LTC6802 based 12 cell slave.

Threads relating to the BMS system begun by Peter Perkins

Moderators: GregsGarage, retepsnikrep

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Wed Sep 19, 2012 10:02 pm

LTC6802-120919.pdf
(97.5 KiB) Downloaded 460 times
More progress, updated schematic with new optos, pull-up for MCLR pin, new 16F1825 pic and MMB pin on U1 tied to Vreg. I think thats all that needs changing.
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Tue Sep 25, 2012 10:39 pm

I've just sent the new gerbers to China. It may take a while as they are going on holiday for a week. :(

Here's some 3D views of the new layout.
LTC6802B_top.jpg
LTC6802B_top.jpg (111.27 KiB) Viewed 8964 times

LTC6802B.jpg
LTC6802B.jpg (99.13 KiB) Viewed 8964 times

The main changes are new optos, new pic, test points labeled better, errors from previous layouts fixed (I hope :shock: ) and surface mount leds changed to a single through hole bi-colour led which can be mounted vertically or horizontally. Also the isolation distances are better on this board.

The schematic has also seen some changes, here's the latest version.
LTC6802B-120924.pdf
(96.03 KiB) Downloaded 450 times
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

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

Re: LTC6802 based 12 cell slave.

Postby retepsnikrep » Thu Sep 27, 2012 2:47 pm

Looks good. I should have used usb connector on my boards :roll:

My A123 boards won't be sent until 9th October :(
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.

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Fri Sep 28, 2012 4:39 pm

retepsnikrep wrote:Looks good. I should have used usb connector on my boards :roll:


I just thought of a potential problem of using usb cables on the original style slaves. The master bus and slave bus are at very different voltage potentials. The slave bus is at pack potential and the usb connectors/cables probably aren't rated for this. I've been able to use them because I've moved the isolation between pack and master completely onto the slave board, so both master and slave bus share a common earth. Just a thought.
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

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

Re: LTC6802 based 12 cell slave.

Postby retepsnikrep » Fri Sep 28, 2012 6:18 pm

Good point, i was just wondering who would be the first to spot that!! LOL
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.

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Wed Oct 03, 2012 9:57 pm

While waiting for the boards, nothing else to do but work on the software. Here's an initial version. It still needs work, mainly in the configuration of the pic but this should give a rough idea of how it will work.
6802Slave_V01_031012.zip
Alpha version, compiles o.k., but not tested.
(4.68 KiB) Downloaded 457 times
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

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

Re: LTC6802 based 12 cell slave.

Postby retepsnikrep » Fri Oct 05, 2012 6:58 am

Code looks interesting, didn't spot any glaring issues. 8)
I haven't used SPI
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.

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Fri Oct 05, 2012 9:14 am

The SPI caused me the most grief, the original optos where to slow so I couldn't use the SHIFTIN/SHIFTOUT commands because the clock signals weren't getting through. These pbp routines have a 10uS pulse for the clock instead of an even square wave. So I tried to use the pics built in hardware spi, only clocking it really slow. After going round and round trying to get it to work I discovered that the 16F886 has a silicon bug that affects the SPI mode I needed to use. :evil:
In the end new optos sorted out the spi bus. I am using the SHIFTIN/SHIFTOUT commands because they are easy, data rate is 50kHz. Using the built in SPI bus would enable 1mHz data rate, but there is no need for that.

The 16F1825 is just relaying commands between the SPI and master/slave buses. I want to add some smarts to the pic so that for example if you ask for the cell voltage register without first starting an A/D conversion it would do the A/D then read the cell voltage register. I should also be able to add multiple 6802s onto a slave making 24, 36 and 48 cell slaves possible on one board. But first I need to get this simple version working. Next week I am going to work on some Maximite master software. I am also mid-butcher of a toaster oven from ALDI to do re-flow soldering of the boards. I'll post some pictures of that soon.
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Tue Oct 30, 2012 11:21 pm

DSC_0052.JPG
DSC_0052.JPG (194.67 KiB) Viewed 8837 times

Latest slave hand soldered. The plan is to make 6 of these up with reflow soldering in my newly butchered toaster oven. :shock:
Lots to do on the software, but I've managed to get the maximite talking to the slave and the 16F1825 talking to the 6802, shame I haven't managed it all at the same time, :( but I'll get there.
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Re: LTC6802 based 12 cell slave.

Postby GregsGarage » Sun Nov 04, 2012 3:44 pm

I've been doing some baking, although the wife wasn't impressed that the results were inedible. :lol:
NearlyThere.jpg
10 boards, just need the RCA and Dsub connectors.
NearlyThere.jpg (229.52 KiB) Viewed 8814 times


Some pictures of how I did it.
StencilJig.jpg
Jig for the stencil made up of old pcbs taped down so board to be stenciled is held securely.
StencilJig.jpg (105.51 KiB) Viewed 8814 times

SolderPaste.jpg
Close-up of paste applied to board.
SolderPaste.jpg (82.37 KiB) Viewed 8814 times

PartsPlaced.jpg
Ready to go into the reflow oven.
PartsPlaced.jpg (114.88 KiB) Viewed 8806 times

Oven.jpg
My hacked Aldi toaster oven with PID controller off of ebay. Temperature set to 250C, takes nearly 4 minutes to reflow a board. I did them one at a time.
Oven.jpg (127.33 KiB) Viewed 8814 times

Reflowed.jpg
SMT parts reflowed. The parts self align once the solder becomes molten, however I did have one cap so far off the pad it jumped to the next pad. Also got solder bridges on 3 of the 6802 chips, probably caused by poor placement and smearing the paste too much, easily reworked though.
Reflowed.jpg (110.51 KiB) Viewed 8814 times


So there you have it. I've ordered the connectors I need to finish the boards. Well that would nearly finish the boards, except for a small goof I made. In Kicad when you run a design rules check in pcbnew, the results are shown in 2 tabs. The first tab shows clearence problems, tracks to close to pins, etc. I got those sorted out, but didn't realize I needed to click on the second tab that shows unconnected pins. I had accidentally deleted 2 vias the connect the 6802 earth pins to cell 1 negative. :evil: So each board needs 2 small holes drilled and the pins connected to the ground plane. Not enough of a problem to warrant new boards for testing. I'll fix the problem and tidy up the silk screen layer some more, as long as I lock the smt components I can use the same stencil.

EDIT: Added PartsPlaced.jpg
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191


Return to “BMS thread”

Who is online

Users browsing this forum: No registered users and 4 guests