Second build sold out. Message will appear here when store is ready for third build ordering.

Major interference. Not local. Any ideas?

HI. New here. Came home today to major interference on my KiwiSDR. Between 18MHz and 24MHz there is a huge batch of interference. I think earlier in the day it looked like it could have been a repeating "signal", every 5 seconds. But it looks random now. I only noticed it today so I couldn't say when it started, other than it was within the last few days. No new bits of kit here. I've tried turning the power off completely to the flat to eliminate something local to me. Also large section at 23KHz to 52KHz. Photos attached.

Kiwi is running 1.271 with a Wellbrook magnetic loop.

Live output here: http://m6gqu.ddns.net:8073
If offline or interference gone, there's a video here:

It's a huge chunk of spectrum that definitely is usually clear on the waterfall. Wondering if anyone else has seen this level of interference before, or have any ideas to what it can be.

Attachments:
https://forum.kiwisdr.com/uploads/Uploader/e8/3ea14e269d47ff9e6ed51d7f747631.jpg
https://forum.kiwisdr.com/uploads/Uploader/af/a5456ee363e188df398d68e026944f.jpg

Comments

  • I have indeed seen interference that has similarities to what your detecting. Listening on your Kiwi near 20 MHz, it appears that there are two different source families which are not synchronous. In addition, their spectra (generally it is really helpful to include the spectrum plot along with a waterfall which shows time but not relative amplitude very well) are different. I see a family of ~65 kHz broader lines that have some of the characteristics of a SMPS. There is another family of around 135 kHz that has narrower, more coherent spectrum which cycles on and off with a period of ~8 seconds and ~70% duty cycle. It could also be a switching device of some kind. Each source might also be something quite different from an SMPS too, I'm just guessing wildly. I see evidence of the 65 kHz family other places as well, such as near 10 MHz.
    Because as a whole it is on the order of half an octave wide, I start to suspect a broad resonance, such as the mismatch loss a dipole shows when operated away from design center. BUt since you are using a broadband unresonant antenna, that might suggest that the coupling mechanism is the source of this filtering.

    I'd not be surprised if the source, whatever it is, were generating pretty fast edges down at fundamental and thereby generating the large family of terms. A BJT SMPS might be doing this. I suggest BJT over FET only because often FET SMPS are run at a higher rate to make filtering easier/cheaper.

    The loop probably does not have as much response at LF/MF so I can't be sure whether a fundamental is visible or not. It could look almost like a CW line there and generate high harmonics with easily noticeable width as we see.
    Can you try some simple experiments such as putting a termination in place of the loop to see if any interference remains (possibly much weaker) and/or rotating the loop to get any more insight into the effective direction of its source?
  • jksjks
    edited March 2019
    There seems to be a couple of issues here. Take a look at the 0 - 250 kHz band segment below. There is a terrible -60 dBm (S9+10) switcher starting at about 37 or 38 kHz. And it is very wide at the fundamental which will make the harmonics even wider. There is also this funny frequency-switching (square wave looking) signal at 67 kHz.

    image

    Now look at this 17 - 19 MHz segment. Harmonics of the square wave looking thing are clearly there as well as what sound like harmonics of the 38 kHz switcher (particularly above 18 MHz). Now you might ask why are there no harmonics of these things in the 1 - 15 MHz range (or why they are so much more heavily attenuated). But I have seen this behavior before where there seems to be selective attenuation of the harmonics. It is extremely interesting that two sources of LF RFI seem to have the same attenuation pattern.

    image

    Does it happen at all hours of the day? I'm wondering if a street light (electronic or otherwise) has gone bad. Does the signal strength change (or even null) if you rotate the loop?

    Attachments:
    https://forum.kiwisdr.com/uploads/Uploader/3f/0db152c1d0f05305fa993cce29dae7.png
    https://forum.kiwisdr.com/uploads/Uploader/d3/faac247dc5e35db4367181d7a1f336.png
    G0LUJ
  • It looks a bit like the RFI I get from my Solar PV installation which has an LF component and also spikes around 20MHz.

    Does it vary during the day / night ?

    Regards,

    Martin - G8JNJ
  • assuming the wellbrook loop gets it power from the RF coax cable (injected power for the preamp), turn off the power to your wellbrook loop antenna and then go on your Kiwi, zoom out to max and see if the noise patterns are still there.
    i had a similar noise issue, when i disconnected the preamp power to my loop antenna, the noise pattern was still there at a high level, i traced it to a poor RF coax connection.
  • To answer some of the comments so far:

    It seems consistent day or night as far as I have witnessed myself. The YouTube video was just after midnight. I'm looking at it now at 14.45hrs and it's the same. At present I cannot rotate the antenna as it is in a small space in the attic, but I am going to try other antenna's and another SDR to see if I get the same signals.

    I've tried unplugging the power to the mag loop and the interference at 15MHz-28MHz is still there. The square wave at 67KHz is also still there with unpowered mag loop. Image attached.

    Next mission id a different antenna to the Kiwi, just to see if the signal is still there. I'll then try the mag loop on an AirSpy etc, do some more digging.

    Attachments:
    https://forum.kiwisdr.com/uploads/Uploader/11/4d7ef0e822a087fabed133c52d493b.jpg
  • Are you sure the BNC is properly terminated at the far end?
    Is the shield fully connected all the way, has the loop slipped against something like a TV antenna?
    Looks like more interference than even I would expect here for a "dead" bit of coaxial cable. there is something of a long wire or "full loop back to other kit" in that to me.
    Stu
  • Well it's not the KiwiSDR or the Mag Loop that's the issue. I've tried an AirSpy and a HackRF using the basic ANT500 telescopic antenna, and the switching signal is everywhere below 73MHz. I don't see it anywhere higher than that, so that's something.

    I'm going to do a bit of wardriving later, and see if I can establish a location or direction.
  • edited March 2019
    im assuming the preamp for the loop is located at the loop and that its powered by DC/AC injection into the coax.
    if the coax connection between the Kiwi and the Loop antenna was 100% secure and isolated, turning off the preamp to the Loop would produce very quiet results on the Kiwi except for very local stations that might come through but RFI due to poor connections/common ground will remain at about the same signal level.
    the fact that the interference persists after turning off the preamp to the loop antenna suggests to me that the coax is "leaky" somewhere and/or that the coax is grounded to "common" ground - loops and the coax connected to them work much better in an isolated environment.
    i had this exact same problem as you, my loop antenna has its preamp at the loop itself, its DC injected through the coax.
    mine turned out to be a combination of a poorly connected shield at the loop and that the coax was connected to "common" ground.
    resolving both fixed the issues.
    results may vary though when it comes to grounding the coax to your loop and what its grounded to.
    you may get favorable results and you may not, i say isolation is best.
  • edited March 2019
    Thanks @elitedata That's good info to have for future problem solving. However I have done some investigating, it's not the KiwiSDR or the mag loop antenna that's creating an issue. The KiwiSDR is actually doing an excellent job of doing what it's supposed to do.

    I did turn off the pre-amp and as you say only a few local signals would get through. I had to pump the waterfall display up significantly to see the same interference without the preamp.

    I've used two separate SDR to eliminate the KiwiSDR as being the problem. Both an AirSpy HF and a HackRF SDR running with a telescopic antenna and not the mag loop. The switching signal interference is still visible using a completely different system to monitor. So I've eliminated the mag loop antenna and the WikiSDR as being the issue.

    I'm trying to find if there's software anywhere that would allow me to monitor a section of the spectrum and simply drive around gathering the data strength, and then have that data plotted onto a heat map. Any ideas?
  • edited March 2019
    Listening to it does give the impression of nasty switchmode gone bad (assuming it started good).
    The duty time is not 100% consistent so unless it is, like previously mentioned, some kind of charging circuit I doubt it is any kind of data.
    The AM noise is actually slight less annoying than my neighbours bad PC PSU here.

    You could have said you tried it on various bits of kit and antenna elitedata went the same route I did.
    I'd look for the highest frequency splatter with the same ~8.5 second beat (or AM noise) then DF that, much easier than low frequency.
    The one next door depending on what case he has it in is good for 170MHz +

    Oh and just in case you are thinking of getting OFCOM involved remember you have to show that it is stopping you doing something, your neighbour could wipe out everything to 10GHz but unless it is possible to prove that it is stopping you doing something you could before don't expect any help.
  • It is stopping me from continuing the hobby of monitoring my own radio signals, and allowing me to look at varying digital signals to decode and investigate and understand, etc. But not sure if that'd be enough for OFCOM help. Not going down that route yet though. It's only day 2 of the interference, and I'm hoping it could just be that a neighbour has plugged in some new bit of kit and that's what's causing it.

    Re trying other kit, I did mention at 2.57pm today I was going to try an AirSpy, and then by 4pm I had done a test, so at 4.14pm I posted that I had tried both an AirSpy and HackRF and that I was continuing to investigate :)

    I'm appreciative of any and all comments. It's always good to have others input and ideas just in case it's something glaringly obvious and I haven't spotted it. So please do accept my thanks for all so far.
  • OK I only mention the OFCOM requirement because I'd assumed legally receiving/decoding ham transmissions would be classed as doing something that might be worth protecting, its not, you have to, for example, have a sked with other hams regularly that is no longer possible with the noise. If you record what ham communication it stops now you have more chance of getting a case opened if it turns out to be some kit the owner won't fix.
    One thing I did was have a scanner in my coat pocket with headphones and body shield it while walking round the houses, doesn't provoke the curtain twitchers quite as much, though I suspect you are in a more enlightened neighbourhood.
  • I've gone for a walk like @Powernumpty suggested, with headphones and a radio tuned into 20MHz. The interference was obvious. Turns out the problem is very local to me. Within 4 houses. So it's something in a neighbours house. I'll be keeping an eye on it, and hoping it goes away. If not I'll contact OFCOM for advice.
  • https://www.ukqrm.org.uk/ may be a better first visit, that and gathering any pattern changes.
    Also on here, other threads I have seen a lot of useful info.

    Took me a while to trace my nasty as the noise induced into my own house wiring blurred the source, also it was mainly on overnight it's a Gaming PC, used to earn a wage streaming (allegedly) US hours and weekends.
    Without repeating it all for the nth time, serious (1200W PSU) noise source hooked up to lots of extended wiring, case with glass sides. I'd sometimes get up in the night, see if his video lights were on and check the SDR noise complete match.
    If you can spot any corelation between visible room lights, cars coming and going and related change in signature (load) you might track it down quickly.

    I had one neighbour with a terrible big Plasma TV, he was a bit of a sports nut and noise was always there duringfootball matches, it was so bad I left him a note saying I might be able to help with some ferrite supression. To help him justify changing it to his missus I added (honestly) that "the plasma could be on it's way out as the noise was getting worse", they never really spoke to me but did change TV not long after. Remember to many average people radio hams are weirdos and they will take some convincing that they have a faulty item, find a frequency that is wiped out by their source then lend them a radio to go round the house turning stuff off. My neighbour did do that for me but even with the "went off with the PC" proof I couldn't believe his PC was that bad a noise source.

    Just checked mine, the source is on, he must be making another video for his youtube channel..
    When it is on you can detect it anywhere on hf and some way above, horrible rasping noise.
  • edited March 2019
    I'd like to present a differing opinion from elitedata who wrote "if the coax connection between the Kiwi and the Loop antenna was 100% secure and isolated, turning off the preamp to the Loop would produce very quiet results on the Kiwi except for very local stations that might come through but RFI due to poor connections/common ground will remain at about the same signal level."

    I don't believe this is correct. If the interference is due to common mode current on the feedline then power-off on the preamp or even a termination out at the preamp may not remove the noise. I've measured the susceptibility to ground currents through the KiwiSDR - in/out on the shield of the SMA and out/in on the LAN, PS end. An isolated, balanced (physically small and low self-capacitance) 0 dBm 50 ohm source injecting current through this path results in ~ -70 dBm response from the Kiwi. The differential (TEM) path into the KiwiSDR is by no means the only way interference can enter and in my experience is often NOT the dominant mechanism for elevated noise floor and QRN.

    It's important to recognize that any line that connects to the kiwi has some impedance to the "rest of the world" through the permittivity of space. If one of these lines couples to a noise source then current in that line can flow through the KiwiSDR and out the antenna (or vice versa). Simple inductive chokes are generally inadequate to drive typical current levels below the detection threshold of the receiver. We have a very sensitive measuring device and there are identifiable paths within the KiwiSDR itself that produce IR drop that ends up appearing as an input to the preamp/ADC. A demonstration of this, the alternate, antenna input that is in parallel with the SMA has very different susceptibility in this regard due to different ground paths through the plating and patterns of the KiwiSDR PCB.
  • edited March 2019
    I've captured the signal as it first started today, just after 6pm. I haven't noticed anything significant at the time the interference appeared. The street lights are still off, the emergency lights in the hallway are already on. There's no spotlights or motion triggered lights that are on. Still can't fathom it out.

    The timing of a 6pm start feels like it could be significant, but it isn't day or night related. When I was looking last week, it was on during the afternoon, and then later at 3am it was off again.

    Now that I have a significant comparison of before-and-after, I may approach OFCOM about it for advice.

    YouTube clip here, interference starts around 3m05:

  • The "lumpiness" of your noise floor even prior to the interference suggests to me that you already have an ingress mechanism and are susceptible. Propagated noise doesn't tend to look like that while CM mechanisms often do. Someone familiar with the up/down VDSL spectra can comment on whether the lines at 8-10 MHz vs. those at 18-24 may be typical.
    If you have susceptibility then you may be seeing near-field to something sourced from a neighbor's house. I would work to reduce susceptibility and, once that is done, if need be survey your yard (DF or "sniff") with a probe, even a portable SWL receiver if you can't make your Kiwi portable.
  • edited March 2019
    I'm going to say this again to hopefully save you time.
    OFCOM don't give a flying flip unless you can show that the issue is stopping you doing some ham legal communication you could before the interference.
    Disruption of receive doesn't count, which was a bit confusing to me as decent RX is of course part of being a ham. I suppose they have their commercial interests to consider ;-).

    The engineer I spoke to obviously has PHB who won't allocate time unless that first criteria is met.
    If you can find an off-duty engineer and furnish him with beer/doughnuts etc that's different.

    To me the video shows mains networking VDSL or something similar start at lower data rate then settle and then shows chunks of data moving, think video buffering.
    Sky Q (spits) comes to mind.
  • Continuing to update this thread, so there is a proper history.

    I did report to OFCOM as it was a quick and easy thing to do online, although I did take your post on board @Powernumpty. They were willing to help if I could specify up to three spot frequencies where the interference is causing me problems. They can't investigate interference that hits across entire bands, as there's so much equipment these days that can be a cause that would do that. Off the top of his head, he couldn't give any suggestions on what it could be. "It could be anything".

    My response is that I will look into which bands for my licence are affected, and see if it inhibits my ability to transmit and receive on several of those bands. If it does, I should report back to them with specific spot frequencies and they will then be able to look into it.

    In the coming weeks i will try and set up a few tests to see how the interference affects TX as well as RX and get that information to them.

    I had noticed over this weekend that whatever it is turns on a few seconds after 6pm on both Saturday and Sunday. The signal is currently missing, and I'm assuming/hoping it will turn on again tonight (Monday) at the same time. At least that could prove it is something automated and timed, and not a random event such as a person turning on a PC.

    Thanks for everyones replies so far.
  • Will be interesting to see if any schedule changes with the clocks going forward.
    The change of signal seemed like streaming or something being "topped up" as the breaks got smaller almost like duty phase but I can't think what would do that only once (?) a day.
    Pond top up, dishwasher (after TV program X finishes), heating, washing machine... strange one.
    Also can't quite align the early lower frequency spread that seemed to melt into the higher peaks (dishwasher fill and cycle?).

    If you can find a spot frequency outside of normal used frequencies you could then use one of the scanner audio recorder programs* (with logging) to record and document the times.
    Having the audio would allow you to check it is the known issue and times logged may provide a clue.

    *Scanrec comes to mind, I haven't downloaded, or used for many years, so check the source.
  • Hi,

    I took a quick listen the cyclic raspy noise reminded me of a mobility scooter charger that I can hear from housing in the village near me.

    So it could be something similar.

    Here's info on a drive by system I made to try and ID VDSL sources.

    https://www.g8jnj.net/vdsl

    I used RTL SDR Scanner https://eartoearoak.com/software/rtlsdr-scanner but it has a lot of limitations.

    An SDR Play RSP 1A and the spectrum analyser software https://www.sdrplay.com/spectrum-analyser/ may be another option

    Regards,

    Martin - G8JNJ
  • @M6GQU If you decide this problem is not local to you after all: Do you live anywhere near a facility that might be employing vacuum/RF-heating wood drying at 13.56 MHz? (ISM band). Your video shows multiple carriers coming on between 8.5 and 10.5 MHz with second harmonics as strong as the fundamental. But also "spread out" as the instability of the fundamental carriers is multiplied. But there are also weaker carriers around 11 to 12.5 MHz. So I wonder if this is a bad oscillator plus leakage from one of those machines.

    I think some of them can run hundreds of kilowatts. Because a vacuum tank is involved it should be RF tight. But there is a feed line involved (probably coaxial) and other stuff that might be leaking. My guess is that the RF source doesn't run continuously as it would create hotspots in the wood but is rather on some sort of thermostat which would account for the timing randomness. Here is a paper that mentions both 13.56 MHz and microwave frequencies being used in modern machines: https://scielo.conicyt.cl/scielo.php?script=sci_arttext&pid=S0718-221X2006000200001 (scroll down a bit)
  • Another RFI source I had at a previous QTH that looked similar was from an elevator VFD motor controller. The timing gave that one away.
  • edited March 2019
    I think I have found the source. I noticed last night a neighbours extractor fan (that was never previously used) is blasting out air. It was on when the interference was there this morning, and now that I'm home I've checked again to find the fan is off and there's no interference.

    Could be coincidence. But, I am going to sync my watch and a laptop record to see if both come on at 6pm this evening.
  • As I suspected, both events occur at the same time. The kiwiSDR signal is delayed by a few seconds as the data is uploaded and back down again.

    An extraction fan of sorts, and then something else kicks in? It's almost like a gas boiler and vent turning on and pumping water around the radiators, and then the gas pilot light kicks in. I should point out that none of us in this block have gas. It's entirely electric.

    This vent on the side of the block of apartments is only at this place on the complex so it's some sort of ventilation for all 4 apartments within the block. But I can't figure out what it's for. Not air con nor oven/cooker extraction.

    So at least I think I've found the problem. I now need to investigate what the heck this is venting, and what is powering it.

  • My neighbours have something that comes on for a short time, I did wonder if it was a bathroom fan but it is only on for a set number of seconds so seems unlikely, then i wondered - macerator?
    You are right beside a soil stack, don't stand under the air outlet too long ;-).
    That intermittent one I can live with as it is pretty bad but will be gone in ~40 seconds not hours.

    That fan + other does seem strange if it does not come on at other times.
    Well done tracking it down, now if the owner/responsible-person is not willing to address it, set up a regular evening sked and should that be affected hand the logs to OFCOM.

    I got so wound up with them, I know what the worst device is (can link to high quality video of the exact device with massive wiring in the owners house), it is stupid the amount of junk that comes off the thing, he actually turned the PC off for me one day and the noise went but, at first I could not believe a PC could sound like an arc welder, did more observations and logs but even with evidence OFCOM want me to jump through justification hoops before they will look.
Sign In or Register to comment.