Weber and Coulomb's MX-5

Post up a thread for your EV. Progress pics, description and assorted alliteration
User avatar
Johny
Senior Member
Posts: 3689
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Weber and Coulomb's MX-5

Post by Johny » Tue, 21 Apr 2015, 15:32

On the other hand I've been top balancing for two years now (daily driver) and have no problems other than faulty chargers.
Weber, I'm not fond of FETs in noisy environments. I've had way more problems with FETs than bipolars in low and medium power applications over the years. I also can't imagine what you would be getting high(ish) voltage issues on a cell-top mounted system like you have.
I agree with your earlier post - go with the highest voltage device (bypass FET) you can fit on the board.

Edit: Added "bypass FET".
Last edited by Johny on Tue, 21 Apr 2015, 06:27, edited 1 time in total.

User avatar
Adverse Effects
Senior Member
Posts: 1018
Joined: Sat, 01 Jan 2011, 03:30
Real Name: Adverse Effects
Location: Brisbane

Weber and Coulomb's MX-5

Post by Adverse Effects » Tue, 21 Apr 2015, 16:20

weber wrote:See the problem yet? Image


sure do Image Image Image

your fixing it right now Image Image Image

if you want X voltage just add 1 or 2 cells to cover it

as you know on the upper and lower states of charge the the voltage changes rapidly with very little Wh added to the pack or removed from the pack so if you just dont go there your voltage WONT change rapidly

the top and bottom 0.2V of usable voltage may only be 5Mt of travel in your car simply just dont go to that part of the charge cycle

User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Tue, 21 Apr 2015, 17:19

The issue is it's all fine if all the cells are doing as they should and for the most part they do but with over 200 cells I'm sure there will be one that doesn't and the chances of having a accumulative discrepancy always grow with time.

But what happens if one cells isn't doing as it should. I think you will find that that Jack R guy actually agrees on the need for High low voltage cell monitoring. Not so much on the need for active balancing though. I don't think he has come to the point of wanting cell level Hi/low detection but at least to the point of wanting to splitting his pack up into small enough series cell segments to monitor that you can detect a discrepancy befor its to late.

I just finished reading a post from a guy with a offrid system with 16 x 400ah blue calb se cells That were initially balanced (to within 10mv) and are charged to conservative levels and cycled much the same. He did a balance charge after 6 months and it took 35ah to bring one low cell up to the others Image Now who knows it could be his cell monitor thats doing that could be a bad cell.

I have one cell that needs about 0.5ah pr 4 weeks or 6ah pr year to keep it in check with the others about 1.5% pr year.

Kurt
Last edited by offgridQLD on Tue, 21 Apr 2015, 10:07, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Tue, 21 Apr 2015, 18:40

But Kurt, you have EV Power Aust. CMUs on your cells don't you? So the Rickardists can blame those, and they may be right. Did the guy with the 9% mismatch have permanent cell-level monitoring of any kind?

But not even shouting can make this item of Rickardist dogma true:
"NOTHING A CAN DRAW FROM ANY CELL OR CELLS UNLESS ITS THE TOTAL PACK NOT JUST SOME"

The fact is that every cell has a very small internal self-discharge current, and unless these are exactly the same for every cell, they will eventually cause imbalance.

But I'm not even considering that, when I try to explain the impossibility of safely charging a 200+ series-cell bottom-balanced LiFePO4 pack without cell-level monitoring. I'm just assuming, as even Jack Rickard mentions, that they don't all have exactly the same capacity.

LiFePO4 has such an extraordinarily flat charge curve between about 80% and 95% SoC that you need to take all cells to, at the very least, 3.4 volts (assuming a C/8h charge rate) just to ensure that you've charged the lowest cell to at least 80%. If you have one cell that has a capacity just 5% lower than the average capacity of all the cells, and you charge a 200S pack to 200 * 3.4 = 680 volts then you will have 199 cells averaging something like 3.395 volts while that one cell is at 100+% being cumulatively damaged at around 4.395 volts on every charge.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

T1 Terry
Senior Member
Posts: 684
Joined: Thu, 30 Sep 2010, 20:11
Real Name: Terry Covill
Location: Mannum SA

Weber and Coulomb's MX-5

Post by T1 Terry » Tue, 21 Apr 2015, 19:21

The other option is to remove the cell balance section and simply monitor the cell voltage. stop charging when any cell reaches 3.6v and recommence charging when all cells are 3.4v or less. The same leads that are used to collect the individual cell voltages can also be used to trick charge the lowest cells to maintain that end voltage of 2v min per cell and forget about top balancing. this is only suitable for EV's of course, house battery packs are not generally drained to 0% SOC, so charging to 3.6v and settling to 3.4v, then resume charging remains, then top up the lowest cells till the average of 3.4v per cell rested is met.

T1 Terry
Green but want to learn

User avatar
zeva
Senior Member
Posts: 421
Joined: Sat, 15 Dec 2007, 05:09
Real Name: Ian Hooper
Location: Australia
MSN: sigmunky@hotmail.com
Contact:

Weber and Coulomb's MX-5

Post by zeva » Tue, 21 Apr 2015, 19:27

I’m with you on this, weber. Cell capacities will never match exactly so when charging a bottom balanced pack there’ll always be one cell running away to 4.2V+ while the others are still around 3.4V, and with a large number of cells it’s not possible for the charger to know the difference between this vs a 0.01V increase on all cells. Top balancing plus BMS is the way to go for large packs IMHO.

Edit: As for the option of going without balancing.. I initially did this with my RX7 - manually top balanced, then just used voltage monitors for protection. But this is where differences in self-discharge become problematic, as over time the pack balance drifts and you end up getting one cell "running away" to 4.2V while the others are still at 3.4V. The BMS will of course protect the cells from damage by stopping the charger, but you start losing capacity because the charger gets forcefully stopped before completing its CV stage.

You also lose redundancy - in an ideal world, the charger should stop itself at an appropriate maximum voltage, and the BMS should only have to intervene if something goes wrong.
Last edited by zeva on Tue, 21 Apr 2015, 09:37, 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

User avatar
Johny
Senior Member
Posts: 3689
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Weber and Coulomb's MX-5

Post by Johny » Tue, 21 Apr 2015, 19:43

I agree Zeva. I get the feeling that Jack thinks that everyone should be out fiddling with the EV at every opportunity. I want my car to be useful to family if I cark it - so i's got to have hands-off system. That means top balance and as much auto-magic correction of cell charge as possible. Don't involve the user.
Like Mexy, I have 192 cell pairs spread all over the car so manual tinkering is not on.

User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Tue, 21 Apr 2015, 19:57

"But Kurt, you have EV Power Aust. CMUs on your cells don't you? So the Rickardists can blame those, and they may be right. Did the guy with the 9% mismatch have permanent cell-level monitoring of any kind?"

He is using the identical EV power CMUs as I have Image

I told him to check the cell interconnects though as (He is that same guy who didn't use terminal past on the aluminium connections) My thinking was perhaps he has issues there. His response was his thinking was he would detect heat when charging at high amps if he had a resistance issue and there is nothing.....I didn't take that conversation any further.


I will see how I go next balance in 4 weeks. I am trying to take note of the data each time to try and understand it more. If I get the same result next time on the same cell (it's looking that way as last time it was the same low cell to but I didn't record the mah needed to bring it to the same level as the others) I guess I could try swapping that CMU around and see if the trend follows it to another cell.

Hey if we all didn't have little niggling issues we would all be bored and looking for a new problem to create Image

Kurt

User avatar
Adverse Effects
Senior Member
Posts: 1018
Joined: Sat, 01 Jan 2011, 03:30
Real Name: Adverse Effects
Location: Brisbane

Weber and Coulomb's MX-5

Post by Adverse Effects » Wed, 22 Apr 2015, 00:41

weber wrote:
But not even shouting can make this item of Rickard ist dogma true:
"NOTHING A CAN DRAW FROM ANY CELL OR CELLS UNLESS ITS THE TOTAL PACK NOT JUST SOME"

not meant as shouting but highlighting and couldn't be bothered stuffing around with font and color

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Wed, 22 Apr 2015, 01:54

offgridQLD wrote:I told him to check the cell interconnects though as (He is that same guy who didn't use terminal past on the aluminium connections) My thinking was perhaps he has issues there.
How do you imagine that bad cell connections could cause charge imbalance? Adverse Effect's highlighted statement is true apart from internal self-discharge and discharge by CMUs. In a series circuit the current, and hence the charge, is the same in all components. Only something connected in parallel with a cell (like its internal leakage resistance or a CMU) can bypass charge around it. A high-R joint is in series with the cells.
I guess I could try swapping that CMU around and see if the trend follows it to another cell.
Good idea. Or just measure its current and compare to its neighbors, as we did with those CMUs in that battery box out of MeXy.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Wed, 22 Apr 2015, 02:39

"How do you imagine that bad cell connections could cause charge imbalance"

I was thinking more if the strap (or in his case copper flat bar) had a bad contact with the terminal post then the CMU would also have a bad contact.

I guess this could effect the measured voltage at some of the CMUs perhaps not shunting current when they should be (reading a lower voltage than the true voltage)

Though thinking it through it's unlikely the contact would be so bad a voltage reading would be off. He had one low cell so more likely to be a drain on the cell from its CMUs or a issue with the cell its self.

Kurt




User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Wed, 22 Apr 2015, 03:04

I just dug out the brochure for the EV power CMUs and its listed as having 2.05mA consumption at 3.3v. with a picture of a fluke 175 measuring it.

Kurt

Last edited by offgridQLD on Tue, 21 Apr 2015, 17:05, edited 1 time in total.

User avatar
jonescg
Senior Member
Posts: 2290
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Weber and Coulomb's MX-5

Post by jonescg » Wed, 22 Apr 2015, 06:10

I have 168 cells in Voltron, and these are lithium cobalt oxide cells, so 3.00 V to 4.20 V. I find when nothing is connected across them, their individual voltages will never drop. This is because I have weeded out the sick and the lame, and no self-discharging cell will find a place in my battery pack!

We did find that a group of 11 cells were being slightly overcharged (by slightly I mean 4.30 V) for short periods before a race. I couldn't work out why, but during the off-season I did a balance charge of the pack and found several below-par solder joints between the cell top and the modules. In this case, the most negative cell connection was poor, meaning the module didn't read the values correctly (hence not alerting us to it earlier) but more relevant to the symptoms, didn't drain power at the same rate as every other module. So whenever it would be sitting idle for weeks, every other cell was drained by a few hundred millivolts except this group of 11. Then we'd charge them all up, and they would go over 4.20 V. I've balance charged the pack twice in one year, and the first time didn't really need it.

So yeah, moral to the story is to sort the cells first and then make sure your monitoring circuits have good, solid connections!
AEVA National Secretary, WA branch vice-chair

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Wed, 22 Apr 2015, 18:40

Chris, two minor points of disagreement.

1. All cells have self-discharge, although it may be less than 1% per month for high quality LiPos, and with the voltage vs SoC curve being so flat it will be difficult to detect over less than a year.
2. 4.30 volts is not a small overcharge. It will almost certainly have caused lithium-plating on the anode. What matters is how long it was at 4.30 volts, or rather how many amp-hours it took at 4.30 volts, compared to its capacity.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Wed, 22 Apr 2015, 18:59

He is using lithium cobalt or similar that have a nominal voltage of 3.7v and a fully charged voltage of 4.2? The imiev charges them to 4.1v .

Or do you think the 4.3v would still be a issue. It's not the best still but I didn't think it was to bad on his cells.

Kurt
Last edited by offgridQLD on Wed, 22 Apr 2015, 09:00, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Wed, 22 Apr 2015, 19:26

offgridQLD wrote: He is using lithium cobalt or similar that have a nominal voltage of 3.7v and a fully charged voltage of 4.2
Yes. I'm aware of that. I have closely followed the design, build and career of Voltron Evo.
Or do you think the 4.3v would still be a issue. It's not the best still but I didn't think it was to bad on his cells.
Yes. 4.3 V is bad for all Lithium cells that have carbon anodes. LiFePO4 just has a lot more headroom between fully-charged and plating lithium.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jonescg
Senior Member
Posts: 2290
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Weber and Coulomb's MX-5

Post by jonescg » Wed, 22 Apr 2015, 19:28

True - they will self discharge very slowly - I have measured cells from 18 months ago and they read 3.83 V. At the time I was sorting them I was culling cells which were below this value, with most being 3.88 V.

I've heard of RC guys charging their LiPos up to 4.5 V per cell and hammering them as hard as they can for a competition. They manage to get about a dozen cycles at this rate before things go puffy Image

They spent about 20 minutes at this SOC, right before a race where we'd pull peaks of 20C but only about 7 C for the most part. Certainly not ideal, but not the end of the world.
AEVA National Secretary, WA branch vice-chair

Pocim
Noobie
Posts: 1
Joined: Fri, 12 Jun 2015, 10:39
Real Name: Pierre-Olivier Cimon
Location: Canada

Weber and Coulomb's MX-5

Post by Pocim » Sun, 14 Jun 2015, 10:54

Hi,

I have an issue with using the CAN module and thought maybe someone here could help me.

When powering the charger I was expecting CAN communication even though the battery pack and the BMS are not connected. Instead, I observe a concave descending ramp. (amplitude is ~100mV so it might be electrical noise)

Is my assumption correct i.e. should the charger send data when powered ? Do I need to initialize the communication via the BMS sending an initialisation message ?

Thanks for you help !

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 14 Jun 2015, 13:36

Hi Pocim. Assuming you're talking about an Elcon or TC charger, then yes it should sent a packet on being powered up. We rely on that in the MX-5.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 18 Oct 2015, 00:13

In case anyone following this thread missed this other thread, here's a link to a Mexy story Coulomb posted back in July. It probably should have been in this thread, but Coulomb couldn't resist an alliterative title.

And in case anyone is wondering, Mexy has just been being driven, rather than worked on. The power steering is still "Armstrong", and I still haven't fixed the clutch that slips near full torque, because I've been too busy doing other things, and she's quite fun to drive even so.

In fact she's so much fun to drive that the nice people at the Queensland Department of Transport and Main Roads recently sent me some nice photos of Mexy, along with some cleverly worded invitations for me to make donations to improve the safety of our roads. Followed by a choice between a 3 month suspension or 12 months good behavior. I chose the 12 months good behavior.

I got 6 of these photos within a week. 3 of them were taken at the same place at different times. There were no tell-tale flashes or obvious vans. Hey, I would have slowed down after they sent me the first photo, but for some reason it took them 3 weeks to send it to me.

Image

So now I have a "sat-nag" that displays the current speed limit and makes a rude noise whenever I go over it.

I won't deny I've been a bad boy, but the new sat-nav has shown me that part of the problem was that I thought my speedo was over-reading by 8%, but it is actually spot-on. Or more-specifically, it only over-reads at highway speeds and is close to correct at 50 km/h. So it's non-linear. It's the old cable-driven spinney-magnet type. I had previously only checked it (against a GPS) at 100 km/h and assumed it was linear. So there's a trap other converters might avoid.

Incidentally, I figure to calibrate a speedo against a GPS requires driving flat and straight at a constant speed for at least 500 meters, with good satellite signals.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
offgridQLD
Senior Member
Posts: 1820
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Sun, 18 Oct 2015, 00:52

Six in one week....Yikes!

Each time you got stung. Was it a just over the limit a little bit without knowing. Or was it very spirited driving?

Is it a new fixed camera. Something to look out for?

Kurt



Last edited by offgridQLD on Sat, 17 Oct 2015, 13:53, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 18 Oct 2015, 01:16

offgridQLD wrote:Each time you got stung. Was it a just over the limit a little bit without knowing. Or was it very spirited driving?
It was a mix of both.
Is it a new fixed camera. Something to look out for?

No. It seems like there was a blitz with a mobile hidden camera in my area.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jonescg
Senior Member
Posts: 2290
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Weber and Coulomb's MX-5

Post by jonescg » Sun, 18 Oct 2015, 02:29

Now I'm wondering if Patrick has a pile of speeding tickets from me on his E-max hooting though camera-enforced intersections at 6 km/h over the limit. Daily... Image

EVs are easy to speed in, I'll give you that.
AEVA National Secretary, WA branch vice-chair

User avatar
weber
Site Admin
Posts: 2456
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sat, 27 Feb 2016, 05:52

Mexy has been going fine since my last post 4 months ago, although I still haven't got around to fixing the clutch, that slips before we reach max torque.

But today she had to limp home to Bardon from The Gap. Bladecar and I set off to go to Mt Nebo, but we only got to The Gap before I felt power being limited. I looked at the fuel gauge -- 75% SoC, looked at the cell stress meter -- zero stress, looked at the temperature gauge -- Woah! Off the scale on the high side. So I pulled over and popped the bonnet. There was plenty of coolant in the reservoir (for water-cooling the WaveSculptor motor controller), and the reservoir was cool. I felt the WaveSculptor heatsink -- ouch!.

So we had to crawl home, with frequent stops to cool down, particularly when going up the hills. I surmise that the water pump has failed, although I haven't actually had a chance to check yet whether perhaps the power to it has merely become disconnected somehow, since I had work to do on one of Bladecar's Blade Electrons, installing a 3 position switch: charger A -- both chargers -- charger B.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Sat, 27 Feb 2016, 07:14

Well is it worth considering using an (AIRCON) system to cool the wave sculpture down to say 5 above zero celcius???

Like a heatpump should.

Then what would the maximum Current rating be with 5kw heatpump.

Say 50kw to motor 55kw from battery???

And you could heat the car in winter !!!!!

Oh your not in melb or tassie or Canada or north China...

Oh what is a Heat Pump.????

So much to teach...

What is energy ... How do you make (it) move up-hill??? (An anology used in high school physics)

Coulomb ... Weber do you guys know Joule ... Or lithium Bromide crystals entropy envelope ???


Why don't you build a Velelfire or ALPHARD or Lexus HYBRID or .... OTHER SIMILAR HYBRID
see :
http://www.hybridcars.com/toyotas-estim ... ts-for-us/

Into a super EV ... PLUGIN-HYBRID

I just wanted a ZEV T series with a 15kw motor and I will be happy.

Enjoy 2016

(edited to clarify my mind Image )
Last edited by 7circle on Sun, 28 Feb 2016, 11:38, edited 1 time in total.

Post Reply