RS2 showing bad L2 Data Quality

Our Reach RS2 Unit is on 2.22.0 Firmware

We have conducted a Logging of RINEX 2.11 Files for later PPK Processing. The Processing failed always with 0% PPK Fix Rate. We have contacted the Vendor of the Solution (Wingtra ONE UAV which carries a Septentrio GNSS PPK Module). This is a follow up of RS2: Configuration for PPK in 3rd Party App (Wingtra/Septentrio)

The Customer Service told us that there was very bad Data Quality on the L2 Band (see Attached Image).

What could explain this? Is our Unit broken? We do get accurate RTK Fix and Position Readings.

The System Simple Report:

[details=“Simple system report”]

app version: 2.22.0-r0
'wifi_status, interface: wlan0':
- wifi_mode: ap
- access_point:
    band: bg
    channel: 1
    password: null
  ip: 192.168.42.1
  is_added: true
  is_connected: true
  mac_address: 6C:21:A2:93:0B:CA
  security: wpa-psk
  ssid: Reach_RS2:5D:0D
  uuid: 72d808b0-9fc8-45e7-8e95-e659ef14174e
base mode:
  base coordinates:
    accumulation: 5
    antenna offset:
      east: '0'
      north: '0'
      up: '0'
    coordinates:
    - 0
    - 0
    - 0
    format: llh
    mode: fix-and-hold
  output:
    enabled: false
    format: rtcm3
    path: tcpsvr://:9000#rtcm3
    type: tcpsvr
  rtcm3 messages:
    '1006':
      enabled: true
      frequency: 0.10000000000000001
    '1074':
      enabled: true
      frequency: 1
    '1084':
      enabled: true
      frequency: 1
    '1094':
      enabled: false
      frequency: 1
    '1124':
      enabled: false
      frequency: 1
bluetooth:
  discoverable: false
  enabled: false
  pin: '***'
constraints:
  lora:
    frequency:
    - - 863000
      - 870000
correction input:
  input2:
    enabled: true
    format: rtcm3
    path: *****:***@www.swipos.ch:2101/MSM_GISGEO_LV95LN02
    send position to base: single
    type: ntripcli
  input3:
    enabled: false
    format: RTCM3
    io_type: tcpsvr
    path: :9028
device: null
logging:
  base:
    format: RINEX
    started: false
    version: '2.11'
  correction:
    format: null
    started: null
  debug: false
  interval: 24
  overwrite: true
  raw:
    format: RINEX
    started: false
    version: '2.11'
  solution:
    format: LLH
    started: false
lora:
  air rate: 2.6000000000000001
  frequency: 868000
  mode: null
  output power: 20
position output:
  output1:
    enabled: false
    format: nmea
    path: bluetooth
    type: bluetooth
  output2:
    enabled: false
    format: llh
    path: :9001
    type: tcpsvr
rtk settings:
  elevation mask angle: 15
  glonass ar mode: 'off'
  gps ar mode: fix-and-hold
  max horizontal acceleration: 1
  max vertical acceleration: 1
  positioning mode: static
  positioning systems:
    compass: true
    galileo: true
    glonass: true
    gps: true
    qzss: true
  snr mask: 35
  update rate: 1
sound: null

[/details]

RTKPlots from the EMLID RTKLib Distro. SNR/MP/EL once for L1 (I would say this looks ok):

But for L2 the SNR is much worse and no L2 Signal is 45dB or better

The Survey was conducted on a flat piece of land, not obstacles or vegetation.

I would also highlight, that Survey points have been recorded very accurately the only Problem is PPK Postprocessing of RINEX Files from the Base.

Hi @Andreas.

After reading your post, I checked my RS2 logs and found the same as you, L2 always under 45db and L1 OK.

I downloaded a Rinex file from 2 different permament stations and opened them with Rtkplot. L2 looked good at 50db aprox.

Firmware version 2.22.2

RS2 L2

Permanent station - L2

Another permanent station L2

Does anyone know if this is ok?

I can say that I’m getting Fix with RS2 and RTK and have post-processed from a M+ mounted on a UAV at a distance of 1km, getting fixes for all camera events, but I don’t know if this L2 SNR means something bad on the RS2.

Hi @Andreas and @Doppler_Uav,

May I ask you to share your raw data logs from the Reach RS2 so that we could also check them?

Sent you by PM @polina.buriak.

Thanks

It is normal for the L2 signal to have lower SNR than L1: https://www.researchgate.net/post/What_is_the_difference_between_SNR_due_to_GPS_L1_signal_and_L2_signal_for_Block_IIR-M_satellites

1 Like

I have received the SNR Details from the Septentrio Module that is in the UAV.

This is the SNR for L1:

This is L2C for the same Satellite:

And this is L2C as seen on the UAV:

Position of the Reach RS2 was on a 2m Mast. Nearest Mobileantenna is > 400m away. There is only a field, a road and Railway tracks nearby.

My RF Equipment was a FrSky Remote Controller (2.4GHz) and a 848Mhz Telemetry Link for the UAV, my Position was more than 5m away from the GNSS Position.

So why is G07 so much worse on L2 Band on the Ground GNSS than on the Airborne Septentrio System?

This is a Picture of the System used (© Wingtra):

Hey there,

Let me clear this out.

As Christian has already pointed out, it’s absolutely normal for the L2 SNR to be lower than the L1 SNR.

Andreas, the issue with the post-processing hardly seems to be connected to the difference between SNR values for L1 and L2 frequencies. To help us determine the cause of the issue, would you mind sharing your files from both base and rover so that we can take a closer look? Could you please specify if the software is capable of processing L2C data only or it requires an L2P signal as well?

Hi @polina.buriak,

It may be the case, that L2 SNR is systematically lower than L1 I do not doubt it. However the UAV has significantly better L2 SNR for the chosen Sat than the Base. I’ll send you all the files via PM.

Maybe I’m wrong, but after doing some research I think that it may be related to the receiver/antenna. I have logs from other L2 GPS receivers and can see that L2 SNR values are pretty similar to L1 (50db), while on the RS2 it’s almost never higher than 40, as I described on my previous post.

Maybe it’s a characteristic of the RS2, and I’m not saying it’s necessarily a bad thing because I prefer having L2 than not having it.

I’m having good results in short distances. I don’t know if there can be any issues or disadvantages when doing PPK for longer baselines (more than 20km) with these L2 SNR values comparing with receivers with higher SRN values.

That claim should be fairly easy to reproduce with a 3rd party antenna on the M2, I would suppose?

What is the elevation angle of that satellite? I’m thinking that a very low angle satellite might have more interference from ground obstacles than what you’d get on a drone that’s above most objects.

What about satellites with high angles? Any significant difference between Emlid/Septentrio?

Here you go. RTKPlot SNR/MP/EL for G07 on L2

And same for L1

G05 for instance was above 60° all the time. Exactly -10dB difference L1 -> L2. Highest elevation was at 65° where it got 39-40dBHz

Hi guys,

According to the log from Andreas’ drone, the L2 quality of its data is not drastically different from the L2 quality of the Reach RS2. I’ll attach the screenshots from both of them so it’s easier to understand.

This is Reach RS2 L2 data:

This is the drone’s L2 data:

For example, you can see that SNR for GPS satellites is above 25 only for G30 (it’s around 30) on the drone. Also, for the GLONASS constellation, there are some cycle slips present, especially for R19, R20, and R21.

Reach RS2’s GPS data SNR is mainly above 30 including with higher values for certain satellites (around 40 for G30, around 35 for G07, etc). GLONASS SNR for the L2 is mainly above 35.

However, this slight difference doesn’t prevent us from getting successful post-processing results. I’ve post-processed the logs from Reach RS2 base and Wingtra drone in RTKPost with an almost 96% fix solution. I’ll attach the screenshot.

I assume that the issue Andreas faced with the Wingtra ONE software is due to the specifics requirements of the software. It’d be great to understand what they are.

Andreas, would it be possible to find out whether the software requires L2P data alongside L2C? Reach RS2 records only L2C signal so this might be the reason for the issues with post-processing.

2 Likes

Andreas,

Just wanted to add that it’d be great to test the software ourselves. Would it be possible for you to introduce me to the Customer Service agent at Wingtra who had taken a look at the logs already?

2 Likes

Hi Polina,

I got in touch with the person regarding your Request. If they agree, I’ll send you the Contact via PM.

1 Like

Hi @wizprod,

I was referring to RS2, not M2. I say that maybe the antenna/receiver on RS2 has that capability of receving L2 with SNR values up to 40db (or a bit more), while other L2 GPS can receive up to 50db.

At least that’s what I can figure from the logs that I have:

RS2

Another L2 GPS

And another L2 GPS

I know, but to prove that it is the antenna and not the ublox chipset, one could mount a 3rd party antenna on the M2.

It would be interesting to see RTKPLOT’s “L2” drop-down changed to “L2C”, so we are comparing apples to apples. As Polina mentioned above, Reach doesn’t use L2P signals, which are probably included in the other two graphs and they may have different SNR characteristics or cutoffs.

4 Likes

Ok @bide, I’ll be back next week. I’ll check that then, and post the results.

1 Like