I suppose it is time to face reality. We have won the battle most completely. And we have lost the war most entirely. I concede defeat reluctantly, petulantly and never with good grace. But it is time to finally admit, I was wrong about the Battery Management Systems and/or Battery Monitoring System issue. And I’m just going to have to live with that.
In the win column is of course the entire top balance/bottom balance issue. In this episode, Jason Horak, the one most severely talented example we have of a man who can make an electric vehicle stop in its tracks at will, demonstrates rather empirically the efficacy of bottom balancing in allowing you to recover ALL your cells after an egregiously unfortunate event with your batteries.
Having had both the Speedster Duh (discharged to 6 volts across 57 cells) and the Toyota pickup (about the same across 48) BOTH recover ALL cells after almost identical mishaps within the past two months, the hypothesis is pretty much working theory at this point. You can protect your cells from overdischarge by bottom balancing.
[jwplayer streamer=”rtmp://s3einxnpkaij93.cloudfront.net/cfx/st/” provider=”rtmp” file=”news071213-iPhone.m4v” hd.file=”news071213-1280.mov” image=”http://media3.ev-tv.me/news071213.jpg” width=”950″ height=”584″ html5_file=”http://media3.ev-tv.me/news071213-1280.mov”]
We also know from about 30 EV fires now that the best way to avoid the corollary of over discharge, overcharge, is to avoid the top balancing BMS systems entirely.
I’ve had a couple of “signposts” epiphanies over the past year I will relate briefly.
The first is Rich Rudman of Manzanita Micro. In a brutish attempt to make a joke, and sport of me and the whole “zero drift” concept, at EVCCON last year he noted that he charges every day, and every day his top balancing BMS gets REALLY hot. His analysis of this is that the cells must drift a LOT to cause all that balancing to be going on, resulting in all that heat. Proof positive that they DO in fact drift….
As I happen to know that the drift of the cells is so small you would need to address it about once every five years, he alerts us to a significant and clearly observable phenomenon nonetheless. He may not come to us from the great NASA layoff, but I have no reason to question his honesty. He either observed the heat or did not and he asserts he does observe it.
They ARE shunting current in an attempt to equalize voltage at the top of charge, and if that works, then why does it have to work again the next day? Why DOES his BMS give off heroic amounts of heat every day if in fact they are top balancing the cells.?????
His hypothesis is that I’m totally wrong on the drift concept. I have to struggle with that. But ok. Maybe. Actually no. We know too much about that to accept.
So the remaining possibility is that he isn’t actually top balancing the cells at all? Now how could that be. The circuit basically measures the voltage of the cells, while charging, and cells that reach a certain voltage, you shunt the current that WOULD go into the cell into a resistor or something so that the energy does NOT go into the cell, while maintaining energy flow into the other lower voltage cells. One by one, all cells will be brought to the SAME voltage, and the same amount of shunting and you will be top balanced.
If this is so, then on the morrow, if you discharge the pack and then recharge it, the cells should arrive at the same voltages at the same time, and very little shunting should occur at all. As they are already top balanced, they should STILL be top balanced and the process terminate very quickly. But it wasn’t terminating. It gave off a lot of heat every day.
So we have two possibilities, either they are coming UNBALANCED during the discharge and charge process, or they are NOT achieving the top balance as designed.
Now we know that we can DISCHARGE all the cells and “balance” them at 2.75 volts. If we then charge them, and DON’T top balance, when we discharge them again, if we do so carefully so that any cell reads 2.75 volts after stabilizing, that we can indeed read very close to 2.75v on ALL the other cells. This is empirical, demonstrable, and repeatable. Jason can repeat it.
So that leaves the possibility that they are not being top balanced AT ALL. What a thought.
The other epiphany was the very excellent battery test graphs that Mr. John Hardy of Surrey UK provides us. In one of these, John refers to the writhing of snakes at the top. Once cell will rise in voltage, while the one next to it actually FALLS – while charging. Then the falling cell will rise, and the rising cell will fall. If you see four or six or eight of these together in different colors, it looks like snakes battling over a cricket.
Understand that this is observed by zooming in in both time and voltage so we are displaying millivolts across seconds or minutes. Pretty big zoom factor. Zoom out and they all rise and level off. Zoom in and you seem them bouncing along at the levelled off place, actually writhing in agony.
Semiconductors can actually measure and respond somewhat more quickly than the human time frame. We can picture the process, rarely the speed. How many times a second is a BMS measuring voltage, and adjusting current levels through the shunt. We think of it as CONTINUOUS. But you can chop the time zone to the degree you want to observe – with test instruments of course.
They THINK or VISUALIZE charging to a certain level, and then shunting at a certain level when it “levels off”. That rather assumes it levels off. But if at the more microscopic level it isn’t level at all, but each cell rising and falling rapidly, what would happen to the resulting shunt current? It too would rise and fall rapidly and in tandem of course.
So why do the cells rise and fall minutely at this “level” we are holding in our Constant Voltage phase.
Let’s go back to the crowded night club analogy I used before. In theory, lithium ions are intercalated into the carbon crystalline matrix. And ideally they occupy every other space in the matrix with an empty space between each ion. In this three dimensianal world, entering ions take the first available seat. But later entering ions, having a like charge, displace them to a seat more interior in the matrix.
So for the first ions entering the matrix, it is analogous to going into the nightclub at 5:15 PM finding an empty room but an open bar. You can just walk in to the room and up to the bar and order a beer.
But as a crowd starts to gather at 8:30 PM, you have to walk around and between people to get to the bar. Some of them move to accommodate you. Your path is zig zag through the crowd to the bar.
Now by 10:30 PM when the club is in full swing, for you to enter the door at all, several people have to move to accommodate your entry, and their movement causes several other people to move as well and this travels like a wave through the crowd.
At maximum copacity, for anyone to enter the room, EVERYBODY has to move a bit to accommodate and the normal distance people are comfortable with between each other is quite diminished. In fact, we are kind of touching people all around us. And it’s hard to move through the crowd at all.
Unlike a gas, this may not be an entirely uniform process. Some parts of the room may be more densely packed than others. And if 10 people near the entrance tire of the constant breeze from the door, and move deeper into the room, there’s a sudden opening where several people can enter unimpeded. But then it is more or less “full again”. Eventually, even the deeper, less desireable locations within the bar will be filled with humanoids, a bit under the influence of ethanol and perhaps even available for other late night activities if you play your cards right.
And so this diffusion action of lithium ions intercalating into the carbon matrix, might just have a varying pressure at the “door” depending on the random position of lithium ions in an actually fractured and broken series of matrixes exhibited by each carbon granule.
If you observed this pressure or voltage at the microscopic level, it would tend to vary or bounce around this level “average” both exceeding it when it was difficult to get in and perhaps falling slightly below it when a clump of ions redistributed making room.
But there is a further interaction. We have a BMS trying to vary the pressure on the cell in reaction to what it reads there by shunting current. And this varying pressure reacts with the diffusion pressure and becomes interactive in some time zone.
I would like to propose that instead of referring to this as top balancing, we more accurately depict it as cell charge masturbation. We are jacking off our battery pack. And while it is not likely to cause pregnancy, it will create a little excitement in the form of heat.
That hypothesis eases my mind with regards to the daily heat generation of the BMS. Of course, semiconductors age and fail rather readily in the presence of heat. In failing, they don’t limit current and we overcharge the batteries. If we overcharge them sufficiently, they will of course burn your car, your garage, your house, or your warehouse to the freaking ground. Or say a Norwegian Ferry.
So how are we winning the battle and losing the war?
We currently live in an upside down Alice in Wonderland world. You can actually have a latino neighborhood watch captain who MENTORS black kids in the area, get in a fracas with a known black ne’re do well who attacks him and beats the crap out of him, characterized as a RACIST who goes out at night and hunts down negroes to murder. In our media he was actually morphed into a white guy – a cracker no less.
A Korean airlines crash where in the media we actually had a blond lady aviation expert explain the relationship between knots – nautical miles per hour and mph, statute miles per hour, exactly BACKWARDS with no comment from anyone else on the program or indeed on television AT ALL.
An intern at the NTSB responded to a San Francisco television stations request with the names of the crew of the airline. Imagine the TV stations’ embarassment when they learned that SUM TING WONG and WI TU LO were NOT the actual names of the crew members in the crash. Indeed HO LEE FUK and BANG DING OW were also madeup names.
And the dumbing down of America and twisted Up id down and down is up surreal presentation of reality gets extended to public policy.
The CALB cells we sell everyday are shipped all over the world for years now. Not ONE has every caused a fire or even heated up in transit. But because of some incidents with “lithium” cells, a term almost certain to cover a multitude of sins and cell types, we have actually crafted and deployed certain Department of Transportation “regulations” involving our batteries. We can no longer send them USPS. We can no longer send them ANYWHERE by air. As shippers of the batteries, we have to undergo DOT “training” and be “certified” to ship cells. We now pay $750 per year for a special Hazmat hotline and have purchased HAZMAT stickers with that hotline number on them. We pay a $30 per shipment surcharge on ALL cell shipments. Actually we don’t pay shit. YOU pay for all of that when purchasing the cells.
We have once again solved a problem without an actual single incidence of its occurrence. The inmates have taken over the asylum. ANd of course YOU are to pay for their totally uninformed indeed moronic attempts to do good and make you safe from everything but a clue.
Alice in Wonderland.
I have spent years of intensely focused effort on devining what these cells actually do do and don’t do in a series of given circumstances. We have actually devised a very simple, inexpsive, elegant and TOTALLY EFFECTIVE way to stop the infernal loosening of cell connections that plagued all electric vehicles. Our braided straps and Nordlock washers, devilishly simple, have just STOPPED this problem. You can tighten a connection and it is tight forever now.
We have the simplest means possible of avoiding the overcharge of battery cells and the connected fires. Just undercharge the cells deliberately. We don’t have any writhing of snakes because we never charge them that fully.
And we’ve developed bottom balancing procedures that I admit are onerous and vexing to perform – ONCE – but appear totally effective at preventing cell death and perhaps even from capacity loss due to overdischarge.
But I’ve heard the most amazing thing from three different viewers in a single month, all pretty serious OEM wannabees. All three acknowledge that as best they can tell, I”ve probably got it right. But it doesn’t matter. They have to have a BMS anyway, or at least CLAIM to have a BMS anyway, or there are liability issues they just cannot appear to have not done something about.
And no doubt they are correct in that line of thinking. In fact, if they keep burning up cars with the BMS’s, I would predict that within two years it will be ILLEGAL to operate an Electric Vehicle without a BMS – to prevent fires – which ironically are caused by the BMS.
IS this too improbable for belief? Certainly not. As I said, you are paying $30 per shipment and for a sTICKER with a HAZMAT hotline on cells that have never burned anything up anywhere!!! ITS THE LAW.
So how are we going to deal with this IN Alice’s house?
When in Rome, do as Alice does. The most ridiculous upside down thing that might spring to mind would be for Jack Rickard to develop a BMS. And so we are going to develop a Battery Management System.
Or is it Battery Monitoring System. One thing we’ve always had going for us is nobody can actually really tell you what BMS stands for with any authority. Or really what it is supposed to do. It’s like a blue elephant gun. As long as it prevents sightings of blue elephants, the details of how it works are not precisely of interest. It’s a gun, and it is VERY effective at completely eliminating BLUE elephants from the neighborhood. In fact, there’s never been a failure. Worldwide. Since the development of the blue elephant gun, there have been NO sightings of blue elephants WORLDWIDE.
So if blue elephant guns happen to burn the occasional house to the freaking ground – well – small price to pay wouldn’t you say? If only ONE CHILD IS SAVED from the horrors of the dreaded rogue blue elephant….well need I say more….
So like the kid on Christmas morning, digging through the pile of horseshit, I’m just certain there has to be a pony in here somewhere. What if, what if we built a BMS that didn’t burn ANYTHING UP AT ALL?????
In April of this year, Sendyne announced a new chip that measures shunt currents with heretofore unparalleled accuracy and manages to measure LARGE currents and SMALL currents both with very good accuracy.
The SFP 100 Current Measurement IC looks like a bit of a breakthrough to me – even though kind of pricey at $27 per chip. We are having some small success with the new 32-bit 84 MHz Arduino Due on both the GEVCU project and my efforts at controlling and configuring the TC Charger. Let’s do some design criteria.
1. Accurately measure ampere hours and kilowatt hours IN to the pack and OUT of the pack. Not JLD404 accuracy. Lab accuracy.
2. Measure pack voltage accurately.
3. Break the pack into four sections measuring the voltage of each of the four sections. We will dynamically compare the voltage of these four quadrants. They should be IDENTICAL plus or minus some small amount and that balance between them should hold for cells charging, cells at rest, and even for cells DISCHARGING at 1000 amperes. If any vary, it indicates we have a problem cell. While driving, we don’t need to know WHICH cell. Operationally it doesn’t matter. We are not going to replace the cell while operating the vehicle. If we know we HAVE a bad cell, we can determine which cell is bad with shop equipment that we don’t have to carry and won’t burn anything up with two hundred sphagetti wires.
4. Let’s measure temperature. Accurately. And let’s say in four places in our pack.
I don’t really want to defeat our NORDLOCKS. So let’s drill out some M8 stainless steel bolts and insert the temperature sensor and the voltage tap IN the bolt so we are actually measuring internal anode temperature at four places in our pack.
5. Provide relays like the JLD404 that can be used to set daughter mode, disable the controller, disable the charger, or whatever you want it to do with regards to ACTION you can take on the knowledge.
6. CAN bus reporting to make information on pack voltage and current available to other devices in the car.
7. Perhaps a bluetooth interface to show pretty dials for instrumentation on tablets, smart phones, or dash screens.
So that’s my entire wish list for a non-incendiary BMS. And using this, you could claim to anyone who cared that your lithium battery pack was fully protected by a Battery Management System of the highest order and sort.
In short, it is entirely possible to both monitor and manage a lithium battery pack with no balancing function at all. The top balance is a ridiculously simple minded adaptation of the lead acid cell thinking that you have to top balance to avoid the damages of cell drift caused by internal discharge. The lithium batteries NEVER DID HAVE AN INTERNAL SHUTTLE MECHANISM to support cell drift at all. The reason I knew there was no cell drift is we lack even a THEORY of cell drift for these cells. No one has ever described how it might work in a rocking chair cell. What I’ve known all along in a dialog with the Rich Rudmans of the world is that their base assertion doesn’t actually have even a theoretical basis in these cells. Aging differences due to temperature would be so slight that it would essentially be irrelevant to operating the pack.
Jeff Southern has offered to help me with this project and I would expect it to take some time to develop. Perhaps other viewers might join in as well. The irony of an EVTV BMS product offering is not lost on me. But there’s no particular rush. We do pretty well with the JLD404 for now. But I think something more advanced may be in order. I just have to come to grips with the oddity that we have won the battle, and indeed lost the war. If asked if we have a BMS, I can see a day where the answer MUST be oh, yes. We have an excellent BMS – safest in the world.
Jason Horak has very generously made a more graphic presentation of EVnetic’s logger.exe program available for your consideration. He calls it the Soliton EV Display and Data Logger. It might save you from his front seat full of various test equipment if you face examining the same issues. And it is likely to be otherwise and variously useful.
I find it ironic watching the Affordable Health Care Act implode on deployment after the democrats successfully shut down all Republican resistance to this heroically expensive mess that promises to end health care in America.
One of the basic premises of this act was to “end discrimination” in health insurance. Previously, if you had a pre-existing condition – meaning you were already sick, health insurers would simply blackball you. Deny you insurance. And then of course they also “discriminated against women” charging higher rates than men on the mere basis that women go to the doctor about five times more often than men.
Incredibly, now they are talking about not only ending these discriminations, but deploying NEW ONES with different rates and programs for the obese and of course for everyone’s whipping boy, the zigarrette smoker.
I have smoked non-filter cigarettes pretty much continously since 1968 and have a very unusual blood cholesteral level of about 125, no serious health problems, and visit a doctor about once every two years for a flight medical and VERY occasionally to have my ears washed out. But they insist it is “only fair” to prevent me from costing you all more money because of my health. More Alice in Wonderland.
I had the idea some years ago of giving smokers at least a quiet symbol of protest against this constant persecution and discrimination that has now become an international sport. Herewith, the Raucher Pin.
We are also going to have 4 inch patches made that are more realistic semblences of the Juden badge. And we are of course considering a line for the obese – another victim group to which I might be able to lay some legitimate claim.
I put this up a couple of days ago So far, over 90% don’t realize its a joke, and about half don’t even have any knowledge of the historical reference. But the pin is real. ORDER TODAY. Remember SUPPLIES ARE LIMITED.
Jack Rickard
Hi Jack & Co..
Perhaps it will be a win win situation, Please write me up to a BMS, this is exactly how I wanted it to work, and perhaps also with Wi – Fi so we can get a little app and Apple user can also get smart instrumentation on their phone. Bluetooth on Apple will not work.
sounds like a great battery monitoring system. Its a more advanced version of the BattBridge from Lee Hart.
Jack, I like the idea of an EVTV BMS, especially as you define it. I’ve always felt a bit disloyal for using the EPS BMS that came with my GBS pack. I did disable the top balancing – it’s a programmable value in the little setup and monitor panel. What it gives me is visibility to each cell in my pack. I guess I’m a pretty poor designer, but all of my cells are kind of hidden away and it would mean a significant amount of disassembly to check them with a meter. I get a certain peace of mind being able to see that the cell voltages and temperatures are reasonably close.
My BMS has been out of production for a while and spare boards are dwindling. The next board failure (and I’ve had a few) will spell the end for this one, so put my name on the waiting list. Your design criteria looks spot-on to me.
Hi Jack,
BMS Battery MONITORING System – Atta-boy Jack now you are sucking diesel ( sorry coulombs).
Before I went at an electric car I decided that I needed to know more about this OBD II stuff so I sent off to the US of A for a “book of instructions” about On-Board Diagnostics. Its a bit cool actually. The OBD II carry on is to monitor the engine performance and ensure that it is operating to minimize environmental pollution. Started in California – now its everywhere. The ‘structions book is called ” OBD II Functions, Monitors & Diagnostic Techniques by Al Santini.
The system runs things called “monitors” to check on various things. There are two types of monitor, continuous monitors and non continuous. For instance on a petrol engine the continuous monitors are usually misfire, component system monitoring and fuel system. The engine has to just start and run to enable these monitors to run.
For an electric car system these could be battery voltage, motor temperature & rpm, auxiliary battery voltage and controller temperature for instance.
The non-continuous monitors have enabling conditions which have to be set before the monitor will run and most have to be run two or three times showing a fault each time before the MIL (malfunction indicator lamp) switches on and a DTC (diagnostic trouble code) shows up.
For an electric car a battery monitor could have several enabling conditions before it would run. Say that the car must have been not used for an hour, the state of charge below 50% and the battery not being charged then the four main pack quadrant voltages are cross checked with each other to pick up an imbalance as you detail above.
Similarly, a frame leak monitor could run perhaps one per day when the vehicle is not being used.
So the logic in the way they have the OBD II system set up is very interesting and an integrated EV Power Control Module EVPCM using a 12V or 24V PLC with a display like Jeff uses would be very interesting.
The beauty of the PLC’s is once you have all the inputs and outputs wired up then any changes are done in the logic not in the hard-wiring. One can see different people coming up with better and better programs for the PLC to implement the EVPCM. Ideally the EVPCM should be able to replace the original PCM on the vehicle and control all the electric power system components such as cooling fans pumps etc. and integrate with the cabin display etc.
PLC’s are available with CAN bus but I have never used that facility on PLCs. I have used them on safety critical items such as hull opening monitoring on a ship and it is crazy how you can massage the software to do things.
Surely if a guy is smoking does he need a badge? If a guy is black… does he? And if God forbid, someone loves eating, do people need to look for a sign that says his feet might not get wet if he stands in the rain? The law will force such people to wear a BMS (Body Mass Sequesterer).
After playing with these cells, I can verify that I see exactly what Jack and John have seen in all of their testing. I want to thank them for this work. It has been invaluable to me….
I can more or less do this now with the PLC in my car. It will accurately measure down to 20ma and also to 20mv. The problem is cost. I am using a PLC and a few industrial voltage/shunt isolation devices that cost about $1,200.00 Not including my little 4″ color touch screen which is about $400.00. So, the parts in my car would total about $1,800.00 but you do get a 4″ programmable color touch screen. However, you have to have a 24vdc supply to run them so you are stuck with two DC-DC converters and the cost is right at $2,000.00. However, Rockwell recently released a PLC (Micro830) that could do all of this for about $1,800.00 with the 4″ display. If you need a system like this imediately, I can provide you with a working program and part numbers on how to do this now. If you can wait, I think you will get a much lower cost Arduino based system.
In my car I bottom balanced to 2.75. I have 36 cells and charged it to 3.5v/cell or 126 volts and shut the charger off. My charger only puts out about 19.4 amps near the top of the charge. (2500 Watt) I do not do any Constant voltage mode. I noticed that two or three of my cells would shoot from 3.5 to nearly 4 volts in about a minute when charging to 126 volts. Obviously these are the weakest cells. So, after identifying my weakest cell, I noted the pack voltage as it hits 3.5 volts. That turned out to be 124 volts. I simply told the PLC to turn the charger off as soon as the pack hit 124 volts. Again the PLC in my car is a very accurate amphr counter. The TOTAL difference in charging between 124 volts and 126 volts is about 0.8 AmpHr.
So by giving up about 200 yards!!! of range I COMPLETELY eliminate the withering of the snakes as John and Jack like to put it…..
I look forward to working on this device….
Jeff,
I had a similar experience with my voltronics cells. I use 38 cells and charge my pack to 132 volts (3.47v per cell) If I try to go higher (133volts) my lowest capacity cell starts to ramp up over 3.8v. For that 1 extra amp hr, better safe than sorry.
Voltronics are similar to Thundersky cells. For these cells, 3.45v/cell is a good number to shoot for.
Also since I am undercharging my cells slightly, I have gotten away with using 38 cells with a HPEVS AC50 curtis1238-7601 system.
Your PLC devices would serve well. But you’ve hit on the two hugish drawbacks. One is cost on another scale and another zip code. And the second, and the one that put me off this equipment, is the 24vdc power requirement. I just don’t want another DC-DC converter in the car. We have more or less constant problems with the 12v DC-DC converters now. I don’t know what I’m doing wrong, but we burn those things up no matter what we do, where they come from , or what kind we use.
I’m not promising a $100 Arduino BMS. But I think we can keep the whole thing well under $1200. And no little series of little boards to connect to each cell like many of the BMS out there. You will have to be able to divide your pack into 3 or 4 equal sized chunks. So 46 cells will be a problem. But 45 or 48 won’t.
Jack,
Why does the division have to be equal. Since we are only comparing voltages we could just as easly compare odd and even number of cells. You will just need a way to enter the number of cells in each pack division….
Actually, There are 12vdc version of these PLC’s. The problem is that the Display and the industrial signal conditioners are usually 24vdc…
However, you do not need a very large 24V DC-DC mine is 50watts and never goes more that two or three degrees above ambient temperature… I did use the dual EVTV power diode to protect both DC-DC converters from reverse current out of the caps. I hope that this will extend their life span.
I kind of like having the two converters. I leave the 50W 24VDC unit up all of the time and only power on the 420W 12V DC-DC converter when I want to power the car. This reduces the parasitic load significantly and still gives me a mostly instant on feature without the need to a battery of switch in the dash at pack voltage.
However I do understand the need for a 12vdc system, but the truth is that you have to use DC-DC converters with the Aduinos as well since they are typically 3.5 – 5vdc. If you went with a 12VDC PLC you would only need a 25 Watt 12V – 24V DC-DC converter to make everything work….
Great idea Jack. How about inventing an EVTV bottom balancing machine to round out the whole safe battery ownership system?
MR .JR
My definition of BMS: =Battery Monitoring System
Lightning strikes,count 1,2,3 its kilometres away ; roll in the thunder.
If you heard thunder; chances are you weren’t hit by lightning.
Full serve gas station, you roll up to pump, run over air hose , a familiar bell rings;n ding ding , and the attendant comes out to service all your automotive needs…
it gets too hot outside …same hose leaks. its made of J.e.l.l.o rubber ; you get no full service.
all above can be intergrated
above= Analogy if you didn’t get
Simon
Jack:
We use two bottom balancing machines. The PowerLab 8 and the Vichay resistor with a contactor and control voltmeter. But when we do all that, I’m kind of standing there.
Aside from it being a good penance and good for the soul, the bottom balancing procedure kind of assures that you can’t really walk off and let it do its thing overnight. That means it’s thing can’t be burning down your house while you sleep. You only need to do it once, and it puts a lot of attention on your cells and your straps and your connections. You can also detect bad cells out of the box, as Anne did.
So I don’t see a bottom balancer. I don’t see wiring to each cell. I think a minimalist “first do no harm” approach to Battery Monitoring/Management is in order. A wee bit at a time, over the course of four years, we by successive approximation learn to actually deal with these cells on their terms. And in doing so, rather than passed along myth and superstition, my hope is that investments in these cells render years and years of faithful service. It’s kind of an “acquirable skill” still situation. Not rocket science. A series of measured and simple steps to maximize cell life.
Jack
I just used a DC Load bank. I used the Hewlett Packard 6050A which allows up to 1,800 watts of total maximum power. They make smaller ones as well. I think you can find them for a few hundred dollars.
Well, if you’re serious about making your own battery monitoring system I think I know someone who knows how to design circuit boards in Eagle and write code for embedded systems. He’s a busy guy though.
In order to meet a common ground point there shall be galvanically separated tion between the four battery voltages as arduino look into. Here, an HCPL-7850 would be good.
I have used it in a speed limit to a small train (trolley) to repair railway tracks with for Danish railway. HCPL-7850 is very Reliable.
http://pdf1.alldatasheet.com/datasheet-pdf/view/119489/HP/HCPL-7850.html
Allan
As palmer_md notes this sounds like the same principle as the batt-bridge – monitor for voltage difference between groups of cells. I like the principle – I had been planning a Lee-Hart batt-bridge to detect cell failures or bloopers (such as accidentally missing a cell when bottom balancing). If this was available I’d use it instead. I have several observations that I’d be interested in others views on:
* 500 + cycles on 8 CA series cells has shown the same result – no voltage drift (if anything they have converged but that is probably experimental error). I have done the first draft of an academic paper on these results because I think they are significant. Most papers I have read extend lead acid thinking and simply assume that cell drift happens. I’d like to publish the paper somewhere it gets noticed but is also open access.(don’t want it sealed in behind a pay wall). Any suggestions gratefully received. More at EVCCON
So I am wholly in favour of no automated balancing
* Voltage monitoring can, in an of itself, add tiny asymmetric parasitic loads which would themselves unbalance the pack over time. I have used op amp voltage followers in my test rig. No doubt the EVTV BMS will take this into account.
* The BMS should not interfere with the normal operation of the charger or the controller, only intervening to shut things down if it detects a serious anomaly. I expect that is the plan too.
* How much in the way of electronics does this actually need? Triggering a MOSFET when a voltage difference reaches a predetermined value sounds like a slight insult to the intelligence of a Due – plus simpler is more reliable as well as cheaper. If the devices were the price of a Big Mac I’d have three
For cell-level connections with the CA series on the test rig I considered drilling and tapping the bolt heads and agree that is ideal, but couldn’t find any off the shelf and ended up using terminals like this: http://www.autoelectricalspares.co.uk/ekmps/shops/millington/images/ring-terminal-6.3-mm-male-non-insulated-single-flat-earth-tag-45-deg-m8-tin-plated-x-10-1860-p%5Bekm%5D300x204%5Bekm%5D.jpg
Do we actually want to call it a BMS or should we coin another term (e.g. ABM – Active Battery Monitor) which if pressed we could with tolerable accuracy describe as a form of BMS? I just wouldn’t want any confusion with the deadly gadgets sold under the BMS label
John:
It’s kind of like the FOAD discussion. The whole point IS to call it a BMS. There are those out there who for various reasons HAVE to have a BMS. But they don’t want to top balance or burn their car up. So what would a “benign” information device look like that could protect your pack, detect early cell problems, and essentially “instrument” your battery pack for voltage, temperature, and current without a maze of spaghetti wiring and power semiconductors look like?
Then we build it and call it a BMS. I do not think the mindless Alice in Wonderland world we live in is going to insist we top balance. They don’t precisely know what a BMS is or looks like, they will just insist we have one. Spending the rest of our life explaining that an ABM is kind of like a BMS isn’t the task we want to sign up to. We’ll just call it a BMS and we mean that as a Battery Monitoring System. That should fill the requirement, while not creating a 5000F torch or a hopeless wiring mess.
Or that’s the plan anyway.
When falling on your own sword, push it all the way in. Just piercing the skin isn’t the point….
Jack RIckard
John:
As a battery tester, it is not lost on me that four voltage measurements, four temperatures, and an accurate coulomb counter, particularly if we included data logging to SD card, might have application OUTSIDE the car as well.
Jack RIckard
John & Jack,
If HAL was one-letter shift of IBM and WNT (Windows NT) is one-letter shift of VMS, then my suggestion is to do a one-letter shift of BMS and name it CNT. We can come up with an explanation for the acronym later, as is customary, but it could for example be Cell Normality Tracker…
I think there is value in appeasing the Queen of Hearts by checking the box that says “BMS Installed”.
But to provide distinction while still placating the ignorant fools, why not term it “Battery Protection System”? After all, it has the word protection- it must be safe.
John,
Around 2002 or 2003 I was involved with doing the software for a LiPo charger for the RC hobby industry. I bought 8 of the 1st generation Kokam cells when they were imported into the US. (There was a brief period where an individual could buy cells before Kokam became OEM centric.) I top balanced them and because nobody had any experience with these cells I tested them as a group of 4 in series. Initially the fully charged resting voltage was off by about 0.006 volts (resting) between the high and low cell. At the end of 30 cycles (at 1C) the difference had reduced by 0.0015 volts. I experienced convergence although I considered it too minor to be useful. More recently I noted this same convergence in the A123 26650 sized cells. So I don’t believe your observed convergence is experimental error. My best guess is that the convergence occurs due to a slight decrease in charge efficiency when the cell gets full. In Jack’s nightclub analogy it would be due to the heat caused when the people brush against each other while moving around the nightclub. The cells that get full first don’t accept additional charge as readily. It is kind of a built in shunt balance at an extremely low level. If top end convergence is true this would have an unfortunate side effect of bottom unbalancing over time the lowest capacity cells. So undercharging a little is pretty critical to maintaining the bottom balance over large numbers of charge cycles. It would also make the idea of needing external shunt balance even more silly than it already is for those who insist on top balance.
I would like to talk more about this at EVCCon in a couple of weeks. Hope to see you all there!
Doug
For the drilled and tapped bolts I use Brass bolts because they are much easier to drill and tap than stainless steel is and they don’t tend to work-harden. At the bottom of my blog post (http://2003gizmo.blogspot.com/2011/07/battery-pack-balance-monitor.html) you can see how I used them to attach my modified Lee Hart Batt-Bridge device and you can see one of the bolts with the hole in it where I used to have BMS boards attached. I originally got these from Black-Sheep Technology (http://black-sheep.us/). They are called BMS mounting hardware. FWIW, I use bronze split washers with these bolts and I haven’t had any loosen on their own. Also, isn’t brass less likely to react with aluminum than stainless steel is?
David D. Nelson
Interesting device Allan. But it requires two power supplies that are ALSO isolated from each other.
And basically that is the plan. I want to use a DC-DC converter to make the whole SHEILD isolated with regards to the power supplies. Then we want to use devices that actually convert the sensor readings on the shield to I2C or serial data and port that to the Arduino.
I see a simple voltage divider for each of the four voltage inputs, all referenced to pack negative. Those read by an A/D and converted to serial stream.
Temperature inputs similarly. High linearity 10k NTC thermistors into an A/D.
And of course our current shunt input with serial output. The Sendyne chip actually does some temperature measurement. I don’t know how easily those could be remote.
CAN can be a separate shield and really so can bluetooth.
Actually Bluetooth BLT can be used with iPhone iPad. That’s all a little early yet but some are working it.
Jack
Actually the Sendyne chip has four local voltage references and one remote. In looking at them I am not sure what make the one “remote”….
I think it is a differential input and the others are single ended….
I would rather use five differential inputs because this keeps the realitive voltage much lower channel to channel and is less subject to noise problems. If you want single ended, you can always tie the commons together at the board.
I like the thermistor and it should fit in a ring terminal type housing…..
I think you will need a way to tell the system how many cells are in each zone. This way you can have packs with cells that are not divisible by four.
I do worry a little about the volt measurement method potentially unbalancing the pack over time….. Hopfully we can find a way to minimize this effect….
Hi Jack
I like the idea of DC / DC converter which enables it to get total galvanically isolated from the battery pack and car chassis.
I can not just see for myself how large the program is and the cycle time, but it go fast with a clock speed at 84 MHz, but I would use the filter on all analog inputs complied to “FILO” number of measurements using in “FILO” I would first determine when I know the cycle time of the program wile running.
I work mostly with Simens PLC S7 300 and you have a constant scan which makes the whole thing becomes easier to work with.
I do not think it’s a good idea to use a PLC as they are often not designed for the environment as a car offers, I am thinking of humidity, vibrations and temperature fluctuations also make that condensation.
An idea was perhaps to use Leaded component, I mean battery pack voltage to avoid leakage currents as I have often had errors on SMD components in coherence with moisture and high voltage SMD is delicate and must at least be painted with epoxy.
Earlier I used NTC sensor, with a round head where one can sight a screw through. However, only 3.7 mm body condition, I have ben looking for one with 8 mm hole spacing but in vain.
http://dk.farnell.com/epcos/b57703m103g/thermistor-ntc/dp/1200456
I hope I can be a help to you …..
I would not do the board without optical isolation of the analog and digital signals….
Optical isolation has a hidden corollary, the power supplies have to be isolated as well. I think everything from Sendyne to pack and perhaps from some opamps and ADC to pack is NOT isolated. But we optically isolate the serial data streams to the Arduno and use
an isolated DC-DC converter to power the ICs on the board.
The Sendyne is simply too sensitive for optical isolation on the input side. They do diagram kind of a complicated filtering and switching
system on the current sensor input. I’m kind of wondering why all that isn’t on the chip.
The Intersil ISL28022 would be much simpler to implement. But I sure like the bragging rights on this Sendyne self calibrating
chip. It will likely require a SPECIAL shunt however. I assume they can point me to some. But the advantage is that it can use an order of magnitude SMALLER shunt resistance, which then has less heating and drifting and so forth.
Over and over again, I’ve found instrumentation that actually UNBALANCES the pack and we do want to avoid that.
Jack Rickard
Hi
I Work on EVs inthe UK especially a series of modified Citroen c1 cars made for about 2 years. http://www.theelectriccarcorporation.co.uk/
They are no more. These used 25 thundersky 160ah cells with a Zivan Charger (HPEV 35 and Curtis) . The BMS not only logs the data on each cell, but cuts back the power if any cell goes too low and controls the charger by some pseudo approximation of current based upon the Cell votages. If any cell drifts outside a given parameter during charge a small resistor allows some shunting , and then shunts after charge to balance all the cells to the lowest votage cell. Then if recharged this slowly over time brings up the lowest votage cell. Top Balancing. The charge finish votage is on the overall pack voltage but if a cell goes high its picked up and the charge ceases. All sound OK, however many of the Thunderskys are now failing, not unusual to see 3 to 6 in a pack of 25 go bad, often sequential serial numbers. ANother issue is now the BMS failures. They have a small circuit in ecch of the channels on the main board, that have a tiny SMD MOSFET that bleeds the current. Being on the input circuit before the volt measuring chip looks like a good idea, however if the related components to the MOSFET go bad this impacts on the voltage reading. They can oscillate between channels, read low read high etc. The other problem is with a few cells going bad, if you watch the thunderskys they all move happily together to about 3.4v then some start to rise rapidly and some take their time. From about 3.5v even the good cells start to get a shift on at 18amps charge with all reaching 3.8+ V in a relatively short time. The Zivans were modified with an opto resitive device to control the charge voltage, but this can be removed and the Zivan set back to manual control. I also purchase a modified charger chip that cuts the charge after 2 hrs of reaching the end of the CC cycle. SO now I have charge control back and set the charge voltage to around 93V this gives a decent state of charge and stops the high flyers from going really bad.
The chips used in the BMS are LTC6802g. These are a good product and can be stacked to get the required number of channels needed as well as giving out a constant data stream that can also be stored.
Using a small resistor as a shunt does a small job, and as the cells age and capacity changes, a bigger shunt would be good, even allowing full charge to be by passed. These devices on the BMS could control a remote bigger shunt, right by the cells, more cable, but at least you could clamp a high flyer and stop it getting bloated.
I love being able to download the data on the individul cells. You can see which is going low and by how much in driving, and you can track which is going adrift during charge. I dont agree that you should only do this in the workshop with workshop tools. USers cant do this themselves, you have to have the Workshop softwar to do it, but being able to findout what went wrong , not what is going wrong is really useful. You can look back over weeks of records of each cell and actually work out whats the problem, not just make an assumption from a snapshot in time.
So my vote is for data logging, A simple way of invoking cutback on the controller to prevent further damage, expandable, and data outputs in an open source manner.
Grumpy_b (In the UK)
Grumpy:
You’re a posterboy for how and why NOT to have a BMS at all.
At this point, we have LOTS of Thundersky’s in LOTS of cars and we don’t lose cells at all. In some cases we have driven them to EXTREMELY low voltage levels (like 6 volts in the Speedster) and recovered all just fine. We changed one Thundersky in Gerry’s Toyota SR5 pickup and it was workign ok – just off capacity. You are actually working those cells to death with that BMS. If you would take the BMS OUT of the car, dig a hole, bury it, then pave over it with gravel and concrete, then put FURNITURE on the concrete with a sign that said NO BMS HERE, NOTHING TO SEE, MOVE ALONG, your battery woes would pretty much end. You don’t need to know the individual cell voltages. You are rubbing the spots off those cells. Quit it. It’s not just unnecessary, you are torturing those cells.
Cease and desist. They don’t deserve it.
Jack Rickard
In my own Smart that had one of these BMS in it, its now gone. But for customers cars its a piece of kit thats woven into the build, even working with the Curtis controller. SO living with it is a partly neccesary evil. Taking out the control of the charger is my concession to getting rid of the BMS but leaving the rest in tact. But its really useful being able to look at the data when something does go wrong. Im not advocating using all of the BMS activities that these cars have, but knowing when a cell has gone west is manifest by a cut back that at least can stop even more damage. When I replace cells I always do bottom balancing, but with a mixed pack of cells and a charger that works on pack voltage, some of the ageing Thunderskys do go high too quickly. If its only one cell thats going high early then you can readily drive the rest onward to fulfillment, and trash the early cell. The monotoring of the BMS ouputs gives you the opportunity of seeing whats going on. Thats really useful. Saves doing a continual tour of the cells with the meter probes. What being able to bypass/ shunt cells does give you is the ability to live with poor cells (thats not an excuse for not replacing them) until you can get the replacements. A load of the users have only the C1s as transport, and so if they are out of action they hhave no transport. Unfortunately these Thunderskys are the early ones and the later ones are a different physical size. All are strapped as per the original delivery and so the containers will only take those cells. So replacement may take a few weeks to organise. Again thats not meant to be a good excuse for having a mass control BMS but an observation.
In Europe the last supplier of CALB cells had a requirement for warranty purposes that a cell monitoring and recording system be used . I do see the point its hugely usefull to see whats been going on. But does require loads of wiring, Not impossible.
The greatest failures of these vehicles (Apart from failing thunderskys) is failing BMS units, they dont read a bank of cells, Fail. They alternate between adjacent cells readings.Fail. They dont discharge/ Balance, result longer term is failure. If left they will balance the pack down to noting. Fail. They dont control the charger properly. Fail.
But at least they record the data so you can see what they are doing.
Grumpy-b
Right now he’s really busy on a really important project. I don’t want to slow that down any.
But yes. We’ll no doubt need adult supervision anyway.
Jack
I’ll probably need all the help i can get…
I do own Eagle, but it is a little out of date….
Have you considered that the Rudman BMS quiescent current during the day introduces imbalance?
If a system is always connected it has to have extremely low drain and in the case of the Tesla Roadster they have certainly failed in that regard with 100Watt continuous drain 24/7 that you can never turn off and that will destroy the battery pack if left unattended for a month.
Rather than have a spaghetti of 100 wires, I’ve been considering monitoring individual cell voltage by placing a opto-mosfet and an AVR at each cell. Basically, you have three bus wires and a ground. Two wires are used for sampling any two node voltages and the other commands pairs of opto-mosfets to be on via a serial chain through each AVR. At the tail end of the sampling bus you have an isolated voltage sensing chip such as the ACPL-C870-000E. This lets you use just 4 wires, and since these wires can follow the main cable it’s much less of a mess. I’ve considered drilling and tapping the bolts as a way of attaching these sense boards to terminals without compromising the mechanical connection. The benefit of this approach is that each board is ~$1 in components and you only need a single expensive isolated sensing chip. The above chip also boasts a 1Gohm input impedance, so the output of each node could be protected with perhaps 1Mohm of resistance in case of accidental shorts without compromising accuracy.
The serial chain commands would include a CRC and the command order would be “all_off”, then “enable nodes N and M”. The master node would keep trying a command until it receives a matching echo. In this way you would avoid accidentally exposing the ACPL to huge voltages.
Of course, if sensing 3-4 total points in the pack is sufficient then this all becomes somewhat superfluous, but I really like the idea of being able to read all of the cells at run-time and without breaking out the multi-meter and going cell-by-cell.
My $0.02
Ok I lied, you need 5 wires. I forgot about 3.3V power for the AVRs.
Number of cells in a string.
The proposed BMS would the number of cells to be dividable by four or allow overlapping measurements.
The inverter in my citystromer is splitting the string in two and compare the voltage across the two halves. If it exceeds more than 2.0V it will command the inverter to “Daughter Mode”. Originally it was made with lead acid in mind, where the two volts would be a “Cell Collapsed Event”.
I think I have written something about this previously on your blog.
If I was to do anything more, I would put in a redundant system, that was pure hardware – being a firmware engineer I distrust anything programmable… 😉 Just a simple op-amp with a voltage divider and a feed back to give a sensible hysteresis. The circuit should be made “fail-safe” so it will disable charging even if the circuit is non-energized.
Just my $0.02.
Pressed “Return” to soon – didn’t want to be anonymous 😉
Hmmm… strange things are happening – My comment just disappeared.’
I get the point and the joke. But, please, comparing a smoker’s feelings of persecution to Nazi Germany? Why not your own design of your pin?
Its your ticket to ride on the train to the health spa. Guaranteed to quit smoking.
Hi Jack,
If you talk or write to these battery manufacturers about cell design placing the vent at or near one corner would allow the cells to be placed in either the traditional direction or on the long edge down with the vent at the top in either configuration.
Zak
Jack, The new batteries you just got for testing sounds like they could be a LiMnFeO4, l read something about a year ago about Dow selling a “drop in” for LiFeO4 batteries that improve energy density by 25% and voltage goes up by 0.5v …or somthing like that
Who is John Galt?
When we eagerly wait for the next showing of EVTV and Jack is not there. He will know.
Perhaps Jack is John Galt.
Gathering the worlds best scientists and engineers to a secret libertarian Shangri-la without Government restrictions or control.
You might have something there.
Jack, Did you send Richard out to LA to gather some Coda parts? http://cmaauctions.com/coda-automotive/
Have to keep an eye out for the auction up here in the SF Bay. There assembly line was in Benicia, CA. That is where the pile of motors and controllers would be.
I guess some of the items at that auction are in Benicia. Maybe the motor and controller OEM bought them back.
How many electric garbage cans does a start up car company need?
What I’ve never understood about any dissipating BMS is how does bleeding off a few watts per cell do anything meaningful at all? I know you can put a frozen Jimmy Dean sausage biscuit on there when you plug your charger in at night and in the morning you’ll have a nice warm biscuit sandwich to eat on your way to work but besides that it’s always seemed pointless to me.
Not to pick on Rich but since he does publish specs about his BMS (kudos for that!) we actually have real numbers to analyse. Let’s say you have what seems to me a typical HPEVS conversion package with a nice 36 cell CALB pack and one of the modest 2500W TC/ElCon type chargers AND you’ve wired up Rich’s latest generation MK3x8 regulators. The documentation for this BMS (http://www.manzanitamicro.com/products?page=shop.product_details&flypage=flypage.tpl&product_id=54&category_id=22) states it can bypass up to 2.5 amps per cell or a max of 13.75 watts per cell. The BMS is adjustable up to 5.5 V which is where the 13.75 watts (2.5*5.5=13.75) comes from. But for our CALB cells that we don’t want to charge above about say 3.6 V this BMS can dissipate only 9 watts per cell. That TC charger shouldn’t have any trouble putting out 2400 watts or 66.7 watts (2400/36=66.7) per cell. So bleeding off 9 watts of the nearly 67 watts going in just doesn’t add up to me. What if you have a bigger charger like the 4000w? Then you’re bleeding off 9 watts while you’re pumping in over 100 watts! It’s completely futile unless you are running a lead acid style equalization charge at something below 2.5 amps. That’s the only way you’d have a chance at top balancing a pack with this type of BMS.
I believe Rich when he says his BMS gets “hot” after every charge but so what. That’s a function of his choice of setting for the charger and the BMS. From what I’ve been told by some actual users is the typical setup for a CALB cell with his BMS would be to start dissipating when a cell reaches 3.5 V, terminate charging if a cell gets above 3.65 V, and the target charging voltage would be 3.6 V per cell. I wonder how “hot” his BMS would get if he just lowered his target voltage to 3.5 V per cell? It also makes me wonder how much marketing is behind his choice of settings. After all it’s not so easy to sell something unless it looks like it’s doing something.
Ignore any BMS that gets hot as a danger to you, your possessions and your vehicle. “Dissipating” is PC talk for an open fire waiting to burn everything down.
No electronics should *ever* get hot in any circumstances or with any excuse.
I have to believe that by lowering target voltage to ~3.5vpc that a lot less heat would be generated. While I have experienced no “drift” in my 24 T-Sky 100AH cells on a motorcycle (other than issues caused by monitoring devices!) I *am* using a top-balancing BMS on another project that is performance oriented. Performance applications that demand max available voltage are the single exception to the no-top-balance rule IMHO.
Jon, the charger is supposed to have been throttled back by the time the BMS starts shunting. The BMS boards I used to use shunted at max 500mA at ~4V. The resistor bank was spread out so it only warmed up slightly. In the end I noticed two things. First, it wasn’t always the same cell which reached full first. Second, the approximate difference in (supposed) SOC was no more than 40 mAh (on a 200Ah pack) between the first cell to reach “full” and the last to reach “full”. After Jack kindly contacted me and pointed out that charging to 4.00V/cell was not helpful and was damaging my cells I did my own testing and determined that charging to much over 3.45V/cell was useless and that the cells didn’t really drift. 2 years after pulling off the BMS boards and I’m still not seeing the drift I was told would happen in a very short time and destroy my pack.
http://www.hybridcars.com/oxis-jump-starts-first-commercialization-of-lithium-sulfur-batteries/
Nice link Mark 🙂
Because you are surround by engineer and has unlimited resources does not mean that reslutatet is correctly
How did have they ever be able to convince anybody to insert Cobalt batterie in a plane?
Amazing……………..
Allan
In addition CALB makes nearly the same test. However, the steel spike must stay inside. They report nothing about smoke development, but when the nail is inside there must be smoke.
Quote from CALB
“Nail Penetration Test”
“After the process of standard charging, shelving the cell for 1 hour at the ambient temperature of 20 ° C ± 5 ° C. Afterwards, penetrating the cell Rapidly with a steel spike of φ3mm ~ φ8mm som Should be perpendicular til accumulator electroplating (the steel spike should stay inside of the accumulator). ”
“No Explosion, No Fire”
As a Leaf driver, I have been disappointed that Nissan has decided to wait until after the 2014 year to even suggest an updated Battery Pack. However, on second though, the more anemic current Leafs they sell will mean a larger market and cheaper used EVs after they go off lease or trade in. What they are really creating is good opportunity in used Leafs.
So, you wait until one is available that needs a battery pack–there should be plenty from Arizona alone–and buy the car. Then buy upgrade components from Jack and install a rebuilt pack. With luck by that time Jack will be selling a high density, Lithium Sulfur Battery Kit, including any necessary software changes. What do you think, Jack? Fat Chance?
Mark:
This article is a mess all the way around. IN the first place, it marks a turning point from the lithium ion battery to the lithium sulphur battery. This is just nonsense. Lithium sulphur batteries are just as much a lithium ion battery as any cell we currently use. They are all lithium ion. Sulphur is just a different cathode material to host lithium ions.
But the 500 cycles of the cells Oxis apparently has developed is rather more on par with lead acid cells that you have to replace every two to three years.
Sulphur has some problems. Oxis clearly hasn’t addressed them. And this article is so woefully lost that I can’t tell if its the author or Oxis that is in a state of “don’t know and don’t know they don’t know”.
Jack Rickard
Jack:
Oh! OK, maybe just wishful thinking on Osis’s behalf.
Yesterday I went to a local EV event. It sure was different then it was last year. This time there were 5 Tesla Model S’s, one of which was a P85+ which I had not seen up this point in time. One of the Model S owners, Andy, also brought his prototype boat. http://www.purewatercraft.com/ He was a younger guy who did well for himself in his earlier IT career and this boat was his new venture. The oil cooled REMY was visible from the outside mounted directly on top of the inboard/outboard out-drive. Because of the motors oil sump a custom 90 degree gearbox transferred the rotation to the vertical shaft of the out-drive. The motor was 220 horsepower continuous with a non Rinehart Motions Systems 220 kW time 2 inverter. He signed an NDA with the inverter builder and could not reveal his source. Oh yes this prototype had a 60kW hour pack @ 720 vdc. A little over the top! Hah?
Mark Yormark
Who can disagree with you. It was just a sales pitch carefully playing words and not absolutely comparing like for like.
In one week I will be heading out towards EVmecca, ie Cape Girardeau MO for the 2013 EVCCon. Of course my car is all apart and I am crunch mode to complete some updates and get it all back together before next Saturdays departure. Come to think of it I did this last year too. How many of you are in crunch mode too? Looking forward to getting a better score on the Dyno this year. I had the motor current limited to 300 amps last year so it was a pathetic 39.91 HP. If I had known there was an award for worst Dyno run I would have turned it down a few more amps just so I could win something.
I am so looking forward to seeing everyone again.
Doug
I am also looking forward to EVCONN this year. The car is mostly together, but I just can’t seem to Finnish the last few details….
I still need to replace the torsion springs because they squeeky a little. I have the 48v contactor and have not installed it. I need to make the cable shield to run down the side of the tunnel and need to make the the lids for the rear battery boxes….etc…
The problem is that is am having so much fun driving it I can’t seem to find the time to Finnish these last little details…
I did a 72 mile drive today. I started out with 178.2 ahr and finished with 30 AmpHr. The drive lasted about 2.5hr. That puts me at about 230-240 watt hours per mile or so. I think the car weighs about 2300-2400 lb so that seems amount right…
The Thing has the aerodynamics of a brick so it does ok up to about 50mph. After that your are sucking down a lot of amps to do 55-65mph….. It is not really a freeway car…. But is a dream on back roads. I love to take drives like this and it is kind of nice knowing that i only use a few bucks worth of power for this indulgence…
Hi Jeff,
Your car is a work of art. I look forward to seeing it. I am almost embarrassed to show my car. You make the rest of us look bad. 🙂
Thanks for posting on your blog how you installed your hall effect/ pedal assembly.
I still have not done mine yet but I will after EVCONN.
Trust me, it does not look as good In person. It is a true ten footer….
David (GizmoEV)
Brass on Aluminium? It’s better than other stuff but…Here is a table of metals, anything with a different number (Voltage) react in some manner. The closer the metals are on this table, the less reaction they have. I guess its best to have a perfect join with nothing between to allow any reaction, i.e. anything that carries or contains oxygen etc. I’d just paint the joint to seal it up 😉
http://www.corrosionist.com/Corros1.gif
Hey Andy & Gizmo,
I totally agree. I’m not an electrical engineer and the EE classes I took years ago in college have probably done nothing more than make me ignorantly comfortable with high power electrical systems. However what’s never set well with me (as an engineer with very limited electrical knowledge) is wiring some little power resistor on a board with 20 AWG leads to a high power battery. I’m erring on the side of caution and following my gut on this one.
Yes assumptions are a big problem! Assumptions made by the designers of critical systems such as a BMS or battery charger are downright terrifying. I’ve observed first hand CALB SE cells in my pack that can reach voltages above 3.6V BEFORE the pack gets to the constant voltage phase. These types of behaviours are (maybe not so) well known and have been documented by John Hardy and others in these blogs and other places.
Not to plug for Jack and EVTV but if you’re new to conversions get you information and your parts from here! You won’t be sorry. I’ve seen and talked to many people who buy charges from suppliers configured to “factory recommended settings” (whatever that means) that grossly overcharge or just plain have the wrong type of charging setup. I’ve seen people who’ve been given bad information and damage their batteries before they even get them in their cars. It’s really heart-breaking to witness.
Jack and the whole EVTV community have done such a fantastic job of getting good information backed by unbiased data and observation out to the community. There is so much high quality information here on this site alone that the debate on many of these issues is over and settled. Jack does say this frequently. Do we just need a little Jack (or maybe a handpicked group of editors) to sift through it all and put together a compendium on all the relevant topics?
Thank you Jon.
Heard the latest? Jacks BMS overcomes all the issues from previous BMS’s. Communicates by laze’r. You switch it on and it just laze…. there…
.
How about this for a fun idea, all invented in my mind. Will save on lots of fuse wire for a black box controller (BMS).
A full pack of 200ah cells but one is 180ah — or the known weakest. Every cell goes in on centre charge, i.e. bottom balance plus half capacity of each cell.
This way the weakest cell hits low or charged full first. Others have 10ah safety each side.
The only monitored cell is this single 180AH which resides in an easily replaceable location along with the circuitry which tells the controller to go turtle if its dropping, the charger to knock off when this cell is full enough.
The rest are sealed in their sarcophagus.
Lets face it, if you cannot fit ‘n forget it. Forget it. 🙂
I believe the 180Ah cell would go into reversal the first time you bottomed out your pack with parasitic loads like we’ve seen on EVTV. The other cells would eat it since they are all in series.
Agreed, I have thought about how to transcribe all his videos so we could do computer searches on what we are looking for.
Hi Jack !
Do you know if the Cells that you didn’t receive is the same chemistry?
Best Regards
/Per Eklund
Some thoughts on the BMS issue. I discovered this site just a few days ago and am slowly mulling over Jack’s experiences as described here. Truly thought-provoking material on this website.
As a student I’ve designed a successful BMS for my university’s solar car project. It is a per-cell monitoring affair which has carried two separate vehicles several thousand real-world road miles without any incident. Of course this is more of a research project than something you can drive to work. Still, my perspective is it’s not *that hard* to do BMS right. Of course it is *very easy* to do it wrong. And in fact a verified BMS is a requirement for any applicant to either the American Solar Challenge or the World Solar Challenge (though for Pb just a multimeter is acceptable…).
Seems to me that ultimately the choice of No BMS/Basic BPS/Complex BMS depends on the endgoal. I think I get Jack’s perspective on leaving out the BMS for these EV conversions. If you know your pack well, and just charge to a conservative voltage before you hit the road, there won’t be any issues. It’s cheaper and avoids all those extra points of failure between cells.
But at the same time I feel the motto “as long as you understand your pack, don’t worry” is a pretty limiting message. At the least it isn’t a path which progresses the state-of-the-art. I don’t think people should be lead to believe that BMS is NEVER necessary, or worse, that advocates are somehow ridiculous. This is just as bad as the blind BMS advocates who lack real-world testing experience
Context context context.
What IF we want a pack which can be partially or fully charged off-grid? What if a vehicle has regen braking? How can you avoid overcharge while “in the field” without some BMS? Should I tell operators to flip this switch to turn off a solar array when the SOC is above X and turn it on again when it is below Y? To avoid using regen braking if SOC is above Z? We’ve had electronic systems to manage that kind of stuff for a long time. So the sleepy college freshman who forgot to check the meter in the last hour won’t start a fire because he’s been regen’ing hard down a steep incline on a full pack. And there are plenty of autonomous applications which must be “deploy and forget” to be useful at all.
Yeah, this is strictly an EV conversion site. But in a broader perspective, as players in the movement to wean ourselves of fossil fuels I think we should not shut down these key parts of the discussion.
Your condecension is only eclipsed by your ignorance of the topic Aaron. A classic case of don’t know and don’t know you don’t know.
We are talking about BMSs that have been routinely presented and sold by suchc as yourself as a solution to PROTECT battery packs, that have instead been the direct cause of multiple horrendous fires worldwide. You have not stumbled on a cell of do it yourself converters that havent’ heard of them. We’ve dealt with ReapSystems, Orion, Elithion and many of those purporting to lead the field – mostly and reiteratively to destruction. We’ve heard from literally HUNDREDS of victoms who have lost their battery packs to this idiocy. Many of our “converts” are in out of the storm after losing cells and looking for answers. We have some very specific techniques for the care and feeding of these cells to make them survive and it is hardly a “know your cells” type of situation.
And collectively we’ve heard it all. Regen overcharge indeed. Total myth. I can absolutely guarantee you have never personally observed any cell damage from regen overcharge anywhere, at anytime, in any place in this universe. But you consider it important enough to design a feature for it in your BMS and list it in this message. Youve HEARD about it and been TOLD about it by the old battery women that pass these mythis and misinformation around like candy. There has NEVER been a cell victim of regenerative braking overcharge.
Solar? How about a charge controller. Works just like a charger. When it reaches a certain voltage it kicks off. It needs no BMS to perform this simple act. We routinely interlock controllers to go into limp mode on a certain pack SOC, and indeed interlock to prevent operation during charging. That’s not a cell level BMS. It is simply not needed and will never be needed. 99% of the fires have been end of charge events with perfectly good chargers perfectly capable of terminating the charge at the proper point, SUBVERTED by a well intentioned BMS instead. The other 1% were BMS’s themselves bursting into flames.
The glaring clue is your assertion that designing a BMS is easy. Don’t know, and don’t know you don’t know. WE have experimented with EXTREMELY innocent looking devices that on further examination proved a fire hazard and a hazard to our pack. BMS’s are the second largest killer of cells in our shop, after me doing deliberate battery tests. Parasitic loads from devices in the tens of milliamperes have killed us.
If you actually lived with an electric car, rather than testing one in a university for “thousands of miles” you too would be aware that there is something wrong in mudville. I would say a serious percentage of our viewers came to us AFTER losing a pack or multiple cells to such well intentioned nonsense – looking for answers. We’ve got em.
But hope springs eternal. It is my intention to design a “non-incendiary” BMS. Unfortunately that will take time, money, effort, and thought, in kind of copious amounts. It is NOT easy. And its not trivial. The very concept of guarding relatively robust parts with relatively failure prone parts is absurd to start with. So we have our work cut out for us. But the persistence of the disinformation you purvey is so virulent, that we pretty much have to develop one in self defense to put the mark on the checklist and prevent you from burning our cars and our battery packs with well intentioned totally ignorant nonsense.
Welcome to EVTV.
Jack Rickard
As to mottos, ours is “let’s get liquored up, play with some high voltage, and go for a drive.” If you don’t get it, you can’t be a viewer here.
We have no such nonsense motto as you quote.
I suppose “don’t jack off my cells with stupid” could be a motto we might consider.
Jack Rickard
Jack–
I have to admit I had at least some intention to goad a reply. I don’t spend time typing out yes-man comments. However when I said that this site was thought-provoking, that was not sarcastic. I *am* a product of the BMS-as-religion world. This site offers evidence for a countering perspective and I am at least not dense enough to ignore it. In fact I appreciate the time you took to form a reply.
The example with regen was admittedly a bad one. No we’ve not had such an issue. These packs are constantly cycling while in operation at average C/3 in either direction. Charging while discharging. The only “issue” we’ve ever faced is the temperature rise from this cycling, and that’s after nearly 8 hours of constant driving. And yes the BMS has a temp cutoff. This is the only theoretical limit to range we’ve seen. We only do forced-air cooling, so far. We don’t use a cell balancer either. Although several other teams do, and they have not exploded either.
My uni has built 12 of these cars over the past 20 years or so. Every one of them had a per-module (module being collection of parallel cells) BMS connection and not a single one caught fire or exploded. You form your arguments on observed fact, as do I. Over its 23 year history I’m only aware of one serious fire for an ASC team (OSU). Most accidents are actually mechanical in nature. And that’s with college students who show up once per week. Not seasoned EV converters.
When I say it’s not that hard to do BMS it is because I didn’t know anything about any of these things before I started the project, I just followed basic iterative design priciples–test, build, test build. But the car was built around the pack. There are no black boxes ordered from unvetted third parties. Aparently most BMSes out there are crap, I wouldn’t know, we always build and verify our own to meet our own goals. No, I’m not hawking my design. It’s for our own application, and I wouldn’t dream of shoehorning it into a 400V pack.
All that being said I hear your argument. I’ve been trying to formulate an argument for cell-level monitoring myself other than the fact that it is mandated in this league. Groping in fact. I’ve read a conference paper on an 88 cell-level balancer that is a magnificent design. But nowhere in the paper to they talk about how much extra capacity it should be squeezing out compared to doing nothing. And the papers it cite regarding the need for balancing are either for Pb or don’t do any experimentation themselves.
Now even if we did pack-level monitoring I would still consider that a BMS. Something needs to disconnect the charger, whatever that is, and something needs to talk to the motor to cut power, and precharge any supercaps in the system, and whatever else needs to be managed/monitored.
As for the motto, well, perhaps it is not your motto then. But it’s something I’ve gleaned from others’ comments.
Well, that’s pretty much what we are discussing. What is the minimum intervention and information necessary that would be useful to someone driving or operating the car – not performing maintenance on it.
What we are coming to is some temperature monitoring, monitoring voltage of 3 or 4 pack segments and comparing them to each other, and coulomb counting.
Why do you think the charger needs anything? All of our chargers terminate quite reliably with no assistance from a BMS. I was thinking of a backup to that, to remove AC from the charger if it did fail to terminate, but we really can’t find an incidence of that happening that the charger wasn’t controlled by a BMS instead of its own logic.
Beyond that, we go to pretty instrumentation and smart phones to display it. How much current voltage amp hours kWh, etc.
Jack