Experimenting with waveguides using the LiteVNA

[joeqsmith]

how do you plan on holding a tall object up on a lapping plate?

Like, the part is difficult to hold on to without tilting.

Most likely, I will do it by hand.  WR90 is fairly large and not a problem to hold onto. 

[coppercone2]

I feel like it’s really uncomfortable to get a rythem going with tall objects in figure 8

[joeqsmith]

Standards were done by hand as well.  We are not talking about a lot of material. 

[coppercone2]

I imagine the guy doing the 4 inch gauge blocks in the kit was NOT a happy camper compared to the guy doing 1/4 inch blocks

[joeqsmith]

I imagine the guy doing the 4 inch gauge blocks in the kit was NOT a happy camper compared to the guy doing 1/4 inch blocks

I would assume gauge blocks are not made by hand and the only unhappy people are when we go to pay for them.

[coppercone2]

I don't really know what standards were made that way, lets just say my fingers are not happy with this, that just comes down to bone anatomy or whatever. I felt cramps after experiments to try to flatten things out like that which are tall

When I 'lapped' a gauge block on a granite plate (it would not wring?, but after a few figure 8's it wrung), the 4 inch one was MOST unpleasant. I think this is OK because its not gonna keep happening. Just super awkward to grip and hold true while in motion. I know they have the stone for it from starrett, I just wanted to see if the kits worth anything at all, apparently so. That experiance made me think any wave guide more then say 2 inches tall would be a pain.

Ima just say it reminds me of standing strait in small boat, but with my fingers.

[joeqsmith]

You started out with concerns on tilt but now it seems flatness.   For the waveguide standards, they need both or at least the offset load would need the flanges surface to be parallel to the shorting face.  In my case the mill is good enough for that followed with a quick lap.   

The standard were all well under 2"  This section is 40mm but I have finished some of the other parts by hand as well that were taller.  Holding it by the flange is no problem for me. 

[coppercone2]

I probobly confuse the terms because I don't work with precision mechanical technologies very much and use whatever intuition I have to try to save time and energy because these projects are mega broad in scope and it never feels like you have enough time and it has not yet resulted in a mega fail.

[coppercone2]

BTW, did you use collared bolts with your wave guide?

I noticed when there is a complex wave guide structure, I have no idea if its aligned properly, I can't look inside, and all the manufacturers are different so trusting the flange to act as a guide maybe is not the best idea, some of them seem a bit off.

I mean the bolt that has a slightly wider non threaded part above the threaded part that should theoretically align the holes like pins would. It must be better.

It might be called a pin bolt.

[joeqsmith]

Shoulder bolts.  I have a used them for some setups but for much of what I have done, I used paper clips.  When I measure the antennas, I did not use shoulder bolts.   So far, the most critical measurements I made were to compare my software's calibration algorithms.     

[coppercone2]

It might be interesting to have super aligned when I am playing with the flex wave guide soon. I got a longer one too so I can see what it does while I bend it and its testing.

And even if you can trust the flange, you can't trust the paint difference between the manufacturers.

I will look for some shoulder bolts.

[joeqsmith]

0.1585 - 0.1590 measured on shoulder
0.160 - 0.161  measured on threads

So at least for the ones I have, the thread's OD is more than the shoulders.  Yours?

[coppercone2]

Oh I don't have any shoulder bolts of that size, I want to buy them. Its just not something I ever desired before

I got some gigantic ones from something in one of the parts boxes, but it was probobly from furniture or something.

That is hardware which is very difficult to stock in a useful way, because how could you know what shoulder depth you desire? I guess if you stack washers or put a pipe it could be kinda modular but its really pretty specific IMO

I asked someone IRL and he said that it will always be that way because they are rolled from steel not cut. SO it probobly means that a custom part is necessary. They don't have a way to actually make the shoulder thicker then the thread if its rolled unless you do some really crazy shit

[coppercone2]

But it is possible to make from brass DIY I think, by brazing three things together, and using a cup washer. So long the pin is the correct dimension it will align everything, then even if you have a crappy offset miniature bolt in the middle, because you drilled it on a drill by eye with a wobbly vise.

So I can get the correct dimension brass tube, cut it to lengh to make a stub, drill it, put a smaller brass bolt in it that sticks out a bit, put it on a brass disk, solder it together, then braze two different ID washers together to make a washer that has a ciruclar flange thing, and it should compress everything. I think I can line up a bunch of these stacks and braze them all on one go on a fire brick if I can't find the correct part. I would probobly make it a square bolt because its easy to cut.


Since there is 4 of them, it should hold

[joeqsmith]

Maybe use your precision cal standards to perform a calibration, then rotate the parts from one extreme to the other and measure the difference.   Try it with thin, thick and no paint. 

***
Maybe forget the cal, and just try it.

[coppercone2]

Got it, so long you ream the wave guide hole by a little bit

my dyslexia kicks in on mcmaster

The precision shoulder bolts are available with a shoulder wider then the thread.
https://www.mcmaster.com/products/shoulder-bolts/shoulder-diameter~5-32/


The major diameter is 6-32 or 5-40 = 0.1250 or 0.138, the shoulder diameter is 0.156

The hole is like 0.145. If I ream them, then the shoulder screw should work. The shoulder height is not super critical because I can put a washer so long its not too short.

I will try it when I get my flex guide before I buy the bolts and start modding things. Irregular wave guide would piss off the 3rd hand market. At least I know I can solve it though. Calmness

[coppercone2]

The other thing I can do is like a assembly aid, to buy pins that fit precisely, put in 4 pins, clamp it, remove the pins 1 by 1 and replace with bolts.

Probably should do that, its what people normally do to align metal thats not armor.  If it was on a battle ship that was going to take a beating it might be a problem but yeah

i need some time off lol, not thinking clearly here

[joeqsmith]

For instance, generic flexible waveguides tend to have poorer RF performance, i.e., insertion loss and VSWR, even though they may solve routing, misalignment and vibration challenges. Nonetheless, if a flexible waveguide is made with the appropriate sized sections for a specific frequency, the insertion loss and VSWR can nearly match a solid waveguide over a very narrow bandwidth.

https://www.microwavejournal.com/articles/print/26932-rf-waveguide-tips-and-facts-you-cant-learn-in-school

Paper on alignment.

https://ntrs.nasa.gov/api/citations/19970023425/downloads/19970023425.pdf

[coppercone2]

Do you know if there is like a flex rating for wave guide?

I ordered a 1 inch, 3 inch and 8 inch piece. The 8 inch piece came in today and its kinda stiff. It will do for the VNA front panel I think, but the 1 inch piece if its the same stiffness would just be too stiff.

And yeah my idea is very close to a bootleg version of that paper! wow

[coppercone2]

For cleaning the flex guide, I used neverdull on the end faces and then ultasonic it in the lab soap powder for 8 min and then rinsed and used clean water for another 8 minutes.

The alconox gets brass much better looking. I did not really want to submerge it but it was looking a bit dirty inside, I hope I did not tarnish the edge of the spiral wrap. I imagine a gentle flexing after bake out will reform whatever metal to metal connector was there. Usually things look electrically better after the ultrasonic cleaner...I assume its like corrosion growing on a knife edge or something that gets pushed out

Now I will bake it for a hour at 75C and try it out

Actually I have no idea if its a metal to metal seal or if its just a solid piece of metal. I figured its made like armored power cable out of ribbons. But I never saw a video.. research time I don't think its built like a vacuum capacitor?

https://www.microwaves101.com/encyclopedias/flexible-waveguide

[coppercone2]

some fucking bnc cable went out behind the 8510c and yeah going back there lol behind that machine infested alien bull, at least nothing shocked me, thought the sweeper broke

but it looks like it all works. I just used the saved cal routine that just the apc-7 to wr62 to wr62 to apc7 to cal, and plugged the wave guide thing in front of it with a short circuit on the end of the attenuator

and it goes from bad VSWR to nice VSWR at max, so it works I guess. I need to play with it more but my friday was ruined because of I think a BNC cable, this is with refleciton measurement, I just did whatever to get away from the basement with pictures to post . I left it connected so I can try to play with doing a cal with the wave guide network attached tommorow

WHen I took the attenuator apart I was not sure what angle is max attenuation and what is minimum, in line is minimum (I figured this out by solving the formula near the diagram with the 4th power cosine or whatever, glad I got the 'polarity' correct. I thought 10% chance some mis understanding would occur that I put it backwards.

[joeqsmith]

Looks like the shelf is up to handling all that weight!   

One of the experiments the students performed was to connect a short to an attenuator and then measure the VSWR with different settings.  Similar to what you appear to be showing.  The students had to run this at a fixed frequency of 10.5GHz and use a slotted line to measure it.  None of this fancy VNA stuff.   

Attached showing using LiteVNA to conduct this experiment.   I used my standards to calibrate at the kit's transition then bolted on the attenuator and short.  You can see how poorly these parts behave but good enough to demonstrate the change in return loss. 

[coppercone2]

is that switch reused or did you find a part number for it?

Its the same design that I Hate from the HP amplifier to push the attenuator button  inside the chassis, but it looks really cool on a wave guide

I guess its just a custom mod for a wafer switch? For this application making the detents nonlinear would be interesting too, so you can have a mechanical switch that follows a geometric pattern of detents

[joeqsmith]

is that switch reused or did you find a part number for it?

Its the same design that I Hate from the HP amplifier to push the attenuator button  inside the chassis, but it looks really cool on a wave guide

I guess its just a custom mod for a wafer switch? For this application making the detents nonlinear would be interesting too, so you can have a mechanical switch that follows a geometric pattern of detents

There's no switch.   Only the transition, attenuator and shorting plate.  Guessing you are confused thinking that slotted  flap-type attenuator is a switch with detents.   Maybe read the following on attenuators.   

https://electronics-club.com/waveguide-attenuator/     

Shown sweeping from 10.9 to 11GHz, adjusting the attenuator in half marks from 0-24dB.   The extender's directional coupler limits running it any higher.   

[joeqsmith]

I think if you wanted to use one of these poor quality attenuators, with say a stepper motor.  Characterize it over the entire range.  You still can't correct for it.  It would be better to start with a good attenuator.  These were are a good choice for training kits,  low cost and durable.