X band cavity design and microwave safety

[coppercone2]

how do you make a reactor in there ? like if you need a condenser, gas flow, adding chemicals, mixer, etc?

Or do you mean measure the result of a chemical reaction.

[ChristofferB]

Well usually it's as simple as mixing reagents in an EPR tube - 3-4 mm dia quartz test tube - but some cavity also have a quarz window so mixtures can be irradiated with UV,  or visible light for the study of photocatalyzed reactions. But I've also seen high pressure hi temp probes, as well as liquid helium cooled cryoprobes.

[coppercone2]

But can you make the geometry so you can study a fed continuous reaction?

[ChristofferB]

You mean like a flow probe so reaction mixture is continuously siphoned from a reaction vessel, through the cavity and back in a loop? Yeah most likely. It is commonly used with many other kinds of spectroscopy at least.

[ChristofferB]

Cavity Coming along nicely! Unsure in which plane to put the coupling screw.

[coppercone2]

You mean like a flow probe so reaction mixture is continuously siphoned from a reaction vessel, through the cavity and back in a loop? Yeah most likely. It is commonly used with many other kinds of spectroscopy at least.

I was thinking more like a reaction occurring inside of it, to study what happens when chemical reactions happen (being careful for the reaction not to cause a gas or precipitate to form (avoid phase change), if liquid-liquid reactions like this even exist?), or to vary the temperature of a fluid loop etc

[ChristofferB]

That is very possible. In fact that's the standard way of doing it.

The curing of epoxy resin proceeds via a radical mechanism, and that reaction should be fairly followable on EPR.

[coppercone2]

thats an interesting way to verify if the epoxy you have is good, see if cure time ends when they say it does. Tensile strength test might not tell you everything (might be something going on the side that it clues you into even if tensile came out good).

[ChristofferB]

One could even set up software to monitor the integral of a range in the EPR spectrum over time, seeing the reaction rate decaying. I'm actually not sure if the polymer industry uses EPR at all, but maybe they should.

[coppercone2]

if you can build it, it seems like a good QC for critical applications where you have it shipped climate controlled and have to keep it in a factory in a refrigerator. Cheap

[StuartA]

EPR is a new one on me, but I'm not telling you how long ago I did my chemistry degree! Wiki says it was first observed back in 1944, but I wonder if it had to wait for better electronics and computing power to become useful?

Just curious, what does an EPR spectrometer cost?

[ChristofferB]

I think it was one of those things that was researched, but not used for stanard analysis for a long time. the electronics was ready almost from the get-go, since all the X-band wrangling could be done with surplus radar parts (of which there was a LOT from the 1950s and onwards). I think the first commercial EPR is from 1959 but I may be wrong.

I'd suspect a new, modern dual-freq. one costs north of 500.000$. The magnet is a ton-heavy iron chunk, with watercooled copper coils run by a DC PSU like an arc welder.

[StuartA]

I think it was one of those things that was researched, but not used for stanard analysis for a long time. the electronics was ready almost from the get-go, since all the X-band wrangling could be done with surplus radar parts (of which there was a LOT from the 1950s and onwards). I think the first commercial EPR is from 1959 but I may be wrong.

I'd suspect a new, modern dual-freq. one costs north of 500.000$. The magnet is a ton-heavy iron chunk, with watercooled copper coils run by a DC PSU like an arc welder.

Thanks for that info. Some time ago, I worked for a company which was exploring the limits of what you could do in terms of 'chemical analysis' by making measurements of permittivity; not a new idea, but it came as a spin-off from some other work.
I seem to recall that there was a German company offering quite a low-cost device (<<$500,000) which followed the curing of polymer formulations by watching the change in its permittivity. Unfortunately, our Management decided to abandon the work after getting just one product on the market (an oil 'analyser').

[LaserSteve]

Sounds like a good use for those X-Band Stellex Yig Oscillators that are all over Ebay. 

Steve

[coppercone2]

it sounds like temperature stability is of the upmost importance with this design.

i think the best option has to be selected here, I am not sure Yig is the best choice. Its good for sweeping but he said multi-frequency response and calibration is extremely difficult.

[LaserSteve]

The Stellex yigs are quite stable, and a YIG/Synthesizer board  pair is ~100-150$. I own a pair.

Steve

[coppercone2]

what does the stability of the various oscillators look like on a theoretical level? yig vs DRO vs CRO

or others.

if there was a table of expected behaviors it would be easy to make a confident decision

[ChristofferB]

I've never heard of YIG's being used for this application. I think the industry (Bruker and Varian, really. only commercial EPR companies AFAIK) went directly from surplus radar tech (Klystron sources, later Gunn oscilators) to DDS.

[coppercone2]

10 GHZ dds for a single frequency application?

you mean like one  of these?
https://www.analog.com/en/about-adi/news-room/press-releases/2017/5-2-2017-analog-devices-wideband-rf-synthesizers-size-reduction-versatility-performance-to-13-6-ghz.html

that one only needs 600MHz clock, 40$ for the IC

*thought you meant AWG

[coppercone2]

I was re-reading the thread and I noticed you have one misconception ,you can silver solder fine with a normal mapp torch, or maybe even propane. You just need to insulate it. Also air-acetylene might be something you are interested in, i.e. smith silver torch (this is a jewelers air-acetylene torch, as opposed to the smith little torch, which is a oxy-acetylene torch. they look similar and are made by the same company but two very diff pieces of equipment). The silver smith torch would be 200$ with regulator + ~$80 for the bottle and 30$ for the fill for a few hours run time, if all gotten new. If you don;'t mind getting a non premium large-size torch, it would come out cheaper used then the 'gold standard' smith set. Add another $100 for the silver rods and flux, cleaning brush, etc. Expect to pay 20% for the chinese knockoffs on ebay, but obviously you need a garage, I think 10cU of acetylene would still be hazardous in an apartment. The newer mapp torches with the 'spiral' flame would help, and you can preheat using a torch then use one of those small oxy-mapp sets for $50 to actually direct silver solder flow. BTW you will be impressed by the fillet size/fluidity of actual high silver solder, it tends to 'plate' things after the gap is filled if there is excess, it does not pool up like solder does. The highest fluidity rating of commonly sold braze is the 56% silver type. So it will look like a 'stain' of pure silver that is very thin, rather then globs you get from excess solder, so long you heat it, think low surface tension fast spreading, higher quality I think, the geometry will be more stable. If you silver plate the whole thing afterwards with the cool-amp powder, it might be not detectable?

And if you want a real number on run time with a 10cubic foot acetylene tank (MC) on a smith torch, I can extrapolate from the manual (and this will be lower if you use insulation)
brass: 2 inch diameter, torch life between 1hr to 30 minutes using a #4 or #3 tip respectively (depends on how fast you need the heat, if you fixture good you can use the smaller tip that has a 10cfh flow rate, instead of the larger 20cfh flow rate). I expect heating time to be 5 minutes or less.




My calculations on fillet size effecting the waveguide were based on microwaves 101 specification of common waveguide dimensions being 5%.

I am thinking about this because the waveguide random pieces I have happen to be the correct frequency range and I have not found a use for them in over four years.

The signal power necessary is under 1W (you wrote the bare gunn oscillator without a amplifier is adequate at 100mW 50ohm)? In that case I have all the equipment other then the apparatus itself. I think I just need to fix one amplifier control card (a few transistors I never bothered ordering, the PCB has been sitting on my stereo for 2 years now).

Does the test experience any power dependence from 100mW-10W ? 

And what is the expected measurement noise level from this equipment? How sensitive does the pickup need to be (LNA requirement)?

[TheUnnamedNewbie]

Sweeping frequency is tough to do with good repeatability. Sweeping a DC electromagnets current with a slow (minutes to hours) ramp gen is the option people has gone for. More recently, short bursts of RF energy and then fourier transforming the decaying echo. EPR is still usually CW swept field.

I'm curious why sweeping frequency is tough to do? The frequency of a signal is one of the things we can measure and control most precisely of all physical properties - we can generate tones in the lab that are accurate to ten decimal places or more quite 'trivially', but we need entire industries and years of expertise to create a voltage or current with have as many significant digits.

I'm not an expert on the topic, but it would seem to me that you can use a coupler (with 20 dB or 30 dB or more attenuation) to take the oscillator frequency and feed it back into a classic PLL architecture or similar, and sweep it with very high frequency, or even just linearly sweep the control voltage and just measure the frequency at the same time as you measure the detected power and store the power-vs-frequency with very high precision?

(it might be easier to just sweep field strength, as I said, I am not familiar with the details of the topic, but the statement that you can't sweep frequency repeatably just seems an odd one to me)

[coppercone2]

Sweeping frequency is tough to do with good repeatability. Sweeping a DC electromagnets current with a slow (minutes to hours) ramp gen is the option people has gone for. More recently, short bursts of RF energy and then fourier transforming the decaying echo. EPR is still usually CW swept field.

I'm curious why sweeping frequency is tough to do? The frequency of a signal is one of the things we can measure and control most precisely of all physical properties - we can generate tones in the lab that are accurate to ten decimal places or more quite 'trivially', but we need entire industries and years of expertise to create a voltage or current with have as many significant digits.

I'm not an expert on the topic, but it would seem to me that you can use a coupler (with 20 dB or 30 dB or more attenuation) to take the oscillator frequency and feed it back into a classic PLL architecture or similar, and sweep it with very high frequency, or even just linearly sweep the control voltage and just measure the frequency at the same time as you measure the detected power and store the power-vs-frequency with very high precision?

(it might be easier to just sweep field strength, as I said, I am not familiar with the details of the topic, but the statement that you can't sweep frequency repeatably just seems an odd one to me)

I guess it needs high power and receiver linearity. You are saying to connect a VNA to it. All I can think of is that the sample itself is nonlinear with power, even if you compensate with measurement. I think low signal VNA test subject is supposed to not be effected by power in terms of its response. The most powerful VNA I have seen myself is 150mW as a special feature, and this is only for a 300MHz unit.

The regime transition point between small signal and large signal S parameters might be much more different with a weird circuit element (i.e. epoxy) then it is with a microwave filter made of microwave components. In literature large-signal S-parameters are treated differently and are considered a more difficult, expensive test. The VNA cheats with measurement by assuming power invariance, which is OK for highly linear stuff and low power levels. And it might be bad for business if you put a whole bunch of clauses in the use of analytic equipment (it causes confusion if you say 'these things are ok with these settings, these settings are required for this, it might give you weird outputs for this', generally a simple measurement tool produces better results and prevents bad data from being published, if someone did not read a manual, and a company is more likely to buy it if they think its bullet proof. They want to say 'we put it in this machine and there are no doubts'. If it has huge adoption and everyone knows about it, then you can start putting complex features in it, but this sounds unheard of.. they want the market calm.


**** microwaves 101 ** just a hunch, not sure.

Types of S-parameters

When we are talking about networks that can be described with S-parameters, we are usually talking about single-frequency networks. Receivers and mixers aren't referred to as having S-parameters, although you can certainly measure the reflection coefficients at each port and refer to these parameters as S-parameters. The trouble comes when you wish to describe the frequency-conversion properties, this is not possible using S-parameters.

Small signal S-parameters are what we are talking about 99% of the time. By small signal, we mean that the signals have only linear effects on the network, small enough so that gain compression or other non-linear effects do not take place. For passive networks, small-signal is all you have to worry about, because they act linearly at any power level (at least until you blow them up).


Large signal S-parameters are more complicated. In this case, the S-matrix will vary depending upon the input signal strength. Measuring and modeling large signal S-parameters will not be described on this page (perhaps we will get into that someday)

Mixed-mode S-parameters refer to a special case of analyzing balanced circuits. We're not going to get into that either!

Pulsed S-parameters are measured on power devices so that an accurate representation is captured before the device heats up. This is a tricky measurement, and not something we're gonna tackle yet.

[Kleinstein]

For the oscillator I would have a look at the local oscillators for satellite TV receivers / LNAs (the part mounted at the dish). AFAIK they are some kind of dielectric resonators, somewhat similar to YIG, but more fixed frequency. The specialty with YIG is that they are tunable over a wide range and still high Q and thus low phase noise.

In more complex systems they may use some kind of DDS for modulation, if they want to do odd thinks like pulse echo and similar things. So the DDS part would be much lower frequency and likely not the usual sine but more like some arb generator / pattern generator.

[coppercone2]

called a DRO, dielectric resonance oscillator, if that is what you mean

https://en.wikipedia.org/wiki/Dielectric_resonator

[ChristofferB]

Whoa, I'm way behind on this thread I can see!

With regards to the solddering: A proper acetylene set is out of the picture. Too expensive and I don't want to store the fuel. The only air-fuel set that's a real choice is this kind with disposible canisters, oxy-propane:



With regards to the swept frequency: I don't think the cavity can be made wideband enough that you wouldn't see a curved spectral baseline, as resonance changes.
I'm unsure. The only thing I know is that in 70 years of magnetic resonance, nobody ever uses swept freq. As I don't know all the theory, and are just eyeballing quite a few aspects of this project, I'd rather not go completely against established systems.

I also think the sample (and nature of the cavity) is quite nonlinear. Not to mention the two 10mm holes I need to drill in the cavity to let in the test tube.

I just got my lock-in amplifier, so signal detection will be trough a WR90 1N23 'crystal' detector mount, fed to the lock-in, the magnetic field (or frequency) modulated at something like 10 KHz.

I'm handing in my MSc thesis in 2 weeks, so I don't have the greatest amount of time for a while, but I'll be back once I make progress!