Home grown BMS ideas !
-
- Senior Member
- Posts: 773
- Joined: Thu, 15 Jan 2009, 18:09
- Real Name: Neville Harlick
- Location: Tauranga NZ
Home grown BMS ideas !
Having now designed and built a complete BMS based around a micro sensing cell voltages, communicating with a master control unit and able to turn on/off shunt resistors for cell balancing, I am appalled at the complexity and, to some extent, cost of it all. And when I look at weber and coulomb's effort, Its even worse! No offense guys, but isn't it really complex? So I thought how to do it better.
What I have come up with is a method of balancing the cell voltages by transferring energy from one to another to equalise all the cell voltages over the whole battery. Then you need only monitor the total voltage since you know they are all equal.
The equalisation process is automatic, continuous and completely transparent. Cost looks like being about $4 - $5 per cell. Efficiency is much, much higher than simply turning on a resistor!
Watch this space.
What I have come up with is a method of balancing the cell voltages by transferring energy from one to another to equalise all the cell voltages over the whole battery. Then you need only monitor the total voltage since you know they are all equal.
The equalisation process is automatic, continuous and completely transparent. Cost looks like being about $4 - $5 per cell. Efficiency is much, much higher than simply turning on a resistor!
Watch this space.
- weber
- Site Admin
- Posts: 2934
- Joined: Fri, 23 Jan 2009, 17:27
- Real Name: Dave Keenan
- Location: Brisbane
- Contact:
Home grown BMS ideas !
I certainly agree that the complexity is horrendous. Just not convinced we can do it with much less.
Let me guess. A small high frequency transformer on each cell with "primary" driven by a push-pull switcher (bidirectional to power) and all secondaries in parallel on a common square-wave bus.
But if the cells have different capacities and they have been balanced at full charge and then you go for a long hard drive where you almost completely discharge the lowest capacity cell, will it be able to shuffle charge fast enough to keep that lowest capacity cell at the same voltage as all the others? And that they therefore all hit bottom at about the same time? How many amps of charge shuffling do you need to guarantee that one cell won't bottom out well before the others, under worst case conditions?
Let me guess. A small high frequency transformer on each cell with "primary" driven by a push-pull switcher (bidirectional to power) and all secondaries in parallel on a common square-wave bus.
But if the cells have different capacities and they have been balanced at full charge and then you go for a long hard drive where you almost completely discharge the lowest capacity cell, will it be able to shuffle charge fast enough to keep that lowest capacity cell at the same voltage as all the others? And that they therefore all hit bottom at about the same time? How many amps of charge shuffling do you need to guarantee that one cell won't bottom out well before the others, under worst case conditions?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
- Johny
- Senior Member
- Posts: 3749
- Joined: Mon, 23 Jun 2008, 16:26
- Real Name: John Wright
- Location: Melbourne
- Contact:
Home grown BMS ideas !
Unless your cell energy transfer system is capable of huge current transfers you will still have to monitor cell voltages for failure at near total discharge of the pack.
Anyway - I'm watching, I'm watching...
Anyway - I'm watching, I'm watching...
-
- Senior Member
- Posts: 492
- Joined: Mon, 01 Sep 2008, 19:27
- Real Name: Bruce Armstrong
- Location: Perth WA
Home grown BMS ideas !
Many of us have thought about this a lot.
Another scheme uses switched capacitors to shuffle charge between cell(n) and cell (n+1).
I did consider this sort of approach, but felt that it would only solve part of the problem.
The issues, as I see them are:
1) prevent over-charging of cells (dramatic results?)
2) prevent over-discharging of cells (terminal for cells)
4) monitor cell capacity (time in bypass/lowest discharge voltage)
4) monitor impedance (weakest cells reduce battery performance)
5) monitor terminal temperature (two terminals)
The simple cell top monitor/equalizers do item 1 and 2.
Items 3-5 are preventative monitoring that allows failing cells or terminal connections to be detected before they cause a systems failure and the vehicle stops.
Items 4 and 5 are by software in the controller.
Cells with different capacities are the sort of thing that is really important, since a single weak cell will reduce the entire battery capacity to that of the cell. I would like to be able to determine which cell it is and be able to detect in the early stages, before it before it becomes an impediment to operation.
I am doing a system with 160+ cells. I feel that automated monitoring is more important here because it will be difficult to manually keep track of so many and they get buried behind other cells and stuff.
As for cell balancing, if very much of this is needed, the battery is in trouble. (a good indicator of a failing cell?) It implies that the cells aren't very uniform and some are significantly more lossy than others. Because (a small amount of ?) cell balancing is conducted each charge cycle, I feel that simple resistive balancing makes most sense.
Any smart scheme is going to be full of microprocessors. I started this design process with a controller per n-cells, but the I/O connections rapidly became a problem, steering it to a cell-top design.
I have elected to also use concentrators that isolate the cells into blocks of 10 or 15 cells according to the battery topology. These allow some level of fault tolerance, permitting cell-top processor faults to be localised to a single block.
I am currently doing a second round design improvement to reduce component count and cost.
Sadly, about 30 components per cell is about as simple as I can see to get a solution that addresses the above points. I think at $5 per cell is about as cheap as it can be made, probably considerably more because of the prototyping costs. Software is probably about 100-150 hours excluding the presentation graphics and user/systems interface.
More reinventing the wheel... A good common design that can be kitted would be useful.
Another scheme uses switched capacitors to shuffle charge between cell(n) and cell (n+1).
I did consider this sort of approach, but felt that it would only solve part of the problem.
The issues, as I see them are:
1) prevent over-charging of cells (dramatic results?)
2) prevent over-discharging of cells (terminal for cells)
4) monitor cell capacity (time in bypass/lowest discharge voltage)
4) monitor impedance (weakest cells reduce battery performance)
5) monitor terminal temperature (two terminals)
The simple cell top monitor/equalizers do item 1 and 2.
Items 3-5 are preventative monitoring that allows failing cells or terminal connections to be detected before they cause a systems failure and the vehicle stops.
Items 4 and 5 are by software in the controller.
Cells with different capacities are the sort of thing that is really important, since a single weak cell will reduce the entire battery capacity to that of the cell. I would like to be able to determine which cell it is and be able to detect in the early stages, before it before it becomes an impediment to operation.
I am doing a system with 160+ cells. I feel that automated monitoring is more important here because it will be difficult to manually keep track of so many and they get buried behind other cells and stuff.
As for cell balancing, if very much of this is needed, the battery is in trouble. (a good indicator of a failing cell?) It implies that the cells aren't very uniform and some are significantly more lossy than others. Because (a small amount of ?) cell balancing is conducted each charge cycle, I feel that simple resistive balancing makes most sense.
Any smart scheme is going to be full of microprocessors. I started this design process with a controller per n-cells, but the I/O connections rapidly became a problem, steering it to a cell-top design.
I have elected to also use concentrators that isolate the cells into blocks of 10 or 15 cells according to the battery topology. These allow some level of fault tolerance, permitting cell-top processor faults to be localised to a single block.
I am currently doing a second round design improvement to reduce component count and cost.
Sadly, about 30 components per cell is about as simple as I can see to get a solution that addresses the above points. I think at $5 per cell is about as cheap as it can be made, probably considerably more because of the prototyping costs. Software is probably about 100-150 hours excluding the presentation graphics and user/systems interface.
More reinventing the wheel... A good common design that can be kitted would be useful.
It's not the end of the world, but I can see it from here.
-
- Senior Member
- Posts: 773
- Joined: Thu, 15 Jan 2009, 18:09
- Real Name: Neville Harlick
- Location: Tauranga NZ
Home grown BMS ideas !
Its an idea I'm playing with at present. Here's what its based on:
This is, of course, a half bridge. If the switches are driven on and off with a 50% duty cycle, the voltage at the mid-point is exactly half the supply voltage, ie half of the two cell voltages in series. If one cell is less than the other, current is taken from the higher one and put into the lower one until they are equal.
I can get 40V, 50A FETs for 80 cents and running at 100 kHz means a quite small inductor.
I'm working out how much current can be handled with reasonable size components.
This is, of course, a half bridge. If the switches are driven on and off with a 50% duty cycle, the voltage at the mid-point is exactly half the supply voltage, ie half of the two cell voltages in series. If one cell is less than the other, current is taken from the higher one and put into the lower one until they are equal.
I can get 40V, 50A FETs for 80 cents and running at 100 kHz means a quite small inductor.
I'm working out how much current can be handled with reasonable size components.
- Johny
- Senior Member
- Posts: 3749
- Joined: Mon, 23 Jun 2008, 16:26
- Real Name: John Wright
- Location: Melbourne
- Contact:
Home grown BMS ideas !
Two problems.
1. I don't think you can do it with reasonable inductor/switching losses. I.E. It will constitute a significant current drain on the cells.
2. It only balances 2 cells WRT each other. Unless you plan to overlap cells.
BUT. It looks interesting and will be easy to prove/disprove.
Go for it.
1. I don't think you can do it with reasonable inductor/switching losses. I.E. It will constitute a significant current drain on the cells.
2. It only balances 2 cells WRT each other. Unless you plan to overlap cells.
BUT. It looks interesting and will be easy to prove/disprove.
Go for it.
-
- Senior Member
- Posts: 773
- Joined: Thu, 03 Jan 2008, 19:04
- Real Name: Malcolm Faed
- Location: Australia
- Contact:
Home grown BMS ideas !
Similar to the switched capacitor idea. One of the difficulties with the capacitors is the low efficiency meant that the charge would not make it very far along the string before dissipating. But then resistors are not known for there efficiency either.
Keep going!
Keep going!
-
- Senior Member
- Posts: 773
- Joined: Thu, 15 Jan 2009, 18:09
- Real Name: Neville Harlick
- Location: Tauranga NZ
Home grown BMS ideas !
Switching and inductor losses should be very low.
And yes, overlapping cells is how I handle the whole battery.
It looks promising with a couple of 12V PbA batteries so far.
And yes, overlapping cells is how I handle the whole battery.
It looks promising with a couple of 12V PbA batteries so far.
- Johny
- Senior Member
- Posts: 3749
- Joined: Mon, 23 Jun 2008, 16:26
- Real Name: John Wright
- Location: Melbourne
- Contact:
Home grown BMS ideas !
Look what I found in the TI document I just posted a link to...
Nice that they have a trademarked buzzword for it.
Nice that they have a trademarked buzzword for it.
-
- Senior Member
- Posts: 773
- Joined: Thu, 15 Jan 2009, 18:09
- Real Name: Neville Harlick
- Location: Tauranga NZ
Home grown BMS ideas !
I got the idea from a paper I saw several months ago. I don't recall where it came from, but I don't think it was TI.
Just goes to show how many different people must be working on this stuff!
Just goes to show how many different people must be working on this stuff!
-
- Senior Member
- Posts: 773
- Joined: Thu, 15 Jan 2009, 18:09
- Real Name: Neville Harlick
- Location: Tauranga NZ
Home grown BMS ideas !
Just in case anyone is interested, I have my 8-cell BMS working now. Managed to get enough s/w for the control unit written while I was away (for the past 6 weeks) to get it to talk to the slave board. Seems to work OK. I am now tidying up the pcb design and will then make 6 x 8-cell boards which is enough for my 45 cell car pack and then see how it works in the car.
I note that my slave board draws about 9 ma when running, which is a little more than I thought it would, but still not bad. (the micro spec says typ 4, max 9 ma) That's .216 AHr per day or 18 AHr per year. About half the self-discharge rate of the cells.
I do need to modify the thing so that I can re-program the slave micros via the serial port (bootstrap s/w to be found and added) rather than using the ISP which is how it is at present. It would be a bit of a hassle having to plug in all those ISP connectors to re-program a bunch of boards!
I note that my slave board draws about 9 ma when running, which is a little more than I thought it would, but still not bad. (the micro spec says typ 4, max 9 ma) That's .216 AHr per day or 18 AHr per year. About half the self-discharge rate of the cells.
I do need to modify the thing so that I can re-program the slave micros via the serial port (bootstrap s/w to be found and added) rather than using the ISP which is how it is at present. It would be a bit of a hassle having to plug in all those ISP connectors to re-program a bunch of boards!
-
- Groupie
- Posts: 194
- Joined: Tue, 18 Sep 2007, 18:07
- Real Name: Rod
- Location: WA, Australia
- Contact:
Home grown BMS ideas !
Hello All,
Believe it or not I have only just read this post. I had no idea Tuarn Brown (acmotor) publicized my emails to him about my patent. It does explain a few things about this forum to me.
The patent was a costly experiment on my part just to give me some time to develop and comercialise the simple BMS concept. IMHO that is what a patent is for, not to raise "a shower of cash" as Johnny Wright puts it. I have no intention of pursuing hobbyists who use the idea. It has cost me more than I ever made out of it. However the Patent Examiner did accept it.
My annoyance with Tuarn stems from the fact that he was at an AEVA meeting in late 2008 when Dave Waplington invited me to spontaneously speak about the then very new BMS system I was developing and trying to sell. I had developed it for my Mighty Boy EV in which I had installed TS LFP cells which were new on the market then.
At this meeting I described in detail how the simple BMS concept worked. Tuarn must have thought it was fair game for publication. However he neglected to mention where he got the idea from.
For those of you who don't know what I am on about the system is explained here:
http://www.ev-power.com.au/-Thundersky- ... stem-.html
Believe it or not I have only just read this post. I had no idea Tuarn Brown (acmotor) publicized my emails to him about my patent. It does explain a few things about this forum to me.
The patent was a costly experiment on my part just to give me some time to develop and comercialise the simple BMS concept. IMHO that is what a patent is for, not to raise "a shower of cash" as Johnny Wright puts it. I have no intention of pursuing hobbyists who use the idea. It has cost me more than I ever made out of it. However the Patent Examiner did accept it.
My annoyance with Tuarn stems from the fact that he was at an AEVA meeting in late 2008 when Dave Waplington invited me to spontaneously speak about the then very new BMS system I was developing and trying to sell. I had developed it for my Mighty Boy EV in which I had installed TS LFP cells which were new on the market then.
At this meeting I described in detail how the simple BMS concept worked. Tuarn must have thought it was fair game for publication. However he neglected to mention where he got the idea from.
For those of you who don't know what I am on about the system is explained here:
http://www.ev-power.com.au/-Thundersky- ... stem-.html
-
- Groupie
- Posts: 194
- Joined: Tue, 18 Sep 2007, 18:07
- Real Name: Rod
- Location: WA, Australia
- Contact:
Home grown BMS ideas !
PS. I meant this post:
acmotor wrote: FYI guys, I received this email.....
(names xxxxxx'd on purpose !)
> Hello xxxxxx,
>
> I refer to this post on the AEVA forum.
> viewtopic.php?t=900
>
> Are you aware that the system defined in it is patented by myself?
>
> Regards, xxxxxxxxx
>
my response......
Hi xxxxxx,
Thanks for the email.
There are several systems described in this thread. Which one are you refering to ?
Please clarify your situation and provide a copy/link to a patent claiming ownership of the IP relating to BMS so I can review same. Clearly I would not wish to see any AEVA people breaching your patent for commercial purposes.
BMS for cells (lithium in particular) seems to me to be defined clearly as OV and UV monitoring with some option to shunt cells. This side of things would not be patentable as these functions are simlpy part of cell manufacturers recommendations.
A particular method, if sufficiently unique or novel may find favour with the patent system (in some countires !)
In my case, the circuit is based on the app notes March 2005 from National Semiconductor in both voltage monitoring and shunt operation and as such no patent can be implied or breached.
http://www.national.com/ds/LM/LM431.pdf refer p8,9
Regards,
xxxxxx
I await further developments and will keep this thread posted !
- acmotor
- Senior Member
- Posts: 3614
- Joined: Thu, 26 Apr 2007, 03:30
- Real Name: Tuarn
- Location: Perth,Australia
Home grown BMS ideas !
I think someone is assuming too much.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
-
- Groupie
- Posts: 194
- Joined: Tue, 18 Sep 2007, 18:07
- Real Name: Rod
- Location: WA, Australia
- Contact:
Home grown BMS ideas !
The fact is Tuarn, it was not your idea or original method and you know it.
I had done more than a years worth of R&D before I spoke of it at that meeting in 2008. I faced criticism that the system would not work and friendly advice that there were many better systems just around the corner. However I persisted to refine it, reduce its size and cost to customers. If I get some income from my efforts I deserve to.
If memory serves me right you had demonstrated a charge shuffling BMS also at the time. I don't see that publicized on the forum.
I had done more than a years worth of R&D before I spoke of it at that meeting in 2008. I faced criticism that the system would not work and friendly advice that there were many better systems just around the corner. However I persisted to refine it, reduce its size and cost to customers. If I get some income from my efforts I deserve to.
If memory serves me right you had demonstrated a charge shuffling BMS also at the time. I don't see that publicized on the forum.
- acmotor
- Senior Member
- Posts: 3614
- Joined: Thu, 26 Apr 2007, 03:30
- Real Name: Tuarn
- Location: Perth,Australia
Home grown BMS ideas !
Oh dear, you are going to look stupid over this !
I'm looking at my invoice from Digikey dated 21/5/2008 for 100 x p/n 296-1348-1-ND (TLV431), the file date on Altium PCB designer 6/6/2008 and the PCBcart delivery note for 250 cell mangenent PCBs on 27/6/2008.
All this after months of experiments with 431's.
e.g.
April 2008
with the digikey part..
Quite honestly, I had no idea of what you were working on. You had kept that secret hadn't you ? hoping to make money out of it !
I can understand the hard work you may have put in. Guess what, I had put in plenty of work too. It is often the way that the person in the next room is working on the same project as you !
Even after you presented a BMS system at an AEVA meeting I still had no idea of your cell top circuit or its key component, the 431 prog zener. You gave no internal detail, why would you, That's not the black box commercial way !
It was not until you sent details of your patent this year that I did a double take at the circuit. Man that looks familiar. Just like National Semiconductor's app. note !
I am sorry that you feel as you do, perhaps seeking a scapegoat for your not having taken out patents outside Australia ! I wish you the best with your business but hope it doesn’t rely on hiding behind patents.
National Semiconductor had suggested the circuit ideas in its 2005 app notes for the LM431. Have you actually looked at those ????? I did, and was inspired to make my BMS circuit. I would think the cell circuit idea you claim as yours came from that app. note as surely you researched the data on the 431.
So, that fact actually is, your cell top circuit is not original.
As far as a complete BMS goes, I still tell people to buy your system.
edit: added some jpegs for clarity.
I'm looking at my invoice from Digikey dated 21/5/2008 for 100 x p/n 296-1348-1-ND (TLV431), the file date on Altium PCB designer 6/6/2008 and the PCBcart delivery note for 250 cell mangenent PCBs on 27/6/2008.
All this after months of experiments with 431's.
e.g.
April 2008
with the digikey part..
Quite honestly, I had no idea of what you were working on. You had kept that secret hadn't you ? hoping to make money out of it !
I can understand the hard work you may have put in. Guess what, I had put in plenty of work too. It is often the way that the person in the next room is working on the same project as you !
Even after you presented a BMS system at an AEVA meeting I still had no idea of your cell top circuit or its key component, the 431 prog zener. You gave no internal detail, why would you, That's not the black box commercial way !
It was not until you sent details of your patent this year that I did a double take at the circuit. Man that looks familiar. Just like National Semiconductor's app. note !
I am sorry that you feel as you do, perhaps seeking a scapegoat for your not having taken out patents outside Australia ! I wish you the best with your business but hope it doesn’t rely on hiding behind patents.
National Semiconductor had suggested the circuit ideas in its 2005 app notes for the LM431. Have you actually looked at those ????? I did, and was inspired to make my BMS circuit. I would think the cell circuit idea you claim as yours came from that app. note as surely you researched the data on the 431.
So, that fact actually is, your cell top circuit is not original.
As far as a complete BMS goes, I still tell people to buy your system.
edit: added some jpegs for clarity.
Last edited by acmotor on Sat, 31 Jul 2010, 16:09, edited 1 time in total.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
-
- Groupie
- Posts: 194
- Joined: Tue, 18 Sep 2007, 18:07
- Real Name: Rod
- Location: WA, Australia
- Contact:
Home grown BMS ideas !
Tuarn,
The circuit is irrelevant. There are a hundred circuits to achieve the same method.
The fact is it is not your original BMS idea or method and you know it. Though given your love of endless circular discussion I don't expect you ever to admit it.
Another fact is you came looking to purchase LFP cells from me months after my early customers.
I am still waiting for you to detail on this forum with schematics and software your own BMS design which I believe was a charge shuffling system.
C'mon, lets hold it up to some critical evaluation, it might be useful to others who want to progress the concept.
The circuit is irrelevant. There are a hundred circuits to achieve the same method.
The fact is it is not your original BMS idea or method and you know it. Though given your love of endless circular discussion I don't expect you ever to admit it.
Another fact is you came looking to purchase LFP cells from me months after my early customers.
I am still waiting for you to detail on this forum with schematics and software your own BMS design which I believe was a charge shuffling system.
C'mon, lets hold it up to some critical evaluation, it might be useful to others who want to progress the concept.
- acmotor
- Senior Member
- Posts: 3614
- Joined: Thu, 26 Apr 2007, 03:30
- Real Name: Tuarn
- Location: Perth,Australia
Home grown BMS ideas !
Ah, now we are getting somewhere !!!
An admission by you that the 431 prog zener circuit for shunt regulation and monitoring is not an original idea of yours (or mine).
Thanks for being honest to us all and admitting that !
Circular discussion ? Hello pot !
So my publishing my version of a cell top circuit has nothing to do with your patent on a BMS as you agree the circuit is not original anyway.
Now, as has been discussed on this forum and others, the nomenclature is rather general for BMS, CMS etc. (it pays to listen to the vibe)
As one contributor correctly pointed out some time back, this thread is about a Cell Management System and in particular a cell monitoring and shunt control circuit. In fact the discussion has been about that very cell circuit and ways to go about OV/UV and shunt.
I have not offered any detail on a 'battery management system'. Just a cell top circuit board that you now admit is not original.
In fact, I didn't notice anything much from any contributor on the 'rest' of a system.
I understand you have a patent for the concept of cell based battery management. Do you hold one in China or the US ? So who are you kidding ? A patent only means you were the FIRST to PATENT.
The 'BMS' that I dabbled with in 2007 and showed at an AEVA meeting including circuits and software in Feb 2008 was a pic-axe based shunt system with opto isolated data and PC readout. This was a complete "Battery Management system" as it was used for 12V SLA BATTERIES.
OV/UV and shunt have always been just the same issues with lead acid as lithium. For crying out loud, don't dispute that too !
Actually, there was nothing really new about that system but I didn't push it as it was too expensive as I planned a 220 cell lithium battery pack.
October 2007
If that is all there is left to put a cork in you, I'll dig the circuit out and post it. (don't laugh at it guys !! )
I should note that I have no commercial interest in manufacturing a BMS system. I am not challenging your patent, just you assumptions as you cast allegation at me !
I also vote for the dissemination of information and advancement of EV technology.
Oil companies will be laughing at the squabbles in the EV world !
Sorry forum readers for this interlude.
An admission by you that the 431 prog zener circuit for shunt regulation and monitoring is not an original idea of yours (or mine).
Thanks for being honest to us all and admitting that !
Circular discussion ? Hello pot !
So my publishing my version of a cell top circuit has nothing to do with your patent on a BMS as you agree the circuit is not original anyway.
Now, as has been discussed on this forum and others, the nomenclature is rather general for BMS, CMS etc. (it pays to listen to the vibe)
As one contributor correctly pointed out some time back, this thread is about a Cell Management System and in particular a cell monitoring and shunt control circuit. In fact the discussion has been about that very cell circuit and ways to go about OV/UV and shunt.
I have not offered any detail on a 'battery management system'. Just a cell top circuit board that you now admit is not original.
In fact, I didn't notice anything much from any contributor on the 'rest' of a system.
I understand you have a patent for the concept of cell based battery management. Do you hold one in China or the US ? So who are you kidding ? A patent only means you were the FIRST to PATENT.
The 'BMS' that I dabbled with in 2007 and showed at an AEVA meeting including circuits and software in Feb 2008 was a pic-axe based shunt system with opto isolated data and PC readout. This was a complete "Battery Management system" as it was used for 12V SLA BATTERIES.
OV/UV and shunt have always been just the same issues with lead acid as lithium. For crying out loud, don't dispute that too !
Actually, there was nothing really new about that system but I didn't push it as it was too expensive as I planned a 220 cell lithium battery pack.
October 2007
If that is all there is left to put a cork in you, I'll dig the circuit out and post it. (don't laugh at it guys !! )
I should note that I have no commercial interest in manufacturing a BMS system. I am not challenging your patent, just you assumptions as you cast allegation at me !
I also vote for the dissemination of information and advancement of EV technology.
Oil companies will be laughing at the squabbles in the EV world !
Sorry forum readers for this interlude.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
-
- Senior Member
- Posts: 492
- Joined: Mon, 01 Sep 2008, 19:27
- Real Name: Bruce Armstrong
- Location: Perth WA
Home grown BMS ideas !
Before too much nore acrimony flies back and forth, I would like to point out that better than a patent is to have a product in the market.
I think that Rod deserves much credit for having stuck his neck out by designing and making something that EV constructors need and pricing such that it's affordable.
The latest versions of the ev-power BMS are tidy, compact and easily installed.
For some of us, we like a challenge and insist on designing our own. I'm doing this because I feel that more information from the battery is desirable and because I hopefully can. I think that this is a rite of passge with 500-600V ACIM systems.
Back on topic:
I have been unhappy with the concept of the fibreglass BMS terminals in the main electrical and stress path and am considering tapping the bolt tops to M4 so that the BMS can be installed onto fully assembled battery sections, preventing it getting killed while tightening the terminals.
Drilling and tapping stainless steel is hard on the tools and operator, so I've been waiting until I see the bolts before committing to this course. there's more than 300 of them.
Another thought was to use SS studs, possibly long set screws with an allen key in the end and double nuts.
I think that Rod deserves much credit for having stuck his neck out by designing and making something that EV constructors need and pricing such that it's affordable.
The latest versions of the ev-power BMS are tidy, compact and easily installed.
For some of us, we like a challenge and insist on designing our own. I'm doing this because I feel that more information from the battery is desirable and because I hopefully can. I think that this is a rite of passge with 500-600V ACIM systems.
Back on topic:
I have been unhappy with the concept of the fibreglass BMS terminals in the main electrical and stress path and am considering tapping the bolt tops to M4 so that the BMS can be installed onto fully assembled battery sections, preventing it getting killed while tightening the terminals.
Drilling and tapping stainless steel is hard on the tools and operator, so I've been waiting until I see the bolts before committing to this course. there's more than 300 of them.
Another thought was to use SS studs, possibly long set screws with an allen key in the end and double nuts.
It's not the end of the world, but I can see it from here.
- weber
- Site Admin
- Posts: 2934
- Joined: Fri, 23 Jan 2009, 17:27
- Real Name: Dave Keenan
- Location: Brisbane
- Contact:
Home grown BMS ideas !
While I agree with acmotor's views on patents, I came across the following wise words from Rod Dilkes on the EVPower web site
(http://www.ev-power.com.au/-Thundersky- ... stem-.html) :
But these wise words have caused Coulomb and I to plan some restructuring of the software in our cell-top boards so that they continue to control their bypassing and their error LED and send an alarm on UV/OV/UT/OT, even if they are not receiving anything on their serial communications input. And of course they should also report any such lack of serial input.
Re the bolting of cell-top boards:
The PCB certainly should not be in the main current path. The PCB should be _on_top_ of any high-current link or cable lug. The link or lug must be directly against the terminal (wire-brushed and with anti-corrosive paste). But I don't think there's any problem with the stress of the bolted joint being through the PCB. This is done all the time inside industrial VF drives. The important thing they do is to use a belleville washer between the bolt head and the PCB.
http://en.wikipedia.org/wiki/Belleville_washer
(http://www.ev-power.com.au/-Thundersky- ... stem-.html) :
I feel the most important words here are "layered on top". But I feel Rod is mistaken when he says this can be done "easily".There is a lot of hype about battery management systems for Lithium Ion batteries. Should a BMS have cell by cell voltage monitoring and reporting to a central display unit? Should cell temperatures be recorded and reported? Should the BMS be able to bypass large charge currents? What about active balancing and a CAN bus interface?
The answer is, NO. These functions are not essential.
Sophisticated BMS’s may provide an impressive list of functions but are prone by their very complexity to problems both in installation and usage.
SIMPLICITY = RELIABILITY.
SIMPLICITY = LOW COST
SIMPLICITY = EASY TO UNDERSTAND
The EV Power BMS is the simplest system in the world that will actually provide balancing and fail safe voltage protection of large format LiFePO4 batteries
Extra monitoring functions can be easily layered on top of a simple BMS. A simple BMS should form the core of any large format LFP battery.
But these wise words have caused Coulomb and I to plan some restructuring of the software in our cell-top boards so that they continue to control their bypassing and their error LED and send an alarm on UV/OV/UT/OT, even if they are not receiving anything on their serial communications input. And of course they should also report any such lack of serial input.
Re the bolting of cell-top boards:
The PCB certainly should not be in the main current path. The PCB should be _on_top_ of any high-current link or cable lug. The link or lug must be directly against the terminal (wire-brushed and with anti-corrosive paste). But I don't think there's any problem with the stress of the bolted joint being through the PCB. This is done all the time inside industrial VF drives. The important thing they do is to use a belleville washer between the bolt head and the PCB.
http://en.wikipedia.org/wiki/Belleville_washer
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
- acmotor
- Senior Member
- Posts: 3614
- Joined: Thu, 26 Apr 2007, 03:30
- Real Name: Tuarn
- Location: Perth,Australia
Home grown BMS ideas !
That's just it guys.
I have always pushed Rod's product and been very proud of an Australian manufacturer !
You know those x times a week conversations with prospective EV convertors where they say they don't want to build or research anything ? I remind them of the off the shelf systems and parts available including a BMS system from ev-power.
To be fair, I describe it as basic but functional, but non the less available and proven.
But can you imagine my reaction when I receive an email this year accusing me of doing "immeasurable damage" to that same business by publishing the circuit diagram at the start of this thread ?
I was compelled to defend myself.
Are we to be gagged from advancing the very technology we strive for ?
I have always pushed Rod's product and been very proud of an Australian manufacturer !
You know those x times a week conversations with prospective EV convertors where they say they don't want to build or research anything ? I remind them of the off the shelf systems and parts available including a BMS system from ev-power.
To be fair, I describe it as basic but functional, but non the less available and proven.
But can you imagine my reaction when I receive an email this year accusing me of doing "immeasurable damage" to that same business by publishing the circuit diagram at the start of this thread ?
I was compelled to defend myself.
Are we to be gagged from advancing the very technology we strive for ?
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
- acmotor
- Senior Member
- Posts: 3614
- Joined: Thu, 26 Apr 2007, 03:30
- Real Name: Tuarn
- Location: Perth,Australia
Home grown BMS ideas !
Oh yeah, that SLA battery BMS.
I had to scan from hard copy as my computer has changed at least 5 times since then and I didn't drag that early stuff along.
Circuit using shunt and voltage / temperature measurement via pic-axe 08 with opto isolated serial data comms.
The pic-axe code. (don't laugh !)
Programming Editor (c) Revolution Education Ltd 1996-2006
Filename: e:\pic axe\bms v4 pulse with temp corr.bas Date:2010-06-22
'SLA battery management system pic-axe 08M 10 bit adc and pwm control ver 4.0 June 2007
'pulsout pin(2),period(0-5000 where 5000 gives
'Input voltage sense is 4:1 ratio i.e. 20V: 5V = 1024 bits :20mV resolution
'So 13.8V float voltage = (13.8/4)/5 x 1024 = 706 bits , 15V gives (15/4)/5 x 1024 = 768 bits '14.5V = 725 bits
'LM35 temperature sensor on input 4, 10mV/deg C
' b9 is still available, all other variables are used
output 2
input 1
input 3
input 4
b8=1 'set battery number here 1 to 50
w5=1 'initialise integrate counters
readadc10 4,w1 'read and scale temperature
bl=wl/2 'b1 is in approx deg C
w6=726-b1 'slope is -20mV 1 deg C so since resolution on battery V is 20mV
'then target (706)-(b1-20) i.e 726-b1 is the corrected float V from 20deg C
premain:
b0=0 'flag to say serial data request has been filled 0=not filled
main:
readadc10 1,w1 'read 10 bit adc pint
if w1 >w6 then lolim '706 bits is 13.8V at 20 deg C
w3=0 'no pwm output needed as battery <13.8V
dec w5 'decrement integral as battery <13.8V if w5<2 or w1 <563 then let w5=1
end if
'limit integral to min of 1
'or if battery voltage gets pulled below 11 V then clear integral 'to allow full recharge time
outfit:
pulsout 2,w3 'pulse out
if pin3=0 then premain 'no request from control system for data
if pin3=1 and b0=0 then gosub tx
'pin3 is opto isolated request line from control system
goto main
lolim:
inc w5 'w5 integrates with time
if w5>2550 then let w5=2550 'integrate limit
end if
w2=w1-w6 '706 bits is 13.8V at 20 deg C so w2 is bits above 13.8V
if w2>19 then lim2 'if >15V then full on straight away
w3=w2"12+w5" 2 'proportional factor plus integral
if w3>5000 then lim2 'limit the output
goto outfit
lim2:
w3=5000 'limit the output to 100% duty cycle goto outfit
tx:
sertxd(#b8," ",#w1," ",#w3," ",#w5," ",#b1,13,10)
'send voltage reading and pwm value as bits 4800 baud
b0=1 'flag to say reading has been sent, don't send again until new request
readadc10 4,w1 'read and scale temperature
b1=w1/2
w6=726-b1 ‘temperature correction as above
return
It all looks a bit lame now compared with some recent micro based offerings.
I had to scan from hard copy as my computer has changed at least 5 times since then and I didn't drag that early stuff along.
Circuit using shunt and voltage / temperature measurement via pic-axe 08 with opto isolated serial data comms.
The pic-axe code. (don't laugh !)
Programming Editor (c) Revolution Education Ltd 1996-2006
Filename: e:\pic axe\bms v4 pulse with temp corr.bas Date:2010-06-22
'SLA battery management system pic-axe 08M 10 bit adc and pwm control ver 4.0 June 2007
'pulsout pin(2),period(0-5000 where 5000 gives
'Input voltage sense is 4:1 ratio i.e. 20V: 5V = 1024 bits :20mV resolution
'So 13.8V float voltage = (13.8/4)/5 x 1024 = 706 bits , 15V gives (15/4)/5 x 1024 = 768 bits '14.5V = 725 bits
'LM35 temperature sensor on input 4, 10mV/deg C
' b9 is still available, all other variables are used
output 2
input 1
input 3
input 4
b8=1 'set battery number here 1 to 50
w5=1 'initialise integrate counters
readadc10 4,w1 'read and scale temperature
bl=wl/2 'b1 is in approx deg C
w6=726-b1 'slope is -20mV 1 deg C so since resolution on battery V is 20mV
'then target (706)-(b1-20) i.e 726-b1 is the corrected float V from 20deg C
premain:
b0=0 'flag to say serial data request has been filled 0=not filled
main:
readadc10 1,w1 'read 10 bit adc pint
if w1 >w6 then lolim '706 bits is 13.8V at 20 deg C
w3=0 'no pwm output needed as battery <13.8V
dec w5 'decrement integral as battery <13.8V if w5<2 or w1 <563 then let w5=1
end if
'limit integral to min of 1
'or if battery voltage gets pulled below 11 V then clear integral 'to allow full recharge time
outfit:
pulsout 2,w3 'pulse out
if pin3=0 then premain 'no request from control system for data
if pin3=1 and b0=0 then gosub tx
'pin3 is opto isolated request line from control system
goto main
lolim:
inc w5 'w5 integrates with time
if w5>2550 then let w5=2550 'integrate limit
end if
w2=w1-w6 '706 bits is 13.8V at 20 deg C so w2 is bits above 13.8V
if w2>19 then lim2 'if >15V then full on straight away
w3=w2"12+w5" 2 'proportional factor plus integral
if w3>5000 then lim2 'limit the output
goto outfit
lim2:
w3=5000 'limit the output to 100% duty cycle goto outfit
tx:
sertxd(#b8," ",#w1," ",#w3," ",#w5," ",#b1,13,10)
'send voltage reading and pwm value as bits 4800 baud
b0=1 'flag to say reading has been sent, don't send again until new request
readadc10 4,w1 'read and scale temperature
b1=w1/2
w6=726-b1 ‘temperature correction as above
return
It all looks a bit lame now compared with some recent micro based offerings.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
-
- Senior Member
- Posts: 492
- Joined: Mon, 01 Sep 2008, 19:27
- Real Name: Bruce Armstrong
- Location: Perth WA
Home grown BMS ideas !
Yes,
I like belleville washers, but they're hard to get. It seems that no matter how I describe them, I get spring washers
It was a thought to make the BMS maintainable (removable) without disassembing the battery string and risking high impedance joints or reverse biasing the PCB.
(Temperature monitoring should find any bad joints before a meltdown occurs)
Hi acmotor,
There'll be plenty of gagging opportunities when you eventually see my monstrosity
I like belleville washers, but they're hard to get. It seems that no matter how I describe them, I get spring washers
It was a thought to make the BMS maintainable (removable) without disassembing the battery string and risking high impedance joints or reverse biasing the PCB.
(Temperature monitoring should find any bad joints before a meltdown occurs)
Hi acmotor,
There'll be plenty of gagging opportunities when you eventually see my monstrosity
It's not the end of the world, but I can see it from here.
- weber
- Site Admin
- Posts: 2934
- Joined: Fri, 23 Jan 2009, 17:27
- Real Name: Dave Keenan
- Location: Brisbane
- Contact:
Home grown BMS ideas !
Not lame. Not so different from ours (which could be a worry ). At least you remembered to put a capacitor on your analog input. How come no one picked us up on that when we published our schematic? Rather than redo the board we'll just stack them on top of the lower divider resistors.
The salespeople may not know what belleville washers are, but you can find them on websites. Google "fasteners" within Australia and then search on "belleville" or the misspelling "bellville" within the site.
I think of Blackwoods.com.au as the Farnell of mechanical engineering (only nowhere near as quick). You have to register and sign in before you get to see prices and place online orders. They have them in M6 stainless for 23c each inc GST for qty 200+.
http://www2.blackwoods.com.au/infoBANKp ... &P=7019643
The salespeople may not know what belleville washers are, but you can find them on websites. Google "fasteners" within Australia and then search on "belleville" or the misspelling "bellville" within the site.
I think of Blackwoods.com.au as the Farnell of mechanical engineering (only nowhere near as quick). You have to register and sign in before you get to see prices and place online orders. They have them in M6 stainless for 23c each inc GST for qty 200+.
http://www2.blackwoods.com.au/infoBANKp ... &P=7019643
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
- Johny
- Senior Member
- Posts: 3749
- Joined: Mon, 23 Jun 2008, 16:26
- Real Name: John Wright
- Location: Melbourne
- Contact:
Home grown BMS ideas !
Dissimilar metals guide. The one that isn't intuitive to me is stainless steel and zinc plated don't mix well.
Interesting that an aluminium post then tin plate (like on the BMS PCB) then copper (as for the inter-cell straps) would be a good combination (IMO).
http://www.engineersedge.com/galvanic_capatability.htm
Interesting that an aluminium post then tin plate (like on the BMS PCB) then copper (as for the inter-cell straps) would be a good combination (IMO).
http://www.engineersedge.com/galvanic_capatability.htm