1

Problem :

I'd like to run the following test routine with an LTE router:

  1. cause a loss of signal, as if the router entered a tunnel
  2. remove the signal obstruction, as if the router just exited said tunnel
  3. check how the router behaves after the signal obstruction is removed (i.e. how well does it re-establish its LTE connection, how long does it take, etc.)
  4. repeat that routine 30 times

What I'm looking for is a system that can perform the signal obstruction mentioned above:

  • reliably (the LTE signal must be completely lost and regained with 100% accuracy)
  • safely (I don't want to damage any part of the router, specifically the RF module)
  • automatically (I'm gonna need to do this tests more than a hundred times) I'm especially concerned about the safety aspect.

I have a background in electronics and physics, but I'm not a "radio guy". I do know that playing with antennas while the modem is running can sometimes cause the amplifier to fry, even though I assume this problem is usually curtailed by TX power limits and automatic impedance/antenna detection. But since I don't know exactly how the modem is made and run, I'd prefer to run a design idea before you guiles.

Here's a function-level sketch of what I think would be perfect for this job: enter image description here

Questions :

  1. Is there a better way (simpler, safer)?
  2. What are the risks of this design?
  3. What should I put as the "dead end"? A impedance-rated plug?

Any answer, thought, or even blurt is appreciated; no need to be thorough or 100% sure.

jwav
  • 111
  • 3
  • 1
    Hi did you look at [this question](https://electronics.stackexchange.com/questions/635817/bad-gsm-network-automated-testing/635824#635824) ? – jonathanjo Sep 28 '22 at 16:22
  • @jonathanjo > I did consider a system that would lower "hats" on antennas. I like the idea of the box opening and closing : if this is reliable and I can find one, this might be worth a test, since it's probably the easiest rig to make. But I do have doubts on the reliability. – jwav Sep 28 '22 at 16:25
  • Sounds like a Faraday cage would do... – Solar Mike Sep 28 '22 at 16:28
  • 1
    FYI: "dead end" is called a dummy load (and for low power levels it's basically just a resistor). – user253751 Sep 28 '22 at 16:37
  • 2
    put the router in a microwave oven and close the door – jsotola Sep 28 '22 at 17:21
  • 1
    Your block diagram is correct, but you don't need to build this yourself, there are commercial devices. Search for "microwave RF switch". Pick one that uses compatible connectors, and get a dummy load with that connector too. Some switches are USB controlled so you can just use software on a PC to cycle the tests. – Mark Leavitt Sep 28 '22 at 17:51
  • @jsotola : might work. Hard to automate, though. – jwav Sep 29 '22 at 16:05
  • @MarkLeavitt : Ok, I'll search for that precise term. No danger of messing up the modem, right ? – jwav Sep 29 '22 at 16:07
  • 1
    As long as you include the 50 ohm dummy load, your router will be fine. Solid state switches are very fast so the switching transient isn't an issue. – Mark Leavitt Sep 29 '22 at 20:11
  • @MarkLeavitt : I'm having a hard time finding an RF switch that doesn't cost hundreds of € and isn't a sketchy unmarked device on AliExpress. Do you have a website or reference recommendation ? I'm getting nowhere on Radiospares, Mouser, DigiKey, Amazon, etc. – jwav Oct 04 '22 at 09:33
  • Ok, I think I've found a reasonably priced and available one : https://www.mikroe.com/rf-switch-click. Will confirm later if this is adequate. – jwav Oct 04 '22 at 12:10
  • @jwav looks tempting, but the datasheet for the RF switching IC shows only ~20 dB isolation. You'll want higher isolation. Hint: eBay. – Mark Leavitt Oct 04 '22 at 17:30
  • @MarkLeavitt > Ach, well spotted. I have a few questions : 1. Is the isolation in dB referring to a power ratio? Or amplitude? I assume the former, but I'd rather ask. 2. Does an isolation of 22 dB (99.4 %) mean "0.6% of the radiated power will leak to the antenna"? If yes, is this really a big deal for LTE? 3. For 700 MHz, the isolation is 27 dB. Would that be enough isolation? 4. Can the switch isolation problem be sorted by inserting a 20~30 dB attenuator between the router and the RF switch? I know that's a lot to ask, but even hints or guesses on some of those questions would help. – jwav Oct 05 '22 at 10:52
  • 1
    The switch isolation needed to cause a loss of LTE connection depends on how far above threshold the cell signal is. If you're in the fringe reception area, 20 dB may be plenty to break the connection; but if you're near the cell tower, you may need 60 dB or more. – Mark Leavitt Oct 05 '22 at 15:36
  • @MarkLeavitt > Ok. What about introducing a 20 dB attenuator, then ? Would that be enough to garantee a loss ? Or a bunch of cheap 10 dB SMA attenuator in series to scale it to as much as 40 dB ? – jwav Oct 05 '22 at 19:31
  • Yes, you can add attenuators in series with the antenna to reduce the signal strength closer to threshold. In fact, you should repeat your test at various RSSI levels; the modem will probably recover more slowly the lower the signal level. – Mark Leavitt Oct 05 '22 at 19:44
  • @MarkLeavitt > Ok, I'll think I'll do that. Thanks a lot for your help, Doc (yeah I read your entire website :) ) – jwav Oct 05 '22 at 19:47

0 Answers0