Mastodawn

Diagnosing unexpectedly bad tx/rx performance

https://lemmy.dbzer0.com/post/65892666

Diagnosing unexpectedly bad tx/rx performance - Divisions by zero

Sorry, it’s my dumb ass again, I’ve been spamming a lot as I learn. So, I just got a rooftop node set up at my gf’s place. It is a RAK 3312 ESP32-S3 module (labelled 3112 for some reason), on a 19007 WisBlock base board, coupled with the RAK1906 environment sensor and RAK12002 RTC. It’s connected to a RAK 5.8dbi outdoor antenna via a male-N to MHF-4 connector I special ordered from a different site. I’m powering it via some 18ga cable down the chimney connected to a RAK 5v power supply, and it has a 2000mah battery alongside it to provide backup power. Everyone uses longfast here. Some pics: ::: spoiler spoiler [https://lemmy.dbzer0.com/pictrs/image/a367088c-c1ec-474f-a3dc-6bf35c241741.webp] [https://lemmy.dbzer0.com/pictrs/image/15234e98-9cc1-4d08-8052-74322e1eaffe.webp] ::: Now, I am having significantly worse TX/RX performance than I had expected out of a nice aerial. My handheld node, which is a TLORA T3-S3 using some crappy ass 6" long china 915mhz antenna from Aliexpress, can see at least one repeater from her driveway. Not with great signal strength since it’s around a ridge, but it CAN see it, pretty reliably. This node I built, can’t see anything. After leaving it on it saw maybe one other node for a little while on bootup, then it disappeared. The reliable hill side node is invisible. Transmitting on LongFast (our area’s default) shows maybe 50-60% of what I send gets out, but basically nothing is ever received. I can sometimes see it’s enviro telemetry from my home node through 1 hop, but I can’t ever talk back to it. Whereas my handheld node can often hear the reliable hill side node and still talk to my home base. I tried reflashing it with both stable and alpha firmwares, and while the alpha firmware did increase it’s stability and fix some other issues with retransmitting and wifi, I noticed while doing that I get a lot of these red messages in the console complaining about an Error -7 when rx’ing: [https://lemmy.dbzer0.com/pictrs/image/529d01d2-976f-4b03-bcde-aab955578868.webp] Before you jump down my throat, yes I know about high gain antennas having a narrow vertical beam width. 35deg on the RAK (17.5 up/down) should be way more than enough to overcome the ~40m height change between this and any other major node at the distances I am dealing with. Anyway, I’m at a loss with this stupid hardware, and I don’t have any other antennas to really test it with because of the worst connector ever known to man (mhf-4). I am open to any ideas for how to diagnose this node’s performance. Is it a bad radio, bad antenna, bad connector I bought? Do I just RMA the radio and see if a new one does any better?

Show thread

Dave.

A few things to work through:

Try running just off the battery for a little bit with the power supply off and unhooked, to see if it’s supply related.

Check the inside both sides of the N connector, I’ve seen bad punching/forming of the connector body leave a little bit of metal swarf in them which can short them.

Check the continuity of, and for a short circuit between the inner and screen conductors of that tiny Lora antenna cable. You might need a needle to get the centre pin on the mhf-4 end.

Check for an open circuit between centre pin and outside of your antenna, there should be no continuity with that antenna design.

Take the antenna from your other device and try it ( u.fl should fit ), just in case you’ve got a mislabeled antenna that’s 868 instead of 915MHz.

Show thread

empireOfLove2 3d ago

Take the antenna from your other device and try it ( u.fl should fit ), just in case you’ve got a mislabeled antenna that’s 868 instead of 915MHz.

My other devices and antennas use SMA, so I can’t swap antennas. However, you just made me realize I do have an N to SMA adapter from my rtl-sdr, so I may try that with my handheld to rule out the radio part.

Thanks, I’ll work thru your list here over the next day or so. Can’t wait to confuse the shit out of the neighbors being up on the roof for hours at a time lmao

Show thread

empireOfLove2 2d ago

So I’ve had a chance to go up on the roof.
I ran it on battery only with the 5v totally disconnected at the wall, no change over 20 minutes. The antenna cable from radio to antenna checks out (1 ohm both inner and outer conductor, perfectly open circuit from inner to outer).
I hooked up my handheld to the antenna and got similarly bad tx performance.
All N connectors look clean, not stretched, and aligned. They are a bit more recessed than you say but I can feel a good connection slip fit when threading them together for about 3-4 threads before tight.

Finally, I ohmed out the antenna plug. There is 4-6 ohms between the center pin and outer casing of the antenna, not an open circuit. This seems to indicate it is the antenna thats bad, ye?

Edit: I took the antenna down and testing it again on the ground, it’s more like 0.8 ohms center to outside. That is a dead short and likely my problems. Thanks!

Show thread

Dave.1d ago

I took the antenna down and testing it again on the ground, it’s more like 0.8 ohms center to outside. That is a dead short and likely my problems. Thanks!

Some antenna designs can be a dead short at DC and present a correct RF load. But the one you have is very likely a collinear design with “stacked” elements, which should be open circuit.

Everything else you described sounds normal. So hopefully this is it!

Show thread

empireOfLove2 1d ago

I hope this is it too, my area kinda needs this node haha. I contacted Rokland support to see I’d they had any details on how to test or confirm the antenna was bad, and I’ll start an RMA once they come back with some details.