Alan B
100 GW
Alan B said:I tried to post this last night but something went down - could not get Picasaweb to finish the download.
Here's OSH Park's rendering of the board, complete with gold plating:
The boards are in the mail!
Ebike Battery Monitoring and Control System
- Thread starter Alan B
- Start date
The boards are in the mail!
Alan B
100 GW
Boards arrived. Looking good.
powersupply
100 W
Congrats!
Hope they work as you expect.
Question: what are the main differences of your design and http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=48961?
From browsing brirefly, you both have BMS function, both have small singular/chaineable boards etc.
I love your data acquisition part BTW. And agree that CellLogs are really cute (and useful).
Would be really great to have a ...one-plug charging from AC... solution.
Thanks!
Hope they work as you expect.
Question: what are the main differences of your design and http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=48961?
From browsing brirefly, you both have BMS function, both have small singular/chaineable boards etc.
I love your data acquisition part BTW. And agree that CellLogs are really cute (and useful).
Would be really great to have a ...one-plug charging from AC... solution.
Thanks!
Alan B
100 GW
Thanks for the encouragement.
As I understand it, David's board is 4x6 or 24 square inches for each 6S "unit". He is planning for much higher balance current, so there will be more heat generated, but balancing will be much faster.
My present design (both our designs are evolving, so any comparison is subject to change) is about 2x2 inches, or 4 square inches per 6S unit. I'm planning to use low balance currents, so there will be little heat generated but balancing will take a lot longer.
I'm finding (as others here on ES have pointed out) that the Lipo batteries require very little balancing. My 18S4P pack has not been balanced in a year and the number of charge cycles is around 250 (I haven't looked at that data recently). I'm seeing 20-30 millivolt differences at worst, so each charge cycle only needs to correct about 0.1 millivolt. The bike is ridden less than one hour per day average, so there are 23 hours available to charge and balance. So gentle balancing seems to be sufficient. The most important thing is not to overcharge, so if the pack begins to get too far out of balance the charge must be stopped when any one cell group gets to the stopping voltage (4.2 volts). After balancing, more charging can take place, so the full capacity can be restored. If this happens on every charge cycle the cell group is weak or possibly failing and should be investigated.
My present design detects voltage from the charger and decides to charge (or not) as the battery pack needs.
As I understand it, David's board is 4x6 or 24 square inches for each 6S "unit". He is planning for much higher balance current, so there will be more heat generated, but balancing will be much faster.
My present design (both our designs are evolving, so any comparison is subject to change) is about 2x2 inches, or 4 square inches per 6S unit. I'm planning to use low balance currents, so there will be little heat generated but balancing will take a lot longer.
I'm finding (as others here on ES have pointed out) that the Lipo batteries require very little balancing. My 18S4P pack has not been balanced in a year and the number of charge cycles is around 250 (I haven't looked at that data recently). I'm seeing 20-30 millivolt differences at worst, so each charge cycle only needs to correct about 0.1 millivolt. The bike is ridden less than one hour per day average, so there are 23 hours available to charge and balance. So gentle balancing seems to be sufficient. The most important thing is not to overcharge, so if the pack begins to get too far out of balance the charge must be stopped when any one cell group gets to the stopping voltage (4.2 volts). After balancing, more charging can take place, so the full capacity can be restored. If this happens on every charge cycle the cell group is weak or possibly failing and should be investigated.
My present design detects voltage from the charger and decides to charge (or not) as the battery pack needs.
dmwahl
10 kW
I'll add a few comments on the differences as well. Size is definitely larger than this project, I'm working on ways to get it smaller than 4"x6", but for something that the user puts together it's not reasonable to expect them to solder surface mount components, so the larger size is the tradeoff. I plan on using it primarily with LiFePO4, which in my experience tends to get out of balance more easily than Lipo. Balance current will theoretically be up to 1.5A, but only with some cooling (fan). With no cooling fan it will be more like 0.6A balance current. If you have more questions about it, feel free to post in the thread, I don't want to hijack Alan's thread 
Alan B
100 GW
Yes, I'm forced to use surface mount parts because the main part, the Intersil chip that does most of the battery interfacing, front end signal multiplexing and level shifting is available only in a surface mount package. So, given that, and the small space available on an ebike, I'm pushing size down a bit. The Intersil chip and micro are both QFN packages which require soldering underneath the part, so no soldering pencil will work, and the pin spacing is about 10 thousandths, so oven, hotplate or hot air soldering is indicated. A few years ago I set myself up with a hot air rework station which consists of an IR board heater and holder, a hot air pencil, and a low power binocular QA inspection microscope, and this is most of what is needed to work on surface mount boards like this (and it cost well under 1k$, less than a decent ebike conversion ). I'm currently looking into a small reflow oven that would solder a few boards at a time. For real production there are any number of assembly houses that would take the job and produce hundreds or thousands of boards if that were needed.
So these boards can be put together at home if you have a bit of equipment, probably the simplest is the hotplate or toaster oven with a reflow controller, and a good magnifier to inspect the work.
). I'm currently looking into a small reflow oven that would solder a few boards at a time. For real production there are any number of assembly houses that would take the job and produce hundreds or thousands of boards if that were needed. So these boards can be put together at home if you have a bit of equipment, probably the simplest is the hotplate or toaster oven with a reflow controller, and a good magnifier to inspect the work.
powersupply
100 W
Alan B & dmwahl,
thank you for your insights, very appreciated!
Love the spirit of your courteous cooperation/competition.
Have to dig deeper into your very interesting creations.
I like dmwahl's one board approach and also Alan B's telemetry.
Then there is the repeatability (self maufacturing) and of course the overall costs.
I've encountered several other approaches, most of which were not finished or are missing something.
Given that my SmartBMS seems too dumb to do a simple equalization in hardware mode, I am very interested in your approaches.
Thanks!
thank you for your insights, very appreciated!
Love the spirit of your courteous cooperation/competition.
Have to dig deeper into your very interesting creations.
I like dmwahl's one board approach and also Alan B's telemetry.
Then there is the repeatability (self maufacturing) and of course the overall costs.
I've encountered several other approaches, most of which were not finished or are missing something.
Given that my SmartBMS seems too dumb to do a simple equalization in hardware mode, I am very interested in your approaches.
Thanks!
Alan B
100 GW
Parts on order, most of them at least. Mouser doesn't have the JST's that I could find. I didn't look hard because I have the 7 pin units already. It took awhile to fill out the BOM, and a few parts went out of stock before I finished so had to re-select those.
There are 46 parts on this board. I need to decide whether to load the board in sections, or go for broke and do it all at once. The parts cost is not all that high, so might just do it all.
There are 46 parts on this board. I need to decide whether to load the board in sections, or go for broke and do it all at once. The parts cost is not all that high, so might just do it all.
Alan B
100 GW
I had to make a decision on the balance current when ordering parts. This design is limited due to the switches in the Intersil chip. It has a 100mA per cell limit and a 200mA per chip total limit, and these are under ideal conditions. I thought about it a lot and decided for the prototype to go with less than 200mA total for all six cells simultaneously, or about 30mA each. This is a good value to start with until the software is reliable, in case it accidentally switches them all on we don't want to damage the hardware. There is a temperature shutdown on the chip as well, so if all six switches get turned on it would probably thermally limit after a time in warm weather.
Even at this low rate, if you leave the bike plugged in most of the time it could balance about half an amp hour per day, and stay safe while doing it. This way the board won't get very warm.
Even at this low rate, if you leave the bike plugged in most of the time it could balance about half an amp hour per day, and stay safe while doing it. This way the board won't get very warm.
powersupply
100 W
Interesting, didn't know about the low equalizing current.
I have the same scenario/needs like you, so slow charge and balancing are fine with me.
Hope your design will work well, especially the telemtry/recording of individual cells is very important at some point.
How often does it happen that more than one or two cells need discharging?
Keep up the great work!
I have the same scenario/needs like you, so slow charge and balancing are fine with me.
Hope your design will work well, especially the telemtry/recording of individual cells is very important at some point.
How often does it happen that more than one or two cells need discharging?
Keep up the great work!
Alan B
100 GW
I have not balanced in 200 cycles. I just want to insure that a software error doesn't damage the board.
Alan B
100 GW
There is likely a way to add an external switch to increase balancing current. Their examples show much lower balancing currents, they were not thinking about our large packs when they designed this chip. But using a $3 chip is really appealing to keep the system costs down.
powersupply
100 W
Amazing, 200 cycles w/o balancing.
What batteries/configuration do you use?
But you do have some sord of monitoring for HVC and LVC?
What batteries/configuration do you use?
But you do have some sord of monitoring for HVC and LVC?
Alan B
100 GW
My Croborg pack is currently 12 each 6S 5 amp hour Turnigy batteries, four in parallel, three in series for 18S 20AH. Cycles are half 50% and half 60%, fairly consistent due to commute run. I have Methods boards which have cell group HVC and LVC and Cycle Analyst set for whole pack LVC. Presently the cell group outputs are not hooked up to do anything, but with only 60% cycles I'm not near running out. I do have three 6S connectors coming out so I can check balance and I do that every few days with a battery medic type 6S readout.
I bulk charge at home with a 2A 18S charger, and bulk charge at work with three 12A LED supplies set up to each do a 6S section.
I did do one deep discharge cycle when the charge cycle shutdown incomplete the previous evening, I shut it down about 2 miles from work, and pushed the bike uphill for a mile and then coasted into work. The battery was not depleted, but was down to about 10%. It still is fairly well balanced, but I do see about 30mV of variation total across the 18 cell groups after about 200 cycles.
I bulk charge at home with a 2A 18S charger, and bulk charge at work with three 12A LED supplies set up to each do a 6S section.
I did do one deep discharge cycle when the charge cycle shutdown incomplete the previous evening, I shut it down about 2 miles from work, and pushed the bike uphill for a mile and then coasted into work. The battery was not depleted, but was down to about 10%. It still is fairly well balanced, but I do see about 30mV of variation total across the 18 cell groups after about 200 cycles.
powersupply
100 W
Thank you so much about this detailed info.
I had read your other threads regarding your bike etc. btw.
Hope the cells of my 4s Turnigy packs are the same as yours, 30mV difference after 200 cycles without balancing is really good.
I read this about LiPos in general, was unsure about the $23 4s LiPo packs from HK.
I shut it down about 2 miles from work, and pushed the bike uphill for a mile.. . True believer!
Why wouldn't you drive it until the LVC kicks in?
I mean that's what they are good for?
Does your controller have regen?
I had read your other threads regarding your bike etc. btw.
Hope the cells of my 4s Turnigy packs are the same as yours, 30mV difference after 200 cycles without balancing is really good.
I read this about LiPos in general, was unsure about the $23 4s LiPo packs from HK.
I shut it down about 2 miles from work, and pushed the bike uphill for a mile.. . True believer!
Why wouldn't you drive it until the LVC kicks in?
I mean that's what they are good for?
Does your controller have regen?
Alan B
100 GW
Yes, I have regen, and love it. Not so much for the returned energy which is really not much, but for the braking going down big hills, and the lack of heating the brake pads and wearing them out. My disk brake pads are a year old and hopefully they will last for many more years because I don't use them very much, mostly just for the last bit of a full stop.
My cell level LVC is presently not hooked to anything, and it is a pretty low value, so I don't really want to hit it. That's one reason I want to build these boards, so I can set my LVC precisely and to a fairly high value, and have confidence about it. This pack cost $800 and a lot of work, don't want to kill it. So I was watching the pack voltage and went as low as I was comfortable going. Being late to work that day was not a big issue, we have flex time and I had my cellphone so I could check my schedule and call to adjust things. I had some uphills that I couldn't pedal up with the limited gearing I have, so it was easier to push on the uphill, pedal on the level, and coast/regen downhill. The last mile is downhill at up to 15% so I arrived at work with more charge than I had when I started pushing.
My cell level LVC is presently not hooked to anything, and it is a pretty low value, so I don't really want to hit it. That's one reason I want to build these boards, so I can set my LVC precisely and to a fairly high value, and have confidence about it. This pack cost $800 and a lot of work, don't want to kill it. So I was watching the pack voltage and went as low as I was comfortable going. Being late to work that day was not a big issue, we have flex time and I had my cellphone so I could check my schedule and call to adjust things. I had some uphills that I couldn't pedal up with the limited gearing I have, so it was easier to push on the uphill, pedal on the level, and coast/regen downhill. The last mile is downhill at up to 15% so I arrived at work with more charge than I had when I started pushing.
2moto
100 W
Alan, any updates on this project. I am curious if you had a chance to power-up the board yet, or any other progress you've made.
Alan B
100 GW
Not yet. Hopefully soon.
I ordered some new solder paste, mine is a bit old, it should arrive soon.
I ordered some new solder paste, mine is a bit old, it should arrive soon.
Alan B
100 GW
Not yet. Working on business taxes at the moment. What fun.
Alan B said:
Thanks for reminding me... I am going to wait another 11 days for that
Are you planning to offer these for sale or only personal use? I am very interested.
There are indeed way too many BMS projects to pick from. Guessing it's from all the different battery form factor requirements? I am going to be using the same types of batteries as you are (hobby style 4-8S LiPo with balance ports) so your project naturally gets my attention.
Not to deter from your project, but do you know of others designed for these types of batteries with balance ports? I am more interested at this point in getting something to work and communicating with my flight controller over SPI or I2C so I can monitor all my cells in flight. This is going on a multirotor. I am also interested in a small footprint.
Believe it or not, it seems the ebike and ecar community has much more focus on their power supply then the flying community. I expected the opposite as you can coast to a stop on the ground without high risk of injury or death. Go figure. Maybe it is just that the industry is too immature at this time.
I really appreciate your feedback and help,
Alex
KD7EEK
Alan B
100 GW
If it works we'll make it available somehow. Not sure exactly how. I expected to have more time to work on this, but have not lately. I have the board and parts, just need to clear the space and time. I think it is pretty important, but there are a number of projects around and some of them should produce useful product. This chip is the only one I've seen that isn't expensive, overly complex, and hard to buy. Seems like a great choice. If someone else wants to try building one of these, I can sell one of my boards. But it is not going to be simple with all the SMT parts, so it is far from a beginner's project.
Alan B said:
I hear you there and I have a lot going on as well. I do not have a flow oven so it would prove difficult. I am going to continue searching around for something that is already built that fits my application. Please do let me know of anything you are aware of.
I am subscribed and wish the best on your progress... I guess it's time for me to start my taxes.
Alex
KD7EEK
Alan B
100 GW
Unfortunately some of this tax stuff has to be done before the end of this month, like 1099's and annual sales taxes.
Good luck in your search.
Good luck in your search.
Hi what's the status of this project? Have you been able to test BMS? If yes - what are the results?
I would love to power my new pack (http://endless-sphere.com/forums/viewtopic.php?f=14&t=56255) with your bms.
I would love to power my new pack (http://endless-sphere.com/forums/viewtopic.php?f=14&t=56255) with your bms.
Similar threads
