Experimenting with waveguides using the LiteVNA

[xrunner]

RG-402 + SMA connectors ordered. 

[joeqsmith]

Don't waste it on that DC stuff... 

Ended up just drilling out the end that came with the case.  Seems sturdy enough.

[xrunner]

Don't waste it on that DC stuff... 

Oh no, I'm going to use it for the FWD and REV RF out of the coupler. I think the short coax cables that I'm using are OK, but I want to try my hand at something new. 

[joeqsmith]

I was thinking < 500MHz was DC and you need a new project.  Maybe some 3D printed waveguides. 

[xrunner]

I was thinking < 500MHz was DC and you need a new project.  Maybe some 3D printed waveguides.

Oh shucks! I had a project that went up to 450 MHz and I'll be damned if you went ahead and defined everything under 500 MHz as DC! I see how you play now. Well, now I'm thinking about laser communications next so we'll see where DC is defined then ... 

[joeqsmith]

Sure, but I don't see how that new coax fits into the picture.   

[joeqsmith]

Thought about adding something like a 9-pin D shell for power and future expansion (maybe allow my software to control the switch).  But ended up with a simple phono jack for power.   Stuck it on the front so all of the connections are on one side.   Needs hardware to mount the SMA to the plate along with new labels. 

[xrunner]

Phono jack for power? Sure as long as you own it it'll be fine. Otherwise better label it or another owner will want to listen to the thing working. 

Photography nice - good presentation setting on the table. Hey what's in the bowl?

[joeqsmith]

I was thinking if I did decide to control the relay, I would swap out the phono jack for a mini DIN.   Mechanically it will fit in that area.   9-pin D would be a lot of wasted space.   I've used phono jacks for power before,  like this amplifier that I used to demonstrate one potential problem when using these low cost VNAs with their squarewave drive.  Note the battery holder is stored with it and the connector is marked.  Even my design scribbles are in the bag.   Just in case after my death my wife holds a big sale and a ham buys it for 5 cents.

The table was made by my grandfather and the bowl and egg shapes inside are Alabaster from Egypt. 

[joeqsmith]

Making a basic checkout, seems no worse for the wear.   So for fun, I connected a decent GORE cable to the converter and attached it to a 23 inch section of that semi-ridged RG402 that I used to wire it with.   At 11.8GHz, I measure about -4.4dB.  No attempt was made to calibrate the system as I was really just interested in a relative measurement.

Next I replaced the semi-rigid cable with one of my unused cables supplied with my V2Plus4.  If you follow the channel, there was segment where my test setup slowly falls off the desk dragging the VNA to the floor by the cables.  These cables are not the ones used.  I had two sets.   These are about 12" long.   Notice I loose about a dB.   

Last, I placed a Weinschel WA3 3dB attenuator in series with the V2Plus4 cable.  This attenuator has a spec'ed frequency range of DC - 12.4GHz.   Like magic, it measures about 3dB lower. 

So the semi-rigid cable you purchased can certainly improve the losses if you decide to go that route but it's a one-shot deal.  Make sure you understand that it is easy to damage, like our friend here learned, and I gave it my best shot at repairing it with no success.   It's also a pain to put together.   

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

I ran a few checks of S11 as well and I think we are ready now for some labels....

[joeqsmith]

Here's a stupid test for the fun of it.  It's my spin on the Jim W6LG videos showing how much loss connectors have in the HF bands.   We all know not to use transmission lines beyond their cutoff but I did say stupid....   
https://www.microwaves101.com/encyclopedias/coax-cutoff-frequency

Tuned the VNA to the lower end of the extender and inserted a SMA thru.   No attempt was made to calibrate the system.  Like before, I just looking at relative numbers.   Measure about 0.5dB. 

Next I replaced the thru with 12 BNC adapters, including two SMA to BNC adapters.   Measured about 2.5dB lower. 

Last stuck in a Midwest Microwave BNC attenuator PN# ATT-0313-06-BNC-10 into the mess.   Measured about 9.3dB.   
https://www.mouser.com/datasheet/2/643/pi-CCS-MW-ATT-0313-06-BNC-10-1312565.pdf

Nothing real useful from our Baby N Connectors.  I did qualify with stupid.   

[joeqsmith]

Been waiting to run this test because the SMA wasn't yet secured to the front panel. 

Connected the extender directly to my microwave counter and set the LiteVNA to 500MHz CW.   The PDRO is at 7.86G (just below the X-band).  We should measure 8.3600000GHz but the counter measures 8.360019671 and it's hunting....... Imagine that...

Connecting the extender to the 10MHz GPS reference,   get the counter measures 8.359999964 and is stable within a count.  The rest of the error is the LiteVNA's internal clock.  Too bad the VNA doesn't have a reference input as some of the experiments I was attempting, the VNA's drift comes into play. 

Good to see it all working again.  Time to clean up the mess..

[joeqsmith]

When I first swapped out the directional coupler for something with better directivity, I plotted how it improved the noise floor.   
https://www.eevblog.com/forum/rf-microwave/experimenting-with-waveguides-using-the-litevna/msg4903016/#msg4903016

You can see there are some areas where you can get much better dynamic range than others.  Attached for example, running some attenuators from 0 to 70dB to get an idea of the linearity.   Finger tight and normalized thru.  Mix of my very low cost Chinese attenuators.    Typically its going to be around 50dB and worse case looks like maybe 40dB assuming 10dB headroom.    Not great by any means but far better than trying to use the harmonics at 12GHz and good enough for my home experimenting. 

[joeqsmith]

I did try repeating an earlier test stacking some 10dB MWM attenuators in series.  These are still just jelly bean parts but a bit tighter tolerance.  Parts were torqued.  For the 3 and 6 dB, I used two other name brand attenuators I had.   This is really about what performance I expect out of it.   

[joeqsmith]

Running a SOSL cal with my homemade standards and measuring the return loss of the terminator supplied with the university microwave trainer.   Swept it from 7.9-12.5G just to get a better idea how it performs.  Not great but cheap. 

Note the three sets of data are from my original PCB construction prior to the new case, and with both the old and new directional coupler.   

[joeqsmith]

Using Dislords firmware, the LiteVNA supports I think unlimited harmonics, or at least way beyond what is useful.   If I wanted to measure S11 directly rather than using the frequency extender,  at least for the X-band and my particular versions of hardware, we can't.   Using the supplied cables, short, open and sorted mini-circuits load to calibrate the VNA from 1M to 9GHz.   Plot showing the return loss with the standard attached.   Even at 9GHz, the return loss is basically 0.

[joeqsmith]

With the Narda waveguide load attached, about the only thing we can see is where the low frequency cutoff is.   Still, below 6GHZ where it was designed for, it has decent performance and I am not aware of anything else in this price performance range.   

[joeqsmith]

Tried to locate a good condition, used N to SMA adapter for port 2.   They are out there but the high costs made them not worth the risk.   At 12GHz, no point in trying the low cost Chinese parts.  Ended up buying a new Hirose  adapter and called it good enough.

https://www.digikey.com/en/products/detail/hirose-electric-co-ltd/HRM-554S-40/5148329

[joeqsmith]

The work of Jagadis Chandra Bose. 

https://www.cv.nrao.edu/~demerson/bose/bose.html

[joeqsmith]

One problem with playing around with used waveguides is finding data sheets.   Most of the cold war military contract houses are long gone, along with all the information for their products.

I've been looking for a isolator with decent isolation specs.  People who are selling such devices are not typically going to have the means to measure them.  So you are basically going on a blind date.   In this case, looks like I may of gotten lucky.   Using the frequency converter with the LiteVNA, I measured an insertion loss of <2dB and isolation > 26dB.  The low quality transitions were clamped on and I doubt very much the insertion loss is this poor.