Feed cuts out on a regular basis

From the last 24 hours (since replacing the FA stick with the old DVB-T stick?
If so; here (I replaced the FA stick with the old DVB-T stick last night (Saturday around 22:00):

ADS-B Signal Level 24 uur DVB-T

As you can see, I did until this morning.

No such file or directory.

Something else though; now the system is telling me undervoltage and throttling has occured.

/opt/vc/bin/vcgencmd get_throttled

gives me:


I did not get these undervoltage/throttling errors before. I didn’t get them in the past, not yesterday but only now. Guess both 5V phone chargers (1A and 3A) have problems, so will order the official RPi power supply.

Hopefully that will solve the problem…

IIRC, they sometimes do supply less then 5V, especially if they have to supply more A. Not problem while charging a device, but the Pi’s want their 5V.

And often they are not made for 24/7 duty, so they do not last that long.

Not what i linked, let me be more clear:

sudo journalctl --no-pager -u dump1090-fa

But yeah if you get undervoltage messages … get a proper power supply.

Maybe I’ve been lucky then; the HTC 1A phone charger had ran the setup for almost 5 years straight without problems. :open_mouth:

Ah, right. Did that and nothing unusual there.

The RPi power supply has been ordered and should arrive this Tuesday. We’ll see…

Actually, I am impressed, saw in another thread that it powered a Pi3, the FA stick and the rtl-sdr adsb-preamp…

It did. And as it turned out; that old phone charger isn’t faulty; I just replaced the other (3A) phone charger with the official RPi power supply and it was working for some minutes but down again.

So to me it comes down to either:

  • Raspberry Pi (hardware failure, doub it, that would cause all sorts of problems I don’t have, except maybe some sort of USB hardware failure)
  • FA orange stick (hardware failure)
  • bias tee injector (hardware failure; power LED is on though)
  • LNA
  • Software issue
  • Some weird external influence on the reception.
    (I’m currently doing this test: Do I Need A Filter?)

I think I have an old 1st-gen RPi model B somewhere; I replaced it with the RPi 3B+ since the old RPi couldn’t handle the increased load, but I may be able to swap the RPi 3 with the old RPi and see if the RPi3 is failing.

Could it be wlan network connectivity? Not sure how that would affect the stats collection, but it could explain why no data is send to FA and you can not access Skyaware, and why it happens during the day (neighbors streaming Netflix etc).

Better try this one. It covers full frequency range of Dongle, and is pretty fast.

Spektrum - How-to Speedily Scan RF Noise in band 24MHz ~ 1800MHz

The results of scans are below:

The scan 1 shows RF noise picked by antenna and processed without any filter , The scans 2 & 3 were done with filters, and show how filters remove this noise.

Scan 1 of 3 - FA Antenna + Generic DVB-T (no internal or external filter)

Thumb-Generic DVB-T


Scan 2 of 3 - FA Antenna + ProStick Plus (Only Internal filter of ProStick Plus. No External filter)


Scan 3 of 3 - FA Antenna + ProStick Plus (with internal filter) + External Filter (FA Light Blue)


There’s no network problems. I can access the Pi just fine, including Skyaware. It’s just not receiving/displaying any aircraft, or very few. And it comes and goes.

Thanks, will look at it tomorrow.

graphs1090 doesn’t depend on the pi having network connectivity and it’s showing the gaps.

Yeah you could absolutely try the old pi.
Could very well be MLAT won’t work or you lose some data.

But it should show consistent reception even if it’w worse than possible reception due to insufficient CPU.
The sd-card should work i believe but i’m not 100%.

Will check some more things tomorrow, maybe indeed the old RPi, the old DVB-T stick and that full frequence range test thing.

As long I don’t lose my streak! :wink: 1811 days and counting. :open_mouth:

By the way; mega thanks to you folks for helping me out so far!

I don’t think you need spectrum testing.
Well actually, install the most recent version of graphs1090, a couple months back i re-added the noise floor graph.
If it goes to -0 that means it’s interference.
Well or the LNA only producing noise i suppose.

I wouldn’t expect the rtl-sdr LNA to be overwhelmed by interference, it’s very unlikely.

I think you’re on to something:

Last 48 hours:

ADS-B Signal Level noise 48 uur

Last 24 hours:

ADS-B Signal Level noise 24 uur

I’m trying to pinpoint when it al started. First the same graph for the last 3 months:

ADS-B Signal Level noise 3 maanden

So once or twice in May. But it really starts going in the middle of July:
(Graph is the last 30 days)

ADS-B Signal Level noise 30 dagen

I haven’t changed anything to the setup in since I installed it this way on December 11, 2020.
The only thing changed is that I installed the case fan on June 26 which lowered the maximum core temp with about 10-15 degrees.

So, anyone a clue what causes the interference, interference is something I really don’t know much about…

Might well be the LNA being defect.

What’s the exact models you use again?

Did it start when you turned on the fan?

If that is the only thing that has changed it is possibly the noise source.


All the hardware I use:

  • RPi 3B+ (connected to network via Wifi)
  • Transparant case with fan
  • FlightAware Pro Stick (orange)
  • Bias tee injector
  • FA antenna

Indoor setup: Raspberry 3 with ADS-B | Iemand91 | Flickr
Outdoor setup: FlightAware antenne with LNA | Iemand91 | Flickr
(LNA is in a waterproof box)

You can see in the above graphs the problems started sometime around middle of July, whereas I installed the fan on June 26.

Perhaps try and run without fan for a few days?

1 Like

How did you connect the fan. On the product page of the case it is connected to two gpio-fans, which cannot be right.

There are power pins in the GPIO header.
Those can be used for a 5 V fan without issue.

But yes, i’d try disconnecting it as well to exclude it as a noise source.

Yes, but in the picture of the manufacturer it is connected to 6 and 7 on the row closer to the CPU, that cannot be right. :thinking: