modifying nanoVNA to a benchtop instrument

[nonius_]

Has anyone made modifications to their nanoVNA to make it act and appear more like a stand-alone VNA (as for example HP-instruments)?

I'm thinking about installing my nanoVNA in a much larger case with BNC-connectors instead of SMA, larger buttons/switches and built-in powersupply for 230Vac, and a large USB-connector. The entire thing controlled by a dedicated laptop (or possibly a tablet). The idea is to turn it into more of a bench piece of equipment, than a portable one, as it is now.

I know this won't improve the spec's of the machine to the HP-level , but I find a major disadvantage of the nanoVNA its small screen, tiny text and finicky buttons and the delicate SMA-connectors. 25 years ago not a problem for me but eyes aren't getting any better with age. Hence why I intend to in the future use the nanoVNA only as a 'black-box', with all control and display done with a laptop.

Most of my work is done on the workbench, not in the field, so portability and battery-power aren't important to me. I intend to have an old laptop running Linux and nanoSaver, fully dedicated to this nanoVNA. I.e., as much a 'stand-alone' system as possible, in the hope it'll still work in 10 years time. (I haven't tried out to connect it to a tablet but that might be another option and a built-in tablet would make it even more stand-alonish with fewer interconnects, as in the case of an external laptop).

In the picture below my initial idea: mount it in a much larger case, behind an aluminium frontpanel (with screen visible) and all controls/inputs made available on the frontplanel.

I've been searching on the internet for ideas from others who have made similar modifications but haven't found anything so far. Plenty of case-mods though, but I want to go a little further. If anyone know of such projects, or has made it themselves, I'd be interested in learning more about it.

[geggi1]

The simplest way would most likely be to do a Black-box type instrument with only the two RF connectors and a USB in the back.
You should use high quality connectors if you work in the Ghz range and use N or TNC connectors because these are better rated over 1Ghz.
If you mostly  are working below 500Mhz BNC with DIY calibration standards are OK.

[Smokey]

You know what would be funny.  Get an old dead actual benchtop network analyzer and put a nanoVNA inside it's case.  Plumb the RF to it's front panel connectors.  LCD in the CRT slot. 

https://www.ebay.com/itm/298369904365

$265 with make an offer for a dead(ish) HP 8753A

[ChristofferB]

I think this is a wonderful idea. You might even be able to up the shielding on the RF parts.

In general, stiff heavy cables and very tiny light instruments/modules is a pain. I assume you want it to be heavy enough to be "shelf stable".

I recently put a RTL -SDR V4 stick in a die-cast box, with a TNC connector for very similar reasons. Mostly I wanted it heavy and stable enough to support antennas and cables, and avoid straining the PC USB port.

I would suggest a large die-cast box. Maybe you can make some feet so it will stand upright, kinda like a digital bench oscilloscope?

Agree with geggi1, get some bulkhead N connectors of a good brand, with short, flexible coax to the nanoVNA board connectors (which you might also want to replace with name brand equivalents, but it probably won't matter a huge deal under 1 GHz).

[tszaboo]

My first though is not to use BNC because it's an outdated unreliable standard. Noise is picked up by it, shielding on the cable is often wrong, and the tolerance on cables is atrocious. Frequency response is often times only a few hundred MHz, no matter what the datasheet will claim.
SMA or N is the way to go for these instruments, 3.5mm above.
And yes, I've thought about this. Also including two 1S4T non reflective RF switch to turn it into a 4 port VNA.

[Randy222]

Is dismantling the VNA allowed?
I can see where you replace the touchscreen with something like 10-11" tablet size touchscreen, and then plumb some connections to the bottom side of the screen, and re-locate the jog wheel? Mount the guts behind the screen with shielding.

That would be a neat retrofit kit, that I would probably buy.

[nonius_]

Is dismantling the VNA allowed?

It is, as far as I'm concerned but I'm afraid my hacking skills aren't good enough to replace the small 2.8" screen with a larger one.

However, another idea that recently came to mind was to use a Raspberry-Pi with a small computer screen - that way it'd be a stand-alone system, without requiring any external hardware. Will have to search around on the internet to find if/how others implemented this. Have so far never tinkered with Raspberry-Pi's myself though.

[dobsonr741]

It's a nice project.

One thing that's changed dramatically in the last couple of years is that you can lean heavily on AI for much of the engineering grunt work. Have it help bring up the Raspberry Pi, write the code, generate test plans, review schematics, troubleshoot issues, and even produce documentation. You don't need to have done this exact kind of project before.

These tools are now good enough that the most important skill is often knowing what you want to build and being able to clearly describe it. If you can articulate the requirements, break the problem into pieces, and validate the results, you'll be surprised how quickly you can get from idea to working hardware.

[Randy222]

I have a few Pi 5's for testing stuff. Using their flasher util and their Ubuntu image, super easy turning the device into a palm size computer. From there it's hats or network based apps. But Pi is flexible.

[radiolistener]

I'm thinking about installing my nanoVNA in a much larger case with BNC-connectors instead of SMA

I would strongly advise against using BNC connectors on a VNA.

The main issue is that BNC connectors do not provide a particularly stable or repeatable RF contact. The bayonet locking mechanism can introduce small variations in the contact geometry, and the center contact is not as mechanically consistent as precision RF connectors. For an oscilloscope this is usually not a major problem, but for a VNA even small impedance discontinuities and contact variations can noticeably affect calibration repeatability and measurement accuracy. So BNC for VNA is not suitable at all.

BNC was never intended as a precision microwave connector. If you are going to the effort of building a larger enclosure, SMA would be a much better choice. For maximum robustness and long-term reliability, N-type connectors are even better.

[nonius_]

[Tszaboo] My first though is not to use BNC because it's an outdated unreliable standard

[radiolistener] I would strongly advise against using BNC connectors on a VNA.

Thanks for the warnings w.r.t. BNC. My problem is that I've standardized almost entirely on BNC, with only very few exceptions that use SMA (the 13GHz DIY frequency counter uses SMA). So just about every time that I use the nanoVNA I install an SMA->BNC plug first. I though I could skip that extra step by simply using BNC on the (future) frontpanel. I figured that since BNC was ok-ish up till 3GHz I could get away with BNC, but apparently not.

[hugen]

You can try our 7-inch liteVNA67.

[rteodor]

You can try our 7-inch liteVNA67.

Is this LiteVNA67 the device that @DiSlord was discussing with joeqsmith over here ?

[radiolistener]

I figured that since BNC was ok-ish up till 3GHz I could get away with BNC, but apparently not.

For VNA measurements, BNC is not "ok-ish", even below 30 MHz. It may be acceptable for rough VSWR estimation below about 500 MHz, but at GHz frequencies the poor repeatability of the connector and random contact variations introduce errors large enough to make it unsuitable even for rough VNA measurements.

Saying that BNC is suitable for a VNA because it works well on an oscilloscope is like saying thin hookup wire is suitable for measuring hundreds of amps because it works perfectly well for measuring voltage.

Again, BNC is perfectly fine for an oscilloscope, but not for a VNA. These are very different applications with very different requirements for
connector.

[Berni]

The reason that BNC connector might pass a 1GHz signal seemingly just fine but provide poor results on a VNA is phase and reflections.

If the VNA is being used as a glorified VSWR meter to check if an antenna works at a single frequency, then you wont really see the cable. The antenna "eats" most of the power and you get some random reflections from signals reflecting back into the antenna anyway, so the plot moves around even as you just walk around the room.

But when you are using a VNA as an actual network analyzer to measure a complex impedance then phase matters a lot. The coax cable must be calibrated out to prevent its delay from messing up the results. That's why we have VNA cal kits for (not because the VNA is so drifty it constantly needs calibration or something), they are a known impedance to let the VNA calculate what the cable does and subtract it away. If the cable changes in any way after calibration then it won't perfectly subtract away anymore and will affect the results.

The higher up in frequency you go the more sensitive to the delay things become as they cause more phase shift, including how far in a connector is. Additionally if you have areas along the cable (such as a connector) that are not precisely 50 Ohm that will cause the signal to reflect back. The VNA works with constant wave RF (not like a TDR) so it doesn't know where these reflections are hence they also add into the final measurement with weird all over the place phase and amplitude. Hence the connector has to to be very well impedance matched and consistent in it. Since VNAs can see easily down below -60dB means only a fraction of a % of reflected power at a connector will put a bunch clearly visible spikes into your measurements.

"But the device i am measuring is BNC!"
No worries use a BNC adapter. but put the adapter at the end of the cable. So you calibrate the end of the cable and then measure your DUT+Adapter together. This way you use better cables like SMA only where it matters (for extending the VNAs measurement plane up to the DUT).

[hugen]

You can try our 7-inch liteVNA67.

Is this LiteVNA67 the device that @DiSlord was discussing with joeqsmith over here ?

Yes

[nonlinearschool]

I love this idea!

 I would also think having a much larger screen, despite the limited resolution.  Maybe even a touch screen, If you can find a way to connect the controls to the touchscreen. My entire career has been finding I have to do some modification to whatever device I acquire.  IN a way, I cannot prevent myself from finding a way to make some practical more useful improvement!

Keep us posted as you progress through the adaptations!

[jord4231]

I've have replaced a nano vna touch screen. The one I ordered was a little smaller and it worked just fine. So I'm sure a larger touch screen would be achievable. I have thought about doing this my self.

[Doctorandus_P]

You could saw some 20 mm high strips of wood or similar and put them directly on the plank to make the slot under the stack of equipment a little bit higher. Then you can slide the VNA in between, so it takes up very little shelf space when it's not in use. You can also make a very thin drawer from just a flat plate of material, and then add a hinge somewhere halfway and mount the VNA (semi) permanently on the hinging section. Then you can pull out the drawer a bit, and fold it downward to use the VNA.