Some "don'ts" for EV building

Technical discussion on converting internal combustion to electric
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

Over the course of our MX-5 build, we've come across some aspects of our design, or specific things that we did, that we later came to realise was a mistake. I'd been thinking of starting a topic titled "How not to build an EV", specific to our MX-5, but it can't be only us that has made mistakes that others could learn from. Also, when people find lessons learned by others, say on other forums, they could put links to those lessons here.

I'm hoping that this topic could be useful to skim through for those starting an EV conversion for the first time, and perhaps even more experienced builders could learn a tip or two. To make it easier to get to the important parts (the lessons learned, if you like), I suggest these simple rules:

1) One post per lesson. If you have several lessons that arise, even if from one design decision or one thing you did, split them into separate posts/messages.

2) Each lesson should have a bold heading, like this text. That makes it easier to find the lessons and skip the inevitable chatter that will arise from these lessons. I don't think it's wise to discourage discussion, it just seems inevitable. Also, I suspect that some of the discussion will end up being as important, if not more so, than the original lesson. The heading should be a short, single sentence that tells the reader what this lesson is about, so if they are about to build a battery box, they can easily skip any lessons about motors, controllers, and so on.

3) End the post with a one or two sentence summary, starting with "Summary:".

4) When quoting a lesson with bold text, please remove the bolding (remove the [ b ] and [ / b ] characters). This should help to make the real lessons stand out better. It's not that hard, you just need to look out for this when quoting.

Bold text is easy to insert: just follow this template, without the extra spaces:

[ b ] M y   h e a d i n g [ / b ]

When the extra spaces are left out, you see this:

My heading

If any extra instructions or suggestions are required, I'll edit this first post. Let's hear those instructive lessons that you learned the hard way!

[ Edit: Added "Summary:" item. ]
Last edited by coulomb on Wed, 22 May 2013, 08:28, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

With regen, don't test with a fully charged pack.

Our MX-5 has ten (!) battery boxes, some of them with only seven cells in them. When we had our first two battery boxes built, we wanted to see the motor running, and even drive the car a little (off-road, just driveways and the like). We didn't want to damage these cells, of course, so we limited the motor controller drive current to 3C (120 A in our case).

This was great; we got to test the drive coupling, shook some rust off the brake discs, it was fun, and it gave our enthusiasm a needed boost.

Unfortunately, now that we have more facilities in place for checking battery performance, we are finding that these two battery boxes, which happen to be right in the middle of the car so they are the hardest to get to, contain at least eight cells that are noticeably higher internal resistance than the others. It looks like we've damaged them to some extent, and of course their damage will reduce the capacity and performance of the whole pack. We had been top balancing these cells, so they would all have been at or very near 100% SOC. We suspect that regeneration has overvoltaged some or all of these cells.

When the car is finished (soon now! Image ), regeneration will cause the motor controller to back off with regeneration current if any cell becomes stressed with over-voltage. In fact, we discovered this while testing this very feature.

So what we should have done with either
1) Disable regeneration, at least until the pack had blown off say 5% SOC. Or
2) Waited until the battery management system (BMS) was fully in control of the motor controller.

In hindsight, option 2) is way less practical than 1); one gets impatient.

Edit: in retrospect, it seems that our 16 weren't killed by regen, but by uncontrolled charging in a solar storage system. See this post. So maybe this isn't such a big "don't". Not a grand way to start this topic!   Image

[ Edit: smallest battery boxes have 7 cells, not 6 ]
Last edited by coulomb on Tue, 21 May 2013, 16:12, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

Don't use highly artificial tests

As a follow-up to the above, we found it somewhat difficult to figure out how to test the interaction between the BMS and the motor controller.

In the end, we found this to work best. We artificially changed the voltage to stress mapping for all cells to be way more restrictive than normal. For example, instead of starring to stress at 2.5 VPC (we're still not sure how low we'll tolerate in the end), we set it to 3.1 VPC. We also "increased the slope" of the "curve" (actually a straight line), so that small sags would provoke high stress, and back off the motor controller. We similarly changed the over-voltage stress to start kicking in at 3.40 VPC, rather than 3.50 VPC.

This allowed us to actually drive the vehicle, under close to real-world conditions. We turned on data logging (a handy feature of our Tritium WS200 controller), and studied the logs after a short test drive. (Stay away from roads, of course, since you don't have as much power to get out of trouble.)

This worked out much better than our other idea of "virtual cells", which worked on real motor current, but artificially produced stress, variable with a potentiometer. We could switch between real and virtual stress with a toggle switch. The trouble with this was that the virtual cells responded to a DC bus current that was lagged by 200 ms, whereas the real cells responded instantly. So when we ended up with jerky performance, and we could not be sure if this was because of our highly artificial testing scheme, or something that we really needed to fix.

Summary: where possible, make tests as real as possible; highly artificial testing may introduce enough uncertainty that the test has little value.
Last edited by coulomb on Wed, 22 May 2013, 08:34, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

Don't use MTA connectors in EVs

Well, at least the tin plated versions. I'm talking about this type of Mass Terminated Assembly:

Image

The idea was to use these to join strings of battery management units; within each string, our "squiggle joins" made the connection automatically. They plug onto cheap, standard 2.54 mm spacing headers, like these:

Image

The attraction is that you just "push" thin wires into the tops of these; there is no need for crimping or soldering or even stripping the insulation off the wires.

We found that they were a reliability nightmare. (Plus, the tool to help insert the wires was ridiculously expensive.) Eventually, we got rid of them all, and soldered (!) wires from one header to the other. We used liberal silicone sealant to strain relief the wires. (That's why I used the exclamation point after the word "soldered" above; usually, soldering wires is a no-no in EVs. Strain relieving, and the very short (25 mm) runs hopefully eliminates this objection).

As I post this, I note that one of the wires that is connected with these connectors is a very high impedance circuit. (This will be the subject of another post.) It may be that these connectors would be tolerable when the circuits aren't so high impedance, but it seems unlikely.

Summmary: this type of connector is probably unsuitable for EVs.

[ Edit: added explanation and image of headers ]
Last edited by coulomb on Wed, 22 May 2013, 08:35, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Some "don'ts" for EV building

Post by BigMouse »

Great idea! This thread should grow in to a very useful reference.
User avatar
acmotor
Senior Member
Posts: 3614
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Some "don'ts" for EV building

Post by acmotor »

Don't use an old bomb to do an EV conversion.

The donor vehicle should represent a reasonable part of the EV build cost so that it doesn't require more work than the conversion itself.
The level of safety and comfort and utility for your intended use should be as appropriate as your budget can manage.
Don't convert a vehicle that you would not normally be prepared to drive.
Converting a bad ICE vehicle to electric results in a bad EV.

Conversions of classics or vintage are often sucessful, however the cost of restoration will be a major factor. These conversions may make good show and shines but no good daily commutes.

So summary:
A free vehcile is the wrong one for an EV conversion !

converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
User avatar
acmotor
Senior Member
Posts: 3614
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Some "don'ts" for EV building

Post by acmotor »

Don't buy the batteries until you have a BMS, mounting, charging and vehicle weights sorted out.

Get a sample, make one module etc.
Time is ticking with the cells and technology is moving on.

Often EV builders start with an emotor or battery or charger even just a contactor and dream on from there building the EV around that item. It is essential to back off to the full picture.

Summary:
Don't buy batteries too soon.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
User avatar
acmotor
Senior Member
Posts: 3614
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Some "don'ts" for EV building

Post by acmotor »

On your first test drive, go around the block 10 times rather than drive 10km off into the distance.

The walk home is shorter.

Summary: none required.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Some "don'ts" for EV building

Post by jonescg »

When selecting a motor, don't be fooled by 'peak' power figures

Peak power ratings of motors are typically determined by highly specific, never-to-be-replicated-again circumstances. And the 'peak' lasts 4 seconds or so. This will typically be at a fairly inefficient part of the motor's range, so lots of heat will be shed, and heat is the natural enemy of electric motors. If the manufacturer quotes a continuous power rating, this is the power level which it has been demonstrated to output all day long. Use this figure as your starting point - anything after this is a bonus.

Summary: Use continuous power ratings to design your build.
AEVA National President, WA branch director.
User avatar
acmotor
Senior Member
Posts: 3614
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Some "don'ts" for EV building

Post by acmotor »

coulomb wrote: With regen, don't test with a fully charged pack.


Regen will always return less to a battery than was taken out to move the vehicle in the first place up to any speed capable of producing regen.... unless your driveway was steep enough and you could go down at a reasonable speed 30+kmph then decelerate as the first movement of the vehicle. You assure us that the test was not on the freeway.
Google street view shows a slope but not enough to get up any speed.Image

A resting 40Ah cell takes some 2Wh to get from 3.3V to 4V in my experience. In a 200 cell pack that is 400Wh. A considerable amount of regen.
Yes, if internal resistance is high on one cell then volatge can rise momentarily but assuming your BMS module didn't fry then it would not have been a damaging excursion.
But, if the internal resistance was that high in the first place then the cell probably already had a problem so regen would show up the symptom but not be the cause.

I think you may be blaming regen for some shelf tired cells ?

Many people have found a few % failure in LiFePO4 cells (as with nearly all chemistries).

Were you able to measure internal resistances before using the cells ?

But yes to your summary. Get the BMS completed and limit controller before testing.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Some "don'ts" for EV building

Post by jonescg »

Make sure you build easy maintenance / access into your vehicle

With Voltron I, I managed to shoe-horn 6 kWh worth of battery into the bike, but there was no way I could access the terminals or even inspect the motor shaft without removing 90% of the bike. For example, a solid 5 hours work was needed to swap a drive sprocket - a fairly common thing to do on a bike.

I have also worked on a few EV car conversions where things like battery sub-packs were tucked away into difficult spaces. If a cell goes bad and it needs replacing - will it be a matter of popping the hood and swapping it out; or will it be a 12 hour expedition on a hoist? If you can bolt your whole battery pack up as one unit, it will simply be a matter of undoing 8 or 10 bolts, disconnecting a few cables and lowering it down for inspection.

Sometimes you just can't avoid it, so try to do everything you can to make it easier on yourself - run balance leads, fused if possible, up to a convenient spot so you can keep an eye on those hard to reach cells. Leave a spot in the engine bay so you can reach down to inspect the brushes without removing 7 hazardous voltage cables and the like.

Summary: Build in features to make maintenance easy and convenient.
AEVA National President, WA branch director.
User avatar
acmotor
Senior Member
Posts: 3614
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Some "don'ts" for EV building

Post by acmotor »

jonescg wrote: When selecting a motor, don't be fooled by 'peak' power figures
...
Summary: Use continuous power ratings to design your build.


I know what you are getting at, can I put my thoughts...

Continuous power should be suitable to cruise at target speed with a headwind and 4 FBs on board.
16kW at 100kmph in 1.5t vehicle with average Cd

10 minute power 2x continuous (Greenmount hill)

1 minute power 3x continuous (passing lane, Bindoon hill)

10 second power ... As much as possible without doing damage Image
(very monitored and controlled by the electronics)
Problem is that you will use the power... quite often.

A motor controller must be mindful of the thermal budget resulting from these and repeated loadings.

The question of low efficiency at peak powers is a major factor in DC systems and due to back EMF, they struggle to achieve peak powers at higher road speeds. No wonder AC rules !

Yes, don't design on just a peak power and expect it all the time.
converted RedSuzi, the first industrial AC induction motor conversion
on to iMiEV MY12 did 114,463km
now Tesla Model 3, 4/2021 MIC pearl white
zeva
Senior Member
Posts: 443
Joined: Sat, 15 Dec 2007, 05:09
Real Name: Ian Hooper
Location: Australia
Contact:

Some "don'ts" for EV building

Post by zeva »

Don't be overambitious

Or accept that your project will end up taking longer, costing more, and with a higher chance of failures.

Simple designs are easier to build, maintain, and are usually more reliable. Unless you enjoy extensive troubleshooting, it is best to use well-proven, off the shelf components from reputable suppliers rather than making your own, or being the guinea pig for new/unknown brands. Size your battery pack just to cover your daily driving needs rather than going for the biggest pack you can squeeze in the car, as it will be cheaper, easier to install, and easier to get through licensing.

I learned these things the hard way Image

Summary: KISS
Last edited by zeva on Sat, 11 May 2013, 10:26, edited 1 time in total.
Ian Hooper
--
"Never doubt that the work of a small group of thoughtful, committed citizens can change the world." - Margaret Mead
http://www.zeva.com.au
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Some "don'ts" for EV building

Post by Renard »

Record wiring details as you go

Seems obvious, but more than once I've had to pull things apart because I forgot sometimes to write down what wire was going where.
Even if you complete a module in one blow, you may have to go back to it for revision.
Renard
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

zeva wrote: Don't be overambitious

Or accept that your project will end up taking longer, costing more, and with a higher chance of failures.

Amen to that, Ian. We have that in spades with our MX-5 conversion. At the time, conversions were often pretty similar: 144 V of sealed lead acid (lithiums were just becoming available), Advanced DC motor, and one of a handful of controllers. There were no affordable AC options.

Now, however, you can get "off-the-shelf AC" (e.g. AC50 or similar, with a Curtis 1238 controller, or for more power a Tritium WS-200 and say an SEW-Eurodrive motor. We decided to innovate in several areas:

1) A high voltage AC conversion using (at the time) an industrial controller
2) Our own BMS; just this sub-project cost us about 3 years
3) Cramming as many lithium cells as possible into a small car; we originally aimed for 228 cells in 8 battery boxes (now 208 cells in 10 boxes).
4) Using the highest continuous power motor money could buy in a relatively small (132 frame) motor.

All of these decisions have cost a lot of time and/or money. For the money Weber has spent, he could now buy a brand new Nissan Leaf and a giant charge point.

We learned a lot, and should hopefully have a car that has a bit more "fun factor", power and range than a Leaf, but at the cost of well over 4 years (mostly at about 1 day per week each, plus many many hours of help from two other helpers, and several other helpers.

We initially thought we could knock these over in about 6 months, perhaps a little more for the first one, making a profit on each one to keep the business running. How far our ambitions exceeded our capabilities.

This lead on nicely to the big lesson learned from the MX-5 (my next post).
Last edited by coulomb on Sat, 11 May 2013, 12:45, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
Simon
Senior Member
Posts: 433
Joined: Sun, 19 Aug 2007, 19:38
Real Name: Simon
Location: Perth WA
Contact:

Some "don'ts" for EV building

Post by Simon »

Don't go overboard on weight reduction taking out the sound deadening stuff


This stuff is smelly and messy and actually reduces road noise. Who knew? Image Hehe
User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

Higher voltage has its limits

Our original goal with the MX-5 was to use an industrial controller, designed for 415 VAC input and output, to over-voltage an industrial AC motor. Such controllers can operate up to around 750 VDC or higher, and we figured we'd want to operate the controller close to the upper voltage limit of the controller so we could get as high an output voltage as possible, to get the most performance out of an off-the-shelf AC motor. So our design goal was 228 cells (730 V nominal using 3.2 VPC nominal, though typical resting voltage (3.3 VPC) would be more like 750 V).

The basic idea is that the higher the voltage, the lower the current for the same power, and so the thinner the cables need to be. Also, with lower currents, a small amount of resistance (say from a tarnished lug) has a lesser effect the lower the current. Hence, higher voltage is better, in that sense.

We knew that 228 cells was a lot, and that 228 individual cell-top BMUs would be a lot of work. But we failed to realise *how much* work it was.

The other thing is that things like fuses start getting really expensive above about 600 VDC. Cable starts getting hard to find, and no commercial BMS (that we are aware of) handles this high a voltage. While it is true that each BMU only sees a total of about 4.0 VDC, it has to be able to cope with the possibility of a cell going open circuit under load, in which case the full pack voltage (minus one cell) appears across the BMU connected to the open circuit cell. We've documented what happens with just 240 VDC, when the BMU isn't designed to handle this.

However, the main thing is the sheer amount of time and effort it takes to make ten custom battery boxes, and all the contactor boxes to go with them. Even finding space for all the contactor boxes, and routing all the cables, is a challenge.

So how high is too high for an EV's voltage? Well, it depends on many factors, many of which vary from case to case. I do think that 600 VDC (maximum ever seen, at the end of charging) is a good upper limit. But that's still a lot of cells, some 160. The current Tritium Wavesculptor 200 controller has a voltage limit of 450 VDC; this is still quite high. I'm thinking about 360 VDC (which we have now, in our 109 cell "half packs"), is pretty reasonable. But as I say, this is really a matter of personal comfort.

For a DC conversion, it really depends on what your motor is capable of. I suppose you could argue that you don't need to allow full pack voltage to the motor (e.g. you could have a 320 V pack and only allow say 180 V maximum to the motor), but I don't like the idea of that. It would be too easy for full pack voltage to end up across the motor, in my humble opinion.

Summary: more is better as far as EV voltage is concerned, but there are limits, and over 600 V is probably not a good idea for most designs. Something around 320-360 V (100-108 cells) is a reasonable figure for AC conversions.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Some "don'ts" for EV building

Post by jonescg »

I agree - 600 V is a great upper limit. However we just need to tell all the inverter manufacturers to listen to our requests. Missing out on up to 30% of the available performance by not going all the way up to 720 V is quite unfortunate, but that's the safety compromise.

I think the 'mid' voltages are a good compromise - up to 360 V top of charge, provided your motor, controller and battery all fall into the sweet spot. That's the delicate balance.
AEVA National President, WA branch director.
User avatar
Johny
Senior Member
Posts: 3749
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Some "don'ts" for EV building

Post by Johny »

Renard wrote: Record wiring details as you go
Seems obvious, but more than once I've had to pull things apart because I forgot sometimes to write down what wire was going where.
Even if you complete a module in one blow, you may have to go back to it for revision.
Agreed - I've had to post-document wiring as well.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Some "don'ts" for EV building

Post by Renard »

coulomb wrote: Higher voltage has its limits

Something around 320-360 V (100-108 cells) is a reasonable figure for AC conversions.


Another small advantage of this voltage is that off-the-shelf DC/DC converters designed to accept 240V AC can be used without modification. An RMS of 240V AC implies a wave peak of 339V. And allowing for a 10% variation, that means a DC bus input of about 370V DC is OK.
Renard
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by weber »

acmotor wrote:
coulomb wrote: With regen, don't test with a fully charged pack.

Regen will always return less to a battery than was taken out to move the vehicle in the first place up to any speed capable of producing regen.... unless your driveway was steep enough and you could go down at a reasonable speed 30+kmph then decelerate as the first movement of the vehicle. You assure us that the test was not on the freeway.
Google street view shows a slope but not enough to get up any speed.Image

A resting 40Ah cell takes some 2Wh to get from 3.3V to 4V in my experience. In a 200 cell pack that is 400Wh. A considerable amount of regen.
I totally agree with the above, and I assure you no significant slopes were involved.
Yes, if internal resistance is high on one cell then volatge can rise momentarily but assuming your BMS module didn't fry then it would not have been a damaging excursion.
But, if the internal resistance was that high in the first place then the cell probably already had a problem so regen would show up the symptom but not be the cause.
You're probably right here too. Is it possible that ordinary LiFePO4 internal resistance, when faced with 3C of regen, might cause momentary rise to damaging voltage? e.g. 40 A cells resting at 3.34 V with 3 mR times 120 A will only go to 3.70 V. So no, I guess not.
I think you may be blaming regen for some shelf tired cells ?
It's definitely not that. All our 228 cells have had equal shelf life, except for 16 that were bought a little later than the others. What we found is that of the 109 cells currently driving the vehicle, the 16 with the highest internal resistance (as measured by voltage sag under load) are all in the only two boxes (54 cells) that we used in earlier testing before they had BMS protection (it was the BMS we were testing). One had failed completely and was bulging despite being clamped, and had evidence of having ejected some electrolyte out of its vent.
Were you able to measure internal resistances before using the cells ?
We have the individual 1 kHz measurements provided by the manufacturer, but failed to record cell serial numbers corresponding to BMU sequence numbers as we assembled boxes.

So now it is seeming more likely that it was overcharging with the 5 amp plug-in charger that did the damage. That was so long ago, and records and memories so poor, that we'll probably never know.

In any case, it's worth emphasising that, unlike Lead-acid cells, Lithiums are happiest and safest when they are 50% charged (although 30% to 70% is probably much the same).
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
unheardofinstruments
Groupie
Posts: 86
Joined: Tue, 26 Jun 2012, 15:08
Real Name: ron Berry
Location: Nimbin
Contact:

Some "don'ts" for EV building

Post by unheardofinstruments »

re. choosing a good donor, I concur. I wish I had been more careful choosing donor/host/roller vehicles, a surprising amount of time and money was used up chasing parts and doing repairs to suspension parts, brakes, wheels, rust, paint, windscreen, etc. and although the price was almost free it wasn't worth it when I could have waited for the right one to come along and gotten on with electric vehicle related tinkering and expenses instead. I could have spent more and maybe sold more useful for ICE only bits and recouped the difference and saved on parts too probably had I not gotten a partially stripped roller thinking it would save me some stripping out what I didn't need. As with building the cheapest way is to scavenge/gather your materials first, then see what you can make. For a cheap simple build unfortunately old cars are easier in some ways as power steering, locks, windows and computers etc aren't so common but with age comes difficulty finding parts and general decay.
An old rule of thumb of projects; take what you think it will cost and triple it, and as far as time time you think it will take, triple it again. ('Then see if you are much closer' I would add) [:s]
One can only plug on and enjoy the ride all the more for the blood sweat and tears and swearing.

re batteries one, how long is a detrimental non used shelf life for LiFePO4?

User avatar
coulomb
Site Admin
Posts: 6357
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by coulomb »

Don't use circuits with high input impedance

Half a megohm of input impedance in any circuit near the motor or controller is likely to receive induced voltages of the order of a tenth of a volt. If this controls the motor power, it can be a significant problem.

Details moved here: Weber and Coulomb's MX-5".

Summary: we failed to consider how sensitive to noise our circuit would become as a result of high resistor values. Perhaps also, if no-one else implements an apparently useful feature, maybe there's a good reason for it.

[ Edit: moved to the MX-5 build thread, to keep detailed technical discussion away from here. ]
Last edited by coulomb on Wed, 22 May 2013, 08:33, edited 1 time in total.
MG ZS EV 2021 April 2021. Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
Patching PIP-4048/5048 inverter-chargers.
If you appreciate my work, you can buy me a coffee.
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by weber »

Some new information has just come to light regarding this question:
weber wrote:
acmotor wrote:I think you may be blaming regen for some shelf tired cells ?
It's definitely not that. ... What we found is that of the 109 cells currently driving the vehicle, the 16 with the highest internal resistance (as measured by voltage sag under load) are all in the only two boxes (54 cells) that we used in earlier testing before they had BMS protection (it was the BMS we were testing). One had failed completely and was bulging despite being clamped, and had evidence of having ejected some electrolyte out of its vent.

Note the number 16 above. I just came across this old post of mine while looking for something else.
viewtopic.php?title=real-world-lifepo4- ... 080#p37060
I had completely forgotten that I had connected 16 of the 40 Ah LiFePO4 cells to my 1.5 kW battery-backed-up grid-feed inverter for a month or so in the last half of 2010 or first half of 2011.
a younger weber wrote:You'd have to use 16 LiFePO4 cells in series so they would get balanced once a month (using the sealed cell setting). ...

It would be so easy to have ... 15 cells around 3.55 volts and one at 4.75 V. Then you have a fire risk from that cell venting its highly flammable electrolyte and possibly being ignited by a failing BMS board (unless it's one of ours). ...

I did try it for a month or so (on a SunProfi-based system) and it seemed OK. I just diverted some cells destined for the EV. But you won't really know for some years, and it's an expensive experiment.
Prescient words? From a fool?

The magic number 16 also appears in this old post from Coulomb.
viewtopic.php?title=renards-bmw&p=39097&t=2921#p39097
a younger coulomb wrote:Some of our cells are actually swelling, just sitting loose on shelves. ... [ Edit: not sure if it was our 16 newer cells from EV Works or not. ]
In email at the time, I told Coulomb it was unlikely to be the 16 newer cells from EV Works as these had already been put into the two boxes destined for either side of the MX-5's diff.

So I think acmotor is right, regen is not to blame. Most likely it is the 16 cells that were used in the PV system for a month or so, where they could be charged at up to 20 amps and there was no ability for the BMS to turn off or throttle back the charge source, and the BMUs could only bypass 1 amp. They were clamped as two sets of 8 at the time, so bulging would not have been apparent.

We have 10 spares, so I guess I should replace the 10 with the highest internal resistance (one has already been done) and move the other 6 to a battery box where they are readily accessible such as the bonnet box or the rollbar box.

So perhaps Coulomb's original Regen "don't" should be annotated accordingly. It was me who mistakenly suggested to Coulomb that uncontrolled regen was the problem. The real "don't" that comes out of this is ... <my next post in this thread>

[Edit: Formatting / spelling / minor clarification]
Last edited by weber on Tue, 21 May 2013, 12:37, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Some "don'ts" for EV building

Post by weber »

Don't ever leave Lithium cells charging without a BMS that can monitor individual cell voltages and automatically disconnect the charging source, or cause it to reduce its charge current to less than the BMS can bypass, when any cell goes above the manufacturer's specified maximum charge voltage.

At the very least you will permanently increase the cell's internal resistance and reduce its capacity. At worst it will vent its highly flammable electrolyte and catch fire.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Post Reply