Let's save this only Roadster left in the country

[emir-t]

Hey folks;

Long story short I have the only Roadster left in our country. 2010 model year EU VIN278 around 60k miles. It has not worked snce the owner washed the car in May 2018, had a little drive and parked the vehicle. Thankfully Gruber has been helping me for the past 4-5 months over e-mails but I seldom get a reply, maybe once a month at best, as they're super busy with their projects.

We repair Model S&X vehicles so I have little to no experience on the Roadster. Thanks to online resources I've been reading a lot and learning but information on Roadster is still quite limited. I figured maybe you guys could help me with this battle.

Problem is that the contactors do not close. So ESS is isolated and does not power APS. Hence to power the vehicle we jump it from the APS posts on rear right. When jumped the screen turns back on, even the IP needles jump when ignition is on but the car of course does not shift. Shows the fault lamp and regen disabled lamp alongside a 110 mile range. The error messages are;

- ID265 BSM: V_batt too low
- ID264: BSM: PEM cable fault
- ID50: BMB: No data fault
- ID258: APS output 2 (PEM) reset
- ID1552 VMS: Not receiving BMB version number (yellow warning)
- ID53 GPS: No data fault(yellow warning)
- ID1561: VMS: Not receiving CPLD version number(yellow warning)
- ID1556: VMS Not receiving PM version number(yellow warning)
- ID43 VMS: APS inhibited due to low brick voltage
- ID885: BMB No Data Fault can't start
- ID888: VMS/PEM key mismatch

Actions taken so far;
- My first contact was Gruber and they told me it was a communications issue. We did check every connector outside of ESS and indeed did find corrosion with few. However that didn't solve it.
- Then we took a buzzer and measured every cable in all harnesses to check if there was a problem. Found that one blue wire wasn't passing through from front to mid harness connector. Pin #15, that was ESS can low as per Gruber. Made another twisted pair wire coupled with pin #14 red as CAN high of ESS, yet it wasn't solved again. Just when I thought we were so close.
- As per Gruber we hooked up an oscilloscope to those wires and saw that CAN bus was functioning properly. We saw square waves. Low was between 0-2.5V and high was between 2.5-4V. The signals indeed were can messages. Resistance between can high and low was fluctuating between 116-132ohm in an energized state so 120ohm should be the resistance when powered off.

Now tomorrow I'll measure all 99 of the bricks from BMBs with probes and see if there is a faulty brick/module indeed. Possibilities in my mind are;

1) There is a dead module/brick. We need to remove and charge it to hopefully recover it, or replace it.
2) I read a thread here about something similar. There was a fuse blown in the front 400VDC controller that blew up the ESS fuse. Worth checking as well after the brick measurements.
3) Some kind of ROM went bad and it needs firmware flashing? (Hope not since all is Tesla proprietary and is impossible to take it to an official Tesla service especially during COVID19)
4) There indeed is a CAN issue, just not with the canbus but with one of the nodes.

I would appreciate your feedbacks and tips. It would be epic for the only Roadster left in our country to be back on the roads. I'm sending a photo of the firmware screen from service menu in case it means something. BMB not having a firmware really does mean communications issue but I did saw the bus function properly so I'm lost. Hopefully dead modules or blown fuses in ESS also somehow affect the comms.

Best;

 

[jason.tv]

when you washed your car seems like you caused the 400v controller to fry. Could of got your 400v controller fixed or upgraded from GS Technology and see if that was the issue before more fire happened.

The 400v controller communicates with the PEM; and if it's blown or not working then it would or should typically cause these error as there is no communication to the PEM.

Once you get the 400v controller working again you can diagnose further problems from there; but if your 400v controller fuse is blown or not working like you said then this must be fixed prior to anything else as a non working controller will cause the car not to charge. Tesla can replace this controller for you I think it's average $1400-$1600

When you say the Problem is that the contactors do not close, do you physically see the contactors or is this an error on the screen?

Have you taken the car to Tesla Service? If I am incorrect or provide wrong information above I do apologize; it's just my best interest or thought process for this situation; as far as the battery I have no clue; if it's dead GS-Tech can prob fix that too.

 

[emir-t]

Thanks for the swift response.

I'll check the inside of the 400V controller first thing in the morning. Is it true that if fuse in it is fried then fuse inside ESS is also fried? Hopefully changing fuses in both fixes it. If not hopefully I can find a spare 400V controller. Still worth checking the brick voltages to make sure pack is fine though IMHO.

Also, is ESS removed via a lift from below of the car like the Model S? In case I need to open up the ESS for the main fuse inside.

 

[ML Auto]

If it has been sitting for 2 1/2 years the ESS is dead. The 110 mile range shown now is when the car was last turned on while the ESS still had enough voltage-it doesn't update unless the car is functional. Check the brick voltages.

 

[emir-t]

If it has been sitting for 2 1/2 years the ESS is dead. The 110 mile range shown now is when the car was last turned on while the ESS still had enough voltage-it doesn't update unless the car is functional. Check the brick voltages.

I will but I've seen Model S packs when left isolated (i.e. contactors open all the time) don't have such self discharger to brick them. So still hopeful. When you consider it was 78% SoC when dead so around 3.9V-4V per cell I want to believe there is hope. If not I can always revive it. Let's just hope 400V controller just has a fuse blown and doesn't need special attention to change or whatnot.

 

[markwj]

I will but I've seen Model S packs when left isolated (i.e. contactors open all the time) don't have such self discharger to brick them. So still hopeful. When you consider it was 78% SoC when dead so around 3.9V-4V per cell I want to believe there is hope. If not I can always revive it. Let's just hope 400V controller just has a fuse blown and doesn't need special attention to change or whatnot.

Design of the roadster is different than Model S. There is significant parasitic drain from the electronic components (most notably the VMS, and coolant pump). A fully charged roadster can drain to zero in a month or so, unless the service disconnect is pulled or APS inhibited.

I concur with ML Auto's suggestion that the first thing to check is the sheet voltages. No point doing anything else until you know the ESS is good.

I also suggest you plug in a CAN bus logger into the diagnostic connector in the front passenger footwell, and check the three main CAN buses for valid traffic.

• CAN1: 1Mbps (instrumentation)
• CAN2: 125Kbps (HVAC, ESS)
• CAN3: 500Kbps (PEM)
• CAN4: 1Mbps (diagnostics)

CAN1, CAN2, and CAN3 should all have visible traffic shown on the diagnostic connector in the passenger footwell.

Regarding your reference to pin #14 and #15 being ESS CAN, which connector are you talking about? At the VMS end, HVAC/ESS CAN is on J2 pins #19 (high) and #4 (low).

 

[emir-t]

Design of the roadster is different than Model S. There is significant parasitic drain from the electronic components (most notably the VMS, and coolant pump). A fully charged roadster can drain to zero in a month or so, unless the service disconnect is pulled or APS inhibited.

I concur with ML Auto's suggestion that the first thing to check is the sheet voltages. No point doing anything else until you know the ESS is good.

I also suggest you plug in a CAN bus logger into the diagnostic connector in the front passenger footwell, and check the three main CAN buses for valid traffic.

• CAN1: 1Mbps (instrumentation)
• CAN2: 125Kbps (HVAC, ESS)
• CAN3: 500Kbps (PEM)
• CAN4: 1Mbps (diagnostics)

CAN1, CAN2, and CAN3 should all have visible traffic shown on the diagnostic connector in the passenger footwell.

Regarding your reference to pin #14 and #15 being ESS CAN, which connector are you talking about? At the VMS end, HVAC/ESS CAN is on J2 pins #19 (high) and #4 (low).

thank you for the response. Connector I’m referring to is the mid to back harness on the right side of the car. A circular connector going in from A pillar bottom right hand side to rear.

I will check sheet voltages today and let you guys know. The design being different than Model S is bad news now. I thought because the car sort of “self inhibited” because of the error I thought there wouldn’t be self discharge.

 

[emir-t]

Bad news. ESS is absolutely dead. @markwj and @ML Auto was right. I don't think this is recoverable. I don't have a lot of experience with batteries but these voltages are absolutely bricked right?

I wish Roadster ESS was like the Model S i.e. contactors absolutely open in case of fault. Either they were not and there was a load on the battery over the years (31 months exactly since it was last charged to 77% SoC) or these self discharge too much compared to Model S. Here are the measurements taken over BMB board connectors.

Also checked the 400V controller fuse and it was fine. No sign of water ingress either. It might be broken though, I'm not sure how to find out it was. @jason.tv Probably 400V controller was the initial issue but ESS is now a bigger issue. Any suggestions?

 

[ML Auto]

Wow, those reading are pretty bad. Worse than ones I've seen from cars that sat over 3 years.

My suggestions are:

1)Find a used ESS

2)Wait for the 3.x battery upgrade from Tesla

3)Sell the car for parts or part it out.

4)Install ICE motor.

 

[gregd]

Long shot, but I'd double check your readings... That they are all down to sub-1v, and pretty evenly so, is kind of surprising. I didn't think lithium-based batteries would self discharge that far, that fast, and anything actively draining them (e.g. the APS) should have dropped out long ago. You might check with Carl Medlock or Peter Gruber to verify your measurement process.

If you power the car from an external 12v source, does the VDS / VMS allow an active reading of the voltages?

Failing that, I agree with @ML Auto , at least #1, and #2. #3 would be a shame, and counter to your desires. Or, you might link up with @ElectricLove and his planned battery retrofit. We're all heading there, at some point in the future.

 

[emir-t]

The fact that they would be almost super healthy if multiplied by 100 is super suspicious. I also don't know if they would discharge so evenly below 1V level. Also having e-mails back and forth with Gruber at this point. I did measure with a simple multimeter set to measure 20VDC. Readings would show as ",XXXV". Like ,354V for brick 9 or ,036 for brick 1. When I took the multimeter setting to 200VDC it would show 0.3 and 0.03 respectively. So they do sound correct but the measurements' pattern does seem suspicious.

 

[ML Auto]

You might be off by one decimal point, but it would still be toast. You are taking readings with the BMB boards removed, correct? You didnt specify which BMB connector you are reading at. Just taking off the top connector and checking at the BMB board wont give you the correct voltage.

 

[chrisro]

The battery is dead, my solution will come spring next year with model S cells. Around 30000eur.

 

[X.l.r.8]

You need to make a board to bypass the BMB before you take any more readings. Right now you are spinning in circles with information which is not conducive to get any advice from. Sorry if that sounds harsh but the assumption at this point is the battery is dead abs rather than. Prove it everyone is giving advice, I thought collectively we would be smarter than this. Yes it probably is unrecoverable, there’s no denying that but for the archives of people reading this it would be helpful to show emir-t how to actually check properly.

something like this will allow you to measure the voltages rather than the BMB. I believe electriclove went a stage further and took the sensible step of adding in-line fuses to enable safe balancing/charging of bricks in the sheets.

 

[emir-t]

Thank you for the explanation. I also saw this board and its github page while researching how to solve this. Yet where I am located it is not easy to print circuit boards nor it is easy to have them shipped out. So to accelerate the process I just taped wires on a long wooden stick used to grill meat, (sish kebaps ironically) removed BMBs and probed the same connector where you would stick this custom PCB in. Pinouts were already there in github pcb info so #24 being the DC negative I probed each 9 positive pins of each brick. Repeated it for 11 sheets for a total of 99 measurements. Is this technically giving me different measurements than this PCB? I doubt it but I'm no expert.

 

[ML Auto]

Thank you for the explanation. I also saw this board and its github page while researching how to solve this. Yet where I am located it is not easy to print circuit boards nor it is easy to have them shipped out. So to accelerate the process I just taped wires on a long wooden stick used to grill meat, (sish kebaps ironically) removed BMBs and probed the same connector where you would stick this custom PCB in. Pinouts were already there in github pcb info so #24 being the DC negative I probed each 9 positive pins of each brick. Repeated it for 11 sheets for a total of 99 measurements. Is this technically giving me different measurements than this PCB? I doubt it but I'm no expert.

I'm not sure you were doing that correctly. There are 10 wires connected to each end of the 9 bricks, the rest are the temp sensors. Testing between two wires gives you one brick voltage, if they are connected to the same brick. If you use one "end" wire as a common, the voltage will increase as you progress towards the other "end". Your voltages dont increase, so I'm not sure what you are reading.

 

[emir-t]

I'm not sure you were doing that correctly. There are 10 wires connected to each end of the 9 bricks, the rest are the temp sensors. Testing between two wires gives you one brick voltage, if they are connected to the same brick. If you use one "end" wire as a common, the voltage will increase as you progress towards the other "end". Your voltages dont increase, so I'm not sure what you are reading.

Quite the contrary they're increasing if you look carefully. They're just so low you actually need to do the math to notice.

 

[ML Auto]

Quite the contrary they're increasing if you look carefully. They're just so low you actually need to do the math to notice.

You are correct. I was looking up and down instead of left to right. So yes, your ESS is junk. Way beyond recovery.