Balancing question
 

I apologize in advance if I'm posting in the wrong forum, or if this has been answered in another place.

I took the battery apart and charged the 20 sub packs over this past winter. I recorded the discharge and charge values. I also recorded the final voltage and then the voltage it was a week to 2 weeks later.

Its time to put it back together and I entered all the numbers into excel. I then realized I have no idea how to determine which battery is good and which is bad. Can someone please help! The numbers to me say they are all good, but I don't believe this to be the case.

Thanks for the help!

I was hoping someone knowledgeable would answer your question because I'm curious too.

In my experience with Roomba batteries, simply knowing the charged and rested voltage along with capacity was not enough to determine which cells were bad. I believe you also need to know what the internal resistance is for each cell, and I'm less clear on how to determine this.

It would be good to know more details, such as what hybrid pack you are dealing with, and what the symptoms are that caused you to dig into them in the first place.

what are your units of charge value?
 

It looks like subpack #3 is bad by your data--all the rest settled down to 8.1-8.2 except that one dropped to 7.5, it's not holding.

i have done cell capacity checks on some batteries, the largest pack being 44 cells of lifepo4 with 100Ahr capacity.

For example i recorded voltage, current, temperature and time while discharging at 30 amps until the voltage reached 2.5V, then recorded same while charging at 25 amps until the voltage reached 3.6. After resting the voltage settled to 3.33-3.34, which is considered the 100% state-of-charge voltage for this chemistry. The cell capacity is calculated by the current multiplied by the time, e.g. 30Ax3.4hrs=102Ahr.

You need to get the specific upper and lower limits for your cell chemistry, the cell capacity rating, the charging procedure, etc to make sure you stay within the box and don't damage them.

There is not a real resistor in the battery that you measure, but a so-called internal 'resistance' can be calculated as the difference between the unloaded and loaded voltage divided by the load current. For example if a 12 V battery drops to 10V with a 100A load, then Ri is 2/100 = 0.02 Ohms. Some people swear by it, but i haven't found it as useful as actual cell capacity, especially when trying to determine expected range.

I have a 03 civic hybrid. When I purchased the car 2 years ago, I was getting 48-50 mpg.

I park my car during the winter months, and last winter I forgot to start the car the entire winter. I think this hurt the battery pack. Sometime during the summer I started to get IMA lights. the light would come and go. but the Umph that the electric assist used to give has weakened. Then the O2 sensors went out, and my mpg dropped to 40. I replaced all 3 o2 sensors this spring and it only brought the MPG to 43.

So I'm trying to rebuild the pack to get my MPG back up and get rid of the IMA light all together.

It is also worth noting that last year the Honda dealership took it upon themselves to update the firmware on my car. And I can't shake the feeling that in doing so it changed my MPG too.


As far as calculating the internal resistance, you say I need a load put on it. Do you have a good way of putting a load on the cells? Or would it not matter just as long as the test is consistent with all cell packs?

Your right about the firmware update. My mileage dropped about 5% after the last recall update. Honda claims this was done to preserve battery life, in other words, the computer will not allow the battery to discharge fully, so this in turn increases the work for the ICE, and decreases MPG.

If it makes the battery last longer, I can live with the decreased MPG.

Im not an expert, but there are plenty of the insightcentral forum that do this for fun. They use a super brain charger to cycle the sticks and record the results over 5 cycles. This gives an amp capacity and in many cases just cycling the stick rehabs it into better shape.

Still the load of the super brain is nothing to what it faxes in real life and a bad stick can pass this procedure with flying colors if it has a high internal resistance. :eek:

I used a super brain 989 for all my reconditioning. I did 4 cycles on most. (I know it should have been more consistant, but from what I read on other forums, 3 was plenty, I just wanted to be extra sure.)



It sounds like I should be set with just replacing stick #3 then. I hope it helps. There feels like theres a lot to do in reconditioning...

Thanks for the help, if no one else replies to this post I'll try to update it in a couple months. (Still have a foot of snow in my backyard so getting the Civic out for another few weeks is not going to happen.)

Yeah, thats fine. Also pair off the sticks good with a bad one so the "taps" are seeing an evenly distributed voltage. If any of the taps sees a lower than collective voltage that triggers a negative recalibration where the car starts recharging for a preset time. Some feed signals to these taps sing a resistor matrix and such to gain more performance out of the system as well as a move to lifepo4 batteries. :eek:

Another dimension you can add to this cycling is to use a laser themerstotat to check the temp while the stick is being cycled.

Typically the first signs of problems with a hybrid is from 1 bad cells. Once that is found the next problems are either another bad cell or an overall worn battery.

Since you got some time on your hands you could try another set of cycles on your sticks. :turtle:

I ended up ordering a new to me stick to replace #3. That should be here next week. Its been at least a month and I checked all the voltages, they are all sitting near 8v. #3is 7.3

Im cuurrently charging all the batteries my question is do I need to let the batteries rest before I check internal resistance or can I use the Peak voltage after a charge, and apply the load?

Nevermind. I suck
 

SO right after I posted the above message, I decided to answer myself and use the voltages that have been at rest for a month, calling that the full or 100% SoC.

Long story short my 7th stick, I hooked up the Super brain with the wrong polarity. Yep fried a transistor. Next step is finding a replacement to put into the board. GRRRR oh well.

I'll keep you posted.

part purchased.

That sucks. :eek:

Interesting enough if you drain the stick to .9 volts or even zero you can in turn reverse polarity charge them and they will even function in that manner. :eek:

Kind of adds a new idea to the old trick of rotating someones monitor screen 90 degrees. Whats better than reversing the polarity of their rechargeable batteries? :thumbup:

What are your specs?
 

i don't know your chemistry or cell arrangement or the rated capacity, but if you measure capacity then you could compare that to the rated spec capacity to determine how worn out they are. Over time they will wear out. Aren't the Leaf batteries supposed to retain ~70-80% capacity over 10 years?

There is no spec on internal resistance--You don't know the internal resistance when they were new--and what can that measurement tell you now, there is nothing to compare against? Plus that value can change depending upon how heavy you load during testing--i don't see any use for it, whereas capacity is a known quantity that you can easily measure and compare and use to determine remaining life, etc.

Lithium cells get permanently damaged if they go into reversal, any talk of reverse charging is voodoo science--probably want to stay away from voodoo

Good enough is a bit subjective... it doesn't have to be 100% new to be usable.

But Here's my 2 bits.

You want the cells / sticks to be as close to each other (balanced) as possible ... close in all aspects : Ah or Wh of discharge capacity ... Resistance ... Self Discharge Rate ... etc.

In order top 4 sorting criteria :

1. Balance of all 20 sticks as close to each other as possible , is better
2. less resistance (Ohms) each stick is better.
3. more capacity (Ah or Wh) each stick is better... whole pack only as good as the weakest cell or stick.
4. less self discharge is better

- - - - - -

Test result values will vary with conditions ... different testing methods and conditions can give different results ... be it Ah or Wh numbers or Ohms of resistance , etc... ( so be as consistent in all testing as you can be ).

to determine resistance ... V=IR --> dV / dI = R
dV = the change in voltage from zero amps to under a load.
dI = the change in current from zero amps to the amps of the load.
R = Ohms of resistance.

A simple method is to use a volt meter before and less than 10 second after the amps change to a steady value ... you can use the superbrain (or any controlled load) to act and the load ( charge amps or discharge amps )... and a basic multimeter to see the change in voltage (dV) before and after that amps change (dI)... more than 10 seconds you are getting a bit too much into the natural voltage change as the battery charges or discharges.

- - - - - - - - - -

If you can verify the following minimums you know you'll be fine ... but plenty of people don't weed out as many as the bellow would.

- - - - - -

I would say if you can't do the 70+ Amp high load test described above ... do the resistance testing as high of Amps as your setup can do... Brand new 6 cell NiMH sticks (depending on temperature and SoC) can have as low as 18mOhms (0.018Ohms) of internal resistance ... the lower your sticks / cells are the better ... but I would personally give an absolute maximum high of about ~43 mOhms per stick ... more than that , and you can use it for something else ... but I wouldn't put it back into a HEV.

Before I destroyed my super brain, I was working on the internal resistance. It would seem all my sub packs were in decent shape. one stick has a capacity of 5400 (which was the lowest) and one stick couldn't hold a charge like the others, but only dropped down to 7.3 volts, the others sat around 8.

From the sounds of it I might be able to have the car back on the road in no time. I ended up getting one stick on ebay in hopes of replacing the one stick that drops to 7.3 volts.

The stick at 5400 might be able to be revived with another cycle or 2.


As far as my measuring of the IR I will try again soon. It seemed my super brain was able to stabilize in 8 seconds at a 6 amp load. and was taking a measurement at 20 seconds. The next time I run this test I will up the discharge to 10 amps and get the voltage measurement within a few seconds after stabilization.

Ian, you have me thinking that I may be doing the capacity ratings wrong. All I did was record the discharge values and charge values taken from the super brain. (excel spreadsheet on the OP) Am I missing something? I have seen others record their time it took to do a discharge and charge as well, but I didn't think those mater. I feel I might be wrong now.

Not wrong ... there is always more depth you can go into ... at a certain point you have to set what you consider to be "good enough" to you ... and if you end up not being satisfied with that ... set the bar higher ... how high you want to set it is entirely up to you... better is always better ... but you might not need it that much "better" ... at a certain point it becomes a question of how much of your time and or money do you want to invest in getting ever higher amounts of "Better".

- - - - - - - - -

There are 2 main aspects of capacity.

#1> How much capacity the stick has.

#2> How much capacity the stick will be able to use in the HEV context.

- - - - - - - - -

Either 1 or 2 above can be recorded in Ah or Wh.

Ultimately energy is energy ... but the recording of the time to transfer that amount of energy can be used to get an idea of the Peukert effect of transferring that quickly.

For example if stick A gives you 5.0 Ah (~36Wh) of discharge capacity from a 10A discharge load ... and stick B gives you the same 5.0Ah (~36Wh) from a 1 Amp discharge load ... if both were equally charged SoC and discharged DoD ( at roughly the same temperature) ... Stick A that can give the same energy to you at a higher rate (amps of current or watts of power) is the better stick.

Because parts of the charge and discharge are not always constant current (Amps) nor is it always constant power (Watts) some people use an overall time method to get a gross ball park idea of how quickly they were able to get that ___ Ah or ___ Wh from the cell.

This additional time data for a gross ballpark of the Peukert effect can be useful , but is not absolutely necessary.

Just like doing Charge / Discharge Cycle efficiency tests for Ah or Wh can be nice ... some people do them ... others don't bother ... etc ... etc ... you can spend days testing or years testing.

- - - - - - - -

The basics just fall back to the top 4:

1. Balance of all 20 sticks as close to each other as possible for each of the 3 bellow.
2. Less resistance (Ohms) each stick is better.
3. More capacity (Ah or Wh) each stick is better... whole pack only as good as the weakest cell or stick.
4. Less self discharge is better

From there the basic guideline is:

If you can't do the 70+ Amp high rate testing ... do the IR testing ... any stick over ~43mOhms should probably be binned (for some other use , not HEV)

- - - - - - - -

Lastly ... when you are done selecting your sticks you will use... and are ready to put it all together back into the car.

Make sure all 20 sticks have a balanced SoC.

There are 3 main methods (in order of my preference):

#1> Use a low (trickle) current (less than 300mA) high voltage grid charger to top off the whole assembled pack ... When you start the car let it sit and idle in neutral until it stops charging on it's own... If it charges on its own at all.

#2> Discharge all 20 sticks before assembling ... let them 'rest' until each stick is within +/- 0.01v of each other ... ... When you start the car let it sit and idle in neutral until it stops charging on it's own... If it charges on its own at all.

#3> Charge all 20 sticks before assembling ... let them 'rest' until each stick is within +/- 0.01v of each other ... ... When you start the car let it sit and idle in neutral until it stops charging on it's own... If it charges on its own at all.

I'm glad I'm going in the right direction.

You said to make sure I have a balanced SoC, I want to build a grid charger that Mike designed. I've just started to do research on it, have my parts list and I'm about to get the supplies, but the small hurdle is the civic fan control. Sounds like it's not as easy as the insight. If you or anyone has input on a grid charger and fan control, or useful links that would be great!

I don't think I can even come close to charging all the sticks within .01v.

But like I said I haven't looked too deeply into my grid charger issue yet, but just looking for input.

Thanks again!!

The reason fall back to the 'usable capacity' thing... and the 'weakest link'.

For example:
If all 20 Sticks were identical resistance , capacity , etc.
But if Stick #1 were at 10% SoC ... the rest of the sticks are at 20% SoC ... the whole pack looses the bottom 10% ... because Stick #1 will bottom out first ... even if the other 19 have charge to give ... it wouldn't be usable to you.

The reverse also happens ... if Stick #1 were at 20% SoC while the other 19 were at 10% SoC ... than the whole pack looses the top 10% as Stick #1 would top out first before the others ... and even if the others still had room to charge ... it wouldn't be usable to you.

Attached is an example of the SoC variation that can creep into a pack over time ... The SoC variation (shown there) was effectively removing about ~2Ah of usable capacity from that pack.

Eventually if a pack get far enough out of balance you can get IMA error lights ... even if all the cells and sticks in the pack are otherwise still usable for resistance, capacity, self discharge , etc.

That kind of high voltage whole pack low amp trickle charger (I think) is the over all easiest / best method of balancing the NiMH pack ... it is also will be useful afterward 2 or 3 times a year as maintenance to correct any small imbalances that might develop.

The fan is needed to dissipate heat ... you can work with hacking the OEM fan (you are correct the Civic version is not as easy as the Gen1 Insight fan) ... you can install your own fan to move the air ... or you want do a low enough amp rate that the heat generated isn't as much of a concern (but it would take longer).

There are a variety out there ... Mike's is a great system (but AFAIK he isn't building new ones anymore) ... Peter in the UK also has a good simple cycler (but AFAIK not a pre-built product) ... EVpowers and Bubblebatteries both have offered balancing services in the past (I don't know if they sell the equipment)... etc... etc.

The +/- 0.01v is a level that you can be certain you will not have any issues.

Like other things ... you will have to decide how good is 'good enough' for you... some people are happy with +/- 0.1V.

The goal is to have all the cells and sticks you have tested and selected all be pretty close to the same SoC.

The difficulty of getting that SoC balance without using a whole pack low amp trickle charger is the self discharge rate of the cells ... See attached Graph I did from previous testing of the 40 sticks I have (other sticks might vary)... the time gap from when you finish the first stick to when you finish the 20th stick there is the potential for a SoC balance to be created ... due to the Self Discharge Rate... in a week or two it won't be a large difference ... but any difference is just effectively removing that % of your usable capacity.

You can always just 'do the best you can' at the stick level now ... and then later use the low amp trickle charger to correct any imbalance that might be there... in the mean time you're just down that % of usable capacity.

What about hooking all sticks in parallel to the hobby charger for a week or so?

I don't see anything wrong with that idea. a lot of wiring would need be done, but as far as I can see that shouldn't be too difficult with the battery pack sitting on my stand. hmmmm

^^
That's pretty much what a grid charger does except you don't have to take the pack out of the car, which is a bear. It also does it in a day and a half so you only lose a weekend.

Wiring, not so much. Build a rack with a bunch of slots to drop the sticks into, wire up a couple of sturdy bus bars, maybe with a few springs to ensure good contact, and then just two wires, one to each bus bar.

Watch your polarity.

The super brain is finally fixed. I finished the internal resistance testing. They are all close to being the same, with an average of .101ohms not spectacular but I'll have to make due.

I decided I'm going to do one more round of discharges and charges just for kicks, this time at 10 amp discharge. The previous cycles were 6.5A.

Thanks for everyone's help

Im currently working on trickle charging the batteries in parallel. Looking at the capacities of my sticks, im trying to figure out the order they should go in.

Does it matter in a hch gen 1 of the order in how they are arranged? Idont see specific voltage taps for where the computer would see specifics. The 3 sticks with wires I thought were the taps but idont see how they are. Now ithink they are for temp. Someone please correct me.

For a gen2 battery pack, I understand taking the weakest and pairing it with the stongest because 2 sticks are bound together.

Would I be best just alternating one strong, one weak, one strong etc...(like a gen2) or would it not matter?

Yeah, the batteries with the extra leads have temperature sensors on them.

The battery box itself has the taps. Peter on the insight form uses a resistor matrix to collectively fool the IMA system into thinking the state of charge is higher or lower to make the car give more, less or no assist while in operation. He also has a computer version as well as a way to replace the battery with lifepo4s.

The car has been driving for a week now. I've put around 450 miles on a tank of gas. When I first installed the battery, the car started right up, no CEL or IMA lights. It took a couple days for the battery gauge to read full. Everything seems to be preforming as it should.

I have a OBD2 tool on its way, I'm trying like crazy to drive conservatively and pump the MPG back to where it belongs. When I'm not paying attention down the highway I'm getting 40-44 mpg @ 70. If I really focus I can get around 48. Who knows, Its all habit forming anyways. We shall see.
I'm also looking into a lower grill block and see if that helps any.

Thanks for everyone's help!

Did you disconnect the starter battery, wait 15 minutes, reconnect it and let the car idle for 15 minutes? Hondas "learn" a lot of stuff during operation and they can learn good as well as bad habits. Make sure you got your radio code if needed before doing this and presets. Dont worry if your ima shows empty, it also has to relearn its soc once reconnected.

I didn't disconnect the battery and wait. The car did idle forever though. And I do have my code.

Like I said, it seems everything is on the up and up though, I'm quite happy with how its running. Sadly the 03 civic will probably never be back up to the 48+ mpg from normal driving. Though I am trying to get it as close as possible.

On a side note: I changed the 4 spark plugs last year (and they were needed). Yesterday I found out that there are another 4 hidden under the plastic engine cover. 2 of those 4 were worn fairly badly. I'm hoping with the new plugs that this will also help me get my fantasy MPG. :thumbup:

Yup, i-dsi. Allows the car to pull more timing and load on lower octane discount fuel.

Just an update. The last tank of gas I worked really hard at driving nicely, getting a good lean burn as much as possible, driving the speed limit or a little less. I was able to achieve 48 mpg on just under 550 miles for the tank. This tank of gas I'm not trying to drive like that, and driving as I have for years. So far I'm hovering around 45 mpg. I'm torn on what to do, because its a lot easier to drive the way I have for years. But theres something satisfying about getting 48+. Thanks for everyone's help.

The Peter Perkins modes allows some manual control over the IMA system. What you do is use the ima system to supplement the power output of lean burn and you abuse this to get epic mpg. Some people go as far as combine a glide session and coast as far as they can on what ima and lean burn will give, then when you hit the bottom of your target speed pulse back up on gas to the upper limit and repeat.

I have an update. Its bitter with a touch of sweet.

The bad: The battery is back to behaving like it did before I charged the cells.
Last week We took a trip from Michigan to South Carolina. The mountains drained the pack, and the recharges, to drain, recharge drain. It lost its home and now is stuck in a whirlwind of recals. I don't know what I'm going to do about it. If anything.

The Good: I've gotten really good at getting into the lean burn of things. I've figured it out and it feels like I can almost force it to do so. The feeling is similar to learning how to drive a stick, before you know it you can do it with very little thought.

So. Should I save my pennies and purchase a new pack? The car has 170,000+ miles and is an 03 manual.

Do you have a grid charger?

I was going to build one, but then I discovered its more difficult to splice into the fan than an insight. And thats as far as I got. I downloaded the schematic for it... does anyone have a link or advice on how to tap into the battery fan?

Throw a wall wart in there, and wire it up to the fan. When you plug in the charger, the wall wart also gets power and the fan turns on. Use diodes if you need to protect anything.

I like your style :thumbup:

Sounds like you got another bad apple or cell in your pack that isnt taking a charge that well. The capacity has to get down to .65 before it generates an ima light or code.

+1 on the grid charging or charging the sticks in parallel again. I have fairly new car and I do the balance thing every 3 months to keep it in excellent health. :thumbup:

Good news
 

Yeah I'm sure I have a bad cell or 2. I need a better discharge system to find which cells have the demons. The grid charger will have to do for the life of this car.

Built and installed my grid charger yesterday. Charged it all night, and unplugged it right before going to work this morning. Time will tell.

Because I took the sticks out this winter, do you think a deep cycle discharge of the whole pack would be necessary? I'm leaning towards no, and just top the battery off every month or as needed.

The charger seems too help, but at the same time doesn't. This weekend I hope to go through the process of doing a discharge using the light bulb method.

I have a question for someone more knowledgeable than myself. Back in May when I was driving through the mountains, I was driving on a highway that had a 4% grade for a hill. This was going down hill. I let the car slow itself down by keeping the car in gear. The charge bars were fully lit. Then while still going down the hill, the IMA light came on, the car jerked and was picking up speed, like there was no more resistance of a charge happening.

I haven't felt the car try to charge since. So I ask, how does the car charge? I understand its the electric motor turning into a generator when not in use. But what tells the battery to charge? The BCM? Is there a fuse or relay that could have blown while I was driving through the mountains?
Thanks

Most likely you have an IMA error (not likely a blown fuse or relay) .. most likely the BCM saw something it didn't like ... sent a signal to disable ... which specific IMA error you have ... go back to the OBDII code reader step to get the code pulled.

You can also repeat the disconnect 12V power , wait , and let idle step ... it won't 'fix' anything ... but it can sometimes let you put off a little while having to do a more detailed fix / service of some part of the IMA system... but it is really just a delay... the error will come back , be it a day or a year latter.

As for the actual questions you asked.

The IMA motor does always spin at the same PRM as the ICE ... but it is the activity of the MDM (Motor Drive Module) that determines if it is assist , Regen , nothing, etc ... it does this by altering the electromagnets in the IMA motor... like any Brushless Permanent Magnet Motor works ... when the electromagnets are working a certain way the IMA motor does not convert any significant rotational mechanical energy to electrical energy ... a different setting for the IMA motor electromagnets and you get 10Amps of assist , a different you get 20Amps of Regen , etc ... etc ... The MDM is the car's 'motor controller'.

The MDM is controlled by the MCM (Motor Control Module) , which is a computer that takes into account readings from various vehicle sensors to 'decide' on the amount of Regen or Assist (or none).

The BCM (Battery Control Module) is a computer that monitors sensors on the battery pack .. it communicates with the MCM ... that communication is part of what the MCM uses to 'decide' how much (if any) assist , regen , etc... If the BCM sends a no-regen signal to the MCM than you will get no regen... if it sends a No-Assist signal to the MCM than you will get no assist... etc.

So nothing tells the battery itself to charge ... instead something (MCM) tells the thing (MDM) that controls the IMA motor itself to alter how it controls the electromagnets ... that control of the electromagnets is what causes the IMA motor to convert rotational mechanical energy into electrical energy ... which just so happens to be connected to the battery ... the battery has no choice about taking the charge or not ... it takes what is sent to it.

well written! I guess then its just my super weak battery pack. Oh well. Once I do the light bulb thing I'll let you guys know how it went.