[yeukfung]

Hi all,

Recently I have come across the Lithium iron phosphate battery(LFP), which is really cool technology to upgrade, its discharge voltage is 12.8 to 13.2 flat, and the bad is cannot overcharge above 14.8V and strangely the new BMW M3 series is already having this options... which charing $1100+ for that battery...





Some chart shows even with heavy stress on the car, the car voltage is keep very stable around 13.8V to 14V, the blue line is the LFP battery alone, and red line is LFP use along with some small super capacitor, which i don't consider it as super capacitor will discharge slowly over time.

here is the normal lead acid voltage chart, same, red line is attaching some super capacitor:


i am assuming the above chart is not cheating and they are doing same stress test, as i don't have equipment to validate but the voltage stability is really different...


Somehow I found that the battery here in hong kong is around $500 with 60ah, so with this budget, I am considering to upgrade, however there is quite a few technical issue I faced. ppl here simply install and enjoy, but i would prefer to make my car ready to install first... so i dig deeper in charing method of our car, which basically it only have lead-acid charging profiles

Our current E-Series is actually having 2 types of charging profile allow to code in CAS, normal lead-acid or AGM type, that covers from around 50ah to 100ah.

However, the charging comparison as stated in F80 BMW technical doc is:
AGM/Lead will normally use 14.2V to charge, but up to 16V (according to charging profile stored in DME to control, i believe)...

http://f80.bimmerpost.com/forums/att...5&d=1406475520

and the latest charging Voltage with LFP in M3 is using flat 14.4V and +- base on temperature....


so, i think in order to properly equip our existing E90 N52 engine to safely use LFP, a 'hack' should do to our existing alternator to 'force' it output the 14.4V, but i have no idea how I can do that...?

I was initially thinking to decode the BMW charging map, say by manually modify the last battery registration km value, and check on the output voltage to see how many voltage will increase per 1000 km, and then do the reset every 1000 km... but i don't know if tool32 can handle that?

any idea/comments would be appreciate.

thanks all for your time

[hassmaschine]

There are parameters for this under the function LBATTCAL:

DME	Name	        Description
MSV70	CW_BATTCOD	Enable hand coding
MSV70	K_APPTBATINI	Enable Application InitValue battery temperature
MSV70	K_BATT_ALFAH	Alpha Tau value battery back
MSV70	K_BATT_ALFAV	Alpha Tau value battery forward
MSV70	K_BATT_BETAV	Beta value battery forward
MSV70	K_BATTHTAUVG	Time constant battery back at VFZG>
MSV70	K_BATTHTAUVK	Time constant battery back at VFZG <
MSV70	K_BATTORT0	Battery mounting location No1
MSV70	K_BATTORT2	Battery mounting location No2
MSV70	K_BATTVTAUVG	Time constant battery front at VFZG>
MSV70	K_BATTVTAUVK	Time constant battery at VFZG <Time-constant battery at VFZG>
MSV70	K_BATVFZGTAU	VFZG switching threshold TAU battery
MSV70	K_DTANSKOR	Intake air model offset outside temperature
MSV70	K_ILMINUMG	integrated Air mass threshold outside temperature model
MSV70	K_LBATTMAXK	Maximum battery charging voltage trunk
MSV70	K_LBATTMAXM	Maximum battery charging voltage
MSV70	K_LBATTMIN	minimalistic battery charging voltage to light
MSV70	K_LBATTTUMG	
MSV70	K_LBATTTUMGBA	
MSV70	K_MCOBATTORT	Coding hand Battery mounting location
MSV70	K_NMINUMG	Speed threshold outside temperature model
MSV70	K_OFTUMGCAN	Offset ambient CAN
MSV70	K_OFULADE	Offset charging characteristic Battery Connection
MSV70	K_PLBATTUMG	Plausibility threshold battery ambient temperature
MSV70	K_RLEITUNGK	Resistance charging line battery trunk
MSV70	K_RLEITUNGM	Resistance charging cable Battery in engine compartment
MSV70	K_TBATINIMAN	Application INIT VALUE battery temperature
MSV70	K_TBATTFTAUV	Time constant 5 * Tau battery Front
MSV70	K_TBATTTAUFH	Time constant battery 5 * Tau behind
MSV70	K_TIMERDLY	Delay Time for starting value calculation
MSV70	K_TUMGINI	Initialization outside temperature model
MSV70	K_TUMGTAU	Time constant 5 * Tau outside temperature
MSV70	K_UBATTNORM	Battery Standard voltage
MSV70	K_UKOMPK	Application value compensation voltage Kaffer room
MSV70	K_UKOMPM	Application value compensation voltage engine compartment
MSV70	K_ULADE_TAU	Battery charging time constant voltage
MSV70	K_VMINUMG	Vehicle speed threshold outside temperature model
MSV70	KL_BATTLSTR	Speed factor battery ambient temperature
MSV70	KL_EXPALH	Function EXP (-ALPHA * tabst_w) back
MSV70	KL_EXPALV	Function EXP (-ALPHA * tabst_w) back
MSV70	KL_EXPBETAV	Function EXP (-BETA * tabst_w)
MSV70	KL_LADEKENNL	Battery charging voltage characteristic
MSV70	S_LBATT1	Application switch1 selection substitute temperature
MSV70	S_LBATT2	Application switch2 selection substitute temperature
MSV70	S_LBATTAPPL

It's definitely possible, but for $1000, I don't see much of a benefit. The vehicles electronics are designed to work with the noisy voltage levels of a standard battery. you could buy 5 normal batteries for that much, which would last you longer than the entire car.

[Biginboca]

hassmaschine the advantage is the 12.5kg weight savings, almost 30lbs. But it’s pricey and will take work to make it uh... work

[yeukfung]

Originally Posted by hassmaschine

There are parameters for this under the function LBATTCAL:

DME	Name	        Description
MSV70	CW_BATTCOD	Enable hand coding
MSV70	K_APPTBATINI	Enable Application InitValue battery temperature
MSV70	K_BATT_ALFAH	Alpha Tau value battery back
MSV70	K_BATT_ALFAV	Alpha Tau value battery forward
MSV70	K_BATT_BETAV	Beta value battery forward
MSV70	K_BATTHTAUVG	Time constant battery back at VFZG>
MSV70	K_BATTHTAUVK	Time constant battery back at VFZG <
MSV70	K_BATTORT0	Battery mounting location No1
MSV70	K_BATTORT2	Battery mounting location No2
MSV70	K_BATTVTAUVG	Time constant battery front at VFZG>
MSV70	K_BATTVTAUVK	Time constant battery at VFZG <Time-constant battery at VFZG>
MSV70	K_BATVFZGTAU	VFZG switching threshold TAU battery
MSV70	K_DTANSKOR	Intake air model offset outside temperature
MSV70	K_ILMINUMG	integrated Air mass threshold outside temperature model
MSV70	K_LBATTMAXK	Maximum battery charging voltage trunk
MSV70	K_LBATTMAXM	Maximum battery charging voltage
MSV70	K_LBATTMIN	minimalistic battery charging voltage to light
MSV70	K_LBATTTUMG	
MSV70	K_LBATTTUMGBA	
MSV70	K_MCOBATTORT	Coding hand Battery mounting location
MSV70	K_NMINUMG	Speed threshold outside temperature model
MSV70	K_OFTUMGCAN	Offset ambient CAN
MSV70	K_OFULADE	Offset charging characteristic Battery Connection
MSV70	K_PLBATTUMG	Plausibility threshold battery ambient temperature
MSV70	K_RLEITUNGK	Resistance charging line battery trunk
MSV70	K_RLEITUNGM	Resistance charging cable Battery in engine compartment
MSV70	K_TBATINIMAN	Application INIT VALUE battery temperature
MSV70	K_TBATTFTAUV	Time constant 5 * Tau battery Front
MSV70	K_TBATTTAUFH	Time constant battery 5 * Tau behind
MSV70	K_TIMERDLY	Delay Time for starting value calculation
MSV70	K_TUMGINI	Initialization outside temperature model
MSV70	K_TUMGTAU	Time constant 5 * Tau outside temperature
MSV70	K_UBATTNORM	Battery Standard voltage
MSV70	K_UKOMPK	Application value compensation voltage Kaffer room
MSV70	K_UKOMPM	Application value compensation voltage engine compartment
MSV70	K_ULADE_TAU	Battery charging time constant voltage
MSV70	K_VMINUMG	Vehicle speed threshold outside temperature model
MSV70	KL_BATTLSTR	Speed factor battery ambient temperature
MSV70	KL_EXPALH	Function EXP (-ALPHA * tabst_w) back
MSV70	KL_EXPALV	Function EXP (-ALPHA * tabst_w) back
MSV70	KL_EXPBETAV	Function EXP (-BETA * tabst_w)
MSV70	KL_LADEKENNL	Battery charging voltage characteristic
MSV70	S_LBATT1	Application switch1 selection substitute temperature
MSV70	S_LBATT2	Application switch2 selection substitute temperature
MSV70	S_LBATTAPPL

It's definitely possible, but for $1000, I don't see much of a benefit. The vehicles electronics are designed to work with the noisy voltage levels of a standard battery. you could buy 5 normal batteries for that much, which would last you longer than the entire car.

Millions thanks for your info, i will dig deeper on the MSV70 command mode

For the LFP battery...... after this post, i do a little bit search, in china taobao with an unknown brand, the design seems originate from taiwan (with overcharge/underchange protection and balancing circuit inside, that's what they told me...) and manufacture in china, with the 60AH LFP battery is just RMB1400... what a unbeatable price and i have ordered one a do some test.

I will post some result later on this thread to compare any difference, probably it arrives next week

[yeukfung]

Regarding to the parameter, i have tried to locate any job having name "LBATTCAL" in MSV70, but cannot find...


wondering is that I should use the FLASH_PARAMETER_SETZEN command to overwrite it? like puting the arg with K_LBATTMAXK;14.4



and i can use FLASH_PARAMETER_LESEN to read out the value first?

another thought is that, if we can even decode the IPO file and retrieve the default parameter from the M3 DME file, then it's even easier to "migrate" the settings from that... but i have no knowledge in reading ipo files...

[hassmaschine]

You cant do that with coding, no such jobs exist. Its in the flash of the DME - it has to be tuned for it.

Also, you can only flash whole blocks of memory, you cant just change one or two bytes. Even if you could the checksum and RSA signature would need to be dealt with. Tool32 has no idea where any of the maps are anyway, it can only do things that are pre-programmed.

The IPO from the M3 will not help, its a completely different DME. There's nothing in there that will let you change these parameters anyway.

I can probably help with tools to do it, but as far as the actual calibrating its all up to you. I'd be cautious with a lithium battery though, you could start a fire especially with bargain chinese cells.

[yeukfung]

Originally Posted by hassmaschine

You cant do that with coding, no such jobs exist. Its in the flash of the DME - it has to be tuned for it.

Also, you can only flash whole blocks of memory, you cant just change one or two bytes. Even if you could the checksum and RSA signature would need to be dealt with. Tool32 has no idea where any of the maps are anyway, it can only do things that are pre-programmed.

The IPO from the M3 will not help, its a completely different DME. There's nothing in there that will let you change these parameters anyway.

I can probably help with tools to do it, but as far as the actual calibrating its all up to you. I'd be cautious with a lithium battery though, you could start a fire especially with bargain chinese cells.

Thanks for your info and will carefully monitor it in first few weeks once installed. The battery isn't Lithium ion, but lithium iron phosphate (LiFePO
4) battery that seems to be quite safe, as least it won't explore like lithium ion battery, but sure everything may goes wrong.
https://en.wikipedia.org/wiki/Lithiu...sphate_battery

I have followed your another post in cloning MSV70 DME and i learnt BDM cloning MSV70 from that, wondering how could I provide you my existing DME file and do the modification, say limiting the charging voltage to 14.4V, the temp different in hong kong isn't really high, from 5 degree to 35 degree throughout a year, so maybe variant 0.1V?

Another strange request is about the ZB file, currently i am using

ZB#7581300

however, when the time i bought this car, it was actually using this ZB
ZB#7568990

not sure if you have some backup or able to tune the DME to same as ZB#7568990?


the story is back to a year ago, when I bought this car and discovered that this car DME has internal flash error, it goes to indy show and replaced a new DME to me with wrong DME ZB, the car runs like a crap, and then I flash with winkfp but unable locate my old ZB, the only way is to upgrade the ZB to latest version on my variant, which is ZB#7581300, but the car just does not seems to run as good as the old ZB number...

this is my full story about a year ago..
http://www.bimmerfest.com/forums/sho...d.php?t=984639

[hassmaschine]

Yeah, I have that file. But it's an old version - I only support the newest version of the DME, but I should have the updated/correct ZB# that matches.

Most likely, the tune you have doesn't have the correct power class for your car. I'd just need your serial # to find it.

edit: actually, 7581300 is the correct new version for your DME. I don't think it's an issue with your calibration, but maybe the power class in the used DME doesn't match your car.

[yeukfung]

Originally Posted by hassmaschine

Yeah, I have that file. But it's an old version - I only support the newest version of the DME, but I should have the updated/correct ZB# that matches.

Most likely, the tune you have doesn't have the correct power class for your car. I'd just need your serial # to find it.

edit: actually, 7581300 is the correct new version for your DME. I don't think it's an issue with your calibration, but maybe the power class in the used DME doesn't match your car.

my VIN is AW49710, with engine variant VB32

[hassmaschine]

was your replacement DME new or used?

[yeukfung]

Originally Posted by hassmaschine

was your replacement DME new or used?

my replacement DME is new from local part store, in virgin state.

at that moment, i have no idea on ECU coding and programming stuff, so I give my car to local indy which he said he could do it

he loaded the ZB#7581278 to the replacement DME as what he told me it's the number provided by BMW HK service center, and the car runs like crap, I asked the indy to switch back to my previous ZB#, he refused to and that leads me to learn winkfp myself.

and I diy flashed ZB#7581300 and the car runs good, so far has been running over 20000km, but just a bit different to what I feel with the ZB#7568990 at the time i bought this car, it pulled much better... maybe it's not related.

[yeukfung]

Hi all,

Some updates on my findings playing with LFP battery after 3000km driving over 1 months...

with some background info:
LFP base voltage: 12.8V, lead acid 12.0V
14.4V constant voltage can charge the LFP to 100%, but with the alternator output, it will only keep the voltage at 14.05V, which means 90-95% of battery capacity can be maintained.

For the LFP I used, it's 60AH, so, only around 55AH is inside the battery, for the charging profile, since AGM is more sensitive to overcharge, I have coded to the battery type to AGM 70AH which is the nearest profile to the 60AH battery..


Some testing
- CCA Testing within this 3000km, it's varies from 1150 to 1230 CCA, it does not rise or drop much, more or less keep stable in this 3000KM journey, while comparing to 90AH AGM, it was around 900CCA to 930CCA

- Internal resistance testing, it's varies from 2.35mOhm to 2.5mOhm , while comparing to AGM, it is around 3.0mOhm to 3.2mOhm

- voltage drop during cold start, the MIN voltage recorded is 11.8V - 11.9V , while i didn't record the AGM, but for 60AH battery, it's very strong

- alternator output test for cold start, I use inpa and deep obd to monitor the generator AMP from cold start to warm up, there is two case:
1) when the battery is full - the max AMP generated from alternator is around 90-100Amp(14.4V) for around 20 seconds, and eventually drop to 20Amp within 1 minutes, and the charge of battery reach 14.05V
2) when battery is NOT full - the max AMP generated from alternator is 170Amp(14.8V) for over 1-2 minutes, and start to drop when battery voltage reaches 13.8V and then eventually goes down to 14.05V after 2 minutes, and then back to 20Amp.

- temperature test, i don't have equipment, just use hand to test several area, I can say... does not feel any heat/warm on the battery, just normal and usual


drawback to my car with IBS:
- IBS is now partially working, as it no longer able to detect battery state of charge when it's in sleep mode, it can still read the battery terminal voltage when the car is running.



====

For the result and feeling after upgrade the battery, it's real strong, the car pickup the rev much quicker, and rev up much quicker than before, especially from 3000 to 5000 rpm.


state of charge when car sitting overnight(over 15 hours, so the surface charge should already consumed):
it reads 13.42V, which is around 90%-95% (indicates the car is charging the battery properly)


For the charging speed compare, I logged some data with the car during cold start for both AGM and LFP battery:

AGM


LFP


comment: LFP charging much faster than normal AGM, maybe due to it's very small internal resistance.


running voltage:
so far, when the car equipped with IBS, it will detect the battery terminal to be 14.05V, and maintain that voltage, so for the Generator Voltage output, 95% is using 14.1 - 14.2V with around 25-35Amp output, and it raise to 14.4V only in cold start and slowly drop back to 14.2 when the battery terminal reachs 14.05V... with my observation...

seems the AGM charging profile with IBS seems compatible with the LFP battery somehow...

[Biginboca]

Originally Posted by yeukfung

For the result and feeling after upgrade the battery, it's real strong, the car pickup the rev much quicker, and rev up much quicker than before, especially from 3000 to 5000 rpm.

Ok this will have to be explained. Why do you think a lithium battery would make the car rev quicker?

The only realized benefit I can imagine to this swap is less weight which means better acceleration and shorter braking (maybe you save 20lbs of weight so not really a measurable, noticeable advantage IMO). And of course the improved gas mileage as a result of the weight savings likely something to the tune of .0001mpg improvement lol

[yeukfung]

Originally Posted by Biginboca

Ok this will have to be explained. Why do you think a lithium battery would make the car rev quicker?

The only realized benefit I can imagine to this swap is less weight which means better acceleration and shorter braking (maybe you save 20lbs of weight so not really a measurable, noticeable advantage IMO). And of course the improved gas mileage as a result of the weight savings likely something to the tune of .0001mpg improvement lol

For the case, i don't put much expectation when swapping to LFP, but that's my feeling so far, more aggressive comparing to AGM...


For the technical data, i don't think i can give you much as i am not expert, but i think it's mainly related to two aspect:
1) internal resistance is lower than AGM referring to it can discharge much faster than AGM, for this case, it's 3.2 - 2.4/2.4 = 33% less in the internal resistance
2) base voltage is the major difference, in which
the AGM/Lead Acid base voltage is 12.0 (full charge is 12.6V)
the LFP base voltage is 12.8 (full charge is 13.4V)

so when on sudden rev up and demand, LFP will provide faster response to current change with much higher supply voltage when alternator regulator may not be fast enough to cope with the sudden current change, so that makes the car runs smoother with better spark? just my pure guess...

[Biginboca]

Originally Posted by yeukfung

For the case, i don't put much expectation when swapping to LFP, but that's my feeling so far, more aggressive comparing to AGM...


For the technical data, i don't think i can give you much as i am not expert, but i think it's mainly related to two aspect:
1) internal resistance is lower than AGM referring to it can discharge much faster than AGM, for this case, it's 3.2 - 2.4/2.4 = 33% less in the internal resistance
2) base voltage is the major difference, in which
the AGM/Lead Acid base voltage is 12.0 (full charge is 12.6V)
the LFP base voltage is 12.8 (full charge is 13.4V)

so when on sudden rev up and demand, LFP will provide faster response to current change with much higher supply voltage when alternator regulator may not be fast enough to cope with the sudden current change, so that makes the car runs smoother with better spark? just my pure guess...

The coils that fire the spark plugs are either on or off, they aren’t drawing more current at full throttle as though they are working harder the deeper you press the throttle.

They fire more frequently at higher rpms, so the hardest demands they see is at redline. Full throttle at 3000 or 5000 rpms is no more demanding on them than cruising steady at those same rpms.

I’m not saying you didn’t feel the car pulled harder. Maybe your car does respond to the weight loss at a level you can feel. But I don’t think a battery can contribute anything to a sensation of more power outside of its lighter weight and 20lbs of weight loss on these cars is the same as adding less than 2hp.

[yeukfung]

Originally Posted by Biginboca

The coils that fire the spark plugs are either on or off, they aren’t drawing more current at full throttle as though they are working harder the deeper you press the throttle.

They fire more frequently at higher rpms, so the hardest demands they see is at redline. Full throttle at 3000 or 5000 rpms is no more demanding on them than cruising steady at those same rpms.

I’m not saying you didn’t feel the car pulled harder. Maybe your car does respond to the weight loss at a level you can feel. But I don’t think a battery can contribute anything to a sensation of more power outside of its lighter weight and 20lbs of weight loss on these cars is the same as adding less than 2hp.

Hi,

your point is well taken

Please correct me if i am wrong or mis concept in calculation in power consumption in ignition coil...

and assume there are few cycles during quick acceleration which the alternator regulator cannot catch up with the sudden demand of the current draw and most of the current is draw from battery for several milli seconds......

with AGM, let's say it's base voltage is 12.6 (assume full)

so the power provided to the spark via coil would be 12.6V * (2Amp (assume each coil draws 2 amp) * 6 (with 6 coil) = 151.2W

if with LFP...
13.4V * 2Amp * 6coil = 160.8W

the different voltage should has some effect to the spark produced?

so far i also don't believe what i feel, ... thus i do so many data logging to see what's actually going on...

if you have some suggestions to what data i should log and try to compare, please let me know... but those data are in milli-seconds level... which most obd diag tools cannot really read in that...

maybe i should use the OTest to read the data out.. but it's windows app which need to connect to computer while driving...

[MILLER335I]

yeukfung, long term update on teh battery? Coding and alternator working ok?

[Steve Sas]

I installed similar LifePo4 12.8v 60aH battery in my 335i.

Engine was replaced from a 135i so the alternator had no LIN interface.

In the beginning everything was ok. Car starts ok and quickly charges to 14v.

But after week in very cold mornings when car was started, the alternator would try and charge up to 15v which tripped the over voltage protection in the battery and the BMW would throw many codes as the battery disconnected with engine running. Very bad.

Anyway seems to be normal function of the alternator internal settings. Got around this buy disconnecting the ignition signal to the alternator. Still charges when car starts but in some other default mode. Voltage slowly rises the 14v.

Anyway all ok since running this way in the last month.

[wayman519]

I think the performance increase the OP thinks he is seeing is the alternator running at lower load on the engine. This can vary based on how bad the original battery was and how much electronics/options the car is running at a given time. I notice my car is bogged down when running all the heated seats at max. I use this trick to determine pulley noises on my engine. I can see the DME controlling the alternator to put less load on the engine as it is primarily programmed for efficiency and the new chemical battery can hold charge for longer and charges quicker but the DME will disengage all auxiliary load on the engine except power steering when power is requested with your foot.

[Steve Sas]

Just an update, car has been running well since.
LiFePo4 seem to be a weight saving option.

[Biginboca]

Just as another option I recently swapped to a 30ah Antigravity Lithium and it was pretty much a plug and play. They recommended I code the car with Carly to the lowest option, 40ah which I did and then it’s just install and run. This battery runs around $700 though, but it has some added features and is very easy to install and live with.

It’s been about 6 weeks now and no issues at all. The battery has a built in jump start from activated from a remote key fob and comes with a Bluetooth monitor that syncs with my phone and sends me state of charge every 24 hours. After 4 days of sitting the car was still at 99% charge with comfort access and all. Best of all this battery cranks harder than my previous 60a/h AGM ever did and weighs 13lbs.

[Noir]

Originally Posted by Biginboca

Just as another option I recently swapped to a 30ah Antigravity Lithium and it was pretty much a plug and play. They recommended I code the car with Carly to the lowest option, 40ah which I did and then it's just install and run. This battery runs around $700 though, but it has some added features and is very easy to install and live with.

It's been about 6 weeks now and no issues at all. The battery has a built in jump start from activated from a remote key fob and comes with a Bluetooth monitor that syncs with my phone and sends me state of charge every 24 hours. After 4 days of sitting the car was still at 99% charge with comfort access and all. Best of all this battery cranks harder than my previous 60a/h AGM ever did and weighs 13lbs.

Sick!