Experimenting with waveguides using the LiteVNA

[joeqsmith]

kinda wonder what CAT rating this probe would have with freon, with oil, and then the proper redesigned one with oil in it.

Sounds like you bought the 6015 probe without reading the manual first as.  Even without the manual, just looking at the probe I would never have come to the conclusion that this was a good choice for AC mains work.   I take it you have little experience in this area.  This probe is very popular and I bet if you used the sight's search engine, you would find everything there is to know about it. 


***
I have an Agilent 33311 that would work to switch between S11 & S21.   Isolation is good but the insertion loss is about a half dB.   I pulled it out and it is marked Agilent, so it would be newer with potentially less wear but its an A.   

https://www.testequipmenthq.com/datasheets/Agilent-33311B-Datasheet.pdf

[coppercone2]

I read the manual it says nothing about how big of an arc it can take before it explodes and shoots hot oil (for the 6015A). Say there was a surge on the power lines.

I assume its meant for lab use like from small transformers and such, since its made by tektronix. I.e. bench top isolated circuits. For a house I assume it would arc over some where in the breaker box if that happened anyway like a spark gap. But past the house I don't think I would trust it.

For the A it does say The inputs are not rated for connection to mains or Category II, III, or IV circuits, but for the 6015 it does not have a cat rating

[coppercone2]

for the oil fill I used silicone oil that I heated in a beaker with a watch glass on top to 150C for like two hours (using immersion thermometer), let it cool down to be manageable then I passed it through a glass frit vacuum filter into a glass flask (washed and baked all of that) because there was some particles in the oil that I noticed after heating and inspecting with bright light, then poured that into the probe that was washed and baked. Not sure if dust got past the watch glass or the oil was contaminated.

It should probobly be dried in a retort but I don't have one (like the old soup & life biology experiment

[joeqsmith]

I read the manual it says nothing about how big of an arc it can take before it explodes and shoots hot oil (for the 6015A). Say there was a surge on the power lines.

I assume its meant for lab use like from small transformers and such, since its made by tektronix. I.e. bench top isolated circuits. For a house I assume it would arc over some where in the breaker box if that happened anyway like a spark gap. But past the house I don't think I would trust it.

For the A it does say The inputs are not rated for connection to mains or Category II, III, or IV circuits, but for the 6015 it does not have a cat rating

We had the freon filled ones when I started out.  There were no CAT ratings back then and so it should be a surprise that the manual does not mention them.   

I have never seen a company say anything about how big of an arc their CAT rated DMM can take before it explodes and shoots hot ....  It's just not something anyone is going to publish.  Again, thinking this is all new to you as I can't think of one time anyone has made such comments about the 6015.  Yes, we have always used them in the labs where I have worked.     

I thought there was a derate curve for the older probes when running them dry.   Personally, if I didn't work with these higher voltages, I would run them dry. 

The later probed we have are filled with a solid, so no maintenance.   I have it's little brother at home that also does not use freon.  I had pushed that probe past the limits several times but eventually made that wideband HV probe to work at these higher voltages. 

Sounds like your interest in HV far exceeds any interest in the thread topic.  You may want to consider starting a new thread in the test equipment area to discuss restoring your probe as I doubt many people are following this particular thread.   You may find a people willing to help out once you make it visible to them. 

[joeqsmith]

Shown with a Narda 3095 installed on my extender.  Added some support for the amplifier and also made a section of semi-ridged to connect the coupler's input port.  I have not yet decided on adding a relay.   

The coupler made a big improvement.  Plots showing the data I had previously collected with the extender compared with the update.  I used the same homemade standards and Agilent X11644A coefficients  to calibrate the VNA to 12.1GHz.   I had only swept to 12GHz before as the performance was so poor.   The mixers only support an IF of 4GHz and as the range is increased beyond 12.1, the performance starts to degrade quickly. 

Decided to roll the dice on used Amphenol connectors rather then trying the Chinese ones.  I was a bit worried they would have just been a waste of time.   

[msat]

Hey guys

I know this is somewhat off-topic but I felt this is the best place other than maybe a new thread to discuss this.

I finally got around to doing a little microwave experimentation. However, since I never jumped on some crappy but cheap-ish ebay WR42 horns for my coax to waveguide adapters when I had the chance, and been somewhat reluctant to 3d print (and paint) some, I instead opted for a couple of ~10GHz gunnplexers. In my recent experimentation, I used one as a transmitter, and then measured the output of the detector diode on the other which is thus used as the receiver. 

As you can hopefully see in the attached image, I printed a mount for them that allows me to adjust the angle on the longitudinal axis from 0 to +/-90 degrees in 22.5 degree increments, with the center of rotation being the center of the gunnplexer.

So in my little experiment, I was getting some baseline voltage reading at the detector when both tx and rx were oriented at the same angle (same polarization), and then measuring the decrease in voltage as they're rotated relative to one another. I wanted to compare and contrast this with what polarizer experiments seem to show.

At 22.5 degree relative angle, the received voltage is 75% of baseline. This makes sense from a trigonometry standpoint. At 45 degrees however, it drops to 25%, which is not expected. These results are consistent regardless of distance between tx and rx, and the way I orient them to get the relative angle differences. Right now I'm thinking these unexpected results are due to my hardware. Soon I'll be modifying some satellite LNBs, specifically because they use round waveguides/horns, to perform the experiment with. However, I was also hoping someone here could duplicate the experiment on their hardware and post their results if they could find the time. Any of you guys willing to do that?

[joeqsmith]

Far more on topic than much of what has been posted.      Walter Lewin had a lecture on this.  Lucky for us, MIT/France could not purge the internet of his content.   In the video, he conducts a demonstration starting at about the 1 hour mark.   

For off topic, I would have liked to have all the details about that sexual harassment case.   I can't imagine anyone who would self-mutilate rather than simply dropping a course and turning off their PC.  I searched her after the fact and while I don't know enough to form an opinion one way or the other over the case, I am glad that we still have access to the content he created.
https://www.govinfo.gov/content/pkg/USCOURTS-mad-1_16-cv-12394/pdf/USCOURTS-mad-1_16-cv-12394-0.pdf

[joeqsmith]

During my short visit on the V2Plus group.io, I had noticed someone asking about using theirs to measure waveguides.  The advice given was IMO poor as obviously anyone following this thread would know.  I attempted to post a comment to them but sadly, they once again censored it and it never showed.   

IMO, people who actually want to discuss the V2Plus4 should just ditch that group.io reflector and use a different group or forum.   

https://nanorfe.com/forum/using-V2-plus-4-Pro-for-waveguide.html

[msat]

I have watched that lecture by Walter in the past, among other ones of his. However, it doesn't provide the data I'm interested in unless I guesstimate it. I've tried to do this on a few other videos with similar demonstrations, but it's unsatisfactory.

I've been wondering if part of my issue the way I'm interpreting the values from the detector diode. I'm merely hooking up a volt meter to it, while the gunnplexer itself has the diode shunted to ground via a 1k resistor and a zener (it was there when I bought it).

Per https://www.qsl.net/in3otd/electronics/power_detectors/power_detectors.html "As known, the diode detector output is proportional to the square of the input voltage/linear with the input power at low power levels and becomes proportional to the peak voltage/square root of input power at higher levels."

I'm still trying to wrap my head around that, and how relevant it is. Like I said before, the results are the same regardless of the distance between the tx and rx (which does have an effect on the baseline voltage reading).

Tangential to this, I got my couple of "Ku" satellite LNBs in the mail, which I ordered new from amazon for ~$10 each. I was a bit premature in my purchase, neglecting to realize such new and cheap parts were highly unlikely to use dielectric resonators or really much of any discrete circuitry, making electronic mods pretty much a no-go. To top it off, naturally many of the passives are among the smallest SMD sizes available. I then thought that *maybe* I could hook it up to my TinySA Ultra and use that to measure received power, but in the almost non-existent spec sheet for the nearly all-in-one LNB IC, it states that it does some automatic gain nonsense, so there goes that idea too  The only usable component may be the waveguide, but then I need to figure out the electronics, and I don't currently have a suitable LO source, mixer or LNA. Maybe I should just go the diode detector route?

[joeqsmith]

You should start a thread for your project.  I would provide details about the goals you have in mind, your background, work you have done so far.    Others may have a similar interest and chime in.   

Problems with using this thread, it long and people don't take the time to read.  It's difficult for readers to locate sub topics.  Then, I started the thread to cover some basics on waveguides.  Anyone reading from the start may loose interest and never discover your project.   

To be clear, I don't mind you posting here but rather just offering a suggestion if you are looking for help on how to increase your exposure. 

[joeqsmith]

Finally got around to mounting the HP33311B relay. 

[xrunner]

We're both proceeding at a snail's pace. 

[joeqsmith]

Slow indeed.  With it being used,  I had to test it first to make sure the performance was still good enough.    I've thought about mounting it all into a box to make it a bit more robust and take up less room.    There are some nice extruded aluminum cases but that coupler and cables would need about 10" depth assuming I let the APC7 connector extend outside the box. 

[joeqsmith]

Roughly $50USD from China.  The faces are a bit thin and I think I would machine a new face for the front with 1/4" plate aluminum.   Something to ponder over the winter.   It's a lot of junk to cram in that small box and I would want some decent bulkhead connectors. 

[joeqsmith]

My plan is to make it a clam shell so the two boards fold up.  A while back I had posted my attempts to repair a damaged semi-ridged cable.  Thread shows the bulkhead connectors I plan to use. 

https://www.eevblog.com/forum/rf-microwave/is-is-possible-to-repair-a-hairline-crack-in-semi-rigid-coax/

Decided against running the the APC7 through the case.    All the RF connectors will be routed to the front,  P1 to the bottom left,  P1extended to the top left.   Single toggle switch will also be mounted on the front to select between S11&S21 maybe in the center, along with the 10MHz reference clock input.    Power connector may be off the rear. 

[xrunner]

Looks very professional so far. Almost like ... you might even know what your doing. 

[joeqsmith]

Made up the bulkhead connectors.  There will be a set of cables in the read that tie the boards together.  Looks like things should fit.   Of course, I get it to look nice, the performance will go out the window.   

[joeqsmith]

Mounted the parts on the lower board and made up cables for it.   Stacking the two boards, there's a good 10mm to spare.    Still debating to route out the APC7 or not.   I'm not liking the idea of stacking all those connectors.  Unlike the ham in the photo who placed several connectors in series,  at 12GHz it may present more of a problem.

[xrunner]

LOL I dispelled that myth myself a few years ago. The HAMs would say "Don't use an adapter in-line because it'll cause a few dB of loss!"

I connected seven misc adapters together and ran some tests. Tests? Tests? How dare you actually try to see what's really happening!

 

Turns out it's all BS advice. At HF seven adapters causes less than 0.1 dB loss, and at 144 MHz around 0.3 dB loss.

They don't say that anymore. 

[joeqsmith]

Looks like that video is 3 years old.  Skimming his video, looks like he may have tested to 30MHz and saw a 0.12dB loss. 

I may put everything together and compare it against the data I already collected. 

[joeqsmith]

Looking over his videos, it appears someone wrote him about it.

[xrunner]

Well, let's put it like this. I used adapters that were in good condition (what other kind would I have?  )

If you connected a few adapters used in the field, like up in the air for 10 years, with no waterproofing, with no care in the world, yea you might get a higher loss. But that's not the fault of the adapter, it's on the HAM's maintenance habits.   

[joeqsmith]

I was imagining someone using RG174 and SMA connectors.  They check the loss with their VNA or other and measure some small loss.    Then they connect it to their 2kW amp...  Should be fine...     I'll just crank up the power to compensate for the losses...

Remember that gold plated PL-259 to N adapter I used as part of my homemade Watt meter testing.   I was seeing a pretty good temperature rise with it even at 300W CW (or what ever I was running).  Wonder when it goes into thermal runaway. 

Anyway, my point was while we are dealing with small signals (<10dBm), the frequencies are in the 8-12GHz, not <30MHz.   

[xrunner]

Joe are you using in your project semi-rigid coax? It all looks like hardline. I think I want to try my hand at making up some semi-rigid lines with SMA connectors. So I wanted to know if you had used that type before. What do you use to cut the ends with? I plan on looking at some tutorials in the next few days.

[joeqsmith]

https://www.fairviewmicrowave.com/rg402-type-.141-semi-rigid-coax-cable-tin-plated-aluminum-fm-sr141altn-st-bulk-p.aspx