Luke's 2001 NB MX-5

Post up a thread for your EV. Progress pics, description and assorted alliteration
User avatar
OzSpider
Groupie
Posts: 47
Joined: Mon, 14 Dec 2020, 10:02
Real Name: Derek
Location: Brisbane, Australia

Re: Luke's 2001 NB MX-5

Post by OzSpider »

Good to see progress. Could you add some photos of the cell to PCB arrangement when you get to that stage?. The 3D model shots are good but I'm not quite imagining it yet.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

OzSpider wrote: Sun, 21 Mar 2021, 16:23 Good to see progress. Could you add some photos of the cell to PCB arrangement when you get to that stage?. The 3D model shots are good but I'm not quite imagining it yet.
Sure thing, I'm kinda there now. I still need to get some heat sinks cut and fold them up but I did assemble my first "test" pack last week. While it's only a single pack it shows how each side will be connected (I hope). Basically four cells form a pack and one side is connected electrically via a the pictured busbar (right) and the other side is attached to the (left) pack only via the PCB for strength. The positive and negative tabs are at opposite ends of the cells.
Pack End.jpeg
Pack End.jpeg (728.9 KiB) Viewed 767 times
Pack End Busbar.jpeg
Pack End Busbar.jpeg (755.23 KiB) Viewed 767 times

The tabs on the cells are a little long so they need a little bit of adjustment to get just right. And given that I'll have 128 in total I most certainly didn't want to do it by eye and hand. So I took one of my spare busbars and made up a jig. With the help of a 4mm punch just to make a divot to later line up a 4.5mm punch and die in my drill press. The jig also allows me to make a straight line where I will cut each tab with tin snips.
Jig.jpeg
Jig.jpeg (1.16 MiB) Viewed 767 times
Jig Result.jpeg
Jig Result.jpeg (869.43 KiB) Viewed 767 times
Press.jpeg
Press.jpeg (576.13 KiB) Viewed 767 times

It's still a little bit of manual work but it ensures all of the measurements and markings are identical.
francisco.shi
Senior Member
Posts: 599
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Luke's 2001 NB MX-5

Post by francisco.shi »

I assume you are punching the holes.
I helped the QUT race team to do their cells. What we did was made a plastic block with a metal insert and a cavity to locate the tab. Then use a drill bit with the back end ground flat on the drill press and punched the holes. It was very accurate neat and quick.
How are you going to clamp the cells? Just be aware that when the cells are clamped they compress a little and the distance between the tabs of adjacent cells will change.
Over 2 or 3 cells is not a big problem but over say 10 or 20 cells you can get a chance of over 10mm.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

The sides of each battery box will actually act as a clamp for the cells.

There is a little give in the spacing between the PCB's and busbars. So my plan is to have the screws lose when I tighten the box ends together and then tighten the busbars.
reecho
Senior Member
Posts: 847
Joined: Sun, 17 Apr 2011, 02:39
Real Name: Richard
Location: Perth WA
Contact:

Re: Luke's 2001 NB MX-5

Post by reecho »

I think @chrisjones has a elegant solution for tab punching... :-)
francisco.shi
Senior Member
Posts: 599
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Luke's 2001 NB MX-5

Post by francisco.shi »

You need to clamp the cells before you put the bus bars. The cells are 11.5mm when compressed. I tried to find info on how much force to use when compressing and the best I found was about 10psi minimum (from some university paper) which equates to about 250kg. What I ended up doing instead was applying enough force until the cells compressed to the 11.5mm specified on the datasheet. When you stack the cells they are considerably more than 11.5mm so you will have to clamp them first then install the bus bars. You also have to hold the cells with the tabs aligned while you clamp them so that the tabs line up. I made a c channel out of plastic and used that to hold the tabs inline while clamping.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

So I've been rather slack in updating thread but plenty has been happening since I last posted.

Here's a quick recap:
  • I've prepared all 133 battery cells to be made into packs (cutting and punching all tabs).
  • My TC 6.6kw charger and 1.8kw DC/DC converter have arrived.
  • My Zero-EV 3kw PTC heater core has arrived.
  • I've finished my custom 90mm2 battery box terminals.
  • Most parts for my battery boxes have arrived (alloy chill plates, delrin sides which act as clamps and composite lids, sides and chill plate caps).
  • I've folded enough chill plates to make the main (largest) battery box.
I've taken photos of progress and will post some up today and get onto the rest in short order...

Some other things have happened also:
  • We sold both of our ICE vehicles (including my ~5.0 second VW Polo Gti :cry: ) and bought a Tesla Model 3. So we're 100% renewable energy with no ongoing emissions. Not to mention transport fuel is now free 8-)
  • I had a Zappi 22kw charger installed and made up a fancy pass through for the charge cable (https://forums.aeva.asn.au/viewtopic.php?f=16&t=7157).
I think that about sums it up, photos to follow...
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

Not long after I prepared all of the batteries my heatsink sheets were finished. I had these waterjet cut as opposed to laser cut as the cost difference was almost 50%, in retrospect I think I'll opt for a "very fine" cut instead of the standard because each sheet required a little bit of cleaning up as they had burs on them. Nothing major but I guess that's where the extra 50% (~$200 goes).
Box of Sheets.jpeg
Box of Sheets.jpeg (444.82 KiB) Viewed 386 times

This was for 35 sheets so a few hours of cleaning up with a fine toothed file. Even though the burs needed removing I most likely would have filed the edges round on laser cut sheets too in order to prevent any punctures of the cells when assembling 🤷‍♂️so a good cost saving I guess.

After successfully making a jig to punch holes in the cell tabs I thought I'd do the same to mark the folds in the aluminium heat sinks. But this turned out to be more effort to get perfect marks than simply marking with a ruler so I only used it on one sheet...
Jig.jpeg
Jig.jpeg (459.98 KiB) Viewed 386 times

These turned out pretty good with a rather cheap manual folding press from eBay. Far cheaper than having the sheets laser cut and then folded by the same workshop which I was quoted almost $500 for. Plus, it's nice to have the tools to use for other things like folding busbars once I get to that.
Heatsink 1.jpeg
Heatsink 1.jpeg (418.52 KiB) Viewed 386 times
Heatsink 2.jpeg
Heatsink 2.jpeg (394.96 KiB) Viewed 386 times
User avatar
jonescg
Senior Member
Posts: 3897
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Luke's 2001 NB MX-5

Post by jonescg »

Definitely cover the inside of the heatsink with Kapton before potting the cells. It's just too easy for a pouch cell to fail up against a sheet of conductive aluminium.
AEVA National Secretary, WA branch chair.
User avatar
Cowardlyduck
Noobie
Posts: 13
Joined: Wed, 15 Apr 2015, 23:02
Location: ACT

Re: Luke's 2001 NB MX-5

Post by Cowardlyduck »

jonescg wrote: Sun, 11 Jul 2021, 17:58 Definitely cover the inside of the heatsink with Kapton before potting the cells. It's just too easy for a pouch cell to fail up against a sheet of conductive aluminium.
The 2 cell modules with these in them from the Chevy Bolt did not have any protection...I'm using 14 of them in my power wall...if it wasn't such a huge hassle to do it now, I would also add some kapton tape for piece of mind even though those aluminium housings were moulded to fit these cells exactly.
DSC_6014 (Small).JPG
DSC_6014 (Small).JPG (260.3 KiB) Viewed 371 times
Also, looks like you've already got the cell connections sorted, but if you're interested an alternative is 3/4" solid aluminium block cut into short lengths. I did it this way because the cell tabs on my cells were really short after they were cut out of Chevy Bolt packs.
DSC_6104 (Small).JPG
DSC_6104 (Small).JPG (212.43 KiB) Viewed 371 times
DSC_6139 (Small).JPG
DSC_6139 (Small).JPG (249.55 KiB) Viewed 371 times
The 3D printed housing for them can be found here:
https://www.thingiverse.com/thing:4437151

The other thing to note is the Chevy Bolts using these cells were recalled. Still not sure what the root cause was for that, but I did have a cell venting under load when I first built my powerwall. I suspect the seams around the edges of the cells may let go under certain conditions so best to take it easy on them.

Cheers
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

jonescg wrote: Sun, 11 Jul 2021, 17:58 Definitely cover the inside of the heatsink with Kapton before potting the cells. It's just too easy for a pouch cell to fail up against a sheet of conductive aluminium.
I've got 60m x 4cm of Polyimide tape on order. Actually I had thought to originally wrap each cell but it looks like that's not common practice? Well not for each cell pack that is. Taping the inside of the heat sinks would be far less time consuming that's for sure but I am still unsure if wrapping individual cells is worth it or not?

Cowardlyduck thanks, but I'm running 45mm x 3mm copper plates between each cell pack. So enough to carry plenty of current. Interestingly though solid aluminium blocks were considered in one of my many many design iterations.
User avatar
jonescg
Senior Member
Posts: 3897
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Luke's 2001 NB MX-5

Post by jonescg »

I usually conclude that the corrosion potential for aluminium is not worth the risk. Half the thickness in copper for the same current, plus copper doesn't expand as much when hot.
As for the polyimide tape, its more about hand assembles packs not having the same tolerance as robot assembled stuff in a Korean factory, so some precautions are worth the effort.
AEVA National Secretary, WA branch chair.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

These arrived not long after I finished the tabs on my battery cells. I went back and forth on design several times and somehow ended up sending a slightly outdated design to the machinist. Luckily though the only real difference was the diameter of the battery terminal holes (14.5mm vs 15mm for the charger terminal and 19.5mm vs 20mm for the power terminals) and the tapper of some of the larger holes needed to be a little deeper to fit the countersunk screws I have. Nothing a little work on a drill press can't fix.
Battery Box Ends 1.jpeg
Battery Box Ends 1.jpeg (586.96 KiB) Viewed 330 times


The battery boxes are designed to have copper busbars with a minimum 95mm2 throughout the system. This includes the connecting cables. Maybe a little overkill as all of the EV suppliers supply 70mm2 but I do plan on some serious abuse and this does leave the door open to a future controller that can handle more current (greater than 750A peak). The issue with this size is finding connectors that can handle that current and that don't cost an arm and a leg. This proved to be pretty fruitless so I decided to go custom, cheaper and more powerful.

The battery terminals are round on the outer edge which sits inside the batter box walls and square on the inside. The idea here is to prevent them being pulled through. On the inside of the box the terminals are located inside a 3d printed nylon brace/bracket with securing screws. I had a local engineering shop cut these terminals down from a 25.2mm round bar to 20mm and 15mm terminals. The brace and box ends had zero tolerance and the terminals were made by hand so once I received them they did require a fair bit of work work with a fine file to get them to fit inside the brace just right.

I also did the drilling/tapping myself as I'm trying to do as much of this work as I can. They turned out "ok" but not perfect. Still completely functional but 🤷‍♂️. Once I've tried fitted everything including the battery cells I'll have these along with all of the busbars tinned to prevent corrosion. And yes, that's a broken tap... Let's chock that up to learning the hard way.
Terminals.jpeg
Terminals.jpeg (496.71 KiB) Viewed 330 times


And some shots fitted into the brace and battery box sides.
Terminal Brace 1.jpeg
Terminal Brace 1.jpeg (671.75 KiB) Viewed 330 times
The dividers didn't fully take into account the tapper on the holes so I drilled them out a wee bit. And once fitted it became apparent that the securing screws for the bracket would make great conductors between the positive and negative terminals if the lugs get too close 🤦🏻‍♂️ This is what happens when you don't model screws in your design. So, I have to get these reprinted.
Terminal Brace 2.jpeg
Terminal Brace 2.jpeg (754.3 KiB) Viewed 330 times
Terminal Brace 3.jpeg
Terminal Brace 3.jpeg (861.22 KiB) Viewed 330 times
Attachments
Terminals Inserted.jpeg
Terminals Inserted.jpeg (776.92 KiB) Viewed 330 times
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

Here's the updated terminal brace design. I have received the updated prints of these and will post these up once I've assembled a battery box.
Updated Terminal Brace.jpg
Updated Terminal Brace.jpg (142.17 KiB) Viewed 330 times
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

If anyone else has broken a few taps this video (https://www.youtube.com/watch?v=GCMjCCDKOes) might be of some help. It's uber dodgy filming that looks like it's right out of the 80's but the content is sound.

The TL;DR is drill and tap your holes using a drill press if you have one. Just do the tapping by hand instead of turning the drill on (i.e. turn it by hand). Sooo simple.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

The latest things to arrive are my chill plates. Instead of following a traditional channel design (which these kinda of are) I ran scores of flow and heat simulations to settle on the most efficient design.

The first cooling design I simulated was based on the Chevy Bolt design I came across in this awesome teardown (https://www.youtube.com/watch?v=ssU2mjiNi_Q&t=495s). It was a little unconventional but featured a twin inlet and single outlet design. Though my simulations didn't prove very fruitful but I am pretty sure this is due to the curved edges on the Y axis which aren't as easy to model/machine. They would lend themselves very well to being pressed to fabricate them very very quickly. Which is why I think they use this design, win win and cheap.
Chevy Bolt without Dividers.jpg
Chevy Bolt without Dividers.jpg (496.86 KiB) Viewed 303 times
Chevy Bolt with Dividers.jpg
Chevy Bolt with Dividers.jpg (494.1 KiB) Viewed 303 times

I then focused my designs on a traditional channel design much like the one used for my controller heatsink. These are much easier to model and cheaper to machine. What I found was it flowed reasonably well but did end up with some hot spots (areas where flow is basically zero) around the bends.
Earlier Iteration.jpg
Earlier Iteration.jpg (507.29 KiB) Viewed 303 times
Later Iteration.jpg
Later Iteration.jpg (494.47 KiB) Viewed 303 times
Final Design.jpg
Final Design.jpg (576.6 KiB) Viewed 303 times
So the solution in the end was a very simple snaking channel design with small dividers after each bend. It does increase the velocity somewhat on one side but ensures that there is also velocity on the opposite which wasn't present in the earlier iterations.

The tool I used to simulate this is SimScale (https://simscale.com).
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

And of course the shiny results...
Chill Plates.jpeg
Chill Plates.jpeg (836.48 KiB) Viewed 303 times
The dividers are a little hard to see but if you zoom in you can make them out.
Chill Plate.jpeg
Chill Plate.jpeg (755.36 KiB) Viewed 303 times


One frustrating thing with these though, the machinist didn't tap the screw holes all the way through (only ~4.5mm of 10-16mm depending on the hole) and I've already bought my screws. So with near on 60 holes for each of the 3 chill plates I am becoming quite adept at tapping :lol:
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

Progress has been really good this last week or so. I finally finished re-tapping the holes in my chill plates, my polyimide Koptan tape arrived (Kapton clone) and I set to wrapping the battery cell heat sinks I folded up a while back. This is only for the first and largest battery (12 cell packs vs 10 in the other two batteries).

The first thing I will say on this is I am REALLY glad I bought a couple of 3M squeegees instead of trying to apply the tape by hand. The polyimide tape is rather difficult to apply perfectly without and folds or bubbles. But with the squeegee and some practise it's really not too hard. That said, it still took a good 15 minutes per heatsink and I've only made 13 of 32 so...
Heatsink 1.jpeg
Heatsink 1.jpeg (704.38 KiB) Viewed 252 times
Heatsink 2.jpeg
Heatsink 2.jpeg (698.37 KiB) Viewed 252 times


And fast forward a little bit, most of the cell packs for the first battery.
Cell Assembly.jpeg
Cell Assembly.jpeg (967.15 KiB) Viewed 252 times

I wanted to trial fit all of the packs into the battery enclosure to ensure that my math was correct and I'd be able to fit them under compression (the battery enclosure is essentially a big clamp after all).
Battery Pack Trial 1.jpeg
Battery Pack Trial 1.jpeg (928.8 KiB) Viewed 252 times

It will be tight, but there should be just enough give to fit the last pack. I was able to compress the packs another ~10-12mm by hand and they need about 20mm. I can always pad them with another 2mm aluminium plate to give it that little bit more (which was what I had in the CAD drawing anyway). Unfortunately I don't have the steel rods for the enclosure clamp (yet another lockdown here in Melbourne) and the 900mm clamps I've ordered still haven't arrived. So I'll have to wait a little bit longer for a complete fit. I've got brackets to fabricate anyway.
Battery Pack Trial 2.jpeg
Battery Pack Trial 2.jpeg (783.21 KiB) Viewed 252 times

One last thing I'm still undecided on is wether to use thermally conductive epoxy between the top/bottom of the cells and the heat sink. It means once they're in they're in, which would be a pain if a cell dies but it will likely yield some minimal gains in cooling perhaps? There really isn't too much of a gap but no gap is better than air for heat transfer I suppose. I'll probably have to clamp each cell pack individually (in my vice perhaps) as the epoxy dries to ensure correct compressed size.
Cell Gap.jpeg
Cell Gap.jpeg (544.47 KiB) Viewed 252 times
User avatar
jonescg
Senior Member
Posts: 3897
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Luke's 2001 NB MX-5

Post by jonescg »

Looking great! And those chill plates are a work of art. Question - how do they seal? In my experience the best option is to have a really positive seam or groove to push an o-ring into, or at least a decent bead of waterproof silicone / polyurethane sealant. Without that a slow leak can still find its way through. The resulting humidity can lead to problems down the track.

I still think some kind of encapsulant between the cells and the aluminium fins is a good idea, if nothing else, just to keep them snug and unable to fret or rub. But air is a great insulator and the cells will benefit from uniform heat dispersion, rather than hot spots. At present the two outside cells will have some decent heat transfer while the inner cells will be warmer. If there was a thermal path between the sides of the cells and the aluminium plate it would result in a far more uniform heat spread.

As far as serviceability goes, if you encapsulate each cell group, and sit them on top of the silicone pad, you would only be up for a four-cell replacement job. That's not bad.
AEVA National Secretary, WA branch chair.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

It really felt like progress event trial assembling these cells and battery, unlike the monotonous work tapping and taping things...

In regards to sealing I'm using a 5mm thick polycarbonate sheet as a cover for the chill plate cooling channels that clamps down using the 40 odd 4mm screws into a 2.4mm nitrile cork gasket of the same shape (i.e. rectangular). This is the same sort of gasket material you'd use for things automotive water pumps and other coolant gaskets. And being rectangular with no cut outs for the channels means there is no extra turbulence in the inlet/outlets.

The reason I'm using a polycarbonate sheet and not aluminium is it will work as a better insulator in that it won't be transferring any motor or road heat into the chill plate. That combined with the nitrile cork should insulate pretty well.

Good points regarding the central two cells, I might have a go at sealing the first cell pack this afternoon and see how that goes. I'll throw some photos and accurate measurements up if I do.
User avatar
jonescg
Senior Member
Posts: 3897
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Luke's 2001 NB MX-5

Post by jonescg »

Polycarbonate does absorb moisture and can actually develop fissures which are exacerbated by the water behind it. Might be better off using acrylic (Perspex) or ABS.
AEVA National Secretary, WA branch chair.
francisco.shi
Senior Member
Posts: 599
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Luke's 2001 NB MX-5

Post by francisco.shi »

Polypropylene will be a really good insulator and wont soak up water except that it is a bit soft and flowy and may sag over time.
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

Damn... Well it shouldn't cost that much. The Polycarbonate panels were pretty cheap to have cut to size. I'll see how Polypropylene goes cost wise.

Better to find out now rather than later!!
necrogt4
Groupie
Posts: 71
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Luke's 2001 NB MX-5

Post by necrogt4 »

I've been meaning to post this up for a couple of weeks, but I started on making packs out of some cells and heat sinks.

My approach was to first mix the thermal epoxy, put it into a squeeze bottle (i.e. a tomato sauce bottle) then neatly and easily make nice beads of epoxy to slide the cells into.
Thermal Epoxy.jpeg
Thermal Epoxy.jpeg (773.39 KiB) Viewed 22 times
Battery Pack 1.jpeg
Battery Pack 1.jpeg (857.89 KiB) Viewed 22 times


But halfway through my first pack it became rather difficult to get the epoxy out so I ended up just using the mixing spade which was far easier and took about the same time.
Battery Pack 2.jpeg
Battery Pack 2.jpeg (681.63 KiB) Viewed 22 times

I did make the mistake of not applying enough pressure / time for the first few packs and they ended up coming out at 49-49.5mm thick (depending on the corner) which is ~1.5mm thicker than I've allowed for being 48mm. Which creates some serious space issues.

So my next couple of packs I got far more serious with compression of the pack while the epoxy dries. This time instead of 60kg and clamping for ~30 minutes I instead used 100kg and clamped overnight. This resulted in 48.5mm thick packs, far far better and completely square which is great.
Battery Pack 4.jpeg
Battery Pack 4.jpeg (1.01 MiB) Viewed 22 times
Battery Pack 3.jpeg
Battery Pack 3.jpeg (645.2 KiB) Viewed 22 times


So after tossing back and forth I've decided to move from 2mm thick heatsinks to 1.6mm thick heatsinks along with a different application of epoxy approach which should give me some wiggle room being <=48mm. I've just got to wait a week or so to have new heatsinks cut, then I need to clean them up, fold them and insulate them with polyimide. Again. :cry:
Post Reply