VNA for beginners

[okw]

Hi,
I have a custom PCB with GPS (9x11mm generic module - external active antenna, plus the option to route it to LTE/4G on-board antenna in case external is missing), 433MHz tranceiver (si4455, PCB trace antenna), Bluetooth (nRF52840, chip antenna) and LTE/4G (nRF9160, PCB mounted antenna, plus IPEX connector to test off board antennas), which I'm trying to tune to achieve better RF performance. But I'm not sure where to start. I've googled a a lot, but can't seem to find the "beginners guide".
Where do I tap in, and which readings am I looking for? And do I use a handheld probe or solder in to not interfere with my hands? And which components do I leave in, and which do I remove for measurements? And which components would move the curve in either direction?
I have a PS100 RF Vector Antenna Analyzer (resistance/reactance/SWR/S11), a Tek oscilloscope (up to 1GHz) and a R&S signal generator (up to 3GHz).
I've attached the schematics for each of the macth/filter circuits. Some might be overkill(?), I'm still quite new in RF designs.

Any help would be greatly appreciated.

[tszaboo]

Read this guide:
https://colinkarpfinger.com/blog/2010/the-dropouts-guide-to-antenna-design/

It's good, and while it might tell you more than you need, it tells you also what you need to start.
For LTE, what's that schematic? You have a bunch of matching circuits for each band?

[okw]

For LTE, what's that schematic? You have a bunch of matching circuits for each band?

Thanks. Yes, it's from the Thingy:91. Nordic told me they needed it to match various bands due to very small sized PCB. I guess in my case, it might not be needed.

[tszaboo]

For LTE, what's that schematic? You have a bunch of matching circuits for each band?

Thanks. Yes, it's from the Thingy:91. Nordic told me they needed it to match various bands due to very small sized PCB. I guess in my case, it might not be needed.

I think I've seen that previously on a Nordic evaluation board. They are right, different bands for 4G make it impossible to match antennas to 50 Ohm on all bands. TBH I never implemented this myself, and sent my boards to match my antennas to the antenna manufacturer for their recommended matching. They have experts who do this, and provide with the recommended part numbers for the BOM, and test reports. Some bands are not 50 Ohm, but the results are often good enough, plus 4G compensates for this mismatch with more power.

[okw]

I see. I'll keep that in mind.
I read the article (very good), but I still have some questions. So I put my connector before the SAW filter at L13/C136 and then after at L27/C163? And since the SAW filter is unbalanced, I need to have a matching circuit on both sides?
And my design is very tight, I can't fit an SMA connector like the one in the article. Can I use an IPEX to SMA cable instead, and just remove the IPEX connector and solder in the wire and shielding? Or could this introduce too much noise?

[mayor]

Take a look here, this is a good resource, and should work even if you don't have much room.

https://www.ti.com/lit/an/swra726/swra726.pdf

[tszaboo]

I see. I'll keep that in mind.
I read the article (very good), but I still have some questions. So I put my connector before the SAW filter at L13/C136 and then after at L27/C163? And since the SAW filter is unbalanced, I need to have a matching circuit on both sides?
And my design is very tight, I can't fit an SMA connector like the one in the article. Can I use an IPEX to SMA cable instead, and just remove the IPEX connector and solder in the wire and shielding? Or could this introduce too much noise?

SAW filters and LNAs for GPS are going to be already 50 Ohms, if they are not, they will tell you in the application circuit.
For VNA measurements up to 4 GHz, IPEX connectors and cables are fine. You can also remove the connector, but I usually try to make it with an IPEX, see later. It's not noise that you are concerned with, it's measurement repeatability, and physical layout. You want to calibrate your VNA on the PCB. Sounds scary but all you need is copper tape, scalpel and 0402 100 Ohm resistors, or 50 Ohm RF resistor. Calibration is very important, if it's not done correctly, the measurements and the matching you are trying to apply will not make sense.
And for the matching, you need typically a capacitor and an inductor sample kit.

So don't try to match circuits that are already 50 Ohm. Or transmission lines. You want to match parts that are variable, antennas in this case.
Don't forget the housing. Drill a hole on it for the cable. This is why I always try to make it with connector.
I usually ask for un-populated PCBs to make the calibration kit, and to test the antennas. Then sample it from production to verify.

[rthorntn]

I just got a LiteVNA, I want to test 915M LoRa antennas.

Do I need to calibrate with different OSL plug connectors for antennas with different connectors, I was testing the little SMA pigtail antennas directly connected but I have N-female and N-male antennas and they have 50CM SMA to N cables?

Thanks.

[tszaboo]

I just got a LiteVNA, I want to test 915M LoRa antennas.

Do I need to calibrate with different OSL plug connectors for antennas with different connectors, I was testing the little SMA pigtail antennas directly connected but I have N-female and N-male antennas and they have 50CM SMA to N cables?

Thanks.

It depends. You can use adapters, but then the adapters are not part of your calibration. What I wouldn't recommend is having any cables that are not part of your calibration. If you are interested if the antenna with the cable works or not, that's usually enough. If you want to do matching then the calibration plane really needs to be at the matching circuit/antenna, otherwise none of your calculation makes sense. You can also add electrical delay to remove adapters.

[rthorntn]

Thanks, do you know where I can read some more about the calibration plane, not too technical, by "do matching" do you mean viewing the antenna SWR of the target frequency?

I'm trying to weed out bad antennas and if possible improve them.  I will probably get to making antennas in the future.

[tszaboo]

Thanks, do you know where I can read some more about the calibration plane, not too technical, by "do matching" do you mean viewing the antenna SWR of the target frequency?

I'm trying to weed out bad antennas and if possible improve them.  I will probably get to making antennas in the future.

Antenna matching means changing the impedance of something, so that the RF is transmitted and not reflected from an antenna. Broadly. Don't use this definition in a test, you are going to fail.

This guy with the Swiss accent can explain calibration:

[A.Z.]

I just got a LiteVNA, I want to test 915M LoRa antennas.

Do I need to calibrate with different OSL plug connectors for antennas with different connectors, I was testing the little SMA pigtail antennas directly connected but I have N-female and N-male antennas and they have 50CM SMA to N cables?

Thanks.

calibrate WITH the cable, just plug the SOL at the end of the cable and run the calibration, that will "exclude" the cable moving the measurement "plan" to the end of it

[rthorntn]

Thanks, so OK, can I just get away with getting an SMA(F) to SMA(M) cable of the same type and length as my other cables (50CM) and using the supplied SMA SOL?

Or should I buy/make N(M) and N(F) SOL to put on the end of my cables, I'm using these 50CM cables to minimise losses and I'll have the LoRa radios in a waterproof case right beside the antenna?

[Kean]

Thanks, so OK, can I just get away with getting an SMA(F) to SMA(M) cable of the same type and length as my other cables (50CM) and using the supplied SMA SOL?

Or should I buy/make N(M) and N(F) SOL to put on the end of my cables, I'm using these 50CM cables to minimise losses and I'll have the LoRa radios in a waterproof case right beside the antenna?

If you are just testing the antennas then in my opinion you may as well just do the SOL at the VNA and measure the antennas and 50cm cables together.  That assumes the cables are good of course.

If you are trying to tweak the matching circuit to the antenna then you need to do things quite differently, generally by connecting a pigtail to a partially populated PCB.

[rthorntn]

Thanks.

I want to use the VNA to find which of my stock antennas have their best SWR at 919.8M which is meshtastic in Oz.  In the future when I build an antenna I want to use the VNA to help me tune the antennas SWR to 919.8M.

That's what I bought the VNA for but I'm a beginner and may have made a mistake, no doubt I'm missing something fundamental like the antenna, cable, connectors and circuitry all make a difference to the SWR at a given frequency, not sure?

Is there a way to do what I want without going DEEP down the rabbit hole?

[Kean]

You can certainly measure the antenna SWR with the LiteVNA or PS100/N1201SA analyzer.  I have both, and am also not an expert.
Every part of the RF path can make a difference, including how tight the connectors are torqued.
But at sub 1-GHz the connectors and 50cm SMA to N cables shouldn't have too big an effect and presumably will be used with the antenna anyway.

Just keep in mind that when you have the analyzer connected to the antenna, they will be continuously transmitting over whatever frequency span you are sweeping.

Not sure what you have planned in terms of tuning those antennas.  Do you mean electrically or mechanically tuning them?

You also mentioned in your original post about custom PCB and matching.  That is a somewhat different topic and process, typically involving swapping capacitors/inductors to improve the match.
In that process, the calibration plane become important and you usually need to do port extension to remove the effect of any pigtail connection.

There are some video on YT covering this topic that I'm sure someone can dig out links for.  In any case, maybe start with some VNA videos by Alan Wolke W2AEW.

[rthorntn]

Thanks, by tuning I mean mechanically.

My bad, I'm not the OP, I just jumped on this thread as the title was perfect.

[bg8npk]

If you want to test some rod antennas (monopole antennas) you may need to be careful about your test environment, it is best to mount the antenna to your equipment and use a cable to connect it out to access the VNA. because your equipment will form the radiator along with the antenna.

Here is an article that talks about this type of antenna measurement  https://deepace.net/learning-from-whip-antenna-vswr-resonance-point-testing/

[A.Z.]

except... if you give the RF a "better path" and use a choke so that your TL won't behave as part of the antenna