Civic hybrid 1st gen IMA won't recognize new battery
 

I got a used pack from the salvage yard since mine wouldn't hold a charge even with the grid charger. I charged the "new" one with the grid charger, held full voltage with no load so I installed it.

Even after pulling fuses and disconnecting the 12v battery to try and reset the system it still refuses to see the fully charged pack.

It did however manage to start itself a few times off the IMA system but not anymore.

Only thing I can think of is maybe what looks like a relay isn't connecting the pack to the system because the bus bars that are always live when the switch is on shows fully charged voltage and I'm reading 60 to over 200v at the main terminals as engine rpm goes from 1500-3500.

I got lucky and managed to find the new pack without a core charge so I could try swapping components. I'm trying to get this running by Monday so any advice would be greatly appreciated, thanks.

I figured it out along with the details on testing the sensor leads from the connectors. Turned out the circuit that runs through the pack along each battery was the problem. The only thing that I can assume about its purpose is to let the computer know if a battery exploded or possibly overheated enough to melt the metal strip shrink-wrapped against the sticks. Should be about 20ohms fyi.

Here's the original pack opened up: http://i.imgur.com/DaovBUp.jpg
And here's the new pack: http://i.imgur.com/PdlO1Y3.jpg

The problem was with the wire wrapped in red plastic that can be seen cut in the first.
My car doesn't like that circuit apparently.

I would be interested if anybody has any ideas why someone would disable what I believe to be a security feature and why it actually needs it that way.



I also figured out how to test the values coming from the three thermocouplers inside the pack from the outside, if anyone's interested I can label the picture I took of the connector.

I narrowed down the pin on the connector for each of the twelve (I'm assuming since their resistance did change with finger temp) thermocouplers on the end plate but you have to have the top off to test those since they terminate on the stick lug next to them.

Honda stopped using the PTC strips sometime around 2003-2004. The wire cut in the PTC circuit is part of the technical bulletin that goes along with an ECU flash for pack replacement.

The PTC strips were pure paranoia. They protect against a massive meltdown, which has never occurred. The 3X thermocouples provide sufficient protection.

Bingo kieth. Honda wasn't sure how this new tech was going to react to people living in the hotter climates like Death Valley, Phoenix Arizona, Las Vegas, etc etc. They went overboard with the security so they made sure nobody was killed. They removed it once it became clear it was pointless and a waste of money.

That's good to know it wasn't some hack job! If they've flashed the ECU does that also mean they did the software update to reduce the IMA power or did they only do that to the second generation?

HCH1 and HCH2 use completely different ECUs/Programs. The "patch" that tweaks the IMA effectiveness only applies to the HCH2. There may or may not be something similar for the HCH1, but I don't know about it.

HCH2 IMA system has 50% more motive power, 10% more total batteries/voltage with 15% less capacity and improved internal resistance. The increase in total motor output taxed the battery system more than anticipated even with the battery pack improvements. The end results were abnormally high battery pack failure and altering the ECU (technically the BCM, I think) programming to be less demanding of the battery for most routine driving. Heavy acceleration will still get the full 20hp of the HCH2 IMA motor provided the battery can deliver it, but for less time due to the more conservative battery management.

I was going to start a new thread, but this topic is so close to mine that I'm going to start by asking here.

2005 Honda Civic Hybrid owner. After a thorough battery reconditioning and a final top off charge on every stick, I've made it to the exciting and worrisome stage of reassembly. I let it sit for 2 weeks to let the voltage stabilize after the final top off charge, installed it, and my gauge cluster is showing an empty battery, while I'm measuring 159.8 VDC at the battery (8V/stick).

Is the computer just not liking the battery because the voltage is higher than the factory parameters allow for? I put a "full" charge into the battery before pulling it out for the reconditioning process, and each stick measured right at 7.80V at the time of disassembly (156.0 VDC for the pack). This makes me assume that perhaps the computer says the maximum voltage is 156. Do I simply need to drain this thing down so my output is not above that?

Because I imagine an investigation into the details of my refurbishment will be precipitated by my first post, here is a little info on my process.

I topped off the battery in car, by using the little trickle charge the IMA at or above 3k rpms trick immediately prior to disassembly, to capture what data I could about what the car considers to be a full charge.

It came out right at 156 VDC for the assembly, and each stick measured right at 7.80 VDC upon disassembly.

I then documented the discharge mAh of each stick during a 10A drain to .8V/cell. The values were pretty pathetic, ranging from 193 to 2748 mAh per stick. A trio or better of 10A-discharge/2A-charge cycles per stick netted me an average charge input of 6642 mAh per stick, with the maximum variation of only 0.04% between sticks.

I monitored and documented the voltage decline over time before cycling each stick again to verify holding capacity. After about 2 weeks, each stick was in the range of 7.95 to 8.00 VDC, and they all discharged right around 5000 mAh. A pretty good looking bunch if I say so myself.

As my final step in the refurbishment process, I quickly went through the entire lot twice in a staggered fashion doing a reflex (996 milliseconds charge @4A, 4 milliseconds @10-16A discharge, every second) top off charge to ensure an even voltage profile across the entire pack. This left the assembly at nearly 170 VDC (>8.3V/stick). I felt that letting it normalize over 2 weeks to <160V (<8V/stick would be smart, as I felt like >20V over the 144V nominal was likely to piss of the battery management computer.

My suspicion at this point is that I have pissed it off, and I'm going to have to load the battery to get it below 156V.

Sorry for the novel, but the forums I regular will publicly hang you if they don't get the full scoop prior to a request for advise.

Will the car start? The car might just need to relearn what the SOC is by driving it a bit.

I was planning on giving it a spin straight to an emissions inspection station, but when the gauge cluster didn't agree with what I know about the battery, I stopped there. I'll go plug everything back in and give it a shot this afternoon. If it starts I'll try to drive it and load the battery a little bit.

Are the P14XX and P1600 codes somrthing I will need to pay someone with a scan tool to reset now that she's got a box full of happy cells?

If the IMA light is out, the car recognizes the battery. The BCM just has no idea of its state of charge as resting voltage does not correlate to state of charge, so it shows zero. If the IMA light is still on, you have other problems.

There are no issues with shoving a fully charged pack in a car. If you can, it's best to let the voltage stabilize over 30-60 minutes (maybe overnight), but two weeks is excessive.

Disconnect 12V for 30 seconds. Reconnect. Start and rev to 3500rpm and hold until you read full bars. You will need to drive it as you will show DTC not ready when you go to emissions, which will be the case no matter how you reset the codes.

Given the atrocious state of imbalance your as-removed testing showed, I expect any improvement will be temporary and you will need regular grid charging/discharging.

Did you confirm the sticks could sustain a 100A load for 10 seconds and remain above 5.4V?

NiMH voltages don't mean a lot. They can give indications, but there is no objective relationship due to other variables.

You have already invested a lot of time in this project. I have done the same over 2 HCH1 packs and 6 HCH2 packs. If you want to maximize the value for your efforts, next time you are looking to re-balance your pack, do it over a weekend with pack grid charging and deep discharging to < 1V (yes, 1V for the entire pack) with 2X 100W bulbs in series. Yes, you will reverse most of the cells. No, it won't hurt them (will happen at current under 1A). Yes, you have already reversed at least 1 cell per stick at far higher current when discharging to 0.8V/cell. I have watched it happen at even lower currents. The speed of weak cell depletion and polarity reversal under high/medium load when discharging to <1V/cell is astounding.

There is a mountain of evidence of others (including me) that have done this with very favorable results. My capacity recovery has been 20-25% for a given load without exception.

Here is a before and after comparison between 2 discharges to 132V on HCH2 packs. First one is following a topping grid charge to peak voltage at 350mA. The discharge was continued to under 2V followed by a topping grid charge to peak voltage at 350mA. The second data set is the same discharge (500W halogen) to 1V/cell:

https://docs.google.com/spreadsheets...it?usp=sharing

Elaboration to help you understand why resting voltages and pack SoC as-removed vs. as-installed doesn't matter much.

The pack monitors 11 voltages: 1 total and 10 stick pairs. It also monitors current.

When assisting, for a given current, the voltages must remain above minimums. These voltages vary with current. When regenerating, for a given current, the voltages must remain below maximums. SoC is computed based on this range of operation and iteratively adjusted. It doesn't mean a lot.

At the end of a positive recal, 100% SoC indicates ONE of the 11 voltages hit max at the charge current.

At the end of a negative recal, 0% or near 0% SoC indicates ONE of the 11 voltages hit min at a given discharge current. Additionally, if any of the 10 stick-pairs deviate by 1.2 volts at any time, it triggers a recal. As long as your stick pairs are within 1.2V, they really don't matter.

I just refurbished an IMA battery for my 2003 civic hybrid. Upon installing it in my car I have the same two codes present: PP1637 and P1569.
Also, the IMA and BATTERY lights are on. If I rev the engine the BATTERY light goes out, which I am aware will happen. The bars of charge are full. We took the car for a ride up a long hill that should have depleted the battery and then refilled it on the way back down. But the car appeared not to recognize the battery because the level of charge never changed and there was no assist.
This battery also had a modification to it that I noticed when dissassembling it. Upon taking off the end where the electronics are attached, I noticed one of the orange wires was no longer connected to a PTC strip. It appeared it was cut. Thinking this was done by a previous owner we reconnected it. I now know that this cut wire was a technical buletin update to the battery so will disconnect later today.

Does anyone have any ideas on why my system did not recognize the refurbished battery? The only other info I can think of sharing right now is that the battery went into the car right after the final top off of each stick. So it went in at 168Volts. I also swapped in around 10 sticks from another battery I recently purchased. All sticks were individually discharged and charged several times prior to assembly.

Report back after you restore that wire to the as-removed condition. PTC strips were eliminated sometime in '04 and BCMs were patched. A patched BCM will not work correctly if the PTC circuit is present, and an unpatched BCM will not work correctly if the PTC circuit is not present or is not correctly spoofed.

Help! I get code p1432 after installing good battery pack!
 

Hello, new guy here. I bought a 2004 civic hybrid with rocketship milleage of 244,375 miles. The IMA and check engine light would come on even after resets with code p1433 battery deterioration. I bought a grid charger and canrged it several times to a peak of 179 volts, which I thought is was high but I don't know. The grid charging helped but when it came time for the battery to discharge and recharge in use the p1433 code would pop up.
I bought a "good" battery from the salvage yard and after installing it I would get an instant IMA light and I could hear the relays witching off. Code P1432 (Cell overheating '04) and code P1433 ( battery pack deteriotation) plus the mandatory check engine 01600.

I took out the "good" battery and set it to the side. but after reading this thread I figure it may have something to do with the cut wires but don't know.

The previous owner had cats and dogs and a blanket of fur formed in the upper row of sticks the original battery so I figure I'd clean it and put it back in since it still charges and assists with the IMA light on and still helps get 38.7 mpgs

With some luck the P1433 was the battery not geting enough air flow. I'll find out when I drive it tomorow.

Check individual cells in the pack
 

Maybe what you need is a way to get into more granular data. The battery is really a pack of batteries. Some in your old pack might have gone bad, and others in the replacement pack. So you need a scan tool that can measure the health of each of the cells. Once you know which are failing, if the battery chemistry is the same, you might be able to build your own "new battery out of the good cells in each pack. Here is a relecant link: https://www.scantool.net/forum/index...;topic=10442.0

I don't know if such a scan tool exists for an old car like a 2004 Civic Hybrind, but this is what one example looks like for the Leaf: https://m.youtube.com/watch?v=DdEyjjt52kU

Of course the failure coukd be something else... don't mistake me for an experienced expert. I'm just a guy on a bulletin board tryna help a little. Good luck.

james

Hi S Keith. Thanks for replying. Here is my updated information. I disconnected the red wire that I had originally found 'cut'. I put the battery back into the car. Upon starting the car there were no lights - no BAT light, no IMA light and no CEL. I let the car idle until the SoC meter was full. I then took the car for a drive. As soon as I called for assist, the CEL and IMA lights came on together. The BAT light did not come back on during my 1/2 hour drive. When I drove, it took the car up a hill and back down twice. Each round trip drained the IMA battery going up and filled it going down. However, I noticed on those round trips, and while driving around town afterwoods, that there was a behavior difference in how the system worked. There really was no assist, as I never saw much on the gauge and could only occasionally perhaps feel the extra umph. Also there was no regenerative braking. Whenever the charge lights would show bars, it was never more than 5. Even when braking going downhill on the hill I spoke of earlier. Finally, the SoC bars would go down and up even though there was no call for assist or for reason to charge. IT would be a slow motion type thing but was occurring.
The good news is that the BAT light is out. Makes it easier to test drive! Next I brought in the Honda Dealer for a discussion. I am curious as to whether the car needs a computer flash. Oh, by the way, the codes I now have are P1600 and P1449. So they changed after disconnecting the red wire from the PTC strips.
Anyway, the dealer said he would be willing to flash the computer. Since I have the carfax from when I bought the car, I checked it to see if the computer had ever been flashed before. There is no record of such an update. The dealer said he would need information regarding what battery I put in the car - such as the year of civic the battery came from, because there were a number of updates to the software. There would be a cost for flashing and he wanted to get the right version of the software so he did not have to charge for a lot of hours.
I explained that I had used the original case from our car but may have put a cover on it from another donor battery. He asked if I would be willing to put the original cover on and I agreed. I also am going to drive the car some more and wait for your reply to see if you have ideas on eliminating the two codes or if you think the flash of the computer is the path to go down.

The computer has been flashed. You confirmed this when you found the the cut PTC circuit upon battery removal. You re-confirmed it when you installed the pack with the circuit restored and related errors were shown. And you re-confirmed it one more time when you restored the PTC circuit to its as-removed state and related codes went away.

P1449 is a failed battery pack. You did not successfully recondition your pack. Unfortunately, too many people are duped into believing this is an easy and usually successful exercise when it usually isn't. Most "reconditioning" descriptions are woefully inadequate to establish that sticks will perform in the car.

Steve

Sounds like I do not need a flash done so I save the 90 dollars there.
Also sounds like I would be wasting time if I try to refurb the cells again. I now have 3 packs and 60 sticks, although some are obviously no good.

Do you suggest dropping the refurb path and purchasing a new battery? The car itself is is very good condition. 90K miles, 5 speed, new brakes and new struts. Really is not a lot more that can go wrong with it. It is more a matter of biting off a $3000. investment in a car that might be worth $4,000.

I would expect a collection of 3 packs could be used to make a single good one, but it's a lengthy process.

If you decide to go new, do NOT buy Honda. Check out Bumblebee Batteries.

Although I was able to obtain amp levels each time I charged each stick during the refurb process, and noted the voltages, I was not sure how to calculate the internal resistance of each stick. Would such information be useful? Can I still do it since I have the data all written down? Is there a formula I can use to do such?
I did contact BumbleBee today and I am looking at a purchase, but if calculating internal resistance for each stick tested can lead to a reconfiguration of my sticks (I have so many to choose from) then I will try that before making the purchase.

Although I was able to obtain amp levels each time I charged each stick during the refurb process, and noted the voltages, I was not sure how to calculate the internal resistance of each stick. Would such information be useful? Can I still do it since I have the data all written down? Is there a formula I can use to do such?
I did contact BumbleBee today and I am looking at a purchase, but if calculating internal resistance for each stick tested can lead to a reconfiguration of my sticks (I have so many to choose from) then I will try that before making the purchase.

Calculating a meaningful IR is difficult. Verifying performance at high current is a validation of acceptable IR.

You've given no details as to what you did or what equipment you have available.

Sticks need to be evaluated for total capacity, IR and self-discharge. There will be top performers, and all others are outliers to some degree or another. You select the best of the bunch, preferably all in the top-performer group and drive with glee.

I had done a lot of reading (at least I thought I did) and followed a procedure that was outlined by 99mpg (http://99mpg.com/projectcars/mimapackwhack/ There were several posts on Ecomodder that corresponded to the procedure and so read that material as well. I set up a chart and then used five Superbrain 977's and one superbrain 989 to test each stick. The tests involved discharging and changing each stick. I discharged at 4 amps and charged at 5 amps. The 989 would allow higher amperage but I thought it would be good to keep the amps all the same. The 977 can only do 4 amp discharge and 5 amp charge. I am now down to 3 of the 977's due to silly mistakes during the procedure.
I did 3 cycles on each stick. If the numbers looked good to me, I did not do additional cycles. Typically, if the charging produced consistent numbers between 6700mA and 7100mA while charging, and within 20% of that during discharge I considered it good. So the discharges were typically around 6000mA.
After the last charge of each stick, I sat them on a shelf for a week. Once the week was up I topped off each stick and note the additional mA as a percent. This was the discharge capacity. Typical sticks put into the pack had charge rates above 6700mA and discharge rates after a week of 13-16 percent.
Since I do not have knowledge on measuring internal resistance I can not do that test. If it is just math I think I can do it but would need a formula.
At one point in the process I became concerned about the bending of the PTC strip tabs because I was doing numerous assembly and disassembles of the sticks into a battery. Now I am not so worried since the PTC strips are out of the loop. It appears if the tabs break off it is of no concern.
The only other piece of equipment I have is a multimeter.

So looking at the numbers above and the procedure I followed, is there anything you would suggest I try prior to or instead of ordering a new battery.; in order to locate 20 usable sticks out of my 60?

That's the original and likely still the best guide. Can I see your data?

Are you saying your 989 is no longer operational? Bummer.

The 989 works fine although I must state I have not used it on higher amp settings. I will assume it can go 10 amps.

I took a photo of the data sheet but it is not very clear. I will redo the sheet and then attach it later today to another reply. Looks like there is a tool for easy image attachments...

S Keith:
I am sending the photos as a private message since I can not figure out how to upload them here.

S Keith:

Apparently I can not make a simple attachment of images. Can I send the two photos of the data to you by way of email?