both you and me. The company may be specifying the paint up to 1GHz because of their measurement ability. Alternatively the filler may be making the paint so poorly conducting that it needs to be thicker than a normal paint thickness to be effective.(Give us a link to the material you are using and we can comment further)

If you use vacuum deposition of the metal then you will not have a problem. If you use foil the critical part is how well you join the sheets. Both faces needs to be really clean and you will want to overlap the sheets by say 2cm to get a low impedance joint.

>Maybe I need to put filters on the power lines to the transformer, as I am using the old transformer, capacitor, diode circuit to feed the magnetron.

Yes! You absolutely need filters on ANY wire going through the shield. Even if the wire is not connected to anything at all, say 1 inch outside the screen and 1 inch inside, the shielding will appear to be completely useless. What you have is a receive antenna on the inside and a transmit antenna on the outside. The shield might just as well not be there in this case.

 

[ssedlmayr]

thank you again. that explains why we thought we had a "ground leak". We used a spectruum analyzer with a rf probe and found that the ground from the power plug was emitting a strong rf signal all the way back to the ground, but when we measured the continuity, it was what it was theoretically was supposed to be for the guage of wire we were using. We also found it on the power leads. It was really baffeling us.

the link to the paint is:

http://www.lessemf.com/paint.html

The silver paint that I used:

http://www.2spi.com/catalog/spec_prep/silver-paint.shtml

Best Regards

 

[logbook]

Changing units gives a lot of scope for errors to creep in. However I will give it a go and you can check the sums.

Copper has a resistivity of 1.7E-8 ohm.metres

The copper paint is 0.3 ohm/sq for 1 mil (=0.001") thickness. This is 0.00762 ohm.mm or 7.62E-6 ohm.metres.

The resistivity of the paint is therefore 448x higher than the pure metal. This means that the skin depth is sqrt(448) times higher, ie 21x higher. Hence the skin depth at 2.45GHz is actually 27µm rather than 1.3µm. So the minimum material thickness needs to be at least 100µm.

It looks like you should try at least four coats to build up the thickness to a suitable level.

 

[ssedlmayr]

I tried coating a new vessel with the silver paint, and have put on three layers. I will try a fourth layer now. When I used the magnetron with it the readings from the microwave detector fluctuated. It started out low (less than .4 mw/cm^2), then it started building up, dropped, built up more, then dropped, and kept this cycling going on from .4 upto over 9. Have you ever seen anything like this? It is almost like the shield on the vessel was acting like a RF capactior, building up a charge, then discharging, building up a larger charge, then discharging, etc. The time in between the cycling seems to be about 1 to 2 secs.

 

[biff44]

Magnetron??? Did you hear a cracking sound each time, like the atmosphere arcing over?

 

[ssedlmayr]

Biff44:

No, no arcing has occurred. I have coated vessels this way before and used them, and the last one lasted nearly three years. However, I let this one run dry without any water in the vessel and the magnetron and the vessel did not turn out so good. This phenomenon that I am describing is something that I have seen with a new vessel design. Before the design was irregular, and now I am using a vessel that is a rather perfect cylinder. I was wondering if I haven't made a wave guide that might be resonating somehow. My prior vessels were irregular cyclinder types made by hand blown glass and joining pieces together, where the later ones are made on a glass machine and are rather uniform.
My question relates to whether a microwave oven, because of its square shape, is a non resonating wave guide that tends to attenuate the rf, where a cylinder becomes a resonating wave guide that acts like an antenna.

What is throwing me is this slow build up in rf energy from the shielding layer that cycles. Is this from the coating not be right, the coating being too small, or from the shape of the vessel?

 

[Higgler]

did you mention what power you are putting into this cavity?
Curious what the purpose is, is it a fun with plasma generation experiment?

If you only trying to shield the outside world from the rf internal to the vessel, then your experiment of building a stainless steel cabinet around the vessel was a 99% answer and you didn't realize it. You needed a little help connecting that metal enclosure to your rf source to prevent leakage at the interface.

If you look at a microwave oven, open one up and remove the little cardboard plate inside, you'll see a rounded proble in a small 2"x4" cavity. This probe/cavity which provides the rf source energy is perfectly contacting the metal box of the oven. I really believe that you didn't have a good contact between your rf source and your stainless steel cabinet hence you saw high leakage levels. RF leaks easily through any small crack opening. Example, if in your microwave oven, the rf source and the oven cabinet had a 0.001" gap for even one inch length, you'd have arcing and high rf leakage. The door on a microwave oven does have a small gap with a magic length of 1/4 wave which helps, seemingly by magic (though not really) to keep rf in. The door gap is far away from the rf source which helps prevents arcing (unless you get some food in it and it bulges a bit).



kch (antenna engineer)

http://www.toyon.com

 

[logbook]

It sounds as if the magnetron is not coupling into the load (water) effectively enough to prevent instability. If the load is inadequate the reflected power will be too great. It sounds as if the magnetron is squegging.

The loss in a rectangular cavity should be similar to the loss in a cylindrical cavity, ie low compared to the load. An irregular cavity will make simple modes more difficult to excite. In a rectangular cavity you should have a good idea where the peak electric field will occur. In an irregular cavity it would be difficult to guess where the field maxima would occur. Try moving the load.

 

[ssedlmayr]

thanks logbook.
However, I cannot move the load due to the design. I can increase the load quite a bit. I saw in another forum you talked about HFSS and modeling a waveguide. Do you think that HFSS is capable of modeling a magnetron antenna output, surrounded by water in a vessel, then coated on the exterior. Would it give me the loading of the water, the VSWR, the attenuation, etc? I know nothing about this package, and was wondering if it was worth buying it, or if there are consultants out there who could model the invention with the software to optimize the output wattage strength and the shielding? It does behave like I am using too much wattage in the magnetron and that the magnetron is squegging. Also, the shielding might be building up a charge on it, then dissapating it, and building up again.

 

[Higgler]

I'll suggest a consultant for any rf analysis, he uses both HFSS and CST Microwave Studio (competing software packages). Jim Reed is his name from Austin Texas. He used to work for HFSS and switched to CST after trying their software. He came to Toyon and gave an excellent 3 day CST course.
Here's a link to his website.

http://home.grandecom.net/~optimaldesigns/index.html

Jim doesn't give me income for referrals, but tell him one beer would be nice (one beer keg).

kevinchiggins

http://www.toyon.com

 

[logbook]

HFSS can certainly model the antenna output, but it is not a simple $200 program. I think the annual licence fee is more like $15000 or more. HFSS and presumably Microwave Office are formidable packages and the learning required to make them work accurately is far from trivial.

Getting a contractor with their own licence and the skill to use the package is much more cost-effective.

 

[CSLJ]

the fluctuation of the microwave leakage detector is caused by the pulse-type working mode of the domestic microwave oven. by simply changing the duty-ratio, the output microwave power is changed.

 

[biff44]

Oh, by all means hire a consultant to analyze it for you! You would spend $50K between buying HFSS and learning how to use it, and your end result will be highly questionable since you need to interpret the data thru experienced eyes!