Hi Paul,
would be super, if you can test it!
Does the splitter on the combo filters the right frequencies out?
Have a good day
Bernhard
Hi Bernhard,
OK, I will order an antenna today. It should arrive around the weekend.
The splitter covers 1200 - 2000 MHz, and performs well at 1545 MHz. (I double-checked this morning!)
Best,
Paul
The Tallysman’s are a bit rich for my tastes
The Beitian BT-208 or Beitian BT-300S are Four Star / Full Frequency. The former is on Amazon for ~$100 USD, or Beitian’s own site for ~$89
https://cdn.shopify.com/s/files/1/0712/ … 1701258018
Hi Clive,
thanks for your input, but I want to getit running, then I will try it with your antenna.
Thanks for all
Bernhard
Hi Paul,
today I got my antenna TW 7972 and I get now RXM-PMP correction datas
“”"13:07:57.736 → New RXM-PMP data received. Message payload length is 528 Bytes.
13:07:57.842 → UBX-RXM-COR: ebno: 7.125 protocol: PMP (SPARTN) errStatus: Error-free msgUsed: Used msgEncrypted: Encrypted msgDecrypted: Successfully decrypted
13:07:58.048 → Lat: 48.8862182 Long: 9.5455725 Height: 335.252 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 576 (mm)
13:07:59.078 → Lat: 48.8862182 Long: 9.5455724 Height: 335.258 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 577 (mm)
13:08:00.087 → Lat: 48.8862181 Long: 9.5455725 Height: 335.276 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 578 (mm)
13:08:01.094 → Lat: 48.8862181 Long: 9.5455725 Height: 335.289 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 579 (mm)
13:08:02.098 → Lat: 48.8862180 Long: 9.5455725 Height: 335.298 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 580 (mm)
13:08:02.792 → New RXM-PMP data received. Message payload length is 528 Bytes.
13:08:02.896 → UBX-RXM-COR: ebno: 6.750 protocol: PMP (SPARTN) errStatus: Error-free msgUsed: Not used msgEncrypted: Encrypted msgDecrypted: Successfully decrypted
13:08:03.146 → Lat: 48.8862180 Long: 9.5455725 Height: 335.305 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 581 (mm)
13:08:04.136 → Lat: 48.8862179 Long: 9.5455725 Height: 335.312 Fix: 3 (3D) Carrier Solution: 0 (None) Horizontal Accuracy Estimate: 581 (mm)
13:08:04.444 → New RXM-PMP data received. Message payload length is 528 Bytes.“”"
But I do not get RTK. as you can see carrier =0.
I use the L Band example 2. Which credetials do I have to put in the secrets.h
I have only the L Band keys put in.
As far I can see is, that the correction datas are encrypted sucessfully in the F9P. Why can the chip not achieve the carrSoln = 2 ?
Do you have an idea?
Thanks
Bernhard
It should have the L-Band keys pulled from ThingStream, currently active one is “319…E21”
You’d want to see some of the messages being flagged as “Used”, not all will be as there are multiple geographic areas described in the stream. But if nothing is taken, that’s going to be a problem.
I’m not sure how the receiver determines if something is decrypted properly, as there isn’t an integrity check, but rather the ability to decode the data instead of being junk/noise
Could add functionality here for a persistent long term “Used” count. That way you could leave it for several minutes and see if anything is working without having to scan all the lines of the logging data.
Hi Bernhard,
I performed some test this morning using different antennas and NEO-D9S/ZED-F9P board combinations.
Antennas:
Beitian BT-800S (this is the antenna I use the most)
Tallysman TW7972 (33-3972-01-01) - using the supplied 100mm ground plane
u-blox ANN-MB-00 - on a 100mm ground plane
RTL-SDR L-Band antenna ( https://www.rtl-sdr.com/product/rtl-sdr … tenna-set/ , https://www.elektor.com/products/rtl-sd … ntenna-set )
I also used another Taoglas antenna to feed just the ZED-F9P - but this one has a poor view of the sky.
Boards:
The original SparkX Combo Board ( https://www.sparkfun.com/products/retired/20167 - this had a different splitter )
NEO-D9S breakout ( https://www.sparkfun.com/products/19390 )
ZED-F9P SMA breakout ( https://www.sparkfun.com/products/16481 )
SparkX splitter ( https://www.sparkfun.com/products/21223 )
I tried to be as systematic as possible. Each antenna was placed on my antenna pole and raised to full height - this provides a full, clear view of the sky - above the house roofline. The cable is ~12m of RG58.
I adapted Example2 so that it would average the “ebno” L-Band signal level (dB, Energy per bit to noise power spectral density ratio), the fecBits (Number of bits corrected by FEC (forward error correction)), and the NAV-SAT cno (dBHz Carrier to noise ratio (signal strength)) for SBAS satellite 123.
The L-Band satellite is very low in the sky for me:
Latitude: 54.9000°
Longitude: -1.4000°
Satellite: 24.8E ALPHASAT
Elevation: 23.1°
Azimuth (true): 149.0°
Azimuth (magn.): 148.9°
Beitian BT-800S
=============
Separate boards with the SparkX Splitter:
fixType: 3 | carrSoln: 2 | hAcc: 58 | ebno: 4.33 | fecBits: 531.4 | cno (SBAS 123): 37.8
NEO-D9S on Beitian BT-800S, ZED-F9P on Taoglas (poor sky view):
fixType: 3 | carrSoln: 1 | hAcc: 106 | ebno: 8.76 | fecBits: 15.8 | cno (SBAS 123): 0.0
(The zero signal for SBAS is expected due to the poor sky view)
Tallysman TW7972
===============
Separate boards with the SparkX Splitter:
fixType: 3 | carrSoln: 0 | hAcc: 540 | cno (SBAS 123): 39.0
No L-Band signal. SBAS is +1.2dB compared to the BT-800S.
NEO-D9S on Tallysman TW7972, ZED-F9P on Taoglas (poor sky view):
fixType: 3 | carrSoln: 1 | hAcc: 115 | ebno: 6.26 | fecBits: 155.7 | cno (SBAS 123): 0.0
L-Band is approx. 2.5dB worse than BT-800S.
ANN-MB-00
==========
Separate boards with the SparkX Splitter:
fixType: 3 | carrSoln: 0 | hAcc: 1838 | cno (SBAS 123): 8.6
No L-Band signal.
NEO-D9S on ANN-MB-00, ZED-F9P on Taoglas (poor sky view):
fixType: 3 | carrSoln: 0 | hAcc: 1232 | cno (SBAS 123): 0.0
No L-Band signal.
RTL-SDR L-Band
=============
The antenna was oriented directly at the L-Band satellite.
Separate boards with the SparkX Splitter:
fixType: 3 | carrSoln: 1 | hAcc: 142 | ebno: 5.00 | fecBits: 365.9 | cno (SBAS 123): 43.0
L-Band is approx. +0.7dB compared to the BT-800S. SBAS is approx. +5dB compared to the BT-800S.
NEO-D9S on RTL-SDR L-Band, ZED-F9P on Taoglas (poor sky view):
fixType: 3 | carrSoln: 1 | hAcc: 90 | ebno: 11.47 | fecBits: 0.4 | cno (SBAS 123): 0.0
L-Band is approx. +3dB compared to the BT-800S.
I think if you give the Tallysman more time, it should achieve at least a float solution. But, unfortunately, it is not the magic fix we were hoping for. It performs worse for me than the BT-800S.
The best solution is to use separate boards and antennas: a NEO-D9S breakout with an RTL-SDR L-Band antenna pointed directly at the satellite; a separate ZED-F9P breakout with a good antenna (the Tallysman should be ideal here). Link NEO UART2 to ZED UART2 if needed - but it is also possible to push the corrections over I2C.
In the USA, I was usually achieving an ebno of 9dB and close to zero fecBits with the ANN-MB-00.
I hope this helps. Please let me know if I can help further or if you would like me to carry out any more tests.
Very best wishes,
Paul
Dear Paul,
now it works and I get a RTK position! 
Thanks to all of you !
I saw in the u- center, that there was a orientation of the antenna. How can I get this?
Have a pleasant day!
Greetings Bernhard
That’s great.
Spookily enough I looked up the view angle for Sunderland this morning.
This is what I’d be shooting for with the PMP, mid 40’s dBHz, low teens Eb/No, approaching zero error / FEC bits.
This was indoors (wood/shingle roof) w/RTL-SDR L-Band orientated optimally. Can get similar numbers with my USB SPARTN-MOUSE
https://raw.githubusercontent.com/cturv … 9s_004.jpg
47.55 dBHz, 13.625 Eb/No, usually probably closer to 45 dBHz
PaulZC:
I hope this helps. Please let me know if I can help further or if you would like me to carry out any more tests.
More tests and real world experiences the better…
As the roll-out and up-take of these services expands, it’s the initial setup and antenna choices that are going to reduce support headaches. Having some “known working” combinations and expectations will always help.
You can’t fix bad placement issues with more expensive antennas or more gain…
Hello, I am using two ZED-FP9 board as a Base station and a Rover. For the testing I am performing, it is required that we move the base station antenna to different locations throughout the process. The only way we have found to establish the new Base station location is to power cycle the base station. That way when it is powered back on, a new survey is completed to get the new location. Is there a way for the general location to be stored and used to speed up the survey when it is run again?
We were told you can typically store the base station’s location information in its configuration or memory. This information can include coordinates (latitude, longitude, and altitude) as well as any relevant geodetic datum. By storing this data, you can avoid the need to conduct a full survey every time the base station is moved to a new location. It can significantly speed up the setup process.
This being said, is there a command needed to store the location or does this happen automatically? Also, is there a different configuration mode where it does not conduct the full survey? I am currently keeping the base station in the Survey-In mode. I have not used the fixed position mode yet but that seems like what would be used here.
Hi piercelatta,
Our u-blox GNSS Arduino library has examples that will help:
Survey-In: https://github.com/sparkfun/SparkFun_u- … CMBase.ino
Static Position: https://github.com/sparkfun/SparkFun_u- … sition.ino
Best wishes,
Paul