Honda Insight Forum banner

1 - 20 of 24 Posts

·
Registered
Joined
·
13 Posts
Discussion Starter #1
Hello to all here. I am a new proud owner of a G1 Manual 191K, and I've been reading and watching stuff on here to get up to speed a bit faster, and have been driving my car and observing. Incredible resource, thanks, I would not have purchased this if I didn't know there was such a wealth of knowledge and passion regarding keeping these going. I live on the Big Island of Hawaii, where there are a number of long hills ( volcanos ), and after driving this home I put it on the grid charger I obtained with the car and charged the IMA batteries for about 16 hours up to about 175V, current draw 256mA. Reset the IMA light (#18 fuse) and pulled the relay, did a battery gauge recalibration (I have a shop manual), In driving around I observe the ASST/CHARGE dance as power is applied or removed, which I assume is correct as I have no experience (why I am posting). IMA light is off. If I go up a large hill I have ASST for a partial time, then the ASST dies and I have the CHARGE bars showing even though I am climbing a hill, which I am assuming is the NEG REGEN that has been mentioned in the forums. (Also the IMA light is now illuminated). Is that true? So for the remainder of the hill I am only using the ICE, lug up the hill, then when on the flats or descending after a time I see the CHARGE bars. I try to engage the IMA and it will not assist. But, after a bit of charge it will. and with enough charging the BATTERY comes up to nearly full. (Seems too short a time to have this happen but what do I know anyway at this point?) IMA light stays illuminated untll I pull the #18 fuse, then is out.
Given the above observation, does this seem normal for a weak battery as I am suspecting? Will discharging and recharging (I'm a retired A&P mechanic and in our world we call this a "deep-cycle") possibly improve this? I have a charger but no discharge method yet, and I guess I'll build a simple device using the lightbulbs I read about.
I don't really care if I sound stupid here, I love this little car and want to do right by it, not damaging anything. It seems totally unmessed with thankfully except for a really crummy stereo! LOL Glad to be part of this forum, I hope to be more succinct in future postings!
 

·
Registered
Joined
·
851 Posts
Welcome! It sounds like one or more of the cells in your pack is losing capacity. When this happens, its voltage will be different from the others and the car will notice. It will try to fix this by charging in the background until the pack is full and all cells have had an opportunity to charge. (The cells that are charged will be overcharging while the others catch up - more on that below). But if the voltage imbalance persists, you're going to get the light.

If you are impatient, then before you rip into the back, you might try using those hills to your advantage and put the pack through a few full charge/discharge cycles and see if it provides any kind of cell conditioning benefit. Get the pack down to the point where it can no longer provide any assist, then get it to the point that it is fully charged. If you find after the second cycle that things are improved over the first, consider repeating maybe two or three more times. If it doesn't help, well, you know the drill.

But as a trained mechanic you may be equipped to attempt to recondition the cells. If you have the place and time, perhaps this would be your first step. I'm working on a conditioner now that's Arduino controlled and will condition every cell to specified voltage limits up to two sticks at a time, and report the final capacity of each cell when complete. (I used darts to puncture the insulating jacket to measure every cell voltage and apply a load if I want to increase the discharge rate or slow the charge rate on individual cells - a compromise to having a dedicated charger and discharger on every cell, which would require more than a dart point to make a sufficiently low resistance contact.) I'm told this might be a fool's errand - that old cells are old cells - but I won't know until I have hard data. The idea of blindly charging and discharging an entire stick or pack without some auto shutdown if one cell reaches a specified max or min concerns me enough to not do it (but plenty here do.)

I have observed actual cell voltages during operation of the car, rather, the voltages of every 12 cells, for a while when I had my car instrumented to display this. Those observations make me comfortable in discharging this way (infrequently) because I think the car's battery monitor will complain well before a cell reverses after it is depleted. I'm not as confident about the charge cycle. I have little data on charging/discharging, but I do have a bunch of sticks, and the original Honda sticks all seem to be able to tolerate overcharging without venting. My non-Honda sticks show signs of having vented. I wonder if the community needs to design a qualification test for candidate replacement cells that involves an overcharging regime that confirms the continued integrity of the battery casing and identifies that the failure point is sufficient.

I also wonder whether those that are overcharging while the others catch up are being insulted; whether these insults are cumulative; whether it is the weakest cells that reach capacity first and then have to tolerate overcharging; and whether this creates a vicious cycle. No data to back this up, though.

If this does not help, and you do decide to tear into it, you will find many opinions and procedures. I disagree with some of them, mainly because I am concerned about potential insult to the cell during overcharge or undercharge. I have a number of other theories, which I'm happy to discuss via private conversation, because I have not done enough testing to confirm them (speculation without sufficient caveats or healthy skepticism sometimes turns into fact).


Again, welcome and what a great place to drive your car! You must schedule a meet-up.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #3
Thanks sean for the reply...and thanks for the detail on the IMA. I'm kinda doing what I think you are suggesting albeit somewhat inadvertently, I have been running up the hills at a partial power, maybe 4-5 bars on the ASST gauge, letting it go to no power and the CHARGE showing like 2-4 bars, and driving on a flat or slightly downhill till I get assist back (it's a PITA to capitalize that!), looking to see if the ability of the IMA to pull the hill longer with a full batt gauge would improve.. I have observed that if I try to use full or almost full power assist the battery will only pull for a very short time, like 30-60 seconds, I'm assuming this isn't normal. I do have another full battery pack that I obtained with the car, as well as a couple of chargers, so I will be diving into the deep cycle process on the stick pairs at some point. I just do not want to irreparably damage something in the process of attempting to observe or "fix" it.
 

·
Registered
Joined
·
851 Posts
I have been running up the hills at a partial power, maybe 4-5 bars on the ASST gauge, letting it go to no power and the CHARGE showing like 2-4 bars, and driving on a flat or slightly downhill till I get assist back
Yeah! This is good. The longer you can do this, the better. I had a real time display of each of the 10 taps on the battery for a 500 mile road trip. The difference between the voltages between the cells grew less and less the longer I drove.

In other words, the longer I drove it, the more balanced it became (until it sat overnight and the cells drifted down in voltage at different rates.)

I have not had this setup installed for a while and need to put it back in (I need to add logging including GPS and time stamping to display on a map.) I've driven it only a couple of times a week and haven't been seeing the recalibration I was seeing before, as if a long rest between long drives has been a good thing????? Without data, I can't say.

The thinking behind cycling between completely discharged and completely charged is to try to put it through conditioning cycles using the car's own monitoring to keep it from doing things like cell reversal.
I do have another full battery pack
This is awesome!!!

My plan is to have a conditioned pack ready to go, and an Arduino looking for a recalibration initiated by the car. If I see this (or my voltage data shows things are drifting) the plan is to swap packs.

Then I will have all the time in the world to recondition the pack that's offline, and no loss of use of the car.

I have tried to do this with one pack and basically it wasn't worth driving the car with the pack missing while working on it. It was too annoying. (others will vigorously disagree.)
I just do not want to irreparably damage something in the process of attempting to observe or "fix" it.
With a second pack, and chargers, and your training, you are going to be in great shape. Looking forward to reading more of your progress.
 

·
Registered
Joined
·
6,985 Posts
...If I go up a large hill I have ASST for a partial time, then the ASST dies and I have the CHARGE bars showing even though I am climbing a hill, which I am assuming is the NEG REGEN that has been mentioned in the forums. (Also the IMA light is now illuminated). Is that true?
What's your "BAT" gauge doing during this time? Does it fall to the bottom? Also, you should get the blink code for your IMA light...

Sounds like you're just reaching empty (of a probably not very charged pack) and the car does a 'forced charge' (not "neg regen," you're probably thinking of "neg recal" - which means empty pack). If your BAT gauge falls to the bottom, that would be a neg recal and if so your IMA light is probably for a P1449-78...

It sounds like one or more of the cells in your pack is losing capacity. When this happens, its voltage will be different from the others and the car will notice. It will try to fix this by charging in the background until the pack is full and all cells have had an opportunity to charge. (The cells that are charged will be overcharging while the others catch up - more on that below).
Pretty sure this isn't quite the way it works. I don't think cells are ever allowed to overcharge in the car, at least not in the strictest sense - where a cell is chock-full 6500mAh but the car continues to charge anyway... I think you're thinking of the dynamics of a grid charge...

hmm, how can I keep this short and sweet?...

In a nut shell: The BCM manages the amount of charge and discharge allowed (total energy usage) based primarily on tap voltages - the pack is considered 'empty' when a low voltage is reached, 'full' when a high voltage is reached. Pretty sure it also uses voltage slope detection as well, both high and low: when a single cell is near empty, the tap voltage will start to drop fast, and when a single cell is near full, tap voltage will increase fast. So, it's an either/or situation, either the BCM detects a steep slope or an absolute low or high voltage, both mean empty or full, whichever comes first.

If the pack's fine - not 'crudded-up', is balanced - you get a lot of usable capacity. If you get a 'neg recal', i.e. pack empty, it'd take a while to charge to 'full'. If the pack's messed up you don't get a lot of capacity - you'd get a neg recal and before say 10% charge can take place one of the taps hits the high voltage 'full' level, or the BCM detects a steep charge voltage slope - and does a 'pos recal' (calls the pack 'full'). If it can't charge more than 10% you get the IMA light and P1449-78...

There might be a compensatory, 'charge more' routine in the BCM. It's been a long time since I've seen it, dealt with it, but I recall in the olden days that I had seen charging behavior that must have been meant to stuff the pack as much as possible, while keeping tap voltages below the 'full' threshold. Basically, after a neg recal, with a problematic pack, I recall the BCM continuing to charge and lowering the charge rate to sub 1 or 2-amp level, and as soon as something like 35% had been added it would pos recal... But that's not typical, not the typical routine...

Um, anyway... Most misbehavior of packs is probably due to imbalance, at least, that's the immediate cause. One tap is near empty, another is near full, that's basically it. The BCM can't or doesn't cram much into a pack when the taps are imbalanced like that. Barring any mysterious 'topping-up routines' that the BCM does or doesn't do, the only thing you can do in the car to fix this is try to cram as much as possible into the pack - at the top. When cells get above something like 70%, charge efficiency decreases. So, if you have an imbalanced pack, repeated excursions at the top of the allowable charge state range can cram more into the lesser-charged cells than the more-charged cells - your proverbial 'catching-up'. But, I'm pretty certain no actual over-charging takes place, even at the cell level...

I think you can achieve some 'stretching' of the allowable charge state range at the bottom as well, but in the car it's virtually impossible without an OBDIIC&C - the BCM makes it too difficult to discharge much 'down there'. Even with an OBDIIC&C it's hard to do. Cramming at the top is easier to do if you have an OBDIIC&C - you can reset the nominal state of charge. Nominal state of charge controls various management behaviors. For instance, you can charge from 75% to 80%, discharge a point to bring tap voltage down, reset nominal SoC to 75% and charge again to 80% - you net say 4% every time. This is a lot more effective than relying on the car's own 'pos recal' behavior, where you just charge to the 72% nominal hang until the BCM pos recals (jumps to 75% nominal, or ~80% on some BCMs), then use assist to drop below 72% again and charge some more until pos recal, and again and again however many times the car allows. I find that this doesn't get you very far.

If you measure tap voltages and discover clear, low voltage outliers, you can achieve balancing by shorting the high taps, at a rate of about 14% per 24 hour period.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #6
eq1...thanks for your input on this. I will need to digest this info before I can even try to ask questions, LOL! But I can ask where I can obtain a OBDIIC&C and get that here to work with. I think I read that Peter makes these in the UK. Are there other options?
 

·
Registered
Joined
·
851 Posts
Pretty sure this isn't quite the way it works. I don't think cells are ever allowed to overcharge in the car
Here is a screen capture from a road test. The top row is the peak voltage seen for each stick pair.

This works out to a pack voltage of 181 volts. The car clearly was pushing a healthy amount of current into those cells to reach that voltage.

I do not think this is good if it happens over and over.

87833
 

·
Registered
Joined
·
851 Posts
Are there other options?
For code reading, clearing, and live data capture, I use an Autel MD802 MaxiDiag Elite "All Systems". This will read the ECUs that most code readers can't see (like the IMA motor control computer, the antilock brake computer, and others, that are not on the OBD2 data line.) Also, it works with pretty much any other OBD2 car so it's money well spent. Mine cost somewhere between $200 and $250, and it has solved problems on three of my cars that normal readers can't see. It can capture live engine or IMA data and graph it in real time. I use this tool often and am happy I bought it.

It will even log live data to SD card but I think you have to buy their program to decode it. I didn't; I wrote a script in Python which I use to decode the log files into CSV which I can then import into Excel.

I bought it a few years ago and since then, a newer version has come out which I can't vouch for.

For seeing stuff in real time, there's Torque Pro on Android that you can use with a Bluetooth OBD2 adapter. I've used this to make a custom "Lean Burn" widget which tells me when I enter lean burn.

Finally, there always is "do it yourself". But I think it's asking for trouble (death?) to mess with the IMA system without knowing what you're doing and putting a lot of thought into it and the safety systems Honda has built in. It concerns me to see photos of IMA systems with wires routed randomly and then leaving the compartment. The capture above was made with a mod that was completely low voltage and contained almost entirely in one module. It used Bluetooth for comm so no wires had to leave the IMA. I used it long enough to capture what I needed to capture and then removed it from the car.

The OBDIIC&C is pretty much the Universal Insight Tool for lots of people who love it. Peter has put his heart into this weird car called the Insight and many people have benefitted from his work. He has shared a lot of information with the forum, and by purchasing this device you are supporting him.

You being an A&P mechanic, I think you are aware of the cockpit concepts of situational awareness and workflow and sterile cockpit and "just fly the plane". This is why I'm happy with my Autel and minimizing what's on the dash.

Best!
 

·
Registered
Joined
·
6,985 Posts
Here is a screen capture from a road test. The top row is the peak voltage seen for each stick pair. This works out to a pack voltage of 181 volts. The car clearly was pushing a healthy amount of current into those cells to reach that voltage. I do not think this is good if it happens over and over.

^ Cool app, I kind of need one of those, or something similar. I'm tired of measuring tap voltages by hand, and entering the data in an excel spreadsheet, by hand...

If those are charge-loaded voltages in your screen shot, I don't think there's a problem with that... Anything under about 186V I wouldn't worry about. That's 1.53V per cell, which is the normal peak voltage during charge - at modest current (around 6 amps). So, in the car, factor in resistance and a higher charge current and 181V is quite a ways off from true full... Pretty sure anything below 186V is conservatively fine. The car's own cutoff is 192V (or rather 19.2V at tap level).

So, in your screen shot, are those tap voltages a snap shot, a moment in time? And during a charge? If so, I'd be thinking "imbalance" - I'm pretty sure tap 4's lower voltage reflects a true, sizable charge imbalance. Or I guess it could be some kind of mismatch, if not simply the amount of charge... I don't see variation like that around those voltages - not unless the taps are imbalanced.

Oh and, BTW, OP, Punatic, my apologies for wonking-out in your new thread... You've probably stumbled into two people who aren't the most practical, or at least conventional, around here when it comes to addressing 'battery problems'. There's a lot of other people who'd give you different, more 'mainstream' advice than we are/do. Keep that in mind.

I'd just like to invite others to add that more practical advice if you have it. And, Punatic, just say the word and I'll try to limit the wonk.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #10
Oh no, on the contrary eq1 I am enjoying diving into this rabbit hole. How I operate is to get a lot of data and eventually the dots start to connect. I appreciate the passion and intelligent discourse. Wonk On!
 

·
Registered
Joined
·
13 Posts
Discussion Starter #11
I just purchased the OBDIIC&C and the IMAC&C P&P from Peter. I figure I can use all the help I can get dealing with the complexity. Plus I plan on keeping this car for the duration. I believe it's the perfect car for here when it's functioning correctly.
 

·
Registered
Joined
·
6,985 Posts
^ I find the OBDIIC&C pretty indispensable. I know I mentioned it earlier as a tool to help deal with the pack, but I wouldn't expect too much, wouldn't count on it. Regardless, I think it's well worth it.
 

·
Registered
2006 Honda Insight MT
Joined
·
618 Posts
Don't forget that even with a new or reconditioned battery on longer grades the battery still is a fairly small battery and will easily deplete itself as the car will try to maintain assist all the way up the grade until the battery needs charging. The car doesn't manage the battery as well on mountain grades as the Prius does.

Downshifting to 2nd or 3rd, or using Peter's IMAC&C can limit the amount of battery assist up the hill, which can be used strategically to just deplete the battery at the top of the mountain and then charge it on the downgrade.
 

·
Registered
Joined
·
6,985 Posts
^ I agree with that. If you're driving over an honest-to-goodness mountain pass/es, for instance, it's often better to just forget assist and regen - bypass with a calpod clutch switch if you can, or use lower gearing at modest throttle. Either that or you really have to be judicious with your usage, timing assist and regen with the natural ups and downs of the terrain...

A decent pack around full car-charged should probably put out about 3.25 Ah, while typical up hill assist might be around 20 amps on average. Calculating total assist time available would look something like this:

3.25Ah/20A=0.1625 hours, or 9 minutes and 45 seconds.

If your up hill battle is approaching that duration or longer, expect a neg recal, forced charge, limited assist.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #15
So a 20A draw relates to what on the assist meter? 9.75 minutes at full power is pretty good and way more than I have currently. When I purchased the car I had to drive back from Kona to the east side, and the most direct way is on the Saddle Road, up between two volcanos. It's a uphill pull I'm guessing about 7 miles sea level to 3500 or so feet. I had assist for maybe 30 seconds. After that I just stayed in 2-3rd gear and took my sweet time.
 

·
Registered
Joined
·
6,985 Posts
^ First, it'd be about 9.75 minutes at 20 amps, not "full power." Full power is about 11kW for 4 seconds, which would be more like 80-90 amps, then it drops to something like 6.4kW, which might be around 40 amps. Pretty sure both of these 'on the meter' are full bars, full meter. So 20 amps might be around half the bars. The meter itself though reflects power in watts, not amps, and it's a theoretical calculation of some sort, not based on actual measured current and volts; I've seen the assist and charge bars moving when there wasn't actually any assist or charge happening...

You can do the math - if 20 amps lasts about 9-10 minutes, 40 would last half that. Etc etc. And again, that's if your pack started near fully car-charged and was fairly decent.

On big hills/mountainous terrain, the main thing you want to avoid is getting background charge/forced charge off the gas engine while you're trying to climb the hill. That can be a real pain. It can be difficult to avoid so best to install a 'calpod' clutch switch (if you have a manual transmission) - it's just a switch installed parallel to the clutch switch. When you press the clutch, i.e. flip the switch, it will disable most of the IMA assist and charging - it doesn't disable the 'forced charge' though, when you've drained the pack entirely, gotten a 'neg recal' (BAT gauge falls to the bottom)... Since you've bought an OBDIIC&C with the IMAC&C you should be able to use that to achieve these sorts of things, I believe.
 

·
Registered
Joined
·
851 Posts
@Punatic

I wanted to follow up and find out how you were doing.

I'm especially interested in knowing whether exercising the IMA between exhaustion and full on the hill improved the battery situation?
 

·
Registered
Joined
·
13 Posts
Discussion Starter #18
Frankly I'm not sure sean...I've been driving and experimenting. Twice I have grid charged the IMA pack, this last time for over 12 hours at least 6 of those were at a constant voltage (as close as I can tell as the charger doesn't delineate lower than volts). Both times I have found that if I baby the assist up a hill I still have something at the top and the IMA light stays off. However after several days it seems even doing this (and with solid regen downhill), depletes the capacity and the IMA light comes on. Interestingly at least to me, is when the IMA light illuminates, the car seems to have more assist capability, the assist/charge dance seems more in tune with what power I'm applying. I don't understand why that happens. But through all of this before or after the IMA illuminates, the battery capacity lights barely change from full lights to maybe 2 or 3 bars down. It seems obvious to me that the battery is being challenged at it's lower limits, ie, the weak cell(s) are inhibiting the discharge function. Does this seem plausible? I've ordered a Prolong deep cycle system to work with this. I can't wait to install and drop all the cells and start a deep cycle process. In the meantime my plan is to drive it with the IMA on and see if the capacity meter drops more. I forget the OBDII code thrown, I have it written down in the garage, I think it may be something like P1104.
 

·
Registered
Joined
·
6,985 Posts
^ I think you need to get your 'blink codes': Misc. RESETS

Some of what you write sounds a little strange. I would have said it sounds like you have some fast self-discharge cells, but I would expect the BAT gauge to drop to the bottom if you did, not only "2 or 3 bars." I guess it could still be fast discharge cells (i.e. some cells near empty despite your grid charging, others near full), only you're not pushing it hard enough to trigger the 'neg recal' (empty behavior).

Get the blink codes, report back.
 
1 - 20 of 24 Posts
Top