Bad Performance with PoE Splitter

FlightAware ADS-B Flight Tracking
1

Hey guys,

I’m using a PoE splitter (5V/4A — it’s a Revodata 2.5G Type-C PoE splitter) to power my setup. The PoE source is a Zyxel GS1200-8HP v2 switch.

Since switching to PoE, I’ve noticed reduced performance:

  • My range has dropped from around 150 nm to about 100 nm
  • I’m seeing less aircraft - about 80 instead of 120

According to the switch’s web interface, the Raspberry Pi is drawing only 5.2 W.

My setup consists of:

  • Raspberry Pi 3B v1.2
  • Nooelec RTL-SDR v5 receiver
  • A cheap active antenna from AliExpress (also powered via USB from the Pi)

Before switching to PoE, I was using a 5V/3A power supply - so even less than what the PoE splitter is rated for.

Any ideas what might be causing the drop in performance?

Thanks :slight_smile:

2

It could be r.f. Noise being generated by the POE splitter.

2 Likes
3

Might be an option.

Do you think a PoE HAT might be better? (of course I’d need a RasPi 4 then)

4

It would be noise from the PoE splitter indeed. Check the log if you are getting undervoltage messages from the Pi, this will result in a decreased range as well. If so see if the range improvises with a regular power supply to eliminate that as a cause.

5

Any switching style power supply like a wall wart or any device or equipment outfitted with a switching power supply will generate serious amounts of RF noise that will degrade your system performance. Advise that you locate any such noise generators as far away from your system as possible. You might consider relocating your current power source or sources as far away from your system components as possible and measure any possible improvements from that relocation.

6

A power problem to an active antenna would certainly degrade performance.
Do you have a link to the antenna?

How are you powering the antenna? - with a Bias-T or does it have it’s own input?

7

Thanks for all the help so far — here’s what I’ve observed testing:

  • Switching back to a normal power supply significantly improves ADS-B performance (more aircraft, longer range).
  • Switching back to PoE reduces both range and number of detected aircraft.
  • I monitored the system for 15 minutes and found no undervoltage or throttling in the logs.
  • I often see aircraft (in the mid and wide range) popping up and disapear again

I’m using this PoE splitter (5V/4A):

CraigWoodThomas Advise that you locate any such noise generators as far away from your system as possible.

Well I’m just building an outdoor Box to place the antenna on the roof. That’s why i want to power it over PoE and the splitter it has to be in the same box. I’m just living in a flat and my landlord does not allow me to place two cables to the roof.

geckoVN A power problem to an active antenna would certainly degrade performance.
Do you have a link to the antenna?
How are you powering the antenna? - with a Bias-T or does it have it’s own input?

I’m using this active antenna:
https://de.aliexpress.com/item/1005007735199789.html?spm=a2g0o.order_list.order_list_main.5.5af45c5fbLcnCK&gatewayAdapt=glo2deu

However, since I only get the problems with the PoE Splitter, it might really be HF interference…

  • Has anyone successfully used this or a similar PoE splitter with an active antenna + RTL-SDR setup?
  • Would using a Raspberry Pi 4B with the official PoE+ HAT give cleaner power and help avoid this issue?
  • Any other tips for reducing interference when the PoE splitter must be close to the SDR and antenna?

The signal isn’t terrible, but the difference compared to a regular power supply is noticeable.

Thanks in advance!

8

@GTIFreak

Browsing this thread may help.

Remote power for Pi

 

9

@GTIFreak

AF Complient PoE

 

10

Thanks for the thread, but I think it doesn’t really help.
My power supply is AF-compliant and it also protects non-POE devices.
And it looks like, the power supply is enough (no throttling etc.).
Also the splitter SHOULD be, at least there is a sticker on it that says “IEEE 802.3af/at”

However, its a cheap splitter from amazon, maybe a better PoE Splitter, a PoE Hat or a ferrite core around the USB cable may help?

11

I haven’t tried the latest PoE HAT for the Pi4 or the one for the Pi5 to feed power to my system.
However I have read reports of RF problems but I don’t know if this was cured when they updated the design to fix the overcurrent problem with the original HAT.
There are some details about the general 5V noise between the original and revised V2 designs here:

1 Like
12

Have you tried that?
When I tried it, I was impressed at the drop in performance.

13

LawrenceHill There are some details about the general 5V noise between the original and revised V2 designs here:

so i might just give the PoE HAT a try. Thanks for the link :slight_smile:

geckoVN When I tried it, I was impressed at the drop in performance.

Really?
A ferrite core around the power supply USB gave you a performance drop?
Sounds strange to me.

14

Sorry - no. I guessed wrong which USB cable you were refering to.
A ferrite core on the Pi - Dongle USB cable caused a drop in performance

Just a thought - how long / how accessable is the UTP cable?

15

A ferrite core on the Pi - Dongle USB cable caused a drop in performance

Hm, even that is strange, I might buy a core and try different solutions :slight_smile:

Just a thought - how long / how accessable is the UTP cable?

For testing its ~10m (33 ft), however, the final cable should be 30-40m (100-130ft).

16

My thought was to try an STP cable and see if it reduces the noise emissions.
Depends if the shield is connected to anything in the splitter.

17

geckoVN
My thought was to try an STP cable and see if it reduces the noise emissions.
Depends if the shield is connected to anything in the splitter.

Do you think that might help? I think, if HF is a problem, it comes from the POE Splitter/ POE HAT itself, right?

18

It sound like it, but I’m thinking that if the PCB in the splitter is connected to the shield of the RJ45. using an STP might reduce the emmission.
I woudn’t put a lot of money on it. but worth a try.

19

geckoVN
I woudn’t put a lot of money on it. but worth a try.

That’s true, nothing to loose :slight_smile:

20

So I did some testing:

  • No difference depending on LAN cable type or length

  • PoE HAT performs slightly better than a PoE splitter

  • When using a metal Raspi case with PoE (either splitter or HAT), performance is significantly worse

  • With the same metal case but powered externally (no PoE), performance is close to the reference setup (no case, no PoE)

  • The antenna wasn’t moved during any of the tests

  • Each setup was tested for 24 hours

Plastic cases would be okay — I don’t really need the metal one. But does anyone have ideas on how to get rid of the PoE-related interference? In the best case with the PoE Hat?