grindz145
1 MW
This is really a sophisticated design. It's refreshing after alot of the other chinese garbage I see on ES all the time. Any plans right now to scale it down for Ebike usage?
Linear Tech 12-cell battery monitor/BMS
- Thread starter mike_t
- Start date
methods
1 GW
Alan B said:
I have one - along with the 3rd optic for a camera mount. I don't use it currently for assembly, sometimes for inspection. At work I have been spoiled by our full face scopes... 40X, super bright light, wide field of vision.... I like to work under those. What I don't like about the stereo scopes is having to have my head in just the right spot. Irritates me. I would rather just wear magnifiers on my face or do it free hand and inspect later.
-methods
methods
1 GW
grindz145 said:
Yes - but it is going to be really hard. I don't like wasting my time with cheap parts.... If I want to accomplish something I pick the best part I can find - the one that excites me - you know what I mean? Like the totally isolated +/- 100A hall-effect current sensor.... that part gets my juices flowing. This makes the parts alone very expensive. Now... If I wanted to just sell a 4S - 24S kit with two slaves, master, and charge control - raw with a thru-hole interface - no relays and low current - we could probably get the price down to something an ebiker would be interested in. This one powers itself on when the charger is attached so it is ebike friendly - no need for a USB connection. I also included an RX input that logs (and uses) the data stream from the CA
I am looking for someone who would like to develop an Android App to mate to my design.... I have a TTL UART output at 9,600 in binary along with a USB output in ASCII. I already have a PALM implementation but who the hell uses those anymore?
-methods
I do, but they're old grayscale units that probably can't run your s/w. :lol:methods said:
grindz145
1 MW
My old Treo 650 is waiting around for an opportunity like this...
Know exactly what you mean though, I like to design cool shit, nuff said. That being said, theres a gaping hole in the ebike offerings as far as a decent, reliable BMS. It's the missing link that makes alot of ebikes DIY and not turn-key. Not that that's necessarily a bad thing...
Know exactly what you mean though, I like to design cool shit, nuff said. That being said, theres a gaping hole in the ebike offerings as far as a decent, reliable BMS. It's the missing link that makes alot of ebikes DIY and not turn-key. Not that that's necessarily a bad thing...
methods
1 GW
I know... I already did the math on an ebike specific BMS and the profit is just not there. The incoming crowd just does not understand the value and is not willing to pay. The problem is that the BMS cost is contrasted to a motor, controller, throttle, and battery that are artificially cheap from China. It is a non-starter until the fantasy days of free stuff from china dry up.
I talk to a lot of nOoBs and trying to explain why they should spend an extra $300 on a quality BMS is like trying to explain to an Eskimo the tender flavors of Hawaiian shave ice.
-methods
I talk to a lot of nOoBs and trying to explain why they should spend an extra $300 on a quality BMS is like trying to explain to an Eskimo the tender flavors of Hawaiian shave ice.
-methods
grindz145
1 MW
Almost have to sell it into the battery, but still who want's to pay more? I think the real selling point would be active balancing. Could save money in the charger. Even so, hard sell.
methods
1 GW
What I dont understand is why people pay like $200 for a 36V 2A charger when they could just buy ten 4.2V 2A DC/DC converters and set up a TC54 LVC array and be done with it.
Now... if you want to charge a 25S pack at 50A... the idea does not really scale.
-methods
Now... if you want to charge a 25S pack at 50A... the idea does not really scale.
-methods
Hyena
10 GW
I actually tried doing this, I scoured ebay and other places and found some 4.2v chargers (for PSP I think) that were supposedly rated at 2 amps. I bought half a dozen to try 6S first and was pissed to find they only output 700ma :x Little 5a meanwells (charging packs 2p) would be ideal but the cost of buying a heap + international shipping gets a bit more costly. I've had fantastic success with parallel charging at the pack level anyway but a better solution would be ideal for noobs.methods said:
Methy, I know you probably posted this earlier in the thread but who is your target audience for these BMS's if not for ebikes ?
Do you plan to sell multiple units to larger EV users ?
I guess they'd also be good for the RC crowd running big outrunners but they'd be a fairly limited market.
Read this thread dude. http://endless-sphere.com/forums/viewtopic.php?f=14&t=22881methods said:
Its what Im working on.
As long as you are sure the voltage is not to hi on any cell you can plug them in and walk away.
I personly think charging on the cell level will be quicker and more efficient then shunting of the excess to heat!
Hyena
10 GW
Sweet, that looks like a better way of doing it.
/me goes off to that thread to read up
/me goes off to that thread to read up
methods
1 GW
Hyena said:
When a small dc/dc is rated at 4.3V 2A what they mean is that it will put out 4.3V and the short circuit current is 2A - not that they will actually put out 2A @ 4.3V. Sounds like yours was even worse though. But - I bet you did not pay much for them. I purchased over 30 different DC/DC converters (12V @ .8A - 2A) and found a HUGE range of quality. You just have to buy all of them, take them apart, and figure out which ones are crap
But again - it does not scale so I am not really interested in it.
Hyena said:
All will be revealed soon. I have been working with a silent partner for a long time and once we get V4.0 squared away we are going to make an announcement. Our prototype work has been targeted at race motorcycles - and that is what drives the need for such aggressive balance currents. Big 25S 80Ah Hobby King packs tend to get a little out of balance when you hammer them hard
If you figure that there is 2 hours available for a charge @ 2A balance current we can correct a 4Ah imbalance in the pack. That is a lot of juice... especially considering that at last years TTXGP more than a few bikes were pushed across the finish line. Every Ah counts. 4Ah * 25S = 360Wh..... On the final lap of a race the rider would be able to run 40KW for 30 seconds on that power.That is the difference between winning and losing!
My secondary target is the DIY EV crowd that I anticipate will be cropping up soon. Every BMS on the market is operating on the FALSE ASSUMPTION that cells will be self discharging at similar rates and therefor need only a small balance current to keep them equal... i.e. they will be new, of the same lot and abuse level, stored in similar environments (i.e. not one under the hood and one in the passenger compartment), etc. The reality is that many, many people will be building cars with salvaged EV packs that have all different levels of self discharge, different chemistries, different capacities, and probably - a Lack of capacity - and therefor people are going to want to charge and balance often and aggressively. If you are running a 100Ah 25A salvage pack in a car you have to opportunity charge at every chance you get and you are going to want to have the pack as balanced as possible.
I know, because I have been driving a car with a 20S 100Ah salvage pack :wink:
And... I have pushed that car home no less than 10 times due to a bank of cells hitting LVC early. 300WH means a lot when you are a mile from home.....
I have gone round and round arguing with some of the other BMS designers who pull up charts and graphs proving that you need only 500mAh of balance current to keep a 200Ah pack balanced. The argument is that the pack can balance 24 hours a day and if you average that over time it will keep any pack balanced. Looks good on paper but it only applies to new batteries in a car that is used infrequently. In any race situation - or any situation where you are cycling your salvage pack 2 or 3 times in a day - the cells become grossly out of balance. At the end of each day Emoli bug was usually at least 300mV out on a few cells. I could care less about dreaming of OEM BMS's in a Ford.... every ford out there will have a BMS. The folks that will need a BMS are those who are doing the DIY thing - and fewer and fewer of them are falling into the Thundersky 2C trap and more and more are repurposing cells from other applications.
There is a 70 year old man down the street from me who has over 6,000 18650 cells in his Bug
But think of this.... Why should you carry around 2,000lbs of batteries with you when you are just going to the store and back??? That is stupid - it wastes energy, and more importantly - it trashes performance! It makes far more sense to carry around a 500lb battery and charge everywhere you go. I have a 100' extension cord in my trunk. When I go to Longs Drug Store I plug that ish in!!! Nearly every store has an outlet out front. You just get one of those little rubber mats to put over your charge cord (tripping hazard) and no one ever questions you. In a 30 minute shopping trip I can draw an easy 700wh out of the wall - which is enough to cover the 2 miles back to my house.
Just a different way of thinking.
Eventually power will be available everywhere (for a price) and I think smaller, lighter, EV's that charge any time they are parked is the way to go.
-methods
P.S. I should no better than to answer posts right after I drink my coffee in the morning
Caffeine induced ramble!
liveforphysics
100 TW
We finally have an EV charge station at work. And I was delighted to see 4 EV charge stations at the airport, each with an EV on them!
There were a couple fred-myers in Seattle that had EV charge stations as well, and a couple parking spots that had EV charge stations rather than parking meters. I'm pretty excited for the future.
There were a couple fred-myers in Seattle that had EV charge stations as well, and a couple parking spots that had EV charge stations rather than parking meters. I'm pretty excited for the future.
GGoodrum
1 MW
I agree with Patrick. Larger packs need more balance current. I too have seen how even the best quality LiPo-based packs can get significant imbalances, if they are hit hard, and/or are drained down closer to the LVC point. It doesn't do much good to be able to charge at higher rates, if it is going to take 4-6 hours to balance the cells. For charge currents up to 15A, or so, I feel at least 1A of balance/shunt current works fine. Above 15A, more would be better, especially for larger capacities.
-- Gary
-- Gary
methods
1 GW
Sound the horn nerds....
Method LTC BMS V4.0 has passed all bench tests and works 100% right out of the box
TO-127 20W shunts packed in tight - 100mm across. These are much smaller than their TO-220 big brothers but they can still handle a ton of power. Those metal pads on the back are electrically isolated so there is no need for Kapton.
Here it is with a makeshift easy-swap heatsink. Notice that the only connection to the slave assembly is an RJ45 and the balance taps. You see that little assembly coming off the JST-XH-7 connectors? That is your best friend if you are testing a BMS. It is a 12S resistor ladder of 150 ohm resistors. They tie into either a 50V source or a 12S pack. They pull a steady 30mA off the pack and can supply enough current to each channel of the LVC to allow it to behave normally - BUT - if the shit hits the fan there is plenty of inline resistance to save you!
The sweet part is that if you enable all shunts you basically put the two ladders in parallel and you can pull the full 2A balance current off the pack. This gives you all the safety of 12 120ohm resistors in series and all the power of 100W balance testing.
Here is a bigger picture. You can see the ladder going to a 12S 5Ah Lipo pack. The blue cable is obviously the RJ45 running over to the master.
View attachment 1
2A balance currents are not for sissies or little girls... each of these little 1"x4" boards dissipates up to 100W at the final stage of charging. Only those who have actually done testing with dissipating large amounts of power in small spaces will understand... Things get very hot, very fast. We are using some predictive algorythems along with some temperature monitoring to throttle back balance current. Of course - by simply duty-cycling we can run at 1A, 500mA, or any current under 2A. The idea is for each of these modules to be bolted to the frame of the vehicle - but a heatsink and fan will work too. I have not done the testing yet to determine how high we can balance with no sinking @ 100F ambiant but I suspect it will be low...
I worked out a new algo for heat management that I think is very clever... Instead of throttling back all channels when things get too hot I can assign *priority* to specific shunts. If one channel is grossly higher than the others that channel can get the full 2A balance current while the others get some smaller ratio that keeps temperatures in check. This maximizes our thermal capacity at any time.
Yep - methods is on top of the world right now..... I have a re-crystallization going in the chemistry lab, the wife is busy sewing with a friend, it is 72F and sunny outside, my BMS works, my side business is picking up, and now I have overcome the hurtles I put up to keep me from quitting my day job.
Methods = Living the Dream
-methods
Method LTC BMS V4.0 has passed all bench tests and works 100% right out of the box
TO-127 20W shunts packed in tight - 100mm across. These are much smaller than their TO-220 big brothers but they can still handle a ton of power. Those metal pads on the back are electrically isolated so there is no need for Kapton.
Here it is with a makeshift easy-swap heatsink. Notice that the only connection to the slave assembly is an RJ45 and the balance taps. You see that little assembly coming off the JST-XH-7 connectors? That is your best friend if you are testing a BMS. It is a 12S resistor ladder of 150 ohm resistors. They tie into either a 50V source or a 12S pack. They pull a steady 30mA off the pack and can supply enough current to each channel of the LVC to allow it to behave normally - BUT - if the shit hits the fan there is plenty of inline resistance to save you!
The sweet part is that if you enable all shunts you basically put the two ladders in parallel and you can pull the full 2A balance current off the pack. This gives you all the safety of 12 120ohm resistors in series and all the power of 100W balance testing.
Here is a bigger picture. You can see the ladder going to a 12S 5Ah Lipo pack. The blue cable is obviously the RJ45 running over to the master.
View attachment 1
2A balance currents are not for sissies or little girls... each of these little 1"x4" boards dissipates up to 100W at the final stage of charging. Only those who have actually done testing with dissipating large amounts of power in small spaces will understand... Things get very hot, very fast. We are using some predictive algorythems along with some temperature monitoring to throttle back balance current. Of course - by simply duty-cycling we can run at 1A, 500mA, or any current under 2A. The idea is for each of these modules to be bolted to the frame of the vehicle - but a heatsink and fan will work too. I have not done the testing yet to determine how high we can balance with no sinking @ 100F ambiant but I suspect it will be low...
I worked out a new algo for heat management that I think is very clever... Instead of throttling back all channels when things get too hot I can assign *priority* to specific shunts. If one channel is grossly higher than the others that channel can get the full 2A balance current while the others get some smaller ratio that keeps temperatures in check. This maximizes our thermal capacity at any time.
Yep - methods is on top of the world right now..... I have a re-crystallization going in the chemistry lab, the wife is busy sewing with a friend, it is 72F and sunny outside, my BMS works, my side business is picking up, and now I have overcome the hurtles I put up to keep me from quitting my day job.
Methods = Living the Dream
-methods
Hyena
10 GW
Awesome 8)methods said:
You should make a high current balancer as a side project with those boards.
Doctorbass
100 GW
Nice work Method!
Will you make a clip showing that great PC interface for programing it?
Will you make a clip showing that great PC interface for programing it?
methods
1 GW
I am totally working over the user interface for V4.0
This time around I am going to leverage the Flash memory on the uController for more custom behavior. The User interface will be capable of flashing firmware updates, changing default settings/levels, and of course - monitoring.
* HVC settings
* LVC settings
* Max charge current
* Temperature compensation
* Retard curves (for motor temp inputs)
Also - this V4.0 brings in the CA data so we are thinking about some higher order functions as well.
Here are the old pics
Today I am going to work on the V4.0 firmware. I have stripped it down to the bone and modularized it. Hopefully by tonight I will have the firmware to the point where it can accept any number of slave units without any hard-coding.
-methods
This time around I am going to leverage the Flash memory on the uController for more custom behavior. The User interface will be capable of flashing firmware updates, changing default settings/levels, and of course - monitoring.
* HVC settings
* LVC settings
* Max charge current
* Temperature compensation
* Retard curves (for motor temp inputs)
Also - this V4.0 brings in the CA data so we are thinking about some higher order functions as well.
Here are the old pics
Today I am going to work on the V4.0 firmware. I have stripped it down to the bone and modularized it. Hopefully by tonight I will have the firmware to the point where it can accept any number of slave units without any hard-coding.
-methods
Alan B
100 GW
Nice! I want a similar feature set on my ebike someday.
methods
1 GW
I have been snapping production assembly pictures as I go. Too lazy to upload them to my computer - but here is one I emailed.
When you are assembling a complicated device in large numbers (there are 45 boards like you see above) it is the little tricks that make all the difference. Trying to brute force the assembly will result in failure. Stopping constantly and "wasting parts" to try and find a better/faster/easier way to do the assembly is crucial. One example would be those 25 pin flat flex cables... each one I install is 50 pins. 5 of those is 250 - so that means I need to solder 250 points, clip 250 points, clean and inspect.
That would be the hard way
The easy way is to cut the pins BEFORE you solder them - either with scissors or a razor blade - and then once you solder them not only do you not have to screw around with the snips - but you end up with a perfectly flush surface instead of a bunch of bumps.
When soldering - instead of trying to solder each of the 250 pins the trick is to wet them with a high quality flux and use a large iron. You "paint them" more than anything - just running the iron down the row. Surface tension will do the rest of the work for you and you will have no bridges.
Most people think surface mount soldering is hard because they dont have the right tools and technique. Once you know how to do it - it becomes much easier and faster than thru-hole soldering.
-methods
When you are assembling a complicated device in large numbers (there are 45 boards like you see above) it is the little tricks that make all the difference. Trying to brute force the assembly will result in failure. Stopping constantly and "wasting parts" to try and find a better/faster/easier way to do the assembly is crucial. One example would be those 25 pin flat flex cables... each one I install is 50 pins. 5 of those is 250 - so that means I need to solder 250 points, clip 250 points, clean and inspect.
That would be the hard way
The easy way is to cut the pins BEFORE you solder them - either with scissors or a razor blade - and then once you solder them not only do you not have to screw around with the snips - but you end up with a perfectly flush surface instead of a bunch of bumps.
When soldering - instead of trying to solder each of the 250 pins the trick is to wet them with a high quality flux and use a large iron. You "paint them" more than anything - just running the iron down the row. Surface tension will do the rest of the work for you and you will have no bridges.
Most people think surface mount soldering is hard because they dont have the right tools and technique. Once you know how to do it - it becomes much easier and faster than thru-hole soldering.
-methods
methods
1 GW
I am sending this one over to Steve for thermal management, packaging, testing, and the inevitable constructive criticism.
-methods
-methods
Jozzer
100 kW
Extremely nice job there. Prettiest PCB's I've seen for a long time!
I'm off to sit by the postbox
Jozz
I'm off to sit by the postbox
Jozz
voicecoils
1 MW
Fantastic stuff!
Congrats on making V4.0 a reality.
I want to see it in action & happen to be flying to the USA tomorrow (3 weeks in Portland + a few days in LA ... Livermore CA's somewhere in between?)
Congrats on making V4.0 a reality.
I want to see it in action & happen to be flying to the USA tomorrow
(3 weeks in Portland + a few days in LA ... Livermore CA's somewhere in between?)
methods
1 GW
voicecoils said:
Just give me your flight number... I will use the 1W blue laser Luke left here to signal you in the plane as you fly over
btw: This BMS was actually Steve's vision from the start. I was just the one lucky enough to get to implement the electronics. :wink: I have learned an incredible amount over the last year and I am now finally in a position to "make my wildest dreams come true" ( In the words of Pedro from Napoleon Dynamite). This BMS may not ever materialize in an Ebike flavor but the lessons we learned here certainly will show up in some ebike offerings very soon.
-methods
