Last edited:
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Caution - Annual Service - Car power-down - GPS issue
- Thread starter jfischer
- Start date
-
- Tags
- clock gps broken gps rollover
DrComputer
TeslaClubLA President
Just an FYI to the Roadster owners in the LA area that need Steve's GPS upgrade cable. My car is currently at Executive Motorwerkz in Lawndale with Javier and his team doing my annual maintenance and a few other repairs (TPMS sensors, alignment, etc). He has the GPS upgrade cable and is upgrading my car while it is there. If you want to perform the upgrade yourself I am happy to give you the cable when I get my car back from him (which should be in the next few days). If you want to have him perform the upgrade is his shop, give him a call (Javier 213-235-5694) and I can leave the cable with him. PM me on here if you are in the LA area and want the cable next.
My cable for doing the update by attaching at the VMS is currently in Los Angeles and has a couple of other destinations in the queue. In the early posts of this thread it sounded like @jfisher may have made a cable to connect at the junction inside the front wheel well. I don't know if any of the Tesla SCs are prepared to do the update -- there have been reports from some that are not.
In a PM @KJA2001 asked for instructions on how to make a GPS firmware update cable like the one I made since there are a few other people in the UK that need the update, including @dpeilow and @KeithBoxer, so it would be a good idea to have one on that side of the pond. I recall a request from Sweden as well.
Accordingly, I have written up the attached note with a list of parts and instructions on making the cable. I assume one of you guys can volunteer to build it, or if it is outside of your skill level, to find someone who can do it.
Accordingly, I have written up the attached note with a list of parts and instructions on making the cable. I assume one of you guys can volunteer to build it, or if it is outside of your skill level, to find someone who can do it.
LOL welcome to the "GPS TIme Travel" family. I can send you the cable or we can meet 1/2 and do it.... its a fairly simple process with the VMS cable that @sncasner designed vs mine that had to pull the front tire as patch the GPS directly.
Tesla I'm not sure if they are doing it now for no charge as part of the "annual". Before when I found the issue for them, they wanted $500 to do the GPS upgrade.
Let me know when you have some time and I can swing your way and help with the upgrade. Funny I had to do mine AGAIN as I think Carl put in a new replacement GPS after my accident.
Mark77a
Member
Awsome ... Love you guys (from the place where a Johnson is the Prime Minister
)
#1038 had no power down issue, but now says Nov 4, 2000 . . . so I guess it's time to queue up for this. Pretty sure date was good last week, but can't swear to that
If your car is not 2.0, is it 1.5 or 2.5? This problem has been observed primarily in 2.0 and 2.5 cars. When I learned that my 1.5 car has the earlier model of the GPS sensor that can't be updated, I surmised that all 1.5 cars might have the older sensor, but there was a 1.5 in Austria that did need the update. Of course, it's possible that the GPS sensor was replaced in that car with the newer 18x unit before the new firmware was developed. We don't know at what point the transition from older to newer sensors occurred, or even if there was a random mix of them in the bin!
If your GPS sensor does need updating but you can't get to the VMS, you might be able to borrow the cable that @jfischer made that connects directly to the sensor's cable behind the left front wheel.
When the cable makes its way back to me, perhaps we can meet.#1038 had no power down issue, but now says Nov 4, 2000 . . . so I guess it's time to queue up for this. Pretty sure date was good last week, but can't swear to that
i'll take you up on that!
and maybe bribe you for help on that brake-light LED into rear-view mirror mod
Well, last night I joined the club. My Roadster was connected to the wall unit (long since done charging) when there were some brief brownouts and outages for a few seconds, several times. When I disconnected from the Clipper Creek, I got the "bing, bing" and there was fault #3001 again (which no one seems to know the meaning of), and I'm at November 7, 2000. Time of day is correct, though. I've been putting off the annual service due to Covid, and am in no rush. Just curious as to whether anyone else who didn't actually power down deliberately had any power interruption or brownouts, or got any error messages as I did. Hard to believe there's no connection, though coincidences do happen.
I am tempted to ask Tesla for an entirely new GPS instead of updating though, as mine "forgets" where it is and I have to do a reset, (though I only use it on rare occasions anyway).
I am tempted to ask Tesla for an entirely new GPS instead of updating though, as mine "forgets" where it is and I have to do a reset, (though I only use it on rare occasions anyway).
Last edited:
i get line undervoltage & extension cord detected faults, but they're not new (UMC plugged into extension)
The offset is an integral number of days, so the time of day is not affected.
My guess is that the charge cable symptoms are just a coincidence. We've seen cars going in for service and not losing the date, and cars not going in for service that do lose the date. So we have not established a solid explanation.
Car Medlock @CM_007 replaced the GPS sensor in @jfischer's car as part of an accident repair and the new sensor did not come with the updated firmware installed, so asking Tesla for a new one might not fix the problem.
i'll take you up on that!
and maybe bribe you for help on that brake-light LED into rear-view mirror mod
I need to understand this brake light led into rear view mod!
I have a 1.5 so probably have an 80% chance of having the GPS sensor that doesn't have the bug.
I was just going to take advantage of the cable being in LA. But for $300+ it's cheaper to but the alternate sensor (if I ever need it)
Did my GPS update yesterday. I constructed the cable to connect to the unit itself using the parts listed in this thread. I'd also ordered a spare unit in the event I wasn't able to get the update to work/something went wrong, etc (that shipped with the newest firmware, 4.40/20191007). In the interest of saving some other folks some time here are some things I discovered.
TL;DR: Updating at the GPS itself is reasonably easy, I used 4.40 from the website (links that the bottom), its working perfectly fine. Just use a normal RS232 adapter of good quality, it'll all work fine, a standard FTDI serial cable is a solid choice. You don't need to build the adapter cable, .1" IDC (Dupont, Breadboard Jumper Wires) female connectors will fit the pins and clear the socket fine. The devices power fine at 5V0 (which is roughly what I measured at the connector for the car), 12V0 seems fine as well, but I suspect you're working their voltage regulators a bit hard.
I have supplies to build another few cables for this update approach if people are interested; I can assemble them and loan them out. Send me a DM. If you're in the area (all 3 upstate NY owners) we can probably arrange a time for me to help you do it.
Having worked with the AmpSeal connectors before (whats used on the VMS), they're not designed for a lot of cycles and get brittle/fragile as they age. So I opted to do the update at the GPS itself.
Access through the wheel well to the connector was pretty easy. With the cable and info I have now I believe I could do the job in 60-90 minutes without issue. I was able to do it without needing to remove the wheel. I turned the wheel full right lock and reached in between the wheel and fender. Removed the light access hatch. Removed the front fasteners on the arch liner and gently bent it back (soft plastic) to the shock. Then removed the plastic front wheel arch bulkhead (several nuts and bolts).
I've had that area apart a few times before, so its possible my connectors were not exactly in the factory locations. I think my stuff is close to factory locations, as the GPS wiring shares the loom with the headlight connectors.
The signaling is 0-5V (which is fine for low voltage serial). This is noted in the data sheet for the unit on page 12:
"The GPS 18x LVC/18x-5Hz interfaces to a serial port. The unit accepts TIA-232-F (RS-232) level inputs and transmits voltage levels from ground to the input voltage, TIA-232-F (RS-232) polarity."
The key language here is that it accepts TIA-232-F (standard serial, in both directions). Given that the unit is designed for low voltage applications, I wanted to see if I could get it to work with a TTL serial adapter. You can, however, because of it being TIA-232-F compatible the signaling in both directions is effectively inverted. As observed by me on my scope and noted in this stack exchange post.
As is noted in that post, if you have a real FTDI based TTL serial adapter you can use FT_PROG on Windows (or one of several Linux/Mac utilities) to reconfigure the device into inverted mode. Doing this I was able to interact with the spare unit I purchased without issue. However, this didn't work reliably in the car (dirty contacts? excessive resistance due to cable length? unclear); I'd get partial NEMA sentences, with occasional garage data in them. Giving up and swapping to a normal serial adapter resolved the issue. Some TTL serial adapters are already configured this way (supposedly, the "DSD TECH SH-U09C5 USB to TTL UART Converter Cable with FTDI Chip Support 5V 3.3V 2.5V 1.8V TTL" is; though I've not validated this).
Assuming your serial cable and power supply have a common ground, you don't need the data ground pin (they're commoned at the device), so you really only need 4 wires (VCC, GND, TXD, RXD). As the unit will run at 5V0 with quite low current requirements, tapping from USB is fine.
Yes, most of the other cables people have made used full voltage serial, so I should have started there, however I wanted to better understand why that choice was made...
As others have observed, you'll want the SNSRXCFG software and the latest firmware. I found the SNSRXCFG software worked better under Windows 10 if run as admin. The update instructions are more or less correct, though I did find under Win10 that I needed to run SNSRXCFG as admin (the Windows install/update detection compatibility laye, will attempt to UAC elevate the updater and fail otherwise).
Thanks for the help! Let me know if folks have questions.
TL;DR: Updating at the GPS itself is reasonably easy, I used 4.40 from the website (links that the bottom), its working perfectly fine. Just use a normal RS232 adapter of good quality, it'll all work fine, a standard FTDI serial cable is a solid choice. You don't need to build the adapter cable, .1" IDC (Dupont, Breadboard Jumper Wires) female connectors will fit the pins and clear the socket fine. The devices power fine at 5V0 (which is roughly what I measured at the connector for the car), 12V0 seems fine as well, but I suspect you're working their voltage regulators a bit hard.
I have supplies to build another few cables for this update approach if people are interested; I can assemble them and loan them out. Send me a DM. If you're in the area (all 3 upstate NY owners) we can probably arrange a time for me to help you do it.
Having worked with the AmpSeal connectors before (whats used on the VMS), they're not designed for a lot of cycles and get brittle/fragile as they age. So I opted to do the update at the GPS itself.
Access through the wheel well to the connector was pretty easy. With the cable and info I have now I believe I could do the job in 60-90 minutes without issue. I was able to do it without needing to remove the wheel. I turned the wheel full right lock and reached in between the wheel and fender. Removed the light access hatch. Removed the front fasteners on the arch liner and gently bent it back (soft plastic) to the shock. Then removed the plastic front wheel arch bulkhead (several nuts and bolts).
I've had that area apart a few times before, so its possible my connectors were not exactly in the factory locations. I think my stuff is close to factory locations, as the GPS wiring shares the loom with the headlight connectors.
The signaling is 0-5V (which is fine for low voltage serial). This is noted in the data sheet for the unit on page 12:
"The GPS 18x LVC/18x-5Hz interfaces to a serial port. The unit accepts TIA-232-F (RS-232) level inputs and transmits voltage levels from ground to the input voltage, TIA-232-F (RS-232) polarity."
The key language here is that it accepts TIA-232-F (standard serial, in both directions). Given that the unit is designed for low voltage applications, I wanted to see if I could get it to work with a TTL serial adapter. You can, however, because of it being TIA-232-F compatible the signaling in both directions is effectively inverted. As observed by me on my scope and noted in this stack exchange post.
As is noted in that post, if you have a real FTDI based TTL serial adapter you can use FT_PROG on Windows (or one of several Linux/Mac utilities) to reconfigure the device into inverted mode. Doing this I was able to interact with the spare unit I purchased without issue. However, this didn't work reliably in the car (dirty contacts? excessive resistance due to cable length? unclear); I'd get partial NEMA sentences, with occasional garage data in them. Giving up and swapping to a normal serial adapter resolved the issue. Some TTL serial adapters are already configured this way (supposedly, the "DSD TECH SH-U09C5 USB to TTL UART Converter Cable with FTDI Chip Support 5V 3.3V 2.5V 1.8V TTL" is; though I've not validated this).
Assuming your serial cable and power supply have a common ground, you don't need the data ground pin (they're commoned at the device), so you really only need 4 wires (VCC, GND, TXD, RXD). As the unit will run at 5V0 with quite low current requirements, tapping from USB is fine.
Yes, most of the other cables people have made used full voltage serial, so I should have started there, however I wanted to better understand why that choice was made...
As others have observed, you'll want the SNSRXCFG software and the latest firmware. I found the SNSRXCFG software worked better under Windows 10 if run as admin. The update instructions are more or less correct, though I did find under Win10 that I needed to run SNSRXCFG as admin (the Windows install/update detection compatibility laye, will attempt to UAC elevate the updater and fail otherwise).
Thanks for the help! Let me know if folks have questions.
Last edited:
Good for you. The more the merrier.
I don't have any other experience with the AMPSEAL connectors, but it seemed pretty solid to me. The pins are solid and thicker diameter than many other connectors. I acknowledge that the free sample I received had one corner broken off, which seemed strange.
That seemed like more trouble than I wanted to undertake, which is why I pursued the connection at the VMS instead. Also I was concerned about the possibility that the GPS connector in that unfriendly environment might fail in the future if I fiddled with it. But going in at the GPS connector instead of the VMS connector avoids having to inhibit the APS and possibly encountering a temporary but frightening comm failure message.
I observed the same signaling with my 'scope looking at the signals at the VMS connector by sticking a sewing pin into the wires before unplugging the cable. The first version of my cable was made with the FTDI adapter and inverted mode. The big advantage of the adapter that it was terminated in a header shell and included a +5V pin.
The second person I loaned my cable, with a 2.x car (mine is 1.5), swore to me that the signaling was bipolar. He ran into all kinds of trouble, exacerbated by freezing temperatures in the garage, and for a while it appeared that the serial adapter was shorting the +5V from the USB port. So I gave up and went with a normal serial adapter with bipolar signaling, but had to hack away the molded plug to get access to the USB power.
Aha, I should add that to the notes I prepared.