[tervic]

Thanks for your input. I will try those delay tests and see what happens.

I looked at that data after I posted it and wondered the same thing - looks like an IMA start. Sure felt like a backup starter deal, but I have been mistaken a time or two (or maybe lots more!) in my life, so....

I'll perform the tests above and report back.

 

[mudder]

Make sure you remove the backup starter relay before further testing... that way you'll know 100% that the backup starter can't work. Keep the relay handy so you can reinstall it if you end up getting IMA-start stranded on the road.

 

[tervic]

Well, I was out running tests earlier, so I did these with the backup starter relay enabled.
100ms, 130ms, 150ms, 180ms and 190ms delays I observed 10 out 10 attempts do an IMA start.
Then I got impatient and jumped to a 220ms delay and recorded 100% backup starts.
At a 195ms delay I observed 20% backup starts, 197ms went to 80% backup starter and 200ms resulted in 100% backup starts.

I realized another data point too. If I got a backup start, shut car off, let it sit for just a short time, say 5 seconds or less (before LiBCM had completely shut down) and tried another start, it would IMA start every time. Let LiBCM reset and if the delay was too long it would perform a backup start.

As I write this I realize this makes sense, if LiBCM hasn't reset, it won't do the beginning loop that includes the delay.

The backup starts would result in an IMA light and check engine light. The code I read was P1648.

I did not see an IMA assist or regen disable while performing these tests.

If I disable the backup starter relay and set a deliberate too long delay, would the car still start on the backup starter?
If not, how would testing various delays without the relay work?

 

[mudder]

Well, I was out running tests earlier, so I did these with the backup starter relay enabled.
100ms, 130ms, 150ms, 180ms and 190ms delays I observed 10 out 10 attempts do an IMA start.
Then I got impatient and jumped to a 220ms delay and recorded 100% backup starts.
At a 195ms delay I observed 20% backup starts, 197ms went to 80% backup starter and 200ms resulted in 100% backup starts.

Excellent data.
It looks like LiBCM has ~196 ms additional headroom in your car, which is certainly enough time for LiBCM to start up... LiBCM takes about 20x less time to startup.

I realized another data point too. If I got a backup start, shut car off, let it sit for just a short time, say 5 seconds or less (before LiBCM had completely shut down) and tried another start, it would IMA start every time. Let LiBCM reset and if the delay was too long it would perform a backup start.

As I write this I realize this makes sense, if LiBCM hasn't reset, it won't do the beginning loop that includes the delay.

Correct: LiBCM is only going to add this delay immediately after the key first turns on. Due to how the OEM wire harness is configured, LiBCM will continue to 'see' that the key is on until the MCM turns the 12V_IMA relay off.

The backup starts would result in an IMA light and check engine light. The code I read was P1648.

That is expected behavior if LiBCM takes too long to send data (e.g. with a 220 ms delay).

I did not see an IMA assist or regen disable while performing these tests.

Are you saying assist and regen still worked even after getting P1648? I wouldn't expect that, but it's good to know.

If I disable the backup starter relay and set a deliberate too long delay, would the car still start on the backup starter?
If not, how would testing various delays without the relay work?

The car certainly cannot use the backup starter with the backup starter relay removed.

With the data you gathered above, I don't see any reason to continue further artifical delay testing... set they delay back to zero ($KEYms=0).

The reason I had proposed removing the backup starter relay is to make it more clear when the IMA system fails to start the car. That is still my recommendation: keep the backup starter relay removed for the next couple weeks so you know for sure that the IMA system isn't starting the car.

...

Also, thanks for your help troubleshooting this issue! Is anyone else seeing this behavior? I think I might have once in the past.

 

[mmdepace]

Happy to report that I landed in Phoenix after 9 days and 10 hours of LiBCM idle in airport parking at 39% SOC. I believe I parked at around 50% SOC. Not very scientific, but nice to know there is nothing to worry about, just as expected.

 

[tervic]

Hi Mudder,

Another delayed backup start last night. Data is attached. I notice approx 3.5 seconds there appears to be a 4 amp draw? That doesn't seem near enough to start motor, or be an attempted IMA start. I have not yet removed the starter delay relay.

Regen and Assist did not show disabled after this.

No IMA light (which illuminated the other evening when the delay was set too long).

I did reset the delay back to 0ms after testing the other day.

PS I'm going to go remove the relay.

 

[mmdepace]

I have just uploaded firmware 0.7.5alpha

Decided to experiment by raising #define CELL_VMAX_GRIDCHARGER by 90mV as I will take it out of gear to prevent regen after my first few pulls of strong assist.

Am going to continue to use #define ONLY_BALANCE_CELLS_WHEN_GRID_CHARGER_PLUGGED_IN as it has worked just fine so far.

Very pleased to see #define WHEN_GRID_CHARGING_COOL_PACK_ABOVE_TEMP 30. I think this may come in to play in the next couple weeks, weather permitting.

 

[Balto]

Happy to report that I landed in Phoenix after 9 days and 10 hours of LiBCM idle in airport parking at 39% SOC. I believe I parked at around 50% SOC. Not very scientific, but nice to know there is nothing to worry about, just as expected.

Yea I recently didn't drive my Insight for 12 days(got one of my projects back on the road and was busy torturing the neighborhood in it). I parked mine at around 70% and it was at 42% 12 days later so not too bad. Also no idea what firmware I have 🤷‍♂️ probably 0.7.3 or something

 

[LowCarbon]

I just built my LiBCM pack. It's sitting on the bench grid charging. When I first plugged it in, about an hour ago, the highest cell was 3.831 and the lowest was 3.563. Now the highest is at 3.858 but the lowest hasn't moved. It's still at 3.563. Is this cause for alarm? Or am I just being impatient?

Never mind. That low voltage is the lowest since power up. Naturally it's not going to change. Hopefully, everything's working as designed. I'll let it charge overnight and put it back in the car and see how it works.

 

[mmdepace]

Balto, I'm remembering back to a couple of weeks ago when upgrading to 0.7.4, there was going to be a system idle power savings benefit in that release. (I think) I suppose you are on 0.7.3 as it seems your idle consumption was a little higher than mine while running 0.7.4. Glad to hear you got another on back on the road. Haven't seen a fellow vehicle out here in the Phoenix area for a couple years now.

 

[Balto]

Balto, I'm remembering back to a couple of weeks ago when upgrading to 0.7.4, there was going to be a system idle power savings benefit in that release. (I think) I suppose you are on 0.7.3 as it seems your idle consumption was a little higher than mine while running 0.7.4. Glad to hear you got another on back on the road. Haven't seen a fellow vehicle out here in the Phoenix area for a couple years now.

Oh no, what I put back on the road was a S80 T6 AWD with the exhaust removed. That's why I have been torturing the neighborhood.

I am however, standing next to another one I just picked up about 30m ago that I'm driving back 8 hours to save.

 

[mudder]

Yea I recently didn't drive my Insight for 12 days(got one of my projects back on the road and was busy torturing the neighborhood in it). I parked mine at around 70% and it was at 42% 12 days later so not too bad. Also no idea what firmware I have 🤷‍♂️ probably 0.7.3 or something

That consumption seems a bit high, but if you're running the latest alpha code, then it's possible the fans came on to cool the pack at some point. Let me know which version you were running while parked for 12 days so I can see if I need to look into it.

Also, as @mmdepace reminded me, v0.7.0 had much higher power consumption. This (intentionally introduced) 'issue' was fixed in v0.7.1.

 

[Balto]

Just got home. Over 1,000 miles driven, but totally worth it. I'll check for you Mudder whenever I wake up. Lol.

 

[mmdepace]

That consumption seems a bit high, but if you're running the latest alpha code, then it's possible the fans came on to cool the pack at some point.

Mudder, I will monitor the possibility of fan on with key off on 0.7.5 today. It'll be 101 degrees today and I will not be driving the car until Monday. I'll pop outside a few times to monitor Voltage and SOC.

 

[Balto]

That consumption seems a bit high, but if you're running the latest alpha code, then it's possible the fans came on to cool the pack at some point. Let me know which version you were running while parked for 12 days so I can see if I need to look into it.

Also, as @mmdepace reminded me, v0.7.0 had much higher power consumption. This (intentionally introduced) 'issue' was fixed in v0.7.1.

V0.7.3.

 

[mudder]

 

[tervic]

Hi Mudder,

A no start this evening, back up starter relay removed. No crank, no IMA light and no limiting on assist or regen.
Data dump attached. Let me know if there is other testing you'd like performed.

Edit: There appears to be a single pulse of assist, 9 amp, 1.6 KW if I'm reading this correctly. I did not observe or hear any sort of engine rotation or attempt to start.

 

[mudder]

Based on the data you've posted, LiBCM shouldn't be doing anything to prevent an IMA start.
-LiBCM is meeting timing during keyON IMA start
-Cell voltage is fine during keyON IMA start (or rather attempted start)
-Battery is charged enough.
-Temperature is high enough.
-The IMA system (ECM+MCM+LiBCM) is at least nominally attempting an IMA start.

...

So if it's not LiBCM, then it's either the MCM, the ECM, or one of the wires between the three computers (or IMA motor).

Next troubleshooting steps:
A: Can you upload a video of you starting the car? Ideally showing the laptop screen you're using to gather data with. I want to see if there are any audible inticators.

B: While gathering data in the future, before starting the car, type $DISP=SCI
This will make LiBCM output the BATTSCI and METSCI data in realtime. Please post that data next time the IMA system won't start the car.

C: If you have an OBDIIC&C:
-are there any silent P-codes (i.e. that don't set the CEL)?
-does LiBCM's spoofed pack voltage (the one in parenthesis, e.g. "xxx(172)" ) match OBDIIC&C's reported pack voltage (e.g. 172 volts)? I think I might have already asked you this, but it all blends together.

D: If you have an oscilloscope, can you probe the following signals during keyON start:
CMDPWR (MCM:C02 - BLU/BLK): 0:5 volt PWM @ 2 kHz
MAMOD2 (MCM:A03 - BLK/YEL): 0:5 volt, L=Assist, H=Regen/standby
MAMOD1 (MCM:C12 - RED/YEL): 0:5 volt, PWM @ 20 kHz
QBATT (MCM: D13 - PNK/PNK): 0:5 volt analog output representing battery SoC (to ECM)

E: If possible, keep the Backup Starter Relay removed (unless you need to reinstall it because you get stranded).

...

It is quite strange that only your car is experiencing this issue... if anybody else is seeing this behavior, please let us know. My hunch is that there's some other issue with your IMA system (that for some reason didn't cause problems with the stock BCM).

I'm certain we can get to the bottom of this once we have all the data. Colorado is too far to drive from Tennessee, so if we can't figure this out remotely, then my next suggestion will be to send you a RevD PCB (next month). Another option is I could send you my development RevC PCB right now.

I'm fine with either option... I just want to inconvenience you as little as possible. I'm sure it's frustrating having a car that won't start correctly half the time. Let me know if/when you're not having fun anymore gathering data. And as always, thanks for providing feedback and data during the beta period.

 

[thibble]

I've been quiet because of work travel. Back in US and just flashed latest pre-release (May7 I think?). Car loves it, no tweaking at all, in with config.h defaults. The 40% current hack isn't tossing a code under full throttle (it was with April 8th and OBDII/IMA C&C). During the down time it did sit on April 8 release with plug in for several weeks happily.

 

[mmdepace]

Watching this with interest due to my years of an ongoing, non-LIBCM related intermittent no-start issue. Occurs very often when its below 45F - almost every morning, but also sometimes when its hot out - once last week. Rarely in the middle. Always only after the car has been sitting for many hours. Brand new 12V batteries and old. My issuse kills both IMA start and 12V start. 12V starter works only with IMA switch off. Up to 5 minutes of frustrating key on and off events before car decides to fire up using IMA. A 12V start would be a treat in these moments. Sometimes pulling IMA fuse under dashboard then replacing lets me start the car. No responses to this are necessary, as I am going to use some of the recommendations above next time I attempt testing on the issue.