Gen II Prius Individual Battery Module Replacement

20 mV

 

20mV/cell?

 

Yes. The idea is to stop the peak detection from operating forcing the charger to input the 7250 mah charge before cutting out.

John

 

-

 

Really not sure where to post this. I'm working with Gen3 modules for a Gen2 Prius, but most of what I've done so far I've learned how to do from reading this thread and the questions I have at this point are really a continuation of what I've seen here. I’ve also posted this to the Gen3 modules in a Gen2 Prius thread because that’s my ultimate goal. If it should go somewhere else or if I'm breaking any posting rules, would someone please let me know? Thank you.

I have read this thread “Gen II Prius Individual Battery Module Replacement”. 40+ pages…took a while. Some parts of it I have read three or four times and even printed some of the posts for future reference. My sincere thanks goes out to John(Britprius), Eric(ericbecky), JeffD(jdenenberg) among many, many others. You seem very knowledgeable and show a sincere desire to help others. You have encouraged me to jump into working on the battery pack of my Prius.

I own a 2005 Gen2 Prius with about 145000 miles. My son has been driving it in Los Angeles for about a year and a half and in early May it developed what was dealer diagnosed as a bad HV battery pack and a bad 12V auxiliary battery. P0a80, c2318, c1253. New battery pack at $3300, new 12V at $300 both prices included installation. I had replaced the 12V with an optima in 2008. At that time there was no optima direct replacement with the “pencil” posts so I bought a kit from someone to make the conversion possible. As luck would have it, the person I had bought the Optima conversion kit from is in Sylmar, CA, a suburb of Los Angeles and only a few miles from where my son lives. He has a service that replaces bad modules in Prius HV batteries and he came to my son’s house to do the repair. Cost was $450 which includes up to 5 module replacements and $40 for each one after the first five. The repair was done, one module was replaced, and the car is now back on the road. I don’t know what the process entailed but it was made very clear to me that this was a shorter term fix rather than a longer term one. 6 months to a year vs two to three years. Seemed to be an honest guy who was concerned that my expectations not get too out of line with reality. Based on his findings, he did tell my son the HV battery probably still has a lot of life left in it, but nothing that can really be guaranteed. Don’t know how he made this assessment. I was just glad for my son’s sake that the Prius was back on the road.

I have been waiting for this battery pack to go bad because I was really looking forward to learning how to replace the old one with a new one. But the car was in L.A. and I am in Scottsdale, AZ so I just didn’t see that happening. But after reading this Prius Chat thread, I would now like to replace bad modules with good ones instead of just replacing the old battery pack with a dealer supplied new one. And if I’m going to be replacing modules, I may as well completely replace all the Gen2 ones with all Gen3 ones.

I can take all the time I want. I have an air conditioned garage to work in, almost a necessity here in the Greater Phoenix, AZ area, and am retired so I can take all summer if I wish. I rebuild old Volvos and am a pretty fair garage mechanic. My budget process works something like: $3300 was the dealer solution less $450 repair and $118 dealer diagnostic leaves quite a balance for a used Gen3 battery pack and whatever chargers, high voltage gloves, insulated nut drivers and anything else that may come up.

So I got the Prius back out to Arizona and bought a used Gen3 battery pack and have been working with the modules for a couple of weeks now. I am using (2) X4-80’s and an “ePowerBox 30A” power supply (540 watts), both by HiTec. I started out with one X4 and quickly realized that it was going to take at least two of them to get this done sometime this lifetime.

I started out doing a load test using an automotive headlight. I used an older DMM to measure amperage and a newer DMM to measure the voltages at the end of the load test and just as soon as the load was removed from the module. Pleas see the photo of the setup I used to do the test. The white switch is just an old wall switch I had in the garage and the headlight is a high/low headlamp from an old Volvo. The load is 55w+35w or 90w total.

I really don’t think I did this part correctly. Capturing the amperage reading was easy. But the voltage readings were moving pretty quickly during the test. I used an iPad timer (one minute) and caught the voltage at the end of the test, “V1-end load” is how it is labeled in the spreadsheet that I have used to try to compile the data collected during the load tests and during the module cycling. At the end of the test I opened the switch to remove the load and then recorded the voltage, “V2-end open”. During the first test, I did’t realize that my new DMM had a hold feature and the voltage readings were at best an approximation. During the second test I was able to press the “hold” button on the DMM as soon as I opened the switch to end the test and the readings were a little more reliable. Having said that, when I looked at the data I collected, there are some readings in the second test that I cannot make sense of while those of the first test seem to make a little more sense.

Also, in post #604 by John(Britprius) of this thread “Gen II Prius Individual Battery Module Replacement”, there seemed to be a bit of a different methodology than the one in the post immediately following, #605 by (uart) and in post #5 by Marc(marcs_carhole) in the thread “Prius Battery Rebalance”. John seemed to suggest reading the voltage used in calculating the internal resistance before the load test and uart and Marc seemed to suggest reading the voltage immediately after the load test. I tried both and got such different results, that I just went with uart and Marc. The results seemed a little more in line with what might be expected.

I would really like some help on this point. The load test process was for me the most problematic aspect of what I’ve been trying to do. And I think it may be one of the most important. I am less than confident in the results of my tests. The first test was done before doing any module Dchg>Chg cycling and the second test was done after the first sets of Dchg>Chg cycles were completed. I’ve tried to lay out the chronology in the way the spreadsheet has the various load tests and series of cycling labeled using the dates showing when things were done. Sorry for how cumbersome the spreadsheet has become, but it seemed like a good idea to try and bring as much information as possible all together in one place. At least it seemed like a good idea at the time.

The Dchg>Chg cycling seemed to be a little more straight forward. I had a tough time tracking down exactly what the program variables should be, but after reading many, many posts, I found that the areas of concern were the current limit, delta peak and capacity cut-off. Thanks to everyone who posted the settings they used. They were really, really helpful.

I decided to stay very conservative and hold the current limit to 2.0A. I did this mostly to keep heat buildup down as heat can be a real problem in the Phoenix area. The first sets of cycles were done during an unusually mild part of May, but I still registered temps of 99F (module #15) during the very last phases of a charging cycle. Adjacent cells were 86F and the garage was at 81F. My evaporative cooler was not on. I turned it on and 15 minutes later the module read 97F and the garage 79F. No effort was made to specifically cool the modules at this point.

Delta peak was left at default which is either 5mV/cell or 7mV/cell. Couldn’t really tell. I believe this setting was responsible for some of the problematic readings encountered during these first sets of cycles. This was changed to 20mV/cell for the later sets of cycles after reading post #634 by John(Britprius). He also suggested a time limit cut-off but I didn’t have clue to what it should be so I just left it off. In later post(s) it was suggested to set this to 720 minutes and that’s probably what I’ll do going forward.

Capacity cutoff was initially set at 7500mAh based mostly on the reasoning of John(Britprius) in his posts covering this setting. I later changed this to 7000mAh at the same time I changed the Delta peak thinking that this 7500mAh may have been too high based on settings used by others that seemed to meet with more success than I was having. If I had it to do again, I would have left this at 7500mAh. I now feel that the improvements I saw in my results were due to the change in Delta peak. In fact, I may have done a disservice to myself by lowering to 7000mAh by not getting all the capacity I could have into the weaker cells within the modules. Don’t really know how to test for this other than by running another set of cycles and I don’t really want to do that unless someone thinks that it could make a significant difference in the capacity I ultimately wind up with.

The last setting that I addressed was the discharge amperage. I had initially set it at 1.0A thinking it would take a little less time than a lower setting. I’m still quite fuzzy on this particular setting and didn’t know if it was adversely affecting my results or not. So I lowered it to 0.7A which seems to be the norm on most of the posted settings if I remember correctly. I have thoroughly given up on doing this process quickly so if things take there own sweet time, that’s just the way it’s going to be.

It may just take all summer, after all.

So I’m now in the final process of re-running cycles on modules 2,4,6,8,10,11,14,17,18,19,21,22,24,26 using the revised settings. Any module that returned a Dchg capacity of less than about 4600mAh after 3 cycles was chosen for a rerun. 4600 is a completely arbitrary number that I just felt right about after looking at the results of the first sets of cycles. And after looking at the results of the latest sets of cycles, I may want to revive that up a little and at the same time change the capacity cut-off back to 7500mAh. I’m a little worried that I may be running too many cycles and somehow affecting the longevity of the modules. But what the heck, this is a learning experience as much as anything else. So reduced longevity at the expense of learning something worthwhile is a good trade. Hopefully, someone with expertise can help guide me a little here. My lowest Dchg capacity of the latest reruns is 5228mAh (module 10). What I don’t know is if this reflects the health of the least healthy cell in the module or if it’s a reflection of something else. I tend to think it’s not a reflection of the least healthy cell in the module. But again, maybe someone can give me a little expert guidance. I would certainly appreciate it.

My last question concerns the return of a Dchg number at the end of cycle 3 that is just slightly lower than the Dscg number of cycle 2. I’m not quite sure how concerned I should be about this. And I can’t even begin to outline any sort of reasoning for it. My impulse is to think it’s important but without knowing its cause or what to do about it, I’m stuck.

Sorry for the long, drawn out post. I hope I’m not violating any rules. But I’ve never posted on any forums before so I’m quite the NOOB at this. It’s just that so much territory is covered during the module “reconditioning” process, it’s hard not to be overwhelmed at times.

And once again, my very heartfelt thanks to everyone who has taken the time to help those of us out who otherwise wouldn’t know where to turn next. You really make a difference.

 

The Dchg>Chg cycles for modules 11,18,23 and 25 finished last night. These 4 are ones that were below the arbitrary 4600mAh number but that I had originally thought were pretty good until I saw the results of some of the others that were returned 5/31,6/01,6/02. The new capacities are higher and in line with what I was hoping to see except for the nagging question of occasionally seeing a slightly lower 3rd cycle.
module 11 from 3945 to 5390 (slightly lower 3rd cycle)
module 18 from 4177 to 5251 (slightly lower 3rd cycle)
module 23 from 4341 to 5288
module 25 from 4571 to 5454
I still don't know why these increases were returned this go around and I still wonder if increasing the capacity cutoff from 7000 to 7500 would be advisable. And what about doing another set of 3 Dchg>Chg cycles on any of the remaining modules that now stand at an amount below 5000mAh? I'm thinking of modules 7,9,12,16,20 and 27. And maybe 15, too?
I mean, really, the summer's just getting started....
Thanks again
John

 

Hi John.
I have not yet had chance to digest all of your work, but one thing I can see is that your test rig has a problem .If I am following it correctly.
The meter on the right in your picture I believe is measuring volts (by following the wires). This meter should not share any wiring with the rest of the circuit. It must be connected directly to the module terminals without any part of it's leads carrying load.
The way it appears to be wired part of the resistance of the wiring will be added to the resistance of the module making the readings high.
I will come back with more observations tomorrow after studying your results.

John.

 

John,
You are exactly correct. I have added another set of leads to go from the DMM on the right, the one measuring the voltage, directly to the posts on the module and leave the other hookup as it is. I was just trying to have only one set of wires to attach to the posts on the modules. Glad you caught this.
Thanks for your help.
John

 

You should find the measured internal resistance now somewhat lower. It is quite likely the resistance of your wiring, and connections was much higher than the module resistance.
Although not given in my post on measuring resistance possibly the best module voltage to do this at is 7.2 volts (the nominal module voltage). The discharge voltage curve at this voltage is almost flat so that any voltage drop from discharge that takes place while taking the readings will be negligible.
Both methods given in the posts are valid, and both have there pros and cons.
Measuring voltage drop adding a load means the voltage continues to drop steadily.
Measuring the the voltage rise suffers the reverse effect as the module recovers voltage steadily rising when the load is removed.

John

 

I did not realize that the internal resistance of the module was so low. I'm now rethinking the use of alligator clips on the end of the DMM leads. I thought the alligator clip connection would suffice but perhaps not.
I equalized the modules last night by placing them in parallel and they now measure 7.991V. I remember the discharge graph you are referring to but don't remember where it is. I'll search for it and try to tell where along the curve this voltage lies, (I do remember the "cliff" aspect of the curve below about 6.0V). Anyway, I'll run a load test today and see what the results are.
Thanks again for your patience and help.
John

 

John. A voltage of 8v represents a charge of over 90%. The quickest graph I could lay my hands on is below, and is for 4 cells so voltages need to multiplied by 1.5 to get to Prius module voltage. You will see that at (1.2 volts per cell) 4.8 volts for the 4 cells the discharge curve is all but flat. At 5.32 volts (4 cells the level your cells are at) it almost fully charged.



John

 

If you measure the voltage at the module terminals, the alligator clips are not involved in the resistance measurement. The DMM has a very high input impedance.

JeffD

 

Thanks, John
I found another graph in post #658 for an 8 cell module. The flat part of that graph seemed to begin about 1.26V/cell or between 7.5V and 7.6V per 6 cell module. So I'm going to run the load tests this morning and see where I stand. If you don't feel that any additional attempts to increase module capacity by running additional Dchg>Chg cycles with the capacity cut-off raised to 7500mAh from 7000mAh that I last used, then perhaps I'm farther along than I thought. Just a reminder, these are Gen3 modules that I'll be transplanting into my Gen2 battery pack.
Or perhaps I'm not really thinking this thing through.... If the results of the load tests show appropriately low internal module resistances and the variance between the modules is not too great, then there's no need to try for additional module capacity? This is all really new territory for me so the concepts are still quite fuzzy around the edges.

John

 

Thank you Jeff.
I appreciate no longer having to obsess on the clips. I see that you are an EE Professor. I just wish I had been able to take your classes. My electrical knowledge is pretty basic. I can trouble shoot automotive 12V circuits with a continuity tester and a simple bulb type power probe. One of the reasons I like rebuilding old Volvos is they are truly simple cars to work on. And I can wire a house or add a 220V sub-panel to a garage easily. I'm comfortable working in a hot main or sub-panel to add circuits and so on. I just try to use the right tools, take my time and be very, very careful. My electronics background is really rudimentary at best. It amounts to being familiar with ohms law and occasionally opening up a device and looking for obviously bad (darkened) resistors or leaking capacitors on a PC board and just replacing that board. I tend to think of terms like impedance and resistance as meaning the same. Maybe one applies to ac circuits more than the other. My reason for going on about this is I don't want anyone to think they are explaining things too simplistically. I can't tell you how many times I've suffered because I pretended to understand more than I really did. I'm totally over that, and now just enjoy adding to my understanding of things. I really value your input.
John

 

Well I put the battery back in today and had a couple of minor hiccups. First after sitting 4-5 weeks the 12 volt was low so after using a jump pack I still could not get the car to start. Finally by inserting the fob in the slot I was able to get the car to start. No lights and no battery issues yet, but apparently my smart key is no longer working correctly. I changed the smart key battery but still no luck and then as I was writing this I thought about the little stinking button under the dash.....now everything is working!!! Thanks for the help and maybe no more issues will appear.

 

I think your capacities are about where they should be, and doing another charge discharge will accomplish little. If you charged to 7500ma you would probably find on the last 4 modules your capacity could be above 6000ma only because you would be fully charging them against a lower charge level previously.
I wish we had some hard data on the charge efficiency of these modules. Most manufactures give a level for NiMH as being under 70% efficient for a full charge cycle, but many things can vary this including temperature, charge rate, and charge level.

John

 

Thanks John. That's good news. I did run another set of load tests and I'll post those as soon as I have them entered in my spreadsheet.
John

 

Here's the results of the latest load test run today, 6/06/15. Hopefully there is a little less operator error in these numbers.
The R-calculated column doesn't throw any red flags that I can see. At least compared to the internal resistances of my current Gen2 HV pack of 0.019/block (0.0095/cell) that Techstream returns. I seem to recall reading somewhere, maybe in this thread, that the numbers reported by Techstream really aren't that reliable/useful? Anyway, non of them is over this amount even though #22 comes close. That does correlate with #22 being the module that I first thought would be bad and have to be replaced.
The voltage drop, Vs-V1, based on Vs, voltage reading taken before the test starts, and V1, the voltage reading reading taken at the very end of the test (before the load is removed from the test circuit), doesn't show (to me at least) any huge number that would point to an obviously bad module. However, V1 is still quite difficult for me to capture. Seems the old hand/eye/brain coordination leaves something to be desired. That's why some of the results are labeled 2nd test and 3rd test. I just blew it on the first try and had to do it again.
So. I'm going to discharge these modules down to about 7.2V or as close as I can get per John's observations in post #889 above. I guess I'll just use the load test circuit for this. Then I'll re-equalize the modules. Then I'll run another load test. Hopefully, practice really does make perfect.
And maybe, just maybe, I'll actually be able to use all of the modules from this originally condemned Gen3 pack in my Gen2. Wouldn't that be great.
John

 

Started discharging module 2 as a test to see how long it would take to discharge to some reasonable level, somewhere between 7.2V and 7.5V. Started at 7.969V, after 5min=7.640V, 10min=7.610V, 15min=7.566V 20min=7.495V. It recovered to 5min=7.700V, 10min=7.710V. I stopped taking readings at this point thinking that the recovery process was close to being over.
So after 30min I effectively had gone from 7.969V to 7.710V. So perhaps after an hour I would be somewhere in the 7.2V to 7.5V range that was my goal. 28 hours for the 28 modules. This evening that seemed like a pretty long time. Maybe I'll feel differently in the morning.
I'm thinking of doubling the load from one headlight to two or finding a higher wattage bulb. This will allow for a faster discharge rate but I don't want to damage anything by doing this. If someone has any advise on this point, I would really appreciate any input.
I also considered using the X4-80 to Dchg to 7.2V but in reading the instruction book that came with it, I couldn't see anyway to then end the cycle at that point. Or even to allow some small Chg rate to be met and then the cycle shut down. Perhaps use the safety timer to shut the cycle down. I could do a little experimentation to come up with the appropriate time limit perhaps. But I'm not terribly familiar with this charger and am hesitant to experiment without some adult supervision.
My impulse here is to play it safe and just use another headlight to increase the discharge rate, do this one module at a time and just spend whatever time it takes to get the job done safely. The DMM measuring the amperage was showing 5.75A during this discharge run. I don't know how large a load to allow. I'm using a 90watt total rated headlight now (rated at 12V). I'm wondering if I can safely increase this to a 150watt (12V) headlight and maybe do this in 45 minutes instead of an hour?
Hopefully the old sub-conscience will come up with something after a good nights sleep.

 

Hello,

This is a great thread and I am amazed at the level of detail and the patience that everyone has for this important work.
I am going through the learning process and I have replaced one of the 28 modules and have had some decent success without balancing but I recently am getting random IBD codes so it is clear that I need to take things apart and properly exercise the batteries.

Here is a summary as I understand it. I will be using two iMAX B6 OEM Battery Balance Chargers.

- Measure and record the individual module voltages. (What are the exact pass/fails numbers to look for here?)
- Discharge each module for the first time. This first discharge may identify questionable modules. (I am not sure it the iMAX B6 or headlight should be used, and I am not sure exactly what the time, current and/or voltage limits should be for this first discharge).
- Charge each module for the first time. (Again, I am not sure of the electrical limits or the best length of time to use for this first charge. Also I think there are warnings or "charge failures" to monitor that may identify questionable modules", if there are, what are they?)
- Discharge each module for the second time. (Are the same parameters used as applied in the first discharge?, Is the charger or headlight used for this step?)
- Charge each module for the second time. (What is done for this step that is different than the first charge?)
- Discharges each module for the third time. (What is done differently?)
- Charge each module for the third time. (What is done differently?)
- Connect the positive and negative terminals in parallel for all 28 modules. (What gauge of wire is used, how long doe the batteries remain connected in this manner, and what electrical measurements are desired during this process?)
- Assuming that any questionable modules have replaced during the above process, re-install the battery pack. (What advice is there for the first drive test and is there a "breaking in" period?)

Please feel free to tear this poorly understood procedure up.

Thank you.