Old BMS General Thread

[retepsnikrep]

Greg

Here is a fully converted version of the Analogue on/off software using the new constants. I'll be interested to see if it runs fast enough at the lowest speed.

www.solarvan.co.uk/bms/SlaveAnalogueOnO ... _v1_13.txt

A measley 55 bytes used out of 256 allowed

[GregsGarage]

Finally got a chance to test the new version of the analogue slave software. I am comparing my current version 1.08 with Peters latest version 1.13.
First thing I noticed is that the new software runs much faster. The old version runs at 250khz and the new software running at 125khz seems quicker to respond to voltage changes! I tried at 31khz, but that takes much to long to respond to voltage changes. The risk here is that a low voltage alarm might not trigger in time and cell voltage will drop so low that the low voltage opto stops conducting. This seems to happen at around 1.25 volts and starts conducting again at around 2.20 volts.

As far as the actual figures go I have made a chart. A note on the equipment I have used. My DC power supply is a £20 0-20 volt 2 amp supply. It has a single knob to adjust voltage, but isn't designed for the fine adjustments required for these measurements. My volt meter cost me a fair amount, about 20 years ago , and was last calibrated Sep 2000! Still it has served me well. So keep all this in mind when looking at the values below.


_________Programmed________________________________
____________Value__________Ver_1.08_________Ver_1.13_
MaxVOptoOn___375____________3.78_____________3.80___
MaxVOptoOff___360____________3.64_____________3.64___
LoadOn________360____________3.65_____________3.63___
LoadOff________355____________3.59_____________3.61__
MinVOptoOff____250____________2.52_____________2.54__
MinVOptoOn____240____________2.40_____________2.44___

So there you have it, the results of my very quick test of the slaves. I will be using the new software as the lower clock speed will lower current consumption as will the new resistor value for R5 (100K from 22K).

Edit: BTW that 3 byte download bug happened again with the new software. I had to copy and paste the main program, without the comments into a new window and then it downloaded o.k.

[retepsnikrep]

Interesting Greg

I think the speed of the voltage drop you envisage in the worst case scenario would be highly unlikely, but it could happen. I think a one second repsonse time should be adequate, that's what the digital Master works at so it seems a good target.

You should be able to get a bit better accuracy with the newer version by experimenting with the values, just add/subtract 1 point at a time from the below calculated values and see how it responds.

symbol MaxVOptoOn = 336
symbol MaxVOptoOff = 350
symbol LoadOn = 350
symbol LoadOff = 356
symbol MinVOptoOn = 526
symbol MinVOptoOff = 504

The constant (63200) we used to calculate the above may be a bit out, there will also be individual variations and temp related fluctuations amongst the lm385 VRef.

The LM385 is only accurate to 1% at the relevant temps.

Edit

Just thinking out loud, and it may not make enough/any difference.

We could re-order the if/then loops so the most important are evaluated first. Also we could cut out the ones that are not reqd after an evaluation has been made.

So for instance if we put the low voltage check at the top, and the cell was under the permitted voltage, it turns on the opto but then goes straight back to the Adc again rather than evaluating all the other stuff which is irrelevant in this situation. Obvious really Might allow the lowest cpu speed as well

In the new version below I have given priority to the activation of the low V and high V opto's, rearranged the other if/then's a bit, added a couple of goto's to jump over needless comparisons. and reduced cpu speed to the minimum 31khz.

I haven't tweaked the constants further as I don't have an adjustable psu to test them with. Greg over to you

You may get the editor text bug sorry about that.

www.solarvan.co.uk/bms/SlaveAnalogueOnO ... _v1_14.txt

[GregsGarage]

Good idea Peter. I will try it out, but it probably won't be until Monday.

Greg

[retepsnikrep]

I'm away from next tuesday until 01/12/08. I'll send off the returned Ver A3 picaxes before I go. Hopefully they will be back by time I get back.

Charger should also be here, so may be able to launch into final install phase. The Master software just needs final tweaking of the proportional charger control routines and then it should be good to go.

I will try and monitor thread whilst away but won't be answering any e-mails. Keep tinkering whilst I'm away.

[mikep_95133]

Hi Gents,

It occurred to me this morning as I was about to install the slaves onto the Lithium cells, that I don't know how to wire the slaves together in analogue mode. I assume the master is not connected. Maybe the slaves are not connected to each other at all?

Same with the digital slaves. Is there a diagram posted somewhere that shows the wiring between the master and digital slaves?

I translated one of Peter's pictures to a diagram for the digital slaves and master. That's what's running on my bench. But for the sake of having a document on those days when it's late at night, and I'm tired, trying to install the hardware.

Mike

EDIT: Whilst you all were sleeping, I tried going one page at a time through all of the posts. Finally on page 8 I read Peter's post: "To implement this the Slave Master Bus optos are connected in parallel as per normal version. The Slave Bus Input optos are also connected in parallel, unlike the normal system where they cascade from one output to the next input."

We are up to 40 pages of posts now. In my experience with open source projects, it makes sense to keep a web based faq for the project so new members or members without any memory (me) can find things again. 8 pages down, 32 pages to go....

[retepsnikrep]

No I don't have a document showing interconnnections. Good Point I'll do one at some stage when I get back.

Analogue Slaves Boards are assembled slightly differently to Digital Slave ones as the opto's are in a different place on the pcb They are not directly interchangeable. Please check the schematics.

Greg can help you more here as he is using the Analogue version at present. I have suggested using IC sockets for the opto's so you can change from Analogue to Digital or vice versa.

www.solarvan.co.uk/bms/Master080908.jpg
www.solarvan.co.uk/bms/DigitalSlave080908.jpg
www.solarvan.co.uk/bms/AnalogueSlave080908.jpg

The Analogue Slaves have all the High V(Master) opto outputs J3 paralled together as a single string. Ditto the Low V(Slave) opto outputs J5. So you have two circuits (4 wires) to interface with. The Master board is not used at all with Analogue Slaves.

The Digital Slaves have all the Master Bus Opto's wired together in a parallel string this connects to the Master Board Master Bus Input J13.

The First Slave in the digital string Has it's Slave Bus input J2 connected to the Master Board Slave Bus output J11. The First Slave in the string has it's Slave Bus output J4 connected to the second Slave's Slave Bus input J2. This continues throughout the string. The Last Slave only has connections to it's Slave bus input J2 and Master bus output J3 as it has no Slave to output too. Hope that makes sense.

Remember the request to send voltage data on the Master bus is triggered by the Master Board sending a signal to the first slave, which then propagates down the string of Slaves being passed from one to the next. Each as it recieves the signal dumps it's voltage data onto the Master Bus. The lack of a signal at any point on the Master Bus is how the Master board flags up a comms/cell problem.

Good idea about the FAQ, but I don't have time to think about doing that until my install is finished. If someone wants to start one then be my guest as long as I can freely contribute to it without jumping through any hoops. Are you thinking of something on wikapedia? or something on here, I've never done one before. How does it work?

[mikep_95133]

The Master board is not used at all with Analogue Slaves.

This is different than the post on page 8. So I take it things have changed and the analog slaves should not and cannot use the master?

A faq is something in my experience that you have a new person to the project read so they don't ask the same questions. A primer.

Also I've found it useful to have an area on the faq that keeps track of the major revisions of software. That way if something needs a previous version of software, a person won't have to read through all of the posts to find out what each revision is capable of.

Something that can be updated easily makes sense. I've used spreadsheets in the past, but something online might be better/faster.

Since this post tells me that analog means no master at all, I'll put them in as analogue's until I get the wiring installed from the 312v battery pack to the cab of the truck.

Do take a real vacation Peter. You are far better for the project with a fresh brain than not. At least that's how it is for me on my projects.

Mike

[retepsnikrep]

Mike

I can't see where on page 8 it mentions Master & Analogue as working together? I might be blind though

The Master board is for Slaves assembled in the digital configuration.

Analogue configured Slaves are stand alone (I suppose I shouldn't call them slaves as Greg pointed out earlier, if they have no Master!), they have a simple on/off opto output which can control some chargers or simply provide a warning alarm about an over/under V condition. How that is implemented is upto to the end user.

Analogue Slaves also have the potential to have a proportional output on the high V opto however I haven't done any real work on the software for this at present though.

Most users will probably utilise the digital system as it provides a much greater range of information and error detection.

If an Analogue configured Slave locks up or fails to react to the cell voltage etc the end user has no means of knowing that or why/which one/cell/slave is affected.

With Digital configured Slaves the Master board can detect data errors and timeouts alerting you to a malfunctioning cell board.

I might start a FAQ after I get back unless you want to start one now, a simple text file is best IMO. I'll probably add it to my first post on page 1 and edit that as things change.

I probably won't post much whilst I'm on holiday for the next two weeks. However this project is very important to me and I've dedicated a huge amount of time and resources to it with the end in sight I'm working hard to make sure it all comes together. I'm waiting with baited breath for my Insight to run with lithium power and hopefully going to push the mpg display well beyond 150mpg

[mikep_95133]

Thanks for the clarity Peter. I just wired the 3 slaves for digital, but programmed them for analog. I removed all 9 opto's from the 3 slaves as there is nothing for them to talk to. Eventually when I get the shielded cable installed from the battery box to the cabin, I'll program them back to digital, with the wiring already done between the slaves.

As for the faq, maybe while you are gone, we can start something with the understanding that it has to have your approval. Maybe theother members if they have any interest could chime in on this faq idea. I don't mind doing it alone as I have to do something to keep everything organized. I spent the last month working on a newly acquired EV and have forgotten where I was with the BMS. So a posted reference somewhere would be of great help.

Your contribution to this project is staggering. This is what I've seen from each of the open source projects I've participated in. So I sure try hard not to be a burden, but a contributor.

Mike