Old BMS Hardware Thread

Threads relating to the BMS system begun by Peter Perkins

Moderators: GregsGarage, retepsnikrep

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Sun Jun 27, 2010 2:35 am

Thats a lot of questions.

Re the dash buttons/LEDs - I've had a look thru the threads and can't find any info on what they all do exactly.

I found the schematic for them ... BMSKeyMatrixLeds.jpg (?) So...

J1/Pin 1 - +5V
J1/Pin 2 - D4 - Watchdog LED - Tells you watchdog is working by flashing (at what rate?) This only illuminates if the watchdog chip signals a timeout error
J1/Pin 3 - D3 - Pulse LED - Tells you Master is working? How? This flashes once a second with picaxe master, more frequently with a pic master
J1/Pin 4 - D2 - Dash LED - Doesn't do anything yet?Works when master in place under specific conditions
J1/Pin 5 - D1 - Dash LED - Ditto Works when master in place under specific conditions
J1/Pin 6 - Switch Common rail(?) What do all these switches do?Control the command system and can reset soc, trip meter etc
J1/Pin 7 - Master Reset Switch(?) Resets whole caboodle? Resets Master chip
J1/Pin 8 - Gnd



JP1 - Master Reset Enable on the Master PCB - is this supposed to be shorted for normal operation or open? Shorted when in use but open when downloading software to master and during testing etc. It allows the watchdog chip to reset the master chip if a timeout occurs

WRT... 'because without an active master chip plugged in, the input to the watchdog pin floats and detects static and stray emi as pulses reseting the watchdog timer'... can I confirm that the input pin is 4? And that it needs to be 'pulled down' to gnd with a 1k resistor?Pin 4 is correct on the watchdog IC however connect the 1-10k resistor to pin 3 of the switch and led connector (pulse led) and then to pin 1 or pin 8 of the same connector

I take it that the dash buttons and leds do not need to be installed to test the Watchdog on its own? And the temp sensors? Both of them?Correct.



For bench testing the system, I thought about using 3 x AA rechargeable cells (in series, obviously) to simulate one lithium cell and then rig 2 or more of these to the same number of slaves. Can you see any issues with this? No issues good idea, but make sure combined cell voltage is within BMS working range 1.75-4.30V

Will I have to remove the PICs from the unused slaves to avoid errors etc? No but you will probably have to reprogram them when i get the software working correctly!!

Connecting the 16 slaves PCBs to the master... I assume Master PCB 'Master Bus', J11 connects to Slave PCB 'Master Data Bus', Conn3 and the other end of the SlavePCB 'Master Data Bus', Conn 4, connects to the Conn3 of the next slave board and so on? Correct, you can stack the boards with PCB standoffs if you wish, but when stacked stand them on there sides in use so heat can rise between the boards and escape

Further, the Master PCB 'Slave Bus', J8 connects to the Slave PCB 'Slave Data Bus In', Conn1 and the Slave PCB 'Slave Data Bus Out', Conn2 daisy chains to the next Slave PCB 'Slave Data Bus In', Conn1 etc...? Will a simple twisted pair of 1/4 of a bit of Cat5 cable do for these connections?yes twisted pair is fine for all this, try and keep things as short as possible of courtse. polarity of the connections is important, so pin 1 of any connector goes to pin 1 of any other

Lastly, the Slave to cell connections... We already discussed fusing these with a 1A, suitably rated, in-line fuse and I gather each slave connects across the cell it is monitoring with adjacent cells sharing the connection thus saving a wired connection per cell?Correct you need a good size wire for the cell connections and keep them as short as practicable, voltage drop in these leads gave us the problem which required the fancy new software as slaves interfere with each other when switching the loads when only one wire is used due to voltage drops in wires. The single cell boards do not suffer from this issue. Remember your high current cell interconnects will also factor into the equation so be methodical and clean every connection so it is spotless and tight


How are your cell blocks being assembled? can you post a photo? are there any long high current connections between seperate cell blocks?
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

martinwinlow
Posts: 79
Joined: Mon Jun 11, 2007 9:35 am
Location: Herts, UK

Re: Old BMS Hardware Thread

Postby martinwinlow » Sun Jun 27, 2010 11:36 am

Hi Peter,

Thanks for that - I can see I should be able to figure out what the dash buttons and LEDs do exactly from the Master txt file.

Pic of my cell layout attached...

I have updated the BMMS page on my site - still needs some re-jigging for clarity and I'll add programming and alert descriptions in due course. Perhaps you would have a read thru and tell me of any errors/make suggestions for further content. Do you mind if I use your PCB pics as mine are not as pretty!

You may notice I put a bit in about maybe using the speed sensor as an RPM sensor instead... which leads onto the very easily implemented possibility of using the controller cutback facility to prevent motor over-speed. Maybe the V3 could have 2 speed sensor i/p's one for each??

I have got the watchdog to work in as much as the buzzer activates after 70 sec or so - the first buzzer was u/s. Not getting any blinking LEDs tho but I'm sure I'll get to the bottom of that soon. With the Master MCU in as well, I get a display that makes some sort of sense - I need to put a few more inputs in place (temp sensors for starters) and rig up a slave board as mentioned before with AA bats to simulate Li ones.

I gather the analogue temp sensors are LM335's not LM35's so while I am awaiting the 335's, can I simulate the 335's presence in the circuit with a resistor just to test things out and stop the temp alarm going off whenever I switch the Master on?

Meanwhile, my van is back on the road after I blew up my new 800A controller - looks like the 100V-240VAC mains to 24VDC PSU that I was using as a DC/DC converter to supply the controllers 24V is allowing the full pack voltage to get thru when first switched on and it's (not surprisingly) frying a couple of optos and a regulator on the controller 'control' PCB. So I am back to using a tap off the pack for the moment - it is the main (only) issue I have the Belktronix system as it causes imbalance in the pack - 250mA worth so an hour of balancing on a full discharge...

Progress, then... !

Regards, Martin.
Regards, Martin Winlow
Herts, UK
http://www.evalbum.com/2092
www.winlow.co.uk

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Sun Jun 27, 2010 11:46 am

Martin Nothing attached?

I gather the analogue temp sensors are LM335's not LM35's so while I am awaiting the 335's, can I simulate the 335's presence in the circuit with a resistor just to test things out and stop the temp alarm going off whenever I switch the Master on?


Assuming you are using the Picaxe master just REM out the analog temp line in the master program and reprogram it.

Code: Select all

   gosub CheckAnalogTemp         ;Gosub Analog Temp routine to collect/display battery temperature


Yes you can use my pics.
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

martinwinlow
Posts: 79
Joined: Mon Jun 11, 2007 9:35 am
Location: Herts, UK

Re: Old BMS Hardware Thread

Postby martinwinlow » Sun Jun 27, 2010 2:18 pm

Cell layout attached (again - I didn't notice the file over-size limit warning).... MW
Attachments
IMGP1578.JPG
IMGP1578.JPG (161.99 KiB) Viewed 4428 times
Last edited by martinwinlow on Mon Jul 05, 2010 8:57 am, edited 1 time in total.

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Sun Jun 27, 2010 3:17 pm

Can't really see much from that can you post a full res pic please, e-mail it me if you need too. A close up of the top of the cells etc and where are the slave boards going?
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Mon Jun 28, 2010 10:20 am

Just a note for those busy with multslave board builds. Please ensure you use some sort of plug and socket between the slave board cell wires and the cells themselves. if you hardwire connections from the slave board direct to the cells you will have no way of cutting power to the slave boards if you need to pull the slave pics to reprogram them. That makes reprograming unsafe and risky for the pics. I used a 50 way D sub connector for mine. I can isolate the pack from the slave boards by unplugging a single connection.

Edit
I've been using this neat 5 way switch for my BMS buttons. All combined in one tiny unit. You can solder the resistors directly to the switch if you are careful and then you have a really neat small unit. Put a little wooden ball on the knob end and look really neat. I'll post a pic of mine later.

Up, down, left, right and center push select. Good for the Master menus.

http://uk.farnell.com/jsp/search/produc ... KU=1435775
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Tue Jun 29, 2010 11:28 am

Some things discovered today.

I found the I2C DS18B20 temp sensors will not work if the varistor VR5 is on the Master V2 board so you have to remove it if you want to use them. :roll:

I do recommend people try the I2C temp sensors as they are much more accurate, you can also use one to calibrate analog sensors if you have any attached by comparing the readings and adjusting the software accordingly. They make a very accurate reference/comparison.

To make the button switch using the recommended multi way button device you need six resistors.
The resistors need to be the same value but can be from 1-10k I used 4k7 as I had more of those.

http://uk.farnell.com/jsp/search/produc ... KU=1435775

Connect a resistor from pin 1 to the + (positive) Master feed (pin1) on the buttons/led connector. Connect another resistor between pin 1 and pin 2, connect another resistor between pin 2 and pin 3, connect another resistor between pin 3 and pin 5 , connect another resistor between pin 5 and pin 6 and then a final resistor between pin 6 and the - (negative) Master feed (pin 8) on the buttons led connector.

Pin 4 the common pin connects to the button input (pin 6) on the Master buttons led connector.

5waybutton.jpg
5waybutton.jpg (34.07 KiB) Viewed 4443 times
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Sun Jul 04, 2010 6:33 pm

I've nearly finished building my 2 new 25 cell slave boards for my 50 x 20ah A123 setup. I'll post some high res pics later this week when completed. I'm experimenting with ommiting a couple of the components and a changed value for another.

At present I am not including the 10uf tantalum capacitor, i'll test this in my insight, if interference becomes an issue i can solder them on. If they work in the insight without these capacitors they will probably work anywhere. Of course you can include them i'm just trying to get the price down by a few pence per slave.

I'm also ommiting the 100k adc ref diode pull up resistor. I have enabled the internal weak pull up in the pic for this pin and will test to see if it is suffcient to give a stable voltage, if it is we can save another resistor and a few pence.

I've also used 1k resistiors on my boards for R4 instead of the 10K on the schematic. This gives a harder pull down and makes it less susceptible to interference.

If these mods work Ok then the next version of the boards can be smaller/cheaper with a reduced component count.
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.

martinwinlow
Posts: 79
Joined: Mon Jun 11, 2007 9:35 am
Location: Herts, UK

Re: Old BMS Hardware Thread

Postby martinwinlow » Tue Jul 06, 2010 10:36 pm

Peter,

An update... I have put together a dash pcb for the leds and switches and connected these to the Master pcb. I am waiting on IC2 ICs but in the meantime have now wired in 2 x LM335's for an1 & 2 and these, with the master board, appear to be working (implying everything else on the master connected thus far is ok too, I suppose).

I have connected one slave board to the master and readied the slave board for 3 temp 'cells' - 3 x AA ni-cads.

I get the green dash LED (3) blinking implying master OK and an alarm 'Cell Error' due to no cells being yet attached. Charge Cutback and Controller Cutback LEDs come on with the alarm - as I gather they should.

So just waiting on connecting some cells to test the rest...

Looking at the dash switch functions, I gather they partly work as follows:-
SW1 = Display Cell Voltages
SW2 = Update Slaves (why? - What does this do?)
SW3 = Dump BMS Data (where too?)
SW4 = Reserved (not used)
SW5 = Exit Sub Menu

Which are the up and down buttons referred to in the txt file?

Anything else I need to know? Perhaps an instructional video on utube would be an idea?

I edited my cell layout photo with a better one, BTW.

Regards, Martin.
Regards, Martin Winlow
Herts, UK
http://www.evalbum.com/2092
www.winlow.co.uk

User avatar
retepsnikrep
Posts: 1387
Joined: Sat May 26, 2007 4:50 pm
Location: North Yorkshire England
Contact:

Re: Old BMS Hardware Thread

Postby retepsnikrep » Wed Jul 07, 2010 5:38 am

Martin

Good progress well done.

Assuming you are using the last/final version of the picaxe master software posted a few days ago make sure the constant (Cells) in the program is set to the number of cells you have/are testing.

Assuming you are waiting for DS18B20 I2C temp sensors to arrive, remember you have to configure the software to use those, and then choose if you want Analog or digital sensors as you can't have both active at same time. I recommend the digital sensors. They should be connected using ultra miniature three core screened cable with the screen connected to GND at the master end only.

The buttons started off as line of single buttons but i then swapped to the multifunction button I described a few posts back. In the program menus you will see it says ULDRC = Up Left Down Right Centre for the multi function button. Just label yours according to which function they activate with ULDRC as appropriate.

The update slave command sends command codes to the slaves. The available commands have been listed in the software thread. The commands should be self explanatory.

Dump data outputs the contents of the eeprom via the remote display unit interface, i havent tested this at all, so it may not work at this time, in fact i don't have the remote unit installed on either of my two systems at present.

I may well do more videos/explanations as time permits.

As an aside my bench test system with one slave attached running all pbpro code for slave and master has been working for last two days with no observed problems.
Regards Peter

Two MK1 Honda Insight's. One running 20ah A123 Lithium pack. One 8ah BetterBattery Nimh pack.
One HCH1 Civic Hybrid running 60ah A123 Lithium pack.


Return to “BMS thread”

Who is online

Users browsing this forum: Majestic-12 [Bot] and 7 guests