Battery Management System project

retepsnikrep

Senior Member
I'm fairly new to the picaxe but not new to basic programming or electronics. I'm an EV and hybrid vehicle enthusiast. My current project is to build a picaxe based bms for some large Lithium cells in a series high voltage (144v 50ah) pack. I'm familiar and happy with all the risks of HV battery projects.

My ideas are as follows.

Use an individual 08M pic on each cell to monitor

1) Cell voltage
2) Cell temperature

the 08M will also control an individual bypass resistor so that when the cell reaches a set voltage say 4.2 it turns on and bypasses a set current.

The individual cell pics should communicate with a host picaxe monitor and display system via an opto isolated bus.

Now the bus is the issue for me. I have ben reading the very interesting slave master serial project on this site, that may be one option.

But I don't need full bi-directional data comms.

I was thinking master pic could send an command signal on 08M input 3 (opto isolated) and all in parallel so the command initiates a data send routine in all the cell pics.

The individual pics could then respond on a simple opto isolated serial interface in sequence using a simple timing loop (different for each pic)


For example

Master sends data request via the opto interface and this triggers a simultaneous interupt on all 08M pics.

Cell 1 transmits identification number and data immediately,
Cell 2 waits X micros seconds and then transmits it's cell ID and data,
Cell 3 ditto and so on.

This data is stored in a simple array/scratch pad until all cells have sent data then master can process as reqd. Any failure to receive a cell's data triggers an alarms as reqd.

So I'm happy with the interupt triggering the data send routine on the 08m pics and the timing, but what about the serial output, how best to construct an opto isolated interface so all can send on same connections.

Anyone offer any help on this? Circuit ideas?

Another issue is that the as each 08m pic's power will be supplied by it's own cell, it has an issue measuring it's own cell V. I was thinking use a 1.25v prescion refrence and then do some simple calcs to take into account fluctuating supply voltage. But this will use another ADC input.

Also as the cell minium voltage is about 2v under heavy load what is the mimimum operating voltage for an 08M device. Should I use a charge pump dc-dc converter on each cell to maintain a steady 5v supply for each pic, this would increase cost considerably. A simple capacitor and diode on the + side of the chip supply isolating it from the drain would maintain the voltage for a few seconds, will that do it? 5v miniature backup power supply capacitors are pretty cheap, what's the running drain of an 08M say at 4v?

Lots of questions, I'm happy with most thing and digesting all the great advice on this site but extra comments are welcome. I'll of course make all my stuff available on the site inc pcb's software etc.

Peter
 

hippy

Technical Support
Staff member
One technique to save having to program each PICAXE with a different delay after interrupt is to daisy chain them so the first sees an interrupt, sends its data on the shared bus then sends an interrupt pulse to the next, carries on doing its normal monitoring.

If every PICAXE delays, that could be many milliseconds for the last to respond. An alternative is to have each PICAXE count down how many interrupt pulses there have been and when their count hits zero they send their data. Each starting with a number one more than they last and they will send in sequence. This minimises the time they are actually hanging around doing nothing. The interrupt itself can check the length of the interrupt pulse so a long pulse can be used to reset each PICAXE to its initial count so kick them into sequencing at turn on or before each interrupt sequence.

Another alternative is to send an interrupt then have the master send an ID indicating which PICAXE should respond. Each PICAXE having it's own ID programmed in. This is probably the solution I'd opt for.

A common bus back to the master should be very straight forward with opto-coupling. On the master have the serial line pulled up and use opto's on each slave which will pull that line down to 0V. As only one slave will transmit at a time there's no risk of data collision.

A PICAXE should work down to 2V but the internal oscillator may become off target affecting the serial baud rate. You might have to adjust the operating frequency with CALIBFREQ on the fly as voltage changes.

I can't see anyway other than using a voltage reference and two ADC channels to self-determine voltage. Other forum inhabitants will undoubtedly have better ideas than I have on this and other aspects of this setup.

Don't forget the 14M is only a little more expensive than an 08M and has much more I/O if needed - Flashing a LED to indicate a fault condition and handy while debugging.
 

manuka

Senior Member
2V???? Make that 2.7V if active! Compared with a 14M, the new 20M,costing ~US$6, has even more I/Os (8 of each-with 4 ADC) is perhaps easier to use.
 

D n T

Senior Member
How many cells do you have

Sorry for my ignorance but how many cells do you have in your pack, I have not played with this before, mine are usually confined to sealed lead acid batteries.
Is each cell being monitored sepearatly or are they grouped into "subpacks"?
 

kevrus

New Member
As lithium ion cells are easily damaged by over voltage, and are also very expensive, I would also incorporate a second method of clamping the charging voltage which didn't rely on the 'software' approach hence hopefully be less vulnerable to problems...just to act as a watchdog/high limit alarm.

I assume you are using Thunder Sky cells imported from China, what motor are you using in your EV, i'm guessing DC series wound ?
 

moxhamj

New Member
This is quite a challenge but I think all the components can be done. Any time a voltage is measured it makes things so much simpler if the picaxe is being run from a regulated 5V. This can be a low quiescent low dropout reg so the current is negligable. I'm just wondering, if each cell is 4V and you have lots in series and the minimum is 2V (is that right)then you need 8V (say) for a 5V reg, can you run each picaxe off the 4 cells above the 0V point for that picaxe. Eg cells start at 1 and go to 70, picaxe 1 is run from the volts between cell 4 and 1, picaxe 2 is between cells 5 and 2 etc. Doesn't work for the 3 picaxes at the top of the stack but works for all the others.

I like the optoisolator idea as the picaxe gets the data into a digital format and that is much easier to pass through an optoisolator than analog.

Still is a lot of picaxes though.

Hybrid vehicles are going to be the future as oil runs out. It is intriguing that so much work on hybrids is being done by enthusiasts rather than by big corporations. It is a bit like at the beginning of the HiFi era - the only way to get a good setup was to do it yourself. These stacks of batteries contain a lot of energy and I gather they tend to catch fire if mismanaged. So a system for measuring volts and temp of cells is going to be quite critical. I suspect as more of us feel the pain at the pump then we may be pondering the issue of measuring battery stacks with more urgency!

I shall sleep on this problem as I have a niggling feeling there might be a cheaper op amp solution.

Another way to get isolated volts to each picaxe might be to use transformer isolation - eg potcores at high frequency. Just need to think about arc crossover voltages for the insulation on the windnig wire.

Maybe divide cells into groups - eg groups of 5 and then for each group of 5 you could use divider resistors (1%) for each of 5 ADCs and a single 14M for that group. I think this might come in at the lowest cost and component count.
 

retepsnikrep

Senior Member
Thanks for the interesting replies so far.

If you want some background on me go here www.solarvan.co.uk.

I have built my own BMS before, but am after something a bit cleverer this time. My own knowledge has improved as well so I'm aiming a bit higher ;)

Yes I probably will be using Thunder Sky cells again, I have used them before, since 2003 in fact, they do not catch fire or explode at the drop of a hat contrary to popular belief if the cell V goes 1mv above max :rolleyes: In fact they are pretty hardy and can be abused a fair bit, one enthusiast I know has some in his own EV and uses no BMS at all!

There will be 40-45 cells in my pack I haven't decided on final number yet.

I like the 14M idea (thanks) and the cascading data trigger idea, I would like to keep all the software in the slave pics the same if possible otherwise software/upgrades/changes will be a nightmare with 40-45 different versions of the slave software. Keep the thoughts comming thanks.

Yes I have an opto analogue solution already using op amps and reference voltages etc using a paralell opto isolated alarm bus for over and under V.

http://agnimotors.com/home/index.php?option=com_content&task=view&id=27&Itemid=39

But again I would like to monitor individual cell V and temp this time with the flexibility of a pic. A central PIC with lcd display can then tie the whole thing up and make a neat info point for the driver.
 
Last edited:

hippy

Technical Support
Staff member
I would like to keep all the software in the slave pics the same if possible otherwise software/upgrades/changes will be a nightmare with 40-45 different versions of the slave software.
Changing the code in forty-odd PICAXE's as you go through development could be an interesting challenge.

Although it is not supported ( and does have potential problems ) you could connect all Download Serial In's together and take Serial Out from just one and program all PICAXE's in parallel. By incrementing a version number for every download you can then trigger them in sequence ( as in normal sampling ) to check what version numbers they return.

An associated issue is getting PICAXE's into download mode. There are situations where a PICAXE can stop seeing downloads ( perhaps exasperated if downloading blind ) and you'd need to reset the device. The M's don't have an external reset and need to have power disconnected then re-connected.

An alternative way of handling each cell could be a large analogue switch which allows a single PIAXE to select the cathode/anode of any individual cell.
 

MFB

Senior Member
On a related subject, please can anyone recommend a link to information on charging low capacity single cell Li-Po batteries? The lowest current charging circuit that I have found is for a 100mA, and I’m using a 30mA/h battery.
 

premelec

Senior Member
Great project! I've given this problem many thoughts over several years; for individual cell monitoring of 50 cells... make a simple voltage to frequency converter or voltage to pulse width converter with very low current draw and an opto isolator to leads to a 50x multiplexer... the V/F could be quite fast - and you need one for each cell.

Another way is simply 50 resistors to battery intersect points with these then multiplexed to two sense resistors at a time which you read the differential voltage from to get the battery cell voltage... a refinement that requires more resistors but could make sense with the high voltages is to use different resistor/divider pairs depending how high on the string you are. Note that the resistor connected to the battery join point serves as a current limiting safety factor to the wire going from it to the lower part of the divider being switched on to make the reading. Note there is no DC isolation in this approach.

I don't know what the $130K_US TeslaMotors 8000 cell [A123 systems cells] battery does for monitoring but I'm pretty sure it isn't individual cell monitoring! :)
 

papaof2

Senior Member
charging lithium cells

On a related subject, please can anyone recommend a link to information on charging low capacity single cell Li-Po batteries? The lowest current charging circuit that I have found is for a 100mA, and I’m using a 30mA/h battery.
Not Li-Po specific, but Battery University has a page on Lithium-Ion charging http://www.batteryuniversity.com/partone-12.htm

Recommendations for Li-Po charging http://manuals.hobbico.com/gpm/gpmp0800-0831-manual-v1_2.pdf

John
 

moxhamj

New Member
Just brainstorming more:

45 batteries with 5 batteries per 14M means there are 9 14M chips, which becomes manageable. Of course, you won't be able to program in the field as the reference 0V might be 100V so all chips could either be socketed with leg2 tied to ground, OR there might be an opto isolator programming circuit you could build. Personally I'd just go for sockets - I've been using this technique on most of my picaxe projects - pull out the chip and put it in a programming board if it needs recoding.

Each 14M would have its own 5V supply via a LP2950 reg. Each chip monitors 5 batteries and their volt referenced to the 14M gound might be 3V, 6V, 9V etc. The first battery at 3V could go straight into an ADC (maybe via a 10k resistor). The second one might be 6V so use a divide by 2 with 2 10ks (1%). The third one might be 9V so to be safe use a divide by 3. Etc. Just work out the max volts the stack could have and check it doesn't go over 5V on the picaxe input. Also, as discussed on another thread, maybe put a capacitor between each ADC input and ground - it smooths the ADC input and gives more reliable readings. At a guess, 0.1uf.

I agree each chip should have the same program. So each chip can't really have a unique ID number as it makes errors in programming more likely if you accidently have two chips numbered "7". There are lots of answers here. I wonder about having a simple master picaxe, one the bigger ones, that has 9 output lines that can go high. Run them through an optoisolator to each 14M and when the line goes high, the 14M sends its data. The master picaxe might end up running other things like an LCD display so it could be easier to go up to a 40 pin picaxe. Or you could use an 18X with daisy chained HC595s to increase the number of outputs.

Re temperature sensing, this doesn't need voltage isolation, but it would make sense to bury lots of sensors in amongst the stack - maybe even one per cell if you want to do it properly.

As an aside, looking through some of the websites around, I didn't realise how advanced electric cars have come with all the home tinkerers. And I'm also surprised how this hasn't gone commercial. Ok, the capital cost is still there but every $1 increase in the cost of oil will decrease the payback time. I can see someone building an EV conversion themselves. Then one for a mate. Then that is the start of a new business.

HTH
 
Last edited:

retepsnikrep

Senior Member
Sample Code

Just been taping away on this project and possibly going to use a 14M Picaxe on each cell (Might use an 08M if I can use the serial out pin to also send the interupt to the next PIC) with a cascade serial opto isolated bus to communicate with a master 18X or similar, thanks for all the ideas.

So my initial/draft slave code, same for each of the PIC's is here. I've just bought a licence for VSM so I'll draw up a quick schematic of a slave module in next few days when I get my code. Yes it uses a few pic's and has some issues, but I want it easily expandable/adaptable for each and every battery technology/cells.

I want to turn on port C0 as an adc input I hope I have done that right.

I have been looking at the ADC voltage measurement thread here when using a 1.25v prescion reference and a variable power supply (the cell) I think I understand it just need to nick the bits of code reqd from the thread and put them into below draft.

http://www.picaxeforum.co.uk/showthread.php?t=8624

`Battery Management System Slave Module
`Based on Picaxe 14M microcontroller
`Original concept, hardware and code by Peter Perkins
`Peter@solarvan.gotadsl.co.uk
`V1.00 280508


`Word Variables 65535 for accuracy
`w0 = Voltage reference 1.25v
`w1 = Cell voltage
`w2 = Cell temperature
`w3 = General maths 16bit variable
`Balancing voltage start is set at >4.2v



start: `Initialise program and variables/interrupts as reqd

let dirsc = %00000000 `Set Port C0 as an extra ADC Input (ADC1)
setint %00000010,%00000010 `Activate interrupt when Input 1 goes high

main:
readadc10 4,w0 `Read the voltage reference into variable w0
readadc10 0,w1 `Read the cell voltage into variable w1
readadc10 1,w2 `Read the cell temperature into variable w2

`Insert Cell V/T calculation and correction routine here

If w1 > 55050 then `Turn on cell balancing load if voltage >4.2v

high 1 `Turn on load resistor
else
low 1 `Turn off load resistor
endif
goto main `Goto main loop



interrupt: `Send cell data to master Picaxe via opto serial bus
sertxd(w0,w1,w2) `Send raw cell/slave data to master, include :-
`V ref, cell V, cell temp

high 2 `Set output 2 high to trigger interrupt on next slave pic

pause 100 `Wait 100 milliseconds to ensure data received OK
`This may be cut or removed depending on hardware response

setint %00000010,%00000010 `Reset the interrupt and wait for another data request.

low 2 `Set output 2 low to reset the next slave pic trigger

return `Return from interrupt to main program.
 
Last edited:

retepsnikrep

Senior Member
I've been working hard on this and am now ready to go to the Pcb creation process. Quite a bit of work to do on the Master Software but it's starting to come together. My 50x40ah Lifepo4 cells should arrive by early August so need it up and running by then.

During bench testing of a prototype Master/Slave I found system could deal with approx 50 slave cells per second at 4mhz, which is fine IMO. Whole thing is expandable and can deal with from 1-255 cells per pack, voltage range 2-5V per cell, Max Pack voltage 650V Max Pack capacity 650ah. ;)

http://www.solarvan.co.uk/Master030708.txt
http://www.solarvan.co.uk/Master030708.bmp (One circuit per system)

http://www.solarvan.co.uk/Slave030708.txt
http://www.solarvan.co.uk/Slave030708.bmp (One circuit per cell)

Thanks for all input so far.
 

retepsnikrep

Senior Member
Just an update on this for the very helpful members on here ;)

I've built up one of my Master Boards (28X1) and 50 Digital Slave (08M) boards now, some others have been built up and tested in Analogue mode by another member of the Battery Vehicle Society in UK and are working. I shall shortly be attaching my Slaves to my pack of 50x40ah Lifepo4 cells for my Honda Insight Battery upgrade. A lengthy blog/thread on the BMS project is here which I update most days.

http://batteryvehiclesociety.org.uk/forums/viewtopic.php?t=1245

The latest posts show where we are.

A couple of pics to wet you appetite.

http://www.solarvan.co.uk/bms/50Slaves.jpg

http://www.solarvan.co.uk/bms/Master&Slaves.jpg

Thanks Peter UK :)
 

moxhamj

New Member
Looking very impressive. I see from the slave schematic you can power up a sensor remotely - I presume this means the sensor is off most of the time (to save power?).

I like the liberal use of optocouplers - these are perfect for translating the info from the top of a high voltage stack.

Your website http://www.batteryvehiclesociety.org.uk/wordpress/ is very impressive. Here in Australia we have lots of political debate about "going green" and reducing dependency on foreign oil. There are a handful of enthusiasts making electric cars in their sheds. There is one guy here in Adelaide who has been trying to import one for years but it keeps failing the crash test. So there are electric scooters on the road but no cars.

But over in the UK, it looks like you have lots of models available. I like the database of recharge points!
 
Last edited:

retepsnikrep

Senior Member
Looking very impressive. I see from the slave schematic you can power up a sensor remotely - I presume this means the sensor is off most of the time (to save power?).
Dr, What part of the schematic are you referring to?

The latest Slave schematic is here

http://www.solarvan.co.uk/bms/Slave120808.jpg

The Slaves can be used in Digital output mode or in a pseudo analogue output mode depending on how you assemble them.
 

ciseco

Senior Member
Hi,

It came to mind, I'm sure I ran into something a year or two back where some guy in south america somewhere was using the basic 1 wire battery sensor 2438 I think, not 2760 (might have my numbers wrong) to do just that.

There was one on each battery monitoring temp and voltage, then these were opto coupled to effectivly one bus and to a PC, no reason why it couldn't be a 28x1 instead. Each one then could be addressed by it's serial number and if one ever died you'd only have to change the address in the 28's code.

Not a suggestion more of a brain dump :)

Miles
________
Plymouth Expresso
 
Last edited:

moxhamj

New Member
Ah, I get it now - U3 is for data input. And power input is directly from that battery.

I presume the slave is sitting in a serin waiting for some data, the master sends a request and the slave loads up the battery, measures the volts and sends the data back back?
 

retepsnikrep

Senior Member
Almost ;)

The slave is in a loop checking the cell V (it's own supply) and operating the balancing load as reqd. When a data input or high signal or interrupt is received via U3 then it transmits the cell voltage data on the Master bus and passes on the data input or high signal or interrupt to the next Slave in the chain via J4 (and U3 on the next Slave)

It's got a few operating modes and different software versions, including a digital mode where serial commands are recieved via U3, or a simple high signal via U3, or an interrupt. The Slave can transmit the recieved data etc to the next Slave in the chain once it has responded as reqd.

In analogue mode U3 is not used but it's moved to U4 where it becomes a second output bus. There's a lot more on the BVS forum thread.
 

retepsnikrep

Senior Member
The Long Day Closes.

Thanks to massive help from Roger Crier a fellow UK Insight owner who drove up from Birmingham today bearing gifts, and spent 8 hrs in my garage with me wiring up 50 Picaxe slaves, I can report the finished assembled cell blocks & bms boards.

A couple of wiring glitches aside all 50 slaves worked first time BMS is also reporting correct pack voltage. Scan rate at 1200baud comms is once per second ish for 50 cells No issues with 50 opto's on Master Bus at present.

I have made no efforts at the moment to include any power saving functions in the Digital Slave software unlike the Slave Analogue On/Off ones which I did for Greg, so the Digital Slaves continous drain is an average of 1ma, or 41days for 1ah consumed.

Roger kindly provided some nice clear vacum formed safety covers for the packs and they set the job off a treat.

The led's looks pretty dancing round the slave boards as well It's like a Mexican wave

www.solarvan.co.uk/InsightTrio.jpg

www.solarvan.co.uk/HalfPack.jpg

www.solarvan.co.uk/FullPack.jpg

Jobs for next three days include adding plugs and sockets for the various data/bcm connections to each cell block.

Installing the Master Display and front to back wiring into the car for the video and control led's switches.

OK just to humour me here's a video of the 50 BMS Picaxe 08M Slave boards talking to the 28X1 Master

http://www.youtube.com/watch?v=JIMPLMPyopY

Probably the most expensive Led Christmas Lights in the world :eek:
 

kevrus

New Member
Very impressive, a well developed project. Slave and master PCBs look very professional, you should be extremely pleased with the result.
I'm interested in electric cars myself and would love to build one, only funds won't allow it, and if they did, I couldn't see myself stretching to lithium ion cells, unless the price came down substantially.
 

nbw

Senior Member
Could you perhaps use an I/O expander like the MAX7300 to give you many more ports without so many chips? You'd need i2C on the face of it, so maybe an 18X - gives you more codespace (if you needed it). A 7300 from memory would give you 20 more ports, so maybe you'd want two of the little cuties.
 

fernando_g

Senior Member
My most sincere congratulations!

While most large automotive corporations are still dragging their feet over electric vehicles, there are enthusiasts like yourself, which on their spare time and with limited funds, are showing them the way!
 

centrex

Senior Member
Hi Retepsnikrep
I have been following the thread on the battery managment system, most impressed.
Are you likly to present any more details on the pcb's and the software.
Regards
Centrex
 

premelec

Senior Member
LTC6802 battery string monitor chip - 12 cells

FWIW a recent IC has been announced specifically addressed to monitoring EV battery strings - 12 cells per chip and chainable... from Linear Techonology linear.com part LTC6802 - a look at their data sheet gives a lot of information on how to do it.... I don't know if they are available yet but it looks good.... more various ICs no doubt will become available as EVs take over! [I'm still a bicyclist but am interested in energy storage battery banks].
 

retepsnikrep

Senior Member
I've seen that chip. It was inevitable that something better would come along as soon as my project was nearly finished. Still someone has actually got to build a functioning cheap BMS using it, so good luck to them.

Peter
 

retepsnikrep

Senior Member
Well I have by BMS installed but have a problem with noise. The Honda Insight car is a noise nightmare. Please have a look at the scope pic from the 28X1 2400 baud serial data bus input. You can see a couple of bits of serial data underneath all the mush but the horrible high voltage spikes are causing various errors.

http://www.solarvan.co.uk/bms/Noise1912.jpg

So I'm after ideas on filtering or pre-conditioning circuits. Experts please!!

Incredibly even with all that noise it's working some of the time!

For info here are the Master & Digital slave circuit diagrams. There are 50 slaves (cells) with the Master bus outputs in parallel all connected to the Master Board master bus input.

www.solarvan.co.uk/bms/Master080908.jpg

www.solarvan.co.uk/bms/DigitalSlave080908.jpg
 

fernando_g

Senior Member
Re: noise, how to minimize it, in order of cost and complexity:

1) Twisted comm wire pairs, with the (-) grounded at a single point (to avoid ground loops).

2) 100 pF caps from each (both + and -) of the sensing leads to chassis ground....SAFETY NOTE: these caps should have a voltage rating of at least twice your steady state voltage plus 200 volts, and be self healing film, not ceramic caps.

3) Clamp-on ferrite beads.

4) Common-mode choke.

5) All of the above.

6) Pray!
 
Last edited:

inglewoodpete

Senior Member
Without trawling the depths of your circuitry, I interpret the noise as being induced into the serial data line after the optocouplers. (Since the optocouplers are a digital device they can't pass added noise, only subtracted noise - if that makes sense).

The noise is fairly regular and comparatively high frequency when compared with the data.

* Regular noise: can you find the source? Screened leads would help here (ensure that current does not flow through the screen).
* Overvoltage and undervoltage noise can be clamped using a zener clamp diode. I suggest a 5.1v zener between the data line and earth.
* Also use a low pass filter. By the look of the trace picture, it should be fairly easy to calculate a break frequency.
 

retepsnikrep

Senior Member
Looking at the spikes in my scope pic earlier anyone good at maths want to give me a few values for the low pass RC filter components. I know the formulae but it's all double dutch equations to me.

http://www.solarvan.co.uk/bms/Noise1912.jpg

You can see the scales for the display.

2v per divison vertical

200us per division horizontal

The spikes are very clear and very short prob <50us , the underlying serial data pulse is about 500us long.
 

BeanieBots

Moderator
Don't want to be the bearer of bad news but I don't think you will be able to get rid of that simply by slapping an RC on it. Besides, the fundamental already has a 'sloppy' off transition which only get worse with an RC.

This looks more like a poor layout problem.
It looks very much like switcher noise. Does the waveform tie up with a switcher of your design or is it from the car's own electrical system?
You get much better results by preventing it from getting in the first place than trying to remove it once there.

If it's on the 0v line, then RC filtering will make it worse. It would simply add AC coupling between your signal and the noise source. Is it even real or only present when the 'scope is connected. (use a ground isolated 'scope to test).
 

retepsnikrep

Senior Member
Thanks to all for the advice and support on this forum. Yes Beanie it's really there and it is the Insight electric motor switching noise. It comes over signal and 0v from Slaves. I'll be reworking my layout with some decent screened cable and some clamping diodes/filtering over next few months.

As a temporary measure I have installed my system in an analog mode with the Master removed for upgrading with filtering etc. Seems to be working well now in my Lithium Honda Insight upgrade. I am using my digital slaves programmed with some analog control code and an anlaog temp master board to control my charger and over/under V alarms.

http://www.solarvan.co.uk/bms/ANALOGUESLAVE_DIGITALPCB_211208_V01.TXT

http://www.solarvan.co.uk/bms/DigitalSlave080908.jpg

http://www.solarvan.co.uk/bms/AnalogueMaster010109.jpg

Went on a 45 mile test run yesterday and pushed mpg upto 112mpg :eek: from my previous norm in the current cold UK winter weather of about 93mpg. I'm hoping to break the 150mpg barrier in decent weather later ;)
 
Last edited:

retepsnikrep

Senior Member
Finally

For those who may have been following the development of my Picaxe powered BMS it is now up and running in digital mode finally.

Full story here

It's a big project split into three threads on the BVS site, but the last few posts on the Software thread contain the latest info.

http://batteryvehiclesociety.org.uk/forums/viewforum.php?f=3&sid=dacf7c87f98738e174176f073c5de74b

Thanks to all here for there help and interest along the way. I couldn't have done it without you.
 
Top