Caution - Annual Service - Car power-down - GPS issue

[slcasner]

Now there is yet another update to the Garmin firmware:

Change History
Changes made from version 4.30 to 4.40:

• Fixed corner case GPS week number rollover for devices using older hardware.

 

[slcasner]

GPS antenna end connector is FHC95
1 Black
2 Red
3 Black
4 Green
5 White
6 n/c

I see VMS connector FHC46 has
1 n/c
2 n/c
3 n/c
4 n/c
5 n/c
6 Red
7 White
8 Green
9 Black
10 n/c

No idea if they are connected, but it seems to match.

Nope. FHC46 goes to the USB connector FHC67. The wires from FHC95 feed into one of the big VMS connectors, FHC47.

 

[gregd]

Maybe because 2.0 has a 12V battery, whereas 1.5 would lose 12V power if the traction battery is disconnected?

Could be The 12v is generated within the ESS comes out on connectors separate from the high voltage. If they go through the PEM, then you are correct. But if they bypass the PEM, then it's possible the GPS could still be powered. I have no clue how it's wired.

 

[slcasner]

I may have to cobble together my own cable (perhaps just with clip leads) before Monday. I have not figured out yet, though, how to do the update from a Mac laptop.

OK, I've cobbled up the connection to the VMS connector and hooked up to a Windows desktop that I moved out to the car. While following the procedure on the download page, I've run into two snags: The SNSRXCFG utility is used to select the update file and then it is supposed to invoke UPDATER.EXE, but the latter does not start. I think initially it was because the Windows protection software asks if I want to allow UPDATER to run, but even after I changed the protection level so Windows won't ask, it still does not start UPDATER. I can run UPDATER then manually, but after connecting to the GPS it says there is no appropriate software for my unit. I suspect that is because it needs the filename passed from SNSRXCFG. Anyone get past this problem?

 

[slcasner]

At startup, my unit reports the following version information, note no 'x' after 18:

$PGRMT,GPS 18LVC - software ver. 3.00,P,P,R,R,P,,25,R*31

Does that mean my early 1.5 has an older GPS unit for which the GPS 18x PC/LVC software version 4.40 is not applicable? Do newer cars report 18x?

--Update--

I was able to compare the technical spec documents for the 18 and 18x products and these indicate that the PGRMT sentence would indicate 18 or 18x, respectively. If Garmin won't provide updated software for the 18 unit then it might be necessary to replace it with an 18x unit in order to operate correctly.

 

[slcasner]

The connections at the VMS are:

Before unplugging the connectors from the VMS, I used my oscilloscope to verify the RX and TX signals and baud rate (4800) by stabbing a sewing pin into the wires. This also allowed me to connect the TX signal from the GPS to RX on my USB serial device to verify that it could receive the GPS data stream correctly.

After unplugging the connectors from the VMS, here is a picture of my cobbled-up connections to the GPS through the VMS connector:

I used some brads with a shaft diameter close to the 1.3mm spec for the AMPSEAL connector. The colors of the clip leads match the wire colors in the diagram above, with yellow substituted for white; in the photo the 'scope is connected to the GPS unit's TX signal. I'm using a little 5V power supply designed for breadboards (ignore the orange wire coming out of the bottom of the power supply with tape on the end).

After hooking everything up, the GPS happily transmits its NMEA data stream and the Windows programs are able to connect to the GPS device and change its mode to accept the download, but I'm stuck with the Update.exe utility reporting "No updates are available for your product."

 

[jfischer]

Nicely done! Figured there was a way to get to it from the VMS connector. Anyone have any idea on that actual connector and where to source it to make a connector?

Not sure as I didnt get that same issue. Maybe 2.5 ones have a newer GPS unit. I did it directly on the unit, but your way should make any difference. I also did the 4.20 software update, I see there are 2 more "fixes" since then, which also might make a difference.


I didnt recall the output of the version but I thought i remembered 18x. Might try your attachment on the vms and see if it works from there.



It updated fine to the newer version without any issues.

Upon connection I saw the same NMEA data stream showing the incorrect date date week entry, after update, it showed the correct one.

 

[slcasner]

Nicely done! Figured there was a way to get to it from the VMS connector. Anyone have any idea on that actual connector and where to source it to make a connector?

From this post I learned that the VMS connectors are TE Connectivity AMPSEAL. The only mating connectors available are intended for PCB mounting, but the best choice for making a cable would be part number 1-776200-1, which is HEADER ASSY, 23 POS VERTICAL, Black (keyed to match the black connector on the VMS), Gold contacts. I've requested a sample from TE Connectivity; also available from DigiKey as part number A128214-ND.

Not sure as I didnt get that same issue. Maybe 2.5 ones have a newer GPS unit. I did it directly on the unit, but your way should make any difference. I also did the 4.20 software update, I see there are 2 more "fixes" since then, which also might make a difference.

Did you use the update procedure at the bottom of the download page, that is, to use the SNSRXCFG.exe configuration utility to select the RNG file and invoke the Updater.exe utility to do the installation? I tried both the 4.20 update that I downloaded earlier when you first started this thread as well as the current 4.40 update. Both failed.

 

[slcasner]

I received the sample connector from TE Connectivity already. That was fast. Strangely, it has one corner broken off and not present in the package, but that won't interfere with usage and the sample was free, so OK.

I plan to buy a USB to RS232 converter with 6' cable and cut into the end with the DE9 connector so I can pick up +5V from the USB. Then I can connect GND, +5, RX and TX directly to the AMPSEAL connector to have a self-contained update adapter. I have at least one friend who needs the update, then this could be passed around.

 

[X.l.r.8]

I have a few of these floating around which I assume will do the same job. I have yet to start playing around with mine yet as it works fine and it’s one of the functions I use a lot

 

[slcasner]

Yes, that might be an alternative and wouldn't require hacking. The pictured device transmits TTL level signals, not RS232 level signals. The instructions for connecting to the GPS to update the firmware call for RS232, but the spec for the GPS sensor itself is this: "The unit accepts TIA-232-F (RS-232) level inputs and transmits voltage levels that swing from zero V (ground) to 5 V TIA-232-F (RS-232) polarity." So presumably the unit would also accept zero V to 5 V, which is exactly what the adapter you've pictured does. Thanks for the suggestion, I'll reconsider my plan.

 

[slcasner]

@X.l.r.8, there may be a problem with using an adapter like the one you showed. These are designed for connection to devices like Arduino and Raspberry Pi that expose logic-level UART signals. The polarity of those signals is opposite to the polarity of RS232 signals because the chips that do the level conversion are inverting. The GPS sensor "transmits voltage levels that swing from zero V (ground) to 5 V TIA-232-F (RS-232) polarity." Note the last word. I don't know if the USB-to-TTL devices are available in both polarities.

I've exchanged emails with Garmin product support and received this interesting tidbit: "The GPS 18 OEM was tested and is not impacted by the GPS rollover that took place in April. If there would have been an issue with the rollover, it would have exhibited itself in April."

That would explain why my early car (VIN 33) still showed the correct date Oct 21 when I reconnected the cables to the VMS yesterday morning after failing to update the GPS firmware. It might also explain why @gregd has not seen incorrect dates. But the 2.5 cars definitely have the problem because they have the newer GPS 18x LVC sensors. A log file taken on a friend's 2.5 on 2019-10-12 was dated 2000-02-27.

Since @jfischer did not see the problem before the annual service caused power to the sensor to be interrupted, the statement from Garmin that the problem "would have exhibited itself in April" may be assuming an application where the sensor is turned on and off.

 

[X.l.r.8]

I’m pretty sure mine would have been turned off during the ground fault issue but I thought maybe I was lucky, but it would make sense as to why it’s still behaving fine.

 

[slcasner]

I found this DTECH FTDI USB to TTL Serial 5V Adapter Cable with 6 Pin 0.1 inch Pitch Female Socket Header UART IC FT232RL Chip, similar to the device @X.l.r.8 pictured, that allows the polarity of the UART signals to be configured in EEPROM. That should fit the bill without any surgery.

 

[coolVariable]

I've exchanged emails with Garmin product support and received this interesting tidbit: "The GPS 18 OEM was tested and is not impacted by the GPS rollover that took place in April. If there would have been an issue with the rollover, it would have exhibited itself in April."

That would explain why my early car (VIN 33) still showed the correct date Oct 21 when I reconnected the cables to the VMS yesterday morning after failing to update the GPS firmware. It might also explain why @gregd has not seen incorrect dates. But the 2.5 cars definitely have the problem because they have the newer GPS 18x LVC sensors. A log file taken on a friend's 2.5 on 2019-10-12 was dated 2000-02-27.

Since @jfischer did not see the problem before the annual service caused power to the sensor to be interrupted, the statement from Garmin that the problem "would have exhibited itself in April" may be assuming an application where the sensor is turned on and off.

That is very interesting.
Has anyone with a 1.5 Roadster encountered the GPS issue? Any way to test the version number of the GPS sensor (apart from removing the front light and looking at the GPS antenna)?


Looks like the "18" version may be this one? Garmin: GPS 18 Updates & Downloads

 

[slcasner]

Any way to test the version number of the GPS sensor (apart from removing the front light and looking at the GPS antenna)?

There is a version number in the data stream that read "18" rather than "18x" on my car, but the serial number is not in the data stream and that is what the Garmin support person wanted in order to verify the version.

Looks like the "18" version may be this one? Garmin: GPS 18 Updates & Downloads

Yes, that is how I figured it. As you can see there, the last firmware update was in 2011.

As shown below, I have now built the adapter cable that I proposed to mate with the cable connected to the VMS module in the passenger footwell. I decided to build it into a little box to make it more robust and to give a better handle when mating and separating the connector. I've tested that the cable works with my 1.5 to receive the data stream from the GPS sensor, but I would like to test the full firmware update with a local car first before shipping it around to others of you who want to do the firmware update. PM me if interested in either case.

To update the firmware in the GPS sensor, you need to use a Windows computer and download both the SNSRXCFG configuration utility and the firmware updater. To use this adapter cable, first plug the USB end into the laptop. The USB-to-RS232 adapter is based on the FTDI FT232R chip for which a driver is readily available and may already be installed. (I tested on Windows 10 which automatically downloaded and installed the driver.) Anyone who just wants to check the sensor version number could use a non-Windows laptop and a terminal emulator program (e.g. screen on Mac or Linux) at 4800 baud. But if you do have the newer "18x" sensor, unplugging the cables from the VMS removes power from the sensor and induces the incorrect date problem.

When you're ready to do the firmware update, unplug the large blue and black plugs from the VMS. For a photo, see this post where wiztecy said he unplugged the black one first, then the blue one. However, the schematic indicates that the blue one provides primary power, so I pulled that one first. In this post @markwj advised being very careful not to bend the pins, but the pins are 1.3mm rods like you can see in the photo. You need to be careful not to break the latch, but I don't think there's much risk of bending the pins. It does require quite a bit of force to separate these watertight connectors.

Then connect this adapter cable to the black plug from the VMS. (The plugs are indexed so only the black one will mate with the adapter.) Then you can follow the instructions at the bottom of the firmware download page to connect with the SNSRXCFG utility and run the update. After that, disconnect the adapter from the black plug and plug the black and blue cables back onto the VMS.

 

[jfischer]

Nicely DONE!

I'd like to know the parts you used and the pinouts. I can kinda gather them from your drawing, but would like to confirm.

outstanding work from the VMS connector.... so much easier!

 

[slcasner]

I'd like to know the parts you used and the pinouts. I can kinda gather them from your drawing, but would like to confirm.

I put the references for the two key parts in previous posts, but here they are again:

1. 23-pin AMPSEAL connector 1-776200-1 from TE Connectivity, available from DigiKey as part number A128214-ND; or the 776200-1 variant with tin rather than gold plating is available through Amazon ("usually ships in 3 to 5 weeks").
   
2. DTECH FTDI USB to TTL Serial 5V Adapter Cable with 6 Pin 0.1 inch Pitch Female Socket Header UART IC FT232RL Chip from Amazon.
   
3. Zulkit Waterproof Plastic Project Box ABS IP65 Electronic Junction box Enclosure Black 3.94 x 2.68 x 1.97 inch (100x68x50mm) (Pack of 1) from Amazon.
   
4. A small perf board, strip of header pins, wire, heat shrink tubing, rubber grommet, cable clamp, and screws from my parts collection.

One could cut off the 6-pin socket from the DTECH adapter and solder the wires directly to the pins of the AMPSEAL connector, but I chose to put in a header strip that the socket could plug onto so that I can repurpose the DTECH later when nobody else needs to borrow this adapter I've made. Also, if you look closely at my photo you can see that I've extracted and rearranged the wires in that socket housing. This was to match up with the pins on a little TTL-to-RS232 adapter I have so I could test against other USB serial port adapters that I use. That pin order also happens to be the same as the order of the signals shown on the VMS schematic. For the wires running between the pin header on the perf board and the AMPSEAL connector I've also chosen the same colors as the those of the corresponding wires in the car's harness just to make it easier for me to keep track. If one were making a similar adapter without rearranging the pins in the DTECH adapter socket, the pinout would be as follows:

6P Socket Pin	DTECH Wire Color	Name	Description	Harness Wire Color	AMPSEAL Pin
1	Black	GND	Device ground supply	Black	2
2	Brown	CTS	Clear to Send	*	*
3	Red	VCC	+5V from USB	Red	17
4	Orange	TXD	Transmit Data	Green	16
5	Yellow	RXD	Receive Data	White	9
6	Green	RTS	Request to Send	*	*

* No connection

The polarity of the RXD and TXD signals in the DTECH adapter must be changed to be inverted from the default polarity as delivered. To do this, on a Windows laptop download the FT_PROG EEPROM programming utility for the FTDI chip and its User Guide. With the DTECH adapter plugged in to a USB port, start with DEVICES > Scan and Parse, then when the Device Tree is populated, navigate to Hardware Specific > Invert RS232 Signals where the Invert TXD and Invert RXD checkboxes can be selected. Then DEVICES > Program will update the chip so it's ready for use.

Having said all this, my intent was to share the adapter I've made with all who need it.

And finally, based on my experiences related in the Communication Problem - need reboot? thread, it might be better to unplug the black connector first, then the blue one, and reconnect blue then black. Or it might be sufficient to just unplug the black one, hook up the adapter to do the firmware update, and then plug the black one back into the VMS. That would be a plus because those connectors are hard to remove.

 

[slcasner]

Moderator: If you know how to eliminate the big blank space above the BBCODE table in my post, please edit.

 

[jfischer]

Those statements are correct with a 2.5 Not sure if power on the 1.x units come from the VMS but I assume so.

one note, its the VMS computer not the VDS getting powered off as you eluded to in the first 2 questions.

1) on 2.x and 2.5 if you pull power on your VMS computer control ( under passenger foot well) then it will reset the GPS unit.

2. You need to pull the plugs ( black and blue) on your VMS and plug it back in to test if you have this issue.

3. and YES, If the date/time is wrong on your VDS after power cycle, then you have this issue.