Honda Insight Forum banner

21 - 36 of 36 Posts

·
Registered
Joined
·
16 Posts
Discussion Starter #21
Eq1 I have done the charge and initial cell level check. I'll check next weekend to see the drop.
 

·
Registered
Joined
·
16 Posts
Discussion Starter #22
With Mudder's Arduino mod you can have lean burn without a battery. It requires a slightly different version of the code. It's documented in his thread.

I had a perfectly working battery, until one winter I had a water leak which shorted a wire under the dash and disabled the IMA system. I found the car a lot less fun to drive, and there was a fuel economy penalty - if I recall it was something like 10-15%. I was pretty unhappy. After winter I pulled the carpet out and fixed the damaged wiring.

Around a year later I intentionally removed my battery and sold it, with the intent to put a 200+ HP Acura motor in. I drove for another year with the battery intentionally removed, and didn't mind it nearly as much. Context sensitive I suppose? In some ways it was nice not to worry about the battery.

Now I have the Acura motor in the car. If I drive carefully I can get 50-60mpg, with mid to high 60's possible with my full bag of tricks. That goes out the window when I put my foot down. Honestly, I can't say I like the car any better or worse now, many thousands of dollars later. It's faster, less refined, less fuel efficient (but still reasonably economical).

~

I feel a stock Insight is a great car, but the battery is a ticking time bomb. Battery replacement costs kill any fuel savings over other economy cars. You do it because you love the car.

A bypassed Insight is most in want of different gearing. The spacing between 1-2-3 is too large, and 4-5 are close enough together that 4th might as well not exist most of the time. However, take the time and money to build a custom transmission and you might as well have just bought a battery. I feel this is also something you do only if you really love the car.

Edit: One thing I really appreciate after my batterectimy is how much extra cargo room I have in the rear.

I think these cars can be really rewarding to those who love them and either a) take care of them, or b) modify them.
I can't imagine what the acceleration is like with that much power. You have a nice sleeper there. I'd assume you had to swap transmission and do alot of wiring as well. I would love to have 3-4 times the power, I'd settle for 40mpg.
 

·
Registered
Joined
·
16 Posts
Discussion Starter #23
I'm using a Hybrid Automotive Prolong grid charger and discharger for the reconditioning process and then made my own battery module load tester by adding a couple wires, a switch and headlight port to the light bulb discharger. I use alligator clips on the end of my grid charger harness to read the voltage and provide a load from the headlight. Here is their website: https://hybridautomotive.com/products/

After doing a bunch of reading, I came up with this approach and so far, so good. I've successfully refurbished several IMA batteries using this method. My first step is to test the voltage at the gray plug to see what the 10 pairs of batteries are showing. That will give you an idea of the problem batteries. Then I recondition the whole IMA battery over roughly 5 days. Charge 24+ hours, discharge to 96v, charge 24+ hours, discharge to 60v and charge 24+ hours. After that, I let the battery sit for 12-24 hours to kinda settle the resting voltage. Then I remove the side covers of the battery, alligator clip the + & - bolts and load test all 20 sticks for 2 minutes each. Resting voltage 12-24 hours after the initial reconditioning is not a good way to determine if you have a bad stick. Unless really bad, they'll all show pretty consistent voltage after a grid charge. However, the 2 minute load test will reveal problem sticks within 2 minutes. I write down the resting voltage, turn on the light bulb for 2 minutes and then write down the voltage after 2 minutes, then subtract the loaded from the resting voltage to see how much each battery drops in voltage. I've seen as low as 1.39v and as high as 3.13v. The lower the better, but the most important thing is trying to keep the sticks consistent across not only all 20 sticks, but the 10 pairs of sticks too. So once I have the 20 sticks complete, I do the same math across all 10 pairs to ensure the resting voltage range is <.2v across all 10 pairs and the loaded voltage range is <.6v. As eq1 has pointed out in other threads, the self-discharge is probably the most important test though. The load test is good if you've done the reconditioning process, but get an IMA light either immediately or within a few days.

eq1 suggests doing a self-discharge test over 1 week. I tested this approach and have found that 3 weeks is a better approach if you want a higher success rate. So after the load testing, I grid charge one more time, wait 12-24 hours and write down the resting voltage for all 20 sticks. 1 week later, I record the resting voltage for all 20 sticks again and repeat for a total of 3 weeks of just sitting on the bench. On one particular pack I was working on, over 1 week, the pack spread was .14v. Week 2 was .18v. But during week 3, I had a stick plummet in self-discharge rate and the spread became 1.37v. If I removed just that one stick, the spread would have been .2v, so I removed 2 more sticks to make the spread .1v over 3 weeks. The tricky part is, you have to do this entire process with spare sticks to ensure the self-discharge rate is the same/close to the rest of the pack you're trying to condition. It takes a month or more to refurbish an IMA battery and is likely why people don't bother.

It's a long, tedious process, but kinda fun. Many of the hours are charging and discharging. Disassembly, load testing & reassembly is about 2 hours, then all the math and strategically placing the replacement sticks, depending on how many, could be another hour. I think if you already have the equipment and sticks, it's worth the effort and time, but to go out and buy it all...I don't know. Might be best to just pay for it to get it done right. After reading through the plethora of knowledge, I have a lot of respect for the folks who figured this all out. I wouldn't have attempted it in the first place. And at first, I thought they made it much more difficult than it needed to be, but after refurbishing a pack that needed 7 sticks replaced, I see why it can certainly be difficult. 1 stick was very easy, 7...not so much.
Thanks for all the info. I may try a full pack discharge/charge cycle or two to see if I can get my cells a little more balanced. You are using your headlight(1) as the load for your two minute load test?
 

·
Registered
Joined
·
6,730 Posts
Eq1 I have done the charge and initial cell level check. I'll check next weekend to see the drop.
If you have time to post your cell voltages after grid charge, your baseline values, I could probably spot...potentially troublesome cells, or at least give you my read on how normal or abnormal they look... If you do, just write them like this, starting with bottom row of pack left to right, next row right to left, top row left to right, where left is based on looking at pack from electronics board side:

EX. 1.432V =1432
stick 1: 1432,1431,1422,1432,1434,1456
stick 2: etc etc...

You don't need to write "stick 1," "stick 2," etc etc...

Also, I assume you did a full grid charge, terminated, and waited an hour to measure cell voltages?

Here's a couple graphs of self-discharge from full, top graph of a whole pack, bottom of 4 sticks. Each dot is a cell. You can see in the top graph how a few cells are outliers by about day 6. That kind of discrepancy will cause problems soon...

85664


Cells for these 4 sticks were all tested for capacity after the self discharge period, self discharge was about the same for all of them. Hours on the x-axis:
85665
 

·
Registered
Joined
·
16 Posts
Discussion Starter #25
Eq1,
I didn't necessarily wait an hour probably around 30 minutes and I'd say it took 20-30 minutes to take readings with some other extra curricular activities I had going on. I should have some time to post the readings tomorrow if I can get enough service to hotspot my laptop.
 

·
Registered
Joined
·
6,730 Posts
^ 30 minutes is fine. Though it's best to get all cells as quick as possible, in one 'swoop', rather than having time gaps between measurements, it's not super critical... There's actually a few things that can make it difficult to get a read on cell voltages just after a grid charge*, so we're already working from a sketchy baseline. We just hope enough pans-out, such as over the week of sitting, that we can feel confident enough that we're able to spot the outliers, or that there are none...

*If you take a closer look at the two charts above, for instance, you'll notice that values in the first chart are a lot lower than those in the second. I'm pretty sure that difference is solely due to overcharge during the grid charge; the 4 sticks in the second chart weren't overcharged at all, the ones in the first chart were overcharged for something like 2 hours... I've been thinking that something like that could make it difficult to read your data, as we won't know which cells were overcharged and which weren't, since you started with a somewhat unknown condition/state pack and probably didn't do a super-precise grid charge...

But, we'll see...
 

·
Registered
Joined
·
16 Posts
Discussion Starter #27
EQ1,
All values will be noted as 1.4xx unless noted 3xx
28,28,07,32,30,20
22,32,33,19,07,32
13,20,26,33,20,07
31,35,28,11,32,32
25,18,31,31,15,20
32,28,27,29,24,27
27,398,25,22,20,26
11,392,391,386,09,26
21,27,06,11,396,23
21,23,26,28,28,27
26,388,15,14,09,15
28,02,393,26,28,29
22,23,15,15,396,01
26,06,17,22,24,18
17,21,23,23,18,20
391,19,23,22,383,25
21,390,02,25,16,21
18,16,388,389,18,23
21,394,22,16,20,26
387,387,396,04,08,393

I agree the process will be different with each person and different equipment. I've never really studied batteries other than running albers tests on emergency banks. An old engineer once told me that you should put a small load for a moment to get the actual battery voltage. Not sure what he called the voltage above nominal. I'm not certain if doing that would give a more accurate account of actual internal discharge rate or not. I suppose the ima battery is being charged at all times and in the case of emergency batteries they would be carrying the load with no charging source.
BTW I'm using a cheap charger rb batteries brand I believe it is. Full pack voltage was 174vdc before I turned it off.
 

·
Registered
Joined
·
6,730 Posts
Here's a tabular summary of your data. How long did you grid charge for? It kind of looks like you you either didn't get a full grid charge or you overcharged quite a few cells, those with voltages under roughly 1.4V. Red values are outside 1.5 standard deviation. In general your values are kind of all over the place... We'll just have to see how the change in voltage looks over the sitting period...

85701


Here's a summary of cell values for a pack of mine, for comparison. Note though that these values are 9 1/2 hours after the grid charge, not 30 minutes, so the overall lower values are at least in part attributable to that longer rest, plus some of the narrower range will be in part due to that longer rest, but not a lot of it...

85703


My spread is 24mV, with an sd of 4mV; yours is 87mV and 15mV. But, in general, actually, your values look pretty close to mine.

If you have time I think you should get cell voltage readings after 5 days, and I think you should stretch the self-discharge period from 7 days to 10 days. I think that will help us get a better read on your pack... I have daily data for this pack, up to day 11, and we can compare the two.


Here's another visualization of the data, I think it's pretty self explanatory:

85704


I'm guessing your low cells are actually the ones that overcharged the most. Your 3 high cells, 1.47V, are likely 'high IR' cells, probably not debilitating but something to keep an eye on or out for.

My grid charge was, in comparison, much more even. I'm guessing the degree of variation in your data reflects the degree of imbalance among your cells. Some cells were completely self discharged and took longer to charge, while some weren't. The result was an uneven amount of overcharge and thus uneveness in the amount of voltage depression that happens after overcharge...

It's actually kind of a new thing to me, to see what a big difference only a little grid charge over charge can make in terms of resulting resting voltages, I think also a reflection of self-discharge/loss of charge in the process. Compare my earlier graphs, the 120 cell one vs. 24 cells that didn't get over-charged...
 

·
Registered
Joined
·
16 Posts
Discussion Starter #29
Your pack looks alot "tighter" than mine for certian. I charged for around 30 hours I would say. I now have 57mV spread. SD is around .01775 according to a calculator I entered the readings into.(Don't have internet or a license for excel).
All values will be noted as 1.3xx unless noted 2xx
39,41,10,49,43,299
06,46,45,05,22,47
293,04,24,48,08,09
36,50,43,21,49,47
18,11,49,49,21,06
47,42,42,46,36,46
47,294,46,45,39,46
299,21,20,19,296,49
23,49,05,08,14,42
38,22,48,49,48,43
44,17,299,298,17,00
46,09,21,45,49,50
11,40,05,00,20,15
49,01,40,49,50,40
37,44,50,48,40,44
11,41,47,46,18,50
43,18,20,49,39,43
42,37,16,24,40,45
43,14,45,37,42,49
20,19,14,02,03,18


Cells 14-20 are actually listed backwards from what you asked for. I noticed this today when I was reading and trying to put them in the correct order for you. They are in the same order as the first set. I appreciate the help, I hate to take up too much time.
 

·
Registered
Joined
·
6,730 Posts
Cells 14-20 are actually listed backwards from what you asked for.
So are you saying that "S14" in my table above is actually "S20," and the rest of that row is flipped like that? If so I'll change that table... It probably doesn't matter too much in your case, but in general temperature makes a difference. Middle cells of sticks, and middle in terms of pack if you're charging the whole pack in its case, can drop voltages relative to others.

Your values above - what's the time period, how long had it been since the grid charge? I'll look at them tomorrow....
 

·
Registered
Joined
·
6,730 Posts
I took a look at your second dataset, and it looks odd. There's a 'discontinuity' in the distribution that doesn't look natural. Are you sure your measurements are reasonably consistent and accurate? Maybe DMM battery was low, or maybe you mixed up some data? Here's what the distribution looks like:

85709


It just doesn't seem likely that that kind of discontinuity would occur naturally. I haven't looked at a ton of data like this though, so I'm not positive...

Also, get back to me on my previous questions.

Also, next time you post data can you type in the values as I described earlier, such as 1432 for 1.432, 1390 for 1.390, etc.? It should make it easier to spot potential data entry errors when you're creating the dataset, and it will eliminate the possibility of me making errors inputting data - because I can simply import it as-is...
 

·
Registered
Joined
·
16 Posts
Discussion Starter #32
Correct in the s14-s20 row flip-flop. The readings were at ~4.5 days. Took a look back at my readings and the data is correct. There were no measurements between 1.324 and 1.336. I think that's a little off just based on odds. Certainly I can type it all in, didn't realize exactly what it was being used for in the beginning. I appreciate the help along the way.
 

·
Registered
2000 Insight
Joined
·
24 Posts
So I have had my 2001 m/t insight for about 16 months give or take. When I purchased it, the owner had the clutch switch wired to use the battery on demand because it was weak. I disabled the battery immediately and started driving my 236 mile daily round trip saving loads on gas. Later on I bought a cheap grid charger, did a few discharge charge cycles, then tested. I loved the extra power and braking assistance but could only get maybe 30-40 minutes of a drive and no hard acceleration before the Ima light would come on. I average usually around 54mpg lately with a commute of around 60 miles round trip. I'd love to have some assist up the hills and be able get as much as I can out of it.

So I'm thinking about doing some charge/discharge cycles with a hobby charger. Question: Is it worth it? Is it really alot of assist? Is it likely I will be able to recondition? I know I'll need some sticks from another pack most likely to build a balanced pack. Should I just go bumblebee? Going that route and spending that much does the gas savings/happiness from extra assist to cost ratio even come close?
Also suspension... Killer3cylinder springs reviews anyone? I'm planning to build a hitch for a bike rack and maybe an ice chest and looking for something to help the rearend. Couldn't hurt as it rides like a tank as is.
Thanks in advance all.
In Arkansas btw
So I have had my 2001 m/t insight for about 16 months give or take. When I purchased it, the owner had the clutch switch wired to use the battery on demand because it was weak. I disabled the battery immediately and started driving my 236 mile daily round trip saving loads on gas. Later on I bought a cheap grid charger, did a few discharge charge cycles, then tested. I loved the extra power and braking assistance but could only get maybe 30-40 minutes of a drive and no hard acceleration before the Ima light would come on. I average usually around 54mpg lately with a commute of around 60 miles round trip. I'd love to have some assist up the hills and be able get as much as I can out of it.

So I'm thinking about doing some charge/discharge cycles with a hobby charger. Question: Is it worth it? Is it really alot of assist? Is it likely I will be able to recondition? I know I'll need some sticks from another pack most likely to build a balanced pack. Should I just go bumblebee? Going that route and spending that much does the gas savings/happiness from extra assist to cost ratio even come close?
Also suspension... Killer3cylinder springs reviews anyone? I'm planning to build a hitch for a bike rack and maybe an ice chest and looking for something to help the rearend. Couldn't hurt as it rides like a tank as is.
Thanks in advance all.
In Arkansas btw
I created an equalizer box that uses 10k resistors to simulate the voltage signal that IMA computer gets. If it reads all equal voltage, the IMA light stays off. You need a spare battery loom to get the plug that connects to IMA module. Works great for me. Been using it for 6 years. Send me an email and I'll send you more details if you want. [email protected]
Also, on rear suspension, a pretty easy modification made my Insight ride like a lexus. I cut 1 1/2" off rear rubber snubbers, (increasing travel before hitting bump stops which are actually overloads, too) and installed a spring spacer in rear coil. You wouldn't know you're in the same car.
 

·
Premium Member
2001 5S "Turbo"
Joined
·
10,568 Posts
Please include your Location in your Profile, as ALL G1 Insighters have done.
 

·
Registered
Joined
·
6,730 Posts
Note: I had to revise the 120 cell self-discharge graph I posted in post #24, there was a couple errors. The X-axis showed time in days since end of grid charge, but the "Day 1" data actually corresponded to only 9.5 hours after the grid charge. Also, the DMM I used measured low by -0.011V vs. a 2.500V reference, so I should have added +0.011V to measured values. But instead I inadvertently subtracted, thus, the values in that graph are 0.022V lower than actual... I also added a value for time zero - end of grid charge (~1.43V, this is just pack end voltage divided by 120 cells, the next value is similar, a little less than 2 hours later). Revised graph below:

85744
 
21 - 36 of 36 Posts
Top