Honda Insight Forum banner
861 - 880 of 945 Posts

·
Linsight Designer
Joined
·
4,144 Posts
Discussion Starter · #861 ·
60S is working:

As I mentioned in the video, I got a couple P1439 codes, which indicate IGBT short circuit (nothing to do with LiBCM). @retepsnikrep has already solved that problem, so I'll need to look into his solution and implement it for 60S LiBCM. Right now I'm off to bed.
 

·
Premium Member
Joined
·
1,373 Posts
With the 60S ( and maybe the 48S), it looks like there is no room for a large battery cooling fan, like with the OEM or EHW5 batteries. Are battery temps going to be a concern?
 

·
Administrator
Joined
·
13,935 Posts
@mudder P1439. I think you have a few options.

1) Cut the FSC pin in the MCM and connect the MCM pcb pin end to 5V sourced from inside MCM. (This is what I do on all my current hacked MCM's) I suggest try this first it's very simple.


This prevents the FSC fail signal setting an IMA code and forcing a reset. The assist / regen will stutter momentarily as the IGBT internal protection shuts it off. This never bothered me, but you might be pushing so high now we need another option..

2) Limit the current/power level to prevent the low rpm overload.
Once the motor rpm gets above about 2000 rpm this problem goes away IIRC. I seem to remember it was due to the different igbt drive signals at low rpm, square v sine?

It can also occur with the engine IMA start cycle as that is very demanding on the igbt. High power into a stopped motor!
Low rpm high current regen was also an issue..

Other solutions/Ideas

a) Don't allow full assist or regen requests when the rpm is below 2000rpm. Use your MIMA and make it aware of engine rpm, then just limit cmdpwr PWM power levels below 2000 rpm sufficiently to prevent the code appearing. Once you get above 2000 rpm then it's ramp up or pedal to the metal and 30kw+ LOL

b) Reduce the current hack level to say 0% or 20% below 2000 rpm. This needs the controllable current hack of course and rpm awareness.

If you do the MCM FSC mod (1 above) you could use the redundant fail signal coming from the MDM to light a dash led or give some LIBCM LCD indication (other than the seat of your pants motor stutter) that the IGBT is overloaded.

I also suggest running the MDM fan on low at all times as a minimum at such high power levels. The MCM doesn't turn the MDM fan on until too late IMO. No point getting a heat cycling stress related IGBT internal temp sensor failure that renders it permanently US.

PS Well done on all the continuing great LIBCM work.
I am watching and reading even on holiday! LOL
 

·
Engine-Off-Coast
Joined
·
2,459 Posts
Strange issue I've seen a few times -- sometimes (rare, maybe once a week) after entering auto-stop even with a good SoC (above 50%) the car will think it has stalled and do the thing where it turns on the brake and battery lights.

Any clues for that one? I've only ever seen it with weak NiMH packs.
 

·
Premium Member
Joined
·
2,950 Posts
Strange issue I've seen a few times -- sometimes (rare, maybe once a week) after entering auto-stop even with a good SoC (above 50%) the car will think it has stalled and do the thing where it turns on the brake and battery lights.

Any clues for that one? I've only ever seen it with weak NiMH packs.
SAME! I thought this was just my car so I ignored it.
 

·
Linsight Designer
Joined
·
4,144 Posts
Discussion Starter · #866 ·
With the 60S ( and maybe the 48S), it looks like there is no room for a large battery cooling fan, like with the OEM or EHW5 batteries. Are battery temps going to be a concern?
LiBCM's onboard fans are able to draw air from the cabin, and will now do so intelligently (i.e. to warm a cold pack if the cabin is wamer; or cool a warm pack if the cabin is cooler). Once this air is inside the IMA bay, there's no longer an air shroud to guide the air, so instead it just flows through the IMA bay... probably mostly around the batteries given that there's a channel on both sides.

Like the 5AhG3 modules, the 47 Ah FoMoCo modules essentially do not heat up due to joule heating... so only ambient temperature is going to affect their temperature. So yes, battery temps would be a bit high if you parked your car on the tarmac in Phoenix for a week, but that's the same issue with any battery in the G1 Insight due to its complete lack of external cooling.
 

·
Linsight Designer
Joined
·
4,144 Posts
Discussion Starter · #867 ·
Strange issue I've seen a few times -- sometimes (rare, maybe once a week) after entering auto-stop even with a good SoC (above 50%) the car will think it has stalled and do the thing where it turns on the brake and battery lights.

Any clues for that one? I've only ever seen it with weak NiMH packs.
There are only two things that can turn the brake+battery lights on:
-the HVDC input to the DCDC converter is too high/low, and/or;
-the MCM pulls GRN/BLK wire to the DCDC low because it's not happy about something in IMA land.

I've never observed this behavior with LiBCM... until the past couple weeks (after installing MuddersMIMA; I need to come up with a better name).
1) Do y'all have a manual IMA controller installed in these cars?
2) Does the car successfully auto-start when you go into 1st gear?
 

·
Linsight Designer
Joined
·
4,144 Posts
Discussion Starter · #868 ·
Cut the FSC pin in the MCM and connect the MCM pcb pin end to 5V sourced from inside MCM. (This is what I do on all my current hacked MCM's) I suggest try this first it's very simple.
Thanks for the explanation. It's always great when someone else has already solved my problems (possibly a decade ago).
I think what I'd want to do is route that signal through a schmidt trigger (or similar delay element), so that if there's a continued short (e.g. longer than five seconds) the MCM can still open the contactor. I'd hate for an actual short to only rely on the fuse to break.

I'll play around with this more later. It would probably be a dead simple PCB soldered inside the MCM.

This prevents the FSC fail signal setting an IMA code and forcing a reset. The assist / regen will stutter momentarily as the IGBT internal protection shuts it off. This never bothered me, but you might be pushing so high now we need another option..
I've definitely felt this stutter a few times now; good to know it's the IGBT driver itself opening the gates (and not the MCM).

Limit the current/power level to prevent the low rpm overload.
Once the motor rpm gets above about 2000 rpm this problem goes away IIRC. I seem to remember it was due to the different igbt drive signals at low rpm, square v sine?
Based on my driving yesterday and today, I agree that this issue only occurs at lower RPM... I was thinking 1500 RPM, but I haven't driven enough to have that kind of resolution yet. Yes, the MCM uses a 6-state trapezoidal drive control strategy at low RPM, and then switches to sine drive at high RPM. I'm super excited to release LiMCM, which will use neither (terrible) strategy... it will always use field oriented control except when starting from a dead standstill (in which case it will use the same trap drive strategy, but only until the motor starts rotating).

It can also occur with the engine IMA start cycle as that is very demanding on the igbt. High power into a stopped motor!
Low rpm high current regen was also an issue..
I definitely notice that IMA starts/stops occur more quickly now. Haven't had any auto-stop-induced P-codes yet.

Don't allow full assist or regen requests when the rpm is below 2000rpm. Use your MIMA and make it aware of engine rpm, then just limit cmdpwr PWM power levels below 2000 rpm sufficiently to prevent the code appearing. Once you get above 2000 rpm then it's ramp up or pedal to the metal and 30kw+ LOL
Another strategy. I'm certainly connecting RPM to the RevA muddersMIMA PCB.

b) Reduce the current hack level to say 0% or 20% below 2000 rpm. This needs the controllable current hack of course and rpm awareness.
Probably not necessary to fix this problem, but yes this would probably work, too.
Honestly with 60S the regen is much less than before (unless the pack is almost completely empty). This makes sense as the PDU voltage is less than the battery voltage except at high RPM.

If you do the MCM FSC mod (1 above) you could use the redundant fail signal coming from the MDM to light a dash led or give some LIBCM LCD indication (other than the seat of your pants motor stutter) that the IGBT is overloaded.
I also suggest running the MDM fan on low at all times as a minimum at such high power levels. The MCM doesn't turn the MDM fan on until too late IMO. No point getting a heat cycling stress related IGBT internal temp sensor failure that renders it permanently US.
Yes I think this is necessary with any MIMA setup... and I still don't know why you'd put a pack this large into a car without MIMA.

PS Well done on all the continuing great LIBCM work.
I am watching and reading even on holiday! LOL
Thanks for the information on how to solve all the various problems I've come across. As always, I'm standing on the shoulders of giants.
 

·
Premium Member
Joined
·
2,950 Posts
There are only two things that can turn the brake+battery lights on:
-the HVDC input to the DCDC converter is too high/low, and/or;
-the MCM pulls GRN/BLK wire to the DCDC low because it's not happy about something in IMA land.

I've never observed this behavior with LiBCM... until the past couple weeks (after installing MuddersMIMA; I need to come up with a better name).
1) Do y'all have a manual IMA controller installed in these cars?
2) Does the car successfully auto-start when you go into 1st gear?
1. I don't.

2. I'm not sure actually... I think it does. It happens so rarely I don't usually pay attention. The times I can remember I was coasting down to make a turn, put it in gear and momentarily noticed the lights on(which turn off very quickly once the engine is going again).
 

·
Premium Member
00 MT
Joined
·
59 Posts
That's happened to me on a normal IMA battery with nothing other than a 40% current hack. I've always assumed it's the higher regen not stopping quickly enough at low RPM as the clutch goes in. At least in my case, it's the whole dash as if the car stalled, not just brake and battery lights. Is this a separate thing?
 

·
Premium Member
Joined
·
2,950 Posts
That's happened to me on a normal IMA battery with nothing other than a 40% current hack. I've always assumed it's the higher regen not stopping quickly enough at low RPM as the clutch goes in. At least in my case, it's the whole dash as if the car stalled, not just brake and battery lights. Is this a separate thing?
I don't have the current hack. I'm completely stock power train wise. Except LiBCM
 

·
Registered
Joined
·
52 Posts
Just did my first firmware update....with a mac! So easy. No problems so far 30 days in. Also it is really easy to plug the car in anytime unlike the older grid chargers. If you're on the fence about this modification/upgrade, it is very worth it. Thanks @mudder
 

·
Engine-Off-Coast
Joined
·
2,459 Posts
Finally swapped the MCM to one that has never been current hacked. Also set LiBCM config to no current hack and reprogrammed LiBCM.

Within 5 miles I got P1440 with blink code 57. So the problem doesn't seem to be particular to my initial MCM.
 

·
Linsight Designer
Joined
·
4,144 Posts
Discussion Starter · #875 ·
Next step is to swap IGBT module. I have a spare if you need one for testing.
While you're in the PDU you might as well swap the QTY3 phase current sensors, too. In fact, I have an entire spare PDU you could use to see if the issue persists.
 

·
Premium Member
Joined
·
219 Posts
Surfacing again, had a major back surgery a couple weeks back, not wrenching for a long time. :(. Walked #1 son through clutch master replacement, Silver is back on road. 4x20 is dead, thought it was pinched cable, but not. Tried a new 4x20 from my hobby box and same. Has power, ground, no data showing. I can't pull the pack for a few months now (Doc says 6+), so I've setup the rpi with arduino monitor scrolling the cell info. The only vampire on the 12V I found was OBDII C&C, which I had left plugged in all the time.
 

·
Premium Member
Joined
·
412 Posts
Hi @thibble! Glad you're back, sorry to hear about the surgery. It's hard to do anything with a bad back, everything transmits forces through there.. I hope you heal quickly and well. I wish I was closer, maybe somebody in local area can give you a hand pulling your pack...
 

·
Premium Member
Joined
·
219 Posts
Hi @thibble! , maybe somebody in local area can give you a hand pulling your pack...
Thank you! Hopefully no needs to pull it anytime soon. The only glitch currently is 4x20, I suspect I ham handed the libcm side of the cable attachment and missed a pin. Can't get to it without a partial pull but taking via the arduino USB works :). It'd be lovely if the top LIdisplay was the one used, but I expect that is reserved for Nextion. Clutch master put me out for a full day though, simply walking someone through it completely wiped me out :(.
 

·
Engine-Off-Coast
Joined
·
2,459 Posts
Y'all someone else really needs to install a Nextion and try it.

With the Nextion you can learn how long it takes your battery to grid charge from a given SoC (whatever it is when you parked it) to full. You also get an accurate assist and regen gauge where the bars actually mean something.

Right now two models are supported:
NX4832K035_011
NX4832T035_011
 

·
Premium Member
Joined
·
219 Posts
Y'all someone else really needs to install a Nextion and try it.

With the Nextion you can learn how long it takes your battery to grid charge from a given SoC (whatever it is when you parked it) to full. You also get an accurate assist and regen gauge where the bars actually mean something.

Right now two models are supported:
NX4832K035_011
NX4832T035_011
Send me connection info. :) quick edit. The one I have is a more expensive 7" NX8048P070-011C-Y. I can get the smaller ones listed as supported.
 
861 - 880 of 945 Posts
Top