ZED-F9P ubx files logging L1 or L2 band per satellite, not both

GPS/GNSS
1

In a sequel to this post, we are logging raw data (UBX-RXM-RAWX and UBX-RXM-SFRBX) from the F9P at 2Hz, however when converting this raw data to RINEX using Emlid Studio, we find that for the vast majority of satellites, only the L1 or L2 band is being logged. We will very rarely see both L1 and L2 logged for a satellite, but typically for only a few observations per file (file length: 2 min). This occurs both when logging through U-Center and a Python script which uses pyubx2 to disable NMEA messages on the USB port of the F9P, enable the appropriate raw messages, and log the raw data from the serial port to 2 minute interval files.

Converted RINEX from U-Center raw data logging:

image
image1920×1032 107 KB

Converted RINEX from Python raw data logging:

image
image1920×1032 87.2 KB

In the above photos, yellow points are observations with the L1 frequency only, purple points are L2 only, and green points have both L1 and L2.

This data was collected using a Calian VSP6337L Triple-band Antenna, which we have had good results with in the past using other receivers.

2

Not sure the imagery is helping me understand the source of the issue. Attach an example .UBX capture from uCenter

3

Here is the ubx file captured from U-Center: https://drive.google.com/file/d/1HLuMM2uQHMD3kCRxjlI9D0x7Sd0on5p_/view?usp=sharing

And here is a ubx file captured from our python script: https://drive.google.com/file/d/1nWQhUd5D1IBZnfeA0GfpbwdALP1KHiRU/view?usp=sharing

4

Honestly, doesn’t look to be getting carrier/doppler lock, and pretty awful CNO (Carrier-to-Noise Ratio). The lock-time is on the right column, if it’s got continuous signal lock this should clamp at round 65K. So Low CNO generally, and discontinuous lock.

image
image729×767 21.9 KB

Try a different antenna, like a ANN-MB or BT-800S
Check the voltage at the antenna end of the feed-line. What type / length of cable are you using? Consider antenna placement. All CNO underwhelming for a 37 dB gain antenna..

5

Here’s an example of what it SHOULD look like where the RF is working properly.

image
image722×924 24.6 KB

Do you have a nearby source of electrical interference? Motors, EMI, RF Transmitters (Bluetooth, LoRa, Cellular, WiFi et al)?

6

Yes, as it was an informal test, I was using the antenna in close proximity to my laptop on a wifi connection and my bluetooth headphones. As for antenna placement, I do believe we had good skyview and clearance from shadowing/multipathing from sounding buildings, but I did not get a photo of the set up.

We used the 10 meter SMA to TNC cable included in the Sparkfun base station kit, as it was our most readily available SMA to TNC cable.

I will repeat this test in a more isolated location, removed from potential EMF interference. I will test with both the original antenna, and the SPK6618H antenna included with the Sparkfun base station kit.

7

I repeated the test this afternoon, setting up on our pier to achieve a clearer skyline and to get out of the range of the building’s wifi. I also turned off bluetooth on my devices. The results across all 3 antennas (the Calian VSP6337L, the Sparkfun SPK6618H, and the Trimble Zephyr 2) showed the receiver receiving an L1 or L2 signal from each satellite, rarely both.

Files can be found here: https://drive.google.com/drive/folders/1e08SdDr5uZh2C4dpMMimsXRsjVR0tMV7?usp=sharing

@clive1 What program are you using to find the CNO?

8

I’m using an in-house dumping tool I have for .UBX files and generating RINEX files.

I’d strongly recommend checking the DC voltage at the ANTENNA end of the cable, I’d hazard that’s much less than the 3.5 VDC the Zephyr needs. The SPK6618H needs at lease 3 VDC

What you’re looking for are CNO numbers closer to 40 dBHz than 30 dBHz, you also want continuous lock

image
image724×365 10.2 KB