Honda Insight Forum banner
1121 - 1140 of 1157 Posts

·
Premium Member
Joined
·
232 Posts
I am much out of the loop. Too much going on. My replaced LIBCM is still doing fine though, running on Natalya LIdisplay fork of 0.9.0n. Looking at a long trip in it in December.
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,122 ·
I'm still not able to update to V0.8.4f, I followed John's latest link he recently posted, I'm including the screenshot this time,
Based on your screenshot, this is an issue where Arduino isn't finding your Mega2560... probably has nothing to do with any specific version you're loading. Sometimes I find that the Arduino program gets "stuck" and the only way to get it working again is to quit the program, unplug USB, relaunch, then replug.

Let me know if you can 100% reproduce only 0.8.4f failing to load and I can look into it more (please post a video).

the odd thing is that I can load the older versions all the way to 8.4 without any issues. see screenshot below.
I've went back to V7.5, V7.8 and V8.4 without any issues, just want to know if anyone else had the same issues or if there is a setting that needs to be changed for this update.
There aren't any settings that would cause this to fail. Maybe try a different computer?
 

·
Premium Member
Joined
·
685 Posts
Based on your screenshot, this is an issue where Arduino isn't finding your Mega2560... probably has nothing to do with any specific version you're loading. Sometimes I find that the Arduino program gets "stuck" and the only way to get it working again is to quit the program, unplug USB, relaunch, then replug.

Let me know if you can 100% reproduce only 0.8.4f failing to load and I can look into it more (please post a video).


There aren't any settings that would cause this to fail. Maybe try a different computer?
Thanks, I restarted the computer and it was able to load it so I guess it was a computer issue.

Thanks again.
 

·
Registered
Joined
·
143 Posts
I understand that the plan is to add a regen cutoff below 0C in the next firmware release, and I'd like to agree this is pretty important as winter is rolling in for those of us in the North East. Furthermore, is there a plan to add charge and discharge current limits based on SoC and Temp? Having worked in the battery systems industry for a little bit now, this is what I am used to seeing for protecting the cells. It seems this would be especially helpful as the temps drop and allowable current should also drop considerably in both charge and discharge. With super low temperatures, I believe the MCM is commanding forced regen to attempt to warm the pack as it is assuming old NiMH chemistry? I saw forced regen at a temperature of -5C the other day, and I am concerned what this is doing to the cells.
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,126 ·
I understand that the plan is to add a regen cutoff below 0C in the next firmware release, and I'd like to agree this is pretty important as winter is rolling in for those of us in the North East.
I will have this feature in the prerelease firmware starting tomorrow.

Furthermore, is there a plan to add charge and discharge current limits based on SoC and Temp? Having worked in the battery systems industry for a little bit now, this is what I am used to seeing for protecting the cells. It seems this would be especially helpful as the temps drop and allowable current should also drop considerably in both charge and discharge.
Yes. There are some difficulties due to how much control LiBCM actually has over this, but worst case LiBCM can always entirely disable regen and/or assist... However, LiBCM doesn't ultimately control the power limits directly, but can indirectly control assist and/or regen levels by spoofing the temperature, SoC, etc. Yes, the firmware will eventually handle this.

With super low temperatures, I believe the MCM is commanding forced regen to attempt to warm the pack as it is assuming old NiMH chemistry? I saw forced regen at a temperature of -5C the other day, and I am concerned what this is doing to the cells.
Right now the LiBCM firmware prevents forced regen by always spoofing a higher temperature than whatever the minimum temperature is that causes forced regen (I don't recall off the top of my head what that temperature is). Obviously this logic contradicts my previous statement slightly, which is one reason I said "there are some difficulties". I'm sure there's a truth table that can get us everything we want, but I haven't sat down and derived it yet.

@Insightful2 if you aren't, already, you should be following this thread: too cold in MN for lithium??

That is the thread where mudder is discussing what he is designing to address the issue.
Once it's shipping, I'm going to strongly suggest installing this battery heater solution in all vehicles that expect to routinely spend days on end below freezing. My goal for this companion product is to heat the batteries up to maybe 15 degC in even the most brutal temperatures. It will be able to heat the batteries while driving; while grid charging; and while the car is sitting unplugged, but only if SoC is above a user-configurable level.
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,127 ·
This is only tangentially related to LiBCM... but MIT just published an article that claims to have discovered the mechanism in which dendrite growth occurs inside solid state lithium batteries. That's a huge deal because one of the major roadblocks in commercializing solid state lithium cells is rapid dendrite growth. Commercializing their findings is likely several years away, but it's a huge accomplishment to finally understand the formation mechanism.

The wild thing is that until now pretty much everyone thought dendrite formation was a side effect of the chemical process, but it appears it's caused entirely by mechanical stress:
The team demonstrated that they could directly manipulate the growth of dendrites simply by applying and releasing pressure, causing the dendrites to zig and zag in perfect alignment with the direction of the force.
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,128 ·
OK for the cold weather folks that don't mind running untested firmware, I've added a quick hack to the pre-release branch that should prevent regen entirely whenever the battery temp drops below 1 degC. I haven't tested this firmware in my car yet, but will once I'm in Wisconsin. If you're already up in the cold give it a try and let me know how it goes. There's no hysteresis, so if your pack rapidly oscillates between 0 and 1 degC you might get some odd regen behavior. This is just a temporary hack until I do the work outlined in the next paragraph.

I wasn't able to add the "turn fan on before measuring intake air temperature" tonight... my existing code (in fan.c) needs a complete rewrite, which will drastically improve my ability to modify it... it's pretty hard to add new features to the spaghetti code that's in there right now. I will rewrite this code this week while I'm in Madison.

Sorry it's not cleaner right now... just ran out of time before driving up to Madison.
 

·
Premium Member
Joined
·
433 Posts
I'll download and try it tonight...
 

·
Premium Member
Joined
·
1,442 Posts
OK for the cold weather folks that don't mind running untested firmware, I've added a quick hack to the pre-release branch that should prevent regen entirely whenever the battery temp drops below 1 degC.
Thanks, Mudder. One question. Does it also prevent grid charging when the battery temp is below 1 deg C?

-Bryan
 

·
Premium Member
Joined
·
433 Posts
@mudder , I just went your link above, downloaded then transferred to car. Appears to have uploaded successfully, please verify that I grabbed the correct cold wx version - 8.4g?
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,132 ·
@mudder , I just went your link above, downloaded then transferred to car. Appears to have uploaded successfully, please verify that I grabbed the correct cold wx version - 8.4g?
v0.8.4g is correct.

...

Thanks, Mudder. One question. Does it also prevent grid charging when the battery temp is below 1 deg C?
It does not, but I can add that, too. Given how low the regen current is, it's probably ok to grid charge below freezing... i will research this now and report back if I find a definitive answer.

...

On the drive up to Madison I finally characterized what's causing the "rapid current fluctuations during assist" issue a few people have experienced... this is where the assist current might be 3 amps one second, then 28 amps, then 6 amps, then 22 amps, etc. This behavior feels similar to "herky jerky", but it's much 'faster'. I used to never see this issue myself, but that's because of how I typically drive the car.

Without boring everyone with the details, the issue occurs when all of the following are true:
-requesting assist, and;
-assist current is less than 30 amps, and;
-engine RPM is slightly higher than the V-tech enable point (I believe this is ~3100 RPM).

When the above conditions are met, I can 100% reproduce this issue. Tomorrow I'm going to have my wife helming the laptop as we drive into Madison, so I can gather the BATTSCI and METSCI packets as I alternate in and out of this mode. With these data snippets I'll be able to figure out which METSCI packet(s) the MCM changes when the above criteria are met. From there it should be as simple as spoofing those packets as required. Honestly it's probably something as simple as LiBCM reading whichever METSCI pack is changing and then spitting it back out on BATTSCI.

...

To make collecting data like this easier, I'm adding a firmware feature that sets a flag in the serial data stream whenever the user pushes a momentary button. We can then use this feature to timestamp when interesting behavior occurs while we're driving, which will make it much easier to troubleshoot issues after-the-fact (i.e. by reviewing the log files). This will be in the next firmware update, too... I'm about halfway done writing it.

It also makes sense to add some sort of continuous datalogger, so that we can continuously log data whenever we're driving. One simple option is to plug a Raspberry Pi into LiBCM's USB port, and then have it constantly write to an SD card. With a little more work, we can skip the Pi altogether by adding an SD card slot to the "FoMoCo Heater PCB", which is similar to the daughterboard heating concept I proposed at the beginning of last week... except that it will have several other 'accessory' features, too:
-SD card slot (using SPI header)
-1500 watt power supply control TRIAC and blocking diode (so I don't have to modify each power supply's internal PCB).
-isolated output to control 1500 watt power supply's output current level (analog or digital TBD).
-I/O for user momentary push button
-Either discrete resistors for forced air heating elements, or screw header for external resistors that bolt to FoMoCo rails.
-temperature sensor (to monitor daughterboard PCB temperature).

...

I am absolutely enjoying both designing and using LiBCM... I certainly don't 'use' it enough. Hence my happiness today as I drove the 60S FoMoCo prototype car 800 miles without stopping to charge. Coupled with the "Mudder's Manual IMA" prototype controller, the IMA motor delivered on average 1 horsepower assist for ten hours! I averaged 61.1 MPG driving fast on the interstates in the cold midwestern air, which is at least 10 MPG better than I normally average on this yearly trip. This is especially surprising given how heavy Squib was: In addition to my wife and our usual luggage, I also had an additional set of FoMoCo modules, which I delivered to @Tynow near Chicago. I estimate we had about 575 pounds in the car (which is about 215 pounds more than it's rated for). I was certainly glad to jettison that extra battery weight in Chicago.

I arrived at our destination with 11% SoC (i.e. 1% left to spare). I can say without hesitation that 'range anxiety' alone will keep me from buying a fully electric car for a long time... at least until charging infrastructure is as ubiquitous as traditional gas stations are. Somehow central Madison only has three free public chargers, which I find severely lacking for a city with this many electric vehicles. We have a long way to go.

As I'm writing this, I'm sitting in the parking lot at an animal hospital charging for at least the next two hours so the forced air heater prototype heater I showed off previously can heat the battery overnight. As mentioned previously, I do not expect this concept to work efficiently, but it's what I have available until I design the actual FoMoCo heater solution.

Thank y'all so much for sharing this experience with me over the past 18 months... this project is proof that "time flies when you're having fun".
 

·
Premium Member
Joined
·
3,142 Posts
v0.8.4g is correct.

...


It does not, but I can add that, too. Given how low the regen current is, it's probably ok to grid charge below freezing... i will research this now and report back if I find a definitive answer.

...

On the drive up to Madison I finally characterized what's causing the "rapid current fluctuations during assist" issue a few people have experienced... this is where the assist current might be 3 amps one second, then 28 amps, then 6 amps, then 22 amps, etc. This behavior feels similar to "herky jerky", but it's much 'faster'. I used to never see this issue myself, but that's because of how I typically drive the car.

Without boring everyone with the details, the issue occurs when all of the following are true:
-requesting assist, and;
-assist current is less than 30 amps, and;
-engine RPM is slightly higher than the V-tech enable point (I believe this is ~3100 RPM).

When the above conditions are met, I can 100% reproduce this issue. Tomorrow I'm going to have my wife helming the laptop as we drive into Madison, so I can gather the BATTSCI and METSCI packets as I alternate in and out of this mode. With these data snippets I'll be able to figure out which METSCI packet(s) the MCM changes when the above criteria are met. From there it should be as simple as spoofing those packets as required. Honestly it's probably something as simple as LiBCM reading whichever METSCI pack is changing and then spitting it back out on BATTSCI.

...

To make collecting data like this easier, I'm adding a firmware feature that sets a flag in the serial data stream whenever the user pushes a momentary button. We can then use this feature to timestamp when interesting behavior occurs while we're driving, which will make it much easier to troubleshoot issues after-the-fact (i.e. by reviewing the log files). This will be in the next firmware update, too... I'm about halfway done writing it.

It also makes sense to add some sort of continuous datalogger, so that we can continuously log data whenever we're driving. One simple option is to plug a Raspberry Pi into LiBCM's USB port, and then have it constantly write to an SD card. With a little more work, we can skip the Pi altogether by adding an SD card slot to the "FoMoCo Heater PCB", which is similar to the daughterboard heating concept I proposed at the beginning of last week... except that it will have several other 'accessory' features, too:
-SD card slot (using SPI header)
-1500 watt power supply control TRIAC and blocking diode (so I don't have to modify each power supply's internal PCB).
-isolated output to control 1500 watt power supply's output current level (analog or digital TBD).
-I/O for user momentary push button
-Either discrete resistors for forced air heating elements, or screw header for external resistors that bolt to FoMoCo rails.
-temperature sensor (to monitor daughterboard PCB temperature).

...

I am absolutely enjoying both designing and using LiBCM... I certainly don't 'use' it enough. Hence my happiness today as I drove the 60S FoMoCo prototype car 800 miles without stopping to charge. Coupled with the "Mudder's Manual IMA" prototype controller, the IMA motor delivered on average 1 horsepower assist for ten hours! I averaged 61.1 MPG driving fast on the interstates in the cold midwestern air, which is at least 10 MPG better than I normally average on this yearly trip. This is especially surprising given how heavy Squib was: In addition to my wife and our usual luggage, I also had an additional set of FoMoCo modules, which I delivered to @Tynow near Chicago. I estimate we had about 575 pounds in the car (which is about 215 pounds more than it's rated for). I was certainly glad to jettison that extra battery weight in Chicago.

I arrived at our destination with 11% SoC (i.e. 1% left to spare). I can say without hesitation that 'range anxiety' alone will keep me from buying a fully electric car for a long time... at least until charging infrastructure is as ubiquitous as traditional gas stations are. Somehow central Madison only has three free public chargers, which I find severely lacking for a city with this many electric vehicles. We have a long way to go.

As I'm writing this, I'm sitting in the parking lot at an animal hospital charging for at least the next two hours so the forced air heater prototype heater I showed off previously can heat the battery overnight. As mentioned previously, I do not expect this concept to work efficiently, but it's what I have available until I design the actual FoMoCo heater solution.

Thank y'all so much for sharing this experience with me over the past 18 months... this project is proof that "time flies when you're having fun".

I don't mean to nit pick you, but you mean Regen fluctuation, not assist right? Also, I told you it feels like herky jerky, that's how I've described it since the beginning :p.

As for the 60S thing... Stop making me want it! I don't have any real options for charging, and as such it's not as practical >:l.
 

·
Premium Member
Joined
·
433 Posts
Well, after at least two weeks of sub-freezing overnight temperatures, I installed the cold wx mod firmware and we have the first night of warm temps!! Battery was sitting at +2C this morning, so no test of the update. Maybe tomorrow morning!
 

·
Registered
Joined
·
10 Posts
I experienced the "herky jerky" for the first time this past Saturday. Also the first time I've driven such a long trip since installing the LiBCM--200 miles whereas I usually drive 40 miles or less. A minor bug. Still loving LiBCM!
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,137 ·
I don't mean to nit pick you, but you mean Regen fluctuation, not assist right? Also, I told you it feels like herky jerky, that's how I've described it since the beginning :p.
I am in fact talking about the assist fluctuation issue.
The regen fluctuation issue is different, and I understand why it happens, too... on my todo list. So much to do.

As for the 60S thing... Stop making me want it! I don't have any real options for charging, and as such it's not as practical >:l.
What did I say that made you want it? If you don't have any real options for charging, then FoMoCo probably isn't for you... just go with the 5AhG3 modules.
 

·
Registered
Joined
·
143 Posts
Once it's shipping, I'm going to strongly suggest installing this battery heater solution in all vehicles that expect to routinely spend days on end below freezing. My goal for this companion product is to heat the batteries up to maybe 15 degC in even the most brutal temperatures. It will be able to heat the batteries while driving; while grid charging; and while the car is sitting unplugged, but only if SoC is above a user-configurable level.
I think having an onboard heating solution is a good idea. However, I do have concern about how much heat we can actually transfer through to the cells using convection through such a small and poor flow path around the modules which is perpendicular to the cells themselves. I am willing to see how this will work practically, though, as my concerns may be illegitimate. Additionally, I've seen this mentioned a few other times, but my understanding is that while not being used, Li-ion cells benefit from being stored at colder temperatures. As the electrolyte becomes more viscous, the side reactions that age a cell slow down, or that is what I've heard. If someone else has different data on this, please let me know. In that light, I think I would only want to heat during use but not during sitting. If it could be heated in a few minutes after starting the car, I think that would be ideal.
 

·
Linsight Designer
Joined
·
4,479 Posts
Discussion Starter · #1,139 ·
I think having an onboard heating solution is a good idea. However, I do have concern about how much heat we can actually transfer through to the cells using convection through such a small and poor flow path around the modules which is perpendicular to the cells themselves. I am willing to see how this will work practically, though, as my concerns may be illegitimate.
To clarify, the PCBs I just ordered for the 5AhG3 solution are not using forced air; the heat is generated directly beside each cell, via QTY9 resistors near the bottom of each cell. Right now I'm playing around with a forced air solution, too, but only because it's what I have available right now and also as a backup in case the first prototypes don't work well. Certainly the most desirable solution is to not run the fans, as this substantially reduces overall efficiency.

Additionally, I've seen this mentioned a few other times, but my understanding is that while not being used, Li-ion cells benefit from being stored at colder temperatures. As the electrolyte becomes more viscous, the side reactions that age a cell slow down, or that is what I've heard. If someone else has different data on this, please let me know.
Lithium cells 'like' to stay in the same environment as people. I've never heard any scholarly article suggest freezing the electrolyte for longevity.

I think I would only want to heat during use but not during sitting. If it could be heated in a few minutes after starting the car, I think that would be ideal.
Heating the module up in a few minutes would not be good for it... you want to slowly change the temperature to allow everything to heat up uniformly.
 

·
Premium Member
Joined
·
433 Posts
Battery temp this morning was -2C. No regenerative charging available until the temp indication was 1C. I did not notice any strange behaviour around the 0C mark...

I did observe that I could get +12 amps assist while battery temp was still -2C. Is this OK or should I be limiting assist until battery temp rises?
 
  • Like
Reactions: mudder and joeaax1j
1121 - 1140 of 1157 Posts
Top