Old BMS General Thread

[wjdennis]

IC, so you would need 70 slaves then in your current configuration.

Sounds a potential nightamre with two Masters, I think it would be better to connect in Parallel/Series if cells are same chemistry/voltage, then use 35 Slaves (One per pair) and One Master. That should work IMHO.

Peter, that becomes more of a logistical and safety problem. Supposing a buddy-paired 200Ah and 90Ah cell, and the 90Ah cell gets an internal short. The 200Ah cell then tries to dump all of its energy into the 90Ah cell, and boom, crackle. So there'd have to be a fuse between each cell to guard against this.

For paralleled strings, let's say now that one of the 90Ah cells again short circuits. Now you have two strings at 3.6V difference, say. So the 200Ah string dumps its load into the 90Ah string. But now you have the internal resistances of all those cells in series to limit the current. So just a few amps passes from one string to the other until they equalize. No big deal.

Here's a paper on paralleling strings: http://www.battcon.com/PapersFinal2002/ ... er2002.pdf

Bill

[retepsnikrep]

If a cell develops a sudden internal short then it has to discharge itself through it's own short which has the same effect. So I would not be worried about that. If it develops a gradual short then it would just pull the other cell down and gradually discharge it. Yes the extra power available from the other cell goes into the duff one but that would show up pretty quickly on the bms. Of course that's up to you. Risk management is personal thing and cost does come into the equation. I'm more than happy for you to buy 70 slaves and 2 Masters if you want to later when they are working but I was trying to save you some money

Just a little update today.

Sent off the order for the Pcb's

Eta two weeks.

Re-named the Software versions to fit in with the Master and Slave boards.

Master Board is V1.00

http://www.solarvan.co.uk/bms/Master120808.jpg

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

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

Slave Board is V1.00

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

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

I know it say V2.00 on the image but it does say V1.00 on the boards

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

Added the interrupt routine back into the digital Slave software to reduce the timing problem. Also added the slow speed power saving code for the Slaves. They are now set to run at 31khz but the Digital Slave cranks back upto 8mhz when the interrupt is detected to send out the data at maximum bus speed. Of course we can adjust Slave cpu speed later one we see how it all works in practise

May tweak the Digital Slave power saving further later and put them to sleep if no data requests recieved for last minute say.

So the current Software versions also re-named are
http://www.solarvan.co.uk/bms/Master090808_v95.txt

Master board watchdog timer
http://www.solarvan.co.uk/bms/Watchdog290708_v04.txt

Still lots of work to do on the Master program

Digital Slave for use with Master
http://www.solarvan.co.uk/bms/SlaveDigi ... _v1_02.txt

Analogue Slave with proportional charger control for use without Master
http://www.solarvan.co.uk/bms/SlaveAnal ... _v1_02.txt

Analogue Slave with On/Off charger control for use without Master
http://www.solarvan.co.uk/bms/SlaveAnal ... _v1_02.txt

I think that's everyting for a bit. All the bits arrived from Farnell as well, I thought about using Mr e-bay cheapo knock of components but decided reliability was more important. I'll do a full cost analysis/price for the project later. But as a taster for the Slaves.

Pcb £1.50
Picaxe 08M £1.50
Other bits and bobs call it £2.00

If the Slave price can be kept under a fiver that's not bad at all. Could be a quid or so cheaper with bulk buying of pcb and pics.

[GregsGarage]

Hi Peter,

Great news that your expecting the pcbs in about a fortnight. I have decided to ditch the Ping boards and go with your slaves in analog mode for now. Will tackle the master later.

You posted a draft parts list 31 July. Is this still correct, or do you have an updated list with maybe the Farnell order codes on it?

Thanks

[retepsnikrep]

You posted a draft parts list 31 July. Is this still correct, or do you have an updated list with maybe the Farnell order codes on it?

Thanks

Greg that would be too easy I haven't got round to that yet.

But I have all the components to make 100 slaves in stock, so I may as well sell you the complete kit to make up how ever many it is you wanted. I bought the Picaxes and optos, voltage refs, resistors, transistors etc in bulk so it will be cheaper for you as well

As you are an earlier beta tester of my system I'm eager for some real life testing results. I will make up a couple of slaves myself first before I send out the pcb's, just in case some horrendous production mistake on the pcb renders it useless. If no smoke comes out I'll send you the parts and pcb's you require. No guarantee whatsoever at this very early stage as I'm sure you understand

The only parts I don't have are the 4 pin dil/dip sockets as I will be soldering my parts to the board, I suggest you use them on the slave for U3 & U4 so you can move the optocoupler as discussed earlier to change from analogue to digital slave. Sockets are not as good as solder for reliability but should be OK. I'll solder mine into Master position and make a couple of seperate analogue slaves just for testing.

[retepsnikrep]

Been looking around at improving display with graphics etc for MK2 and came across this.

http://www.tvterminal.de/index.html#english

Very clever and can communicate with Master very easily.
This demo video shows some very nice features of the chip. Provides a ps2 keyboard interface as well.

http://www.tvterminal.de/tvtcolor_demo_fbas_336x256.mpg

Nice to have the coloured cell bargraph display back agian. Now this needs a VGA monitor but can be used with another chip to give RCA and SVHS video out as well.

http://www.tvterminal.de/TVT_composite_AD724.gif

There is also a simpler cheaper black and white single chip version which interfaces direct to RCA monitor.

http://www.tvterminal.de/tvt_mobi_2.html

Anyway I have ordered one of each to play with

MK2 is many month away though but it's food for thought

[retepsnikrep]

Removed BMS Configuration routines from Master program and started a seperate program for Testing and Configuration.

I wasn't running out of room, but it was getting a bit untidy.

Master is now much reduced in size and stacks of space for whizzy graphs etc later!!

Once initial system is assembled, user can load Master Board with testing software and connect up the pack slaves on the bench. Cells don't need to be interconnected for main power at this stage so it's safer.

Configuration/Testing software limited at the minute, only performs two tests so far.

1) Measure cell voltage sensed by attached single slave. (Good for testing single units and I will probably add some tweaking options to this.)

Adjust V ref calibration/constant etc

2) Data transmission test sends various codes through Slave chain and detects/compares message received back by Master. Could be used for decent burn in testing.

Less than two weeks to go before real testing starts in earnest

I've decided to just upload a zip file now of all the software when I update. So here it is.

www.solarvan.co.uk/bms/BmsSoftware180808.zip

[retepsnikrep]

They have arrived

After a bit of faffing around with courier Delivered personally by David Cowperthwaite in the end Pleasant chap.

50 x 40ah Cells Lifepo4

According to test data on each cell.

Average IR is about 1milliohm.
All measure about 3.3V
Measured capacity according to labels varies betwen 42-45ah.

Need to fuss them up now, can't wait to get stuck into building up the pcb's and seeing if it's going to work

Got to find space for them all in the Insight as well

www.solarvan.co.uk/bms/50x40ahCells.jpg

Supplied Interconnect's appear to be tinned copper I'll see how they hold up under a decent load before I finally decide wether to use them or get some decent braided ones made.

The supplied Aluminium screws I suggest are not up to the job but that's a minor issue, as I will be using threaded stud to allow bms boards to be mounted anyway.

Now wish I had pcb's to work on as can't do much until I have built and tested a few.

[dillond666]

I,ve been messing about with the Linux version of Eagle Pcb and have knocked up a new schematic.

http://www.dillond1.pwp.blueyonder.co.uk/agmslave.png

I have used a 14 volt varistor to protect against transients and have used diodes 1,2 and three to tweak the varistor voltage up a little as they have a 0.5-0.6 voltage drop each when forward biased. This means no more than 15.8v should ever reach the potential divider and the potentiometer can be set to ensure that the picaxe never gets more than 5v at the analogue input. If this is a poor idea please could somebody say so?
There is a jumper to ensure that programming doesn't pulse the opto.
I have used the improved serial connection method as I will be using the readadc10 command as well as the debug command (for calibrating).
There are three screw terminals to allow connection to the battery and to the external load resistor.
I have used ethernet connectors for the data in and data out.

I,ve got the pcb layout done too but I am not too confident as regards masks etc. (do I really need them?)I had hoped to do a single sided board and use wire links where necessary but the only way I can seem to get it to work is by using via's (whatever they are ) and a two layer pcb.
I hope to get the Gerber files sorted soon and maybe send off to the man in China for a few boards.

http://www.dillond1.pwp.blueyonder.co.uk/agmpcb.png

If anybody notices any schoolboy errors please feel free to comment

[retepsnikrep]

Keep at it Dillond nice to see more ideas.

[retepsnikrep]

Offering up the cells to the Hindu god of electric conversions (Ideas for the name gratefully accepted ) I though 'Riva' or 'Eva'

Seems to fit as planned 5 rows of 10 cells, like the proverbial glove in the back of the Insight in the spare wheel well.

Have to lose the spare wheel or carry it elsewhere, but now carry a can of the get you home rubber gloop, never mind, also lose the weight of the spare which is a few kg so not all bad

Interesting with original NIMH battery fitted, suspension with Matiz rear springs is fine with the extra 70kgs of li-fepo4 in the back. Measured from floor through centre of rear wheel to joint at top of wheel cover cowling 655mm. Still about 25kg to take out later when old battery is removed.

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

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

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

To level up the bottom of the wheel well and provide a flat base for the cells I have cut a piece of lightweight plastic foamalux 3mm sheet to size and I will use expanding builders foam underneath to fill the voids and provide support to the sheet on top. I though about pouring plastic resin in the back so it could find it's own level with gravity, but that's a bit too permanent at least the foam can be removed reasonably easily and it's a lot lighter!

There is enough clearance for the cells underneath the IMU compartment and the whole lot will be covered with a sheet of 3mm clear lexan anyway to allow viewing and insulate the top teminals. They will be strapped into 5 or 10 cells blocks and then held down with appropriate bolts.

I thought long and hard about trying to put the new cells into the battery compartment in place of the nimh pack, but there is just not enough room. Plus you then have to find a way of mounting the various components that attach to the oem pack. As I have a spare empty battery case I will simply use that in the original position to allow mounting of all the original components and extend the relevant BMC cables and power cables to reach the new pack.

I also think the weight distribution is a bit better with the new layout as that will help to counteract the weight of a single driver and all the cells on the right of the car

I've been looking at the voltage specs and am hoping I will be able to get away with simply tying the bcm subpack monitor leads onto the appropriate point in the new pack, basically every 5 cells. The total unloaded pack voltage is about 167v which also looks good. Be interesting to see the voltage sag under the 100A max drain later.

The other good thing about the choosen layout is I can assemble, test and install the pack before I even open the battery compartment which should mean I can be off the road for the minimum amount of time later.

Peter