Recover GPS-RTK2 from Connection Issues?

SparkFun GPS-RTK2 u-blox ZED-F9P Connection Issues

Setup:

  • OS: Windows 10
  • Connection: USB-C via u-center

I connected the GPS-RTK2 to u-center via USB-C, and everything seemed fine initially. Then I added an OpenLog SD card reader using Qwiic, but noticed it was only writing blank files to the SD card—strange.

I tried initializing OpenLog with a sketch from the example library in Arduino IDE but got no results. I adjusted settings in u-center, especially messages to/from UART1, experimenting with various configurations based on forum advice and my growing understanding of the system. Then, I tried uploading an example sketch to write RTCM data to the SD card, along with other similar sketches.

Issues: At this point, u-center started having trouble connecting to the COM ports. Opening the menu takes a long time, where originally it popped right up. It seemed the ports were always busy, but I couldn’t find anything else using it—Arduino IDE included, though it might be holding a reference somewhere in Windows. Switching USB ports or rebooting would sometimes resolve the issue, but now I can rarely connect. When I do, the top bar shows a green handshake, but the bottom bar stays grey, and there’s no activity in the UI. u-center crashes when I attempt any action.

Additionally, even outside of device connections, u-center seems laggy. I’ve reinstalled it multiple times, removed drivers, and rebooted without improvement.

To troubleshoot further, I purchased a Serial Basic, soldered pins onto the RTCM correction pins on the GPS-RTK2 (UART2?), and connected it via USB-C using a brand new cable. It shows up as USB-Serial CH340 on COM15. I confirmed the connection with a loopback test in PuTTY, but when I connect it to the GPS-RTK2 and open u-center, I face the same “connected but no info” status. Disappointingly, I get no response from the GPS-RTK2 through the Serial Basic.

The firmware update tool also fails to connect, whether through USB-C or Serial Basic. It repeatedly tries to poll the version but eventually fails, and the update utility crashes with an exit code (2).

I have pulled line low on RST pin on GPS-RTK2 more than once, and while it does reset, it doesn’t do anything meaningful and the problems remain.

The GPS-RTK2 PPS light is blinking once every 4s. I believe in the past, it was blinking every 2s. I have not been able to find posts on PPS blinking at a frequency that is not once per s.

Summary: I’ve read countless documents and forum posts, but I’m stuck. What could I be doing wrong, and is there a way to recover from this? Any advice would be greatly appreciated. Thanks for your time.

Do you have any win11 machines around that you can test with to rule it out as a potential issue?

On Windows 11, unfortunately I get the same results.

To GPS-RTK2 cannot connect to com port
Via Serial Basic I get the connection up top, but no handshake icon on bottom bar, no action.

Thanks for the suggestion. I need to find a way to update my i9900 its ineligible :angry:

Edit: On windows 11, the lag in u-center is gone, however it is of course another machine.

bumping this along

Perhaps someone can confirm to me instead that it is indeed possible to disallow through configuration communication on both uart ports to a point where the reset functionality does not bring it back to a workable state, and the board is in fact ‘bricked’ as they say and is now garbage - would just be nice to hear that you can in fact ruin these boards so I can move on.

In addition - any comment on the successful connection but no handshake activity on the lower bar. This is very common to read about online but no answers as to what causes it or how to restore settings from this state.

Thanks

If I’ve got Bluetooth enabled, I will see u-center lag for 4-5 seconds when I hit the com port drop down box:

image

image

image

This is normal behavior in my mind. This is u-center pinging each COM port to see what is active. If Bluetooth is turned off, and those ports are gone, then u-center is very quick to display the dropdown of COM ports. This behavior doesn’t affect the ZED, and is just an artifact of the u-center interface.

Once connected to a ZED, with an antenna attached, you should begin to see all sorts of data roll by including a 1 second blink of the PPS LED. You report 4 seconds - I have never seen this, nor can I explain it.

I would start from the basics. I assume you’ve got the u-blox driver installed and can see the associated COM port. From there, I would use u-center to verify communication and the version of firmware loaded on the ZED:

image

Above, firmware is HPG v1.32. If you’ve gotten this far, I would do a factory reset by using the CFG-CFG command below:

Assuming you’ve got the externally attached USB to serial board on UART 1 of the ZED, it should be outputting NMEA at 38400 (once you’ve factory reset things), config shown below:
image

Thanks for your detailed reply, and I do apologize for the delay in my response.

I never saw any delay when starting to work with this unit, however after adjusting MSG settings I began to see it - but I will assume for now this isn’t my issue!

I have tried without and with driver. When it worked initially, having ublox installed, it would show as USB Serial Device. I do not see it under Sensor at this time. After installing the driver, which I could not find on u-blox site for my zed-f9p-002-B, but found posted by clive1 as extracted from Win WHQL ~2yr ago*, it shows under Ports as ‘u-blox GNSS Receiver’ - still no ‘Sensor’ tab in Device Manager.

Neither configuration changes anything visible in u-center (24.05 which worked originally out of the box so to speak). I am connected in the top bar, but no activity on the bottom bar, no data at all showing on screen - in fact the UI does not change when connected.

Same behaviour when connecting through the serial adapter. PPS light every 4 seconds, nothing else happens. I have attempted to power the board separately as I read the power via serial adapter is too little for the gps-rtk2 (who knows if that is true, seems odd to me) but no joy there either.

The most frustrating part of this is that the reset functionality does not reset the board. I really don’t understand that.

If you have any more ideas I would be open to them. I read a post where someone de-soldiered the f9p from the gps-rtk2 and reset the f9p with jumpers - I don’t know if 1) that’s worth my time and 2) if I would be good enough to do that soldier work.

Edit: Reset via CFG-CFG:
When “connected” (even though I see no data or ui change) I can still open Configuration View. When I click to CFG-CFG the Send button is illuminated, however any action, such as Revert to default Config, and pressing Send causes the software to hang. Any click on the UI during this hang hard crashes u-center.

Thanks again for your reply above.

*u-blox Portal

Shouldn’t need to have to do a bunch of soldering, the -SAFE pin can be pulled to ground to get the unit into recovery mode.
However this won’t fix electrical damage, issues with USB connectors, cables or USB Host ports, nor driver stack issues on the host PC.
The 1PPS at 1/4 Hz does seem a little odd.
SAFEBOOT allows you to update the receiver and clear configuration (Erase Chip), if you’d broken the setting so bad it’s unresponsive. In SAFEBOOT the UART1 should output at 9600, normally 38400 baud, and should output a bunch of $GxTXT Sentences reporting what’s happening at startup.
Generally speaking you don’t want Microsoft’s sensor drivers and other clutter interfering with the receiver, much better to have direct COM port access either directly via the ZED-F9P’s native USB, or a USB to CMOS Serial dongle on UART1.
On the electrical side you could check voltages with a multi-meter, the TX pins should normally be high (3.3V), as should the center pin for the RF input.
Avoid shorting the board on conductive surfaces.

With or without the u-blox driver installed, booting with Safe low when connecting with onboard usb-C I receive pps light solid yellow but in device manager, I receive an unknown Device descriptor request failed entry

When connecting via serial adapter booting safe low I see the normal serial com port connection, PPS light solid and normal serial com port entry.

but still no data: I have tried pushing firmware update after booting safe low w serial adapter, and am met with these logs. I have tried this various times.

It seems, there is no flash found?

Flash: 0 ; Flash Detection Timed out.

Perhaps, my soldier is weak. However, my device manager is not flicking during the process as it does when connecting or disconnecting the device.

Many, many thanks for your time. I will keep working on this, however it is not of grave concern.

Logs:
Setup pipes

Process launched

-p STDIO -b 9600:9600:9600 --no-fis 1 -C 1 --up-ram 1 -s 1 -t 1 -v 1 “C:\Users\Tom\Downloads\UBX_F9_100_HPG_150_ZED_F9P.82afa0b888dab694869a5098bf3a9f0b.bin”

----------CMD line arguments-----------

Image file: C:\Users\Tom\Downloads\UBX_F9_100_HPG_150_ZED_F9P.82afa0b888dab694869a5098bf3a9f0b.bin
Flash:
Fis: flash.xml
Port: STDIO
Baudrates: 9600/9600/9600
Safeboot: 1
Reset: 1
AutoBaud: 0
Verbose: 1
Erase all: 1
Erase only: 0
Training sequence: 1

Chip erase: 1

Merging FIS: 1
Update RAM: 1
Use USB alt: 0

Sorry Clive I sent this rather prematurely.

The voltage on both TX2 pins is 3.5v.
The 5v measures 5.3 and the 3v3 measures 5.1v.

I assume RF is only accessible if I remove the seated antenna?

I am left wondering - am I using the wrong u-center version, wrong firmware? I tried flashing the chip with Erase All: 1 as my logs above show, but perhaps I did it incorrectly, or there is another process.

Again this part of the log really gets me 1) the flash size 0 - although I have no idea what I am talking about here and 2) that it tries and reaches various stages of downloading to ram.

Choosing not to download to ram results in failure.
Thanks

Hi @tomg ,

That’s very worrying…

I suspect you’re trying to connect using a 5V “FTDI” cable, instead of a 3.3V version? If the cable is feeding 5V power into the ZED breakout board via the UART2 header, that would explain all sorts of badness…

I hope this helps,
Paul

1 Like

Thanks Paul - sorry I don’t know what I was on about yesterday.

With the power through the gps-rtk2, voltage is consistent with the pins
5 - 5.2v

3 - 3.41 (both sides)

With the power through the serial basic (which shows both 3.3 and 5v on back surrounding a jumper)

5 - 3.4

3 - 3.4

I take it the jumper is to set it to 5v, so I am using it correctly.

Apologies for the confusion and thanks for jumping in.

Hi @tomg ,

No worries. Those voltages look correct.

We have a couple of Serial Basic products. Both default to 3.3V, which is what you need for the GPS-RTK2.

Best,
Paul