CAN handling requirements for Master in a BMS

[methods]

So...

Sitting here in the morning with an empty stomach and a grouchy outlook....
(as opposed to viewing the problem with a wide angle lens)

It looks like I need to review the system architecture and estimate how much more it will cost to support funny business.... then eliminate said funny-business.

Funny business examples:

Contactors that run off System Voltage.
You see these a lot in Golf Carts and other systems which were originally 48V max and slowly grew to big for their britches. It results in non COTS components that, ... unless they are adopted... will have LEGACY SUPPLY ISSUES once the manufacturer stops requesting them from the supplier. (READ THAT AGAIN - THAT IS NOT WHAT WE ARE AFTER - COTS IS HOW WE ROLL)

You also see this in "all in one" solutions... like the Sevcon controller... where the controller over-reaches into other sub-systems and assumes it is the only important component in the system. No Isolated CAN, all peripherals run off 1/2 Batt Voltage referenced to Ground... It controls the contactor.... to play nice with controllers like this requires either duplication of componentry or severely encumbering system integration. We do not wish to duplicate components where possible and we certainly do not want to "build into" other sub-systems if we can help it.

Component-centric design is a major focal point...
What component does your design revolve around?
Whatever component it is... that component ends up over-reaching and ... doing things poorly... that a dedicated subsystem would do better.

In some designs I have reviewed multiple components end up being overly centric... in an attempt to avoid needing other subsystems. The Controller and BMS may reach too far... resulting in all sorts of sorrow including scaling issues.

Maintaining strict boundaries in a system architecture may increase costs and complexity a bit... but the payoff comes later.

Where we are going is standardizing these interfaces so offerings from multiple companies can play nice together.
The OPPOSITE of this is compacting a system further and further... until the motor is packed full of controller and the battery is forced to be near the controller... this does not scale

Standard communications (CAN is what it will be)
Standards of isolation (overly burdensome in a low voltage application but obvious in a high voltage)
Agreement on best practices for critical safety systems (this is where you focus super high quality standards on a very small part of the system so you can go willy-nilly in other parts of the system)

Ok... eggs on the table...

-methods

 

[methods]

Some food for thought

* The controller may think it is in charge of the contactor but ... eh... it really should not be. The Master from the BMS should be in charge of the contactor. Allowing the controller to own contactor responsibility causes a classic "lock in" situation.. where system dependencies on behaviors which should not exist develop. For this reason... any time I spec a component (like RoboteQ controller) I strongly urge people not to exercise the superfluous built in features. Pretend they are not there... and maintain COMPATIBILITY with other market offerings... and dont get locked in.

* It may be tempting to save money by packing the BMS into a monolithic design. Eliminating tons of connectors, wiring, isolators, .... and sometimes IPXX issues. Doing this is just a mop um bop um game as new problems crop up. Example: Running a monolithic BMS (to avoid duplication of slave devices and associated costs) results in needing to support harnessing of balance taps) HARNESSING OF BALANCE TAPS IS A FAIL IDEA. There are hundreds of lines... they are extremely sensitive to corrosion, they do not support distributed systems... they do not save money. They are... a dead end. You want a distributed bms where the slave module is LOW COST and BUILT INTO modules (you choose how large your module is ) where the only thing leaving a module is a positive wire, a negative wire, and a CAN pair. These modules are either inside of an over-arching sealed container (which also houses your master) or they are individually sealed... and the master is sealed elsewhere. In the example of a hot-swap pack like Zero does... that complicates things a bit... but I would simply add 2 more wires and stuff a contactor inside... or solve it mechanically with the connector would be even better.

* Interconnect count vs battery chunk size is absolutely critical. The larger you go with a battery chunk size (say 6S... or 14S... or 28S...) the harder it is for people to architect around your battery! Do you know how many times I have tried to utilize a Zero cell box and had to recommend a different product? The manufacturers of the form factor do offer a variant in 2P14S, but it is not in mass production so it is not really an option. As cell count drops, capacity goes up, according to target scale, to avoid unnecessary parallel connections (addressing interconnects...) so sweet spots are hit by reducing the number of cells in series and increasing the number of cells in parallel. An example to think through would be the Zero Motorcycle which uses 4pcs of 28S hooked in parallel.

What if we made that 4pcs of 7S in series???

Doing this would require a distributed BMS
Doing this would eliminate about 100 balance interconnects which will be the root cause of future failure
Doing this would result in being able to use non-rigid high power buss interconnects (rigid high current interconnects will be the cause of failure later at end of life after lots of shake shake)
Doing this would allow a battery to be shaped in a more flexible way... and it would NOT require an overall external case... because it would not be bussing balance interconnects
Doing this would make the cell boxes viable for other efforts... which is in the interest of the manufacturer (Farasis) and NOT in the interest of Zero (supporting competition)

Ah... competition... impeding others in the most subtle ways to get ahead in the race either intentionally or instinctualy.

It can also be Ego....
Some times someone has to grab the flag and run out in front of the crowd and set the new standard. This is honorable... but doing that is very risky... to succeed... long term... you need to have made decisions which run the full circuit of the race series... not just the turns of the first race in the series.

Not trying to bag here... not Monday morning quarterbacking...
I am trying to decide who to align my efforts with and who to work against.

My goal, as always, is to support the ENTIRE electric revolution as best I can and not just any one company that wants to dominate.... that is... unless they dominate in a dominant way... like Tesla. I could tow their rope 100% all day long... but then I know things about that company which are not public and it makes me respect them at a fundamental level. You can just call it lust for musk if you like.

ah... breakfast is over... Time to go to the whiteboard and work out the architecture based on the excellent single requirement we got from home skillet. That should be enough to severely impact the existing design which is up there now.

Remember: Dont fear IP, lawyers, and trolls.

Develop Requirements (even if I have to solicit them from you)
Architect to those requirements
Down select to the architecture based on Schedule, Budget, and availability (i.e. go with COTS thats cheap and in production already)

Nobody,,, and I mean NOBODY... owns the IDEA of building a power system. It is so fundamental as to be like... owning the idea of eating and shitting. Everybody needs to eat and shit. Nobody can eat for everyone and shit for everyone... therefor everyone has the right to build their own power system as they can and where they can.

This is not a time to tread lightly.

-methods

 

[methods]

Off to take a shower and take the boy to the Boardwalk (before the masses arrive...) but here is a closure:

* I already know how I would architect a power system and I have prior art on it going back to the mid 2000's

* I want to architect a system which will lend itself to a broad spectrum of applications... where it can still meet a motorcycles needs but where it could also meet a trains needs

* Money comes from only a few directions... think of it as UP or DOWN. If it comes from below... the idea is to OPEN new markets. If it comes from above the idea is to CAPTURE existing markets.

* This industry is in its absolute infancy... like the gold rush where spikes are in the sand... yea a few people have brought out the big equipment and started mining some of the richest ore... but that is just the start. There is infinite room for growth and the market will grow in a way nobody has even thought of - reference SMART PHONES... a trillion dollar industry.

* The sunk cost we have accumulated to date is a drip in the bucket. If you are worried about $100M dollars... your stuck. This is a poker game with a $100M buy in and infinite re-buy. If you are sitting on a 2-7 off suit and the flop is King King King... sticking with that hand has a very low probability of paying out. Throw those cards away and let the dealer take your ante. Place your bets on cards you know will win pots... like King Ace suited or a solid pair.

* We are so early in the game that nobody wants to poo-poo the progress we have made. Totally understandable... especially after the electric car was killed and we all were so shafted by the ICE car companies, non-rechargeable battery companies, oil companies... those are powerful entities who have had great success laying the shaft to use for a very long time. BUT... (but...) the pendulum has passed the point of no return and the future has been written. There is absolutely NO WAY that smart bets on Electric Power are going to bust. Its just impossible... and proof of that is in every major ICE company rolling out cars that are actually viable... and not just to meet minimum requirements.

* The big win is coming and nobody owns the race. This is not Nascar - its a free for all - like a race across the desert where you can run anything you brought.

off to triple S... shit, shave, and shower.

-methods

 

[methods]

Sunk Cost defined as: Significant spend into qualification and refinement of a system which does not scale epic. I.E. turd polish gone to waste

Every major speculator is looking for the mega jackpot. The 10X x 10X x 10X
To get there they know they are going to lose 19 out of 20 bets... but the big payout covers the losses
No big investor is looking for a steady 10% return... they can get that with the wall-street boys and crooked market tactics.

The big payout only comes from something which can dominate multiple sectors.
Dominating multiple sectors requires architecture which is compliant with others investments... which means leveraging economy of scale... which equates to C.O.T.S.
Any attempt to put a harness on a horse that does not fit... will get bucked

I suspect you have no idea what the frock I am talking about... but there are about 30 people who do.... and that's enough for me.
I am not a strange babble'er... I am beating around the bush... and getting straight to the punch... at the same time.... in a form of prose which I enjoy.

Now all I need is a big cardboard sign, a milk crate, a megaphone, a bunch of silver jewelry, and I am set to go downtown. :?

-methods

 

[marcos]

Most of us don't care about the prose. We're here because the technical discussion. Speaking of which, how do you plan to keep power consumption low? I hear many can bus nodes, how are you going to wake them up from sleep? Anything above 500uA consumption is a disaster.

 

[methods]

Most of us don't care about the prose. We're here because the technical discussion. Speaking of which, how do you plan to keep power consumption low? I hear many can bus nodes, how are you going to wake them up from sleep? Anything above 500uA consumption is a disaster.

Another excellent question and the issue I was beating my head against a few months ago.
The answer is *dont care*

As stated my slave devices are running on isoSPI and they have a low power consideration.
My master and associated peripherals are running off of 12V aux... and any consumption is 3 or 6 orders of magnitude lower than system average so it is hands down dont care.

Thats the primary purpose of having an AUX or boot-strap battery... as with all large scale designs.

When the system is off... the contactor is open... and nothing is happening or should happen. System is shut down and stowed away.

Datalog... some want it while dormant. Short of sounding a isolation failure alarm.. I dont see the point... but if that were required I would run it off of a dedicated onboard low power DC-DC and wake only once an hour for ms.

-methods

 

[methods]

A word about datalogging while dormant:

* While in the system configuration datalog can be "paid for" by system AUX battery to the tune of microwatts over time. The Master can wake for a millisecond once an hour and do this for years.

* While isolated from the system some have claimed that it is a requirement to have a datalog. I have yet to see a valid argument for this.



Lets say there is a fire... datalog destroyed, no help

Lets say there is electrolysis... or something that triggers some sort of isolation fault... ok - who will it be reported to? Will sounding a buzzer help in a container between China and the US? Would it even be heard while in a cardboard box on a shelf with a thousand others 10 feet from your desk? An LED would surely be no help.

I would store batteries on a shelf with a CAN tether that constantly monitored them... that tether could supply external power... but nobody is doing it... so no requirement.

When a battery is not in a system configuration where is it? Out on the front lawn getting hit with sprinklers? Arguably... if it is outdoors we dont really care if it goes up... to some degree. IF it is in doors... and it was wet when it was brought in... and it is a non IP67 battery... well... that was a stupid move by the consumer yes? Dont know what logging that would accomplish other than maybe giving one a timeline... because it either catches fire or it does not. If it does not... the error is found on next attempt to system integrate. If it does catch fire datalog is lost.

Great cost for a 1% chance of a return that could possibly be useful.

I see that in the future... batteries will be stored in a MONITORED CONDITION... meaning hooked up to some kind of backbone.
When they are being shipped or once they are out in service... eh... best to avoid the complexity is my take on it.
From production to consumer they will presumably be in a dry and good condition.
Once in the hands of the consumer there is a pamphlet that says "This is how to care for your battery"

If they short it out with a fork, drop it in the ocean, run it over with a tractor, shoot it with a rifle, or otherwise do something damaging to it... (like drive it on a misty morning lol...)... well... that falls into the category of abuse and miss-use... which datalogging would not really help with warranty or anything else.

The greatest concern is often warranty... and any time the contactor is closed the datalog is going... unless they dig in and bypass.

-methods

 

[velias]

There are a lot of medications besides lithium for treating mania.

 

[methods]

Says the guy with 5 posts who works phone support for Microsoft

Try developing products from scratch and taking them to market with no team, design review, etc.
Until you do... save your wise ass comments. As stated this is where I do my brainstorming.

I am not here to spoon feed you interesting tidbits... that's for the re-hashers of information to do.

If you already do develop products... its probably on a team.... and when you are ready to step up and do it alone... you will understand.
Must be nice collecting a salary... I remember what its like.

Now shush.

-methods

 

[methods]

And to spoon feed you more free ideas...

Right now we are debating putting Precharge on the master or making it remote on a private CAN bus.

Removing PreCharge from the Master removes *all high voltage* from the Master board... and this is OBVIOUSLY the direction one wants to go... especially when considering Manufacturing Test and the complexities of EOL testing on a BMS Master (which obviously I am an expert at since I produce EOL testers for Tesla).

So... it costs more per unit to break up PreCharge from Master due to the added CAN drivers, Isolators, co-processor, connectors, wiring, etc.
A little savings is found in the complexity of test - especially in lower volume - not at all in higher volume...
But a lot more savings can be found by a small company that CAN SELL MODULES compatible with all other systems.

Want to read current? CAN module
Want to PreCharge? CAN module
Want to close a contactor CAN module
Want to read isolated voltages? CAN module

If we have a shelf, and on that shelf are lego's... and everybody uses the same legos... what happens?
Economy of scale
Intercompatibilty of parts
Lower cost for EVERYONE
More parts we can use in the junk yard
Easier to diagnose, repair, swap parts
Generic parts
A range of high quality to low quality parts
Easy duplication... high and low range current sensors...

The idea of building monolithic is penny wise and pound foolish.

So... the goal is to drive the cost of the CAN down to $0

Twisted Pair
Driver Pair
Isolation Pair
Co-Processor Pair

We want to budget $20 to split any part of the system off into CAN.
Thats full produced cost... end to end... idea to cardboard box
Then $10... Then $5...

Until it costs nothing.

Thats half the argument.
Of course the CAN buss has to be completely private.
It can be relatively slow
Some of it can be non-isolated
Power draw is being dropped off the table - done with low power designs - only where needed... which is inside of a Battery module or brick - wherever the slave is installed. Other than that... its 12V power and its on.

-methods

 

[methods]

And in closing...

Instead of being a wise as... think about this for a moment...


Why are the junkyards full of highly complex and useless subsystems?
Economics of course.
Capitalism
Competition
Greed

What if the junk yards were full of no more than 10,000 unique pieces.
What if instead of trying to save $5/unit by compacting 10lbs of shit into a 5lb bag... Manufacturers actually tried to do whats best for the world and the people... and they worked together to be compatible and modular?

Serious.

This is the kind of dialog that goes on in a National Laboratory (but cant happen because we are busy building Nukes)

Do you know how easy that is?
If all agree it COSTS LESS

Yes... it leaves the market WIDE OPEN for competitors to build lower cost generic versions.
That is reality - nobody gets rich. Sorry.

Until profiteers let go of the idea that they have the privilege to get rich at the cost of waste and others becoming poor we will make no progress as a species.

Real simple shit.

Anyone wants to argue that point then I win. Your only available arguments are logical fallacies.

Two guys...
One thread...
Skimming free material...
Bitching about it.

Becoming done with this forum.

-methods

 

[marcos]

Pro-tip. Its called CAN BUS, not sure where you got the buss thing.

 

[methods]

Lol..

BUS and BUSS

Google says BUSS is a kiss... so there you go brother... I am blowing you a kiss.

Notice how I switch back and forth between BUS and BUSS totally randomly.
That's a signature of my writing (near total disregard for the rules of English).

It does not matter if you are after the technical content.
Here is some. Enjoy.

https://endless-sphere.com/forums/viewtopic.php?f=10&t=87548&start=50#p1315156

Now please direct me to some technical content of interest so that I can learn something.

-methods

 

[methods]

AH,.... here is your last post.
Good technical content.

https://endless-sphere.com/forums/viewtopic.php?f=30&t=90149&p=1314781#p1314781

Thanks marcos

 

[marcos]

Ya, sorry. Its one of my issues, I cant finish reading an article if the units or terms are not right, even if english is not my native language.

I already designed and produced can bus cell balancers, I can't give you much details but I try to make yourself ask some important questions. Like the nodes isolation, power consumption, how to identify each node in a pack, 120 ohm impedance controlled wiring harness cost, how do you tolerate a wiring fault. I can sell you this ready to run, but I think you want a diy approach. Also I dont promise keeping up here, as I said long non technical reading bores me.

 

[methods]

see how we catch fish?

One could quickly determine if Marcos is the real deal.
I bet he is

...

When you are looking for help it is near impossible to find.
Poking around in a forum like this can often result in access to super specialized charters.

...

So who is currently employing marcos, how much does he make, do we need him?


-methods