Luke's 2001 NB MX-5
Re: Luke's 2001 NB MX-5
I've learned plenty of automotive skills by modifying cars starting with bolt ons then engine swaps followed by more and more adventurous projects. I'm a software engineer by trade so the engineering side comes fairly easily but in all honesty if you break projects down to their smallest parts it's all pretty simple. And there's plenty of great information + parts online.
A lot of my project is off the shelf parts (motor, gearbox adaptor, motor mounts etc). If done again I'd have these fabricated from my own designs. Though for batteries you're going to have to assemble something as there's nothing off the shelf. Things like Leaf, Model S or even i-Pace batteries do most of the hard work for you there. The trick here is the more you do yourself the lighter and more powerful you can go with your space (I've managed a 33kwh battery for instance).
That's a shame about the Zero-EV MX5, it's still listed on their site though? It would have been a great starting point, though I think they cut a few corners that left the end product a little on the heavy side and not as well balanced as it could be. That said, it looked to be a drop in kit using all of the original mounting points so you can only do so much there.
I'm not sure about building another MX5, I've got an E87 BMW to build after this. Though I'm more than happy to share parts suppliers, lessons learned and technical drawings! Feel free to PM me.
A lot of my project is off the shelf parts (motor, gearbox adaptor, motor mounts etc). If done again I'd have these fabricated from my own designs. Though for batteries you're going to have to assemble something as there's nothing off the shelf. Things like Leaf, Model S or even i-Pace batteries do most of the hard work for you there. The trick here is the more you do yourself the lighter and more powerful you can go with your space (I've managed a 33kwh battery for instance).
That's a shame about the Zero-EV MX5, it's still listed on their site though? It would have been a great starting point, though I think they cut a few corners that left the end product a little on the heavy side and not as well balanced as it could be. That said, it looked to be a drop in kit using all of the original mounting points so you can only do so much there.
I'm not sure about building another MX5, I've got an E87 BMW to build after this. Though I'm more than happy to share parts suppliers, lessons learned and technical drawings! Feel free to PM me.
- OzSpider
- Groupie
- Posts: 132
- Joined: Mon, 14 Dec 2020, 10:02
- Real Name: Derek
- Location: Brisbane, Australia
Re: Luke's 2001 NB MX-5
Luke, While you are answering questions about the build, who did your source the Al flywheel and clutch upgrade through? Was the flywheel custom made for the hyper9?
Re: Luke's 2001 NB MX-5
The flywheel is a forged aluminium flywheel made by Fidanza who are US based. I purchased through Fab9Tuning (https://fab9tuning.com/fidanza-forged-a ... -flywheel/) and comes in at 4.1kg. The beauty is this is just an off the shelf item and the ring gear can be removed to save that little bit of extra weight.
The clutch is a cushion button Exedy unit that I purchased through a local clutch/brake supplier.
Depending on how often I have to change gears I _may_ opt for a direct (clutchless) coupler between the motor and gearbox which will save ~10kg of rotating mass. But I won't know until I get the car moving.
The clutch is a cushion button Exedy unit that I purchased through a local clutch/brake supplier.
Depending on how often I have to change gears I _may_ opt for a direct (clutchless) coupler between the motor and gearbox which will save ~10kg of rotating mass. But I won't know until I get the car moving.
Re: Luke's 2001 NB MX-5
I've been slacking a bit with updates, again... Though things have been progressing smoothly.
A few weeks back I started fabricating some 25mm3 alloy blocks that need to be welded to my chill plates that help the sides act as clamps to keep the batteries under pressure. They can be seen in this render: Unfortunately the aluminium supply shop I use couldn't cut the 25mm x 25mm bar at 25mm so I had to do this myself. I ended up making about 30 of these and after cutting they really need some work to get them looking right. From this:
Then to this after much filing:
The drilled and finally tapped:
A few weeks back I started fabricating some 25mm3 alloy blocks that need to be welded to my chill plates that help the sides act as clamps to keep the batteries under pressure. They can be seen in this render: Unfortunately the aluminium supply shop I use couldn't cut the 25mm x 25mm bar at 25mm so I had to do this myself. I ended up making about 30 of these and after cutting they really need some work to get them looking right. From this:
Then to this after much filing:
The drilled and finally tapped:
-
- Senior Member
- Posts: 826
- Joined: Mon, 17 Sep 2018, 16:30
- Real Name: Francisco
- Location: Brisbane, Strathpine
Re: Luke's 2001 NB MX-5
You need to spray WD40 on the saw and don't cut too fast.
If you cut too fast the tooth fills up and can not take more material. This cause the material in the tooth to rub and stick.
The WD40 will also help to keep the cut material from sticking to the saw.
If you do it right the cut will look very nice and won't need any extra work.
Also get the saw with the smallest teeth you can get.
If you cut too fast the tooth fills up and can not take more material. This cause the material in the tooth to rub and stick.
The WD40 will also help to keep the cut material from sticking to the saw.
If you do it right the cut will look very nice and won't need any extra work.
Also get the saw with the smallest teeth you can get.
Re: Luke's 2001 NB MX-5
The disc I've currently got is a grinding style cutter. But from what I'm told that's not really the best for aluminium. I'll grab a toothed saw next time I go to the toy store.
Nice tip on the WD-40 though. I don't think that'd work on a grinding style disc though?!
Nice tip on the WD-40 though. I don't think that'd work on a grinding style disc though?!
-
- Senior Member
- Posts: 826
- Joined: Mon, 17 Sep 2018, 16:30
- Real Name: Francisco
- Location: Brisbane, Strathpine
Re: Luke's 2001 NB MX-5
A cutting disk will not cut aluminium properly. You need a saw.
This is the finish you can get with a carbide tipped saw from Bunnings (already a bit blunt) and WD40. You want to get as small teeth as possible. This saw can not cut steel.
It will be good for copper and aluminum plastic wood and soft stuff.
This is the finish you can get with a carbide tipped saw from Bunnings (already a bit blunt) and WD40. You want to get as small teeth as possible. This saw can not cut steel.
It will be good for copper and aluminum plastic wood and soft stuff.
Re: Luke's 2001 NB MX-5
Well, another lesson in the right tools for the job. If only I knew that beforehand
Those cuts look way better!
Those cuts look way better!
-
- Senior Member
- Posts: 826
- Joined: Mon, 17 Sep 2018, 16:30
- Real Name: Francisco
- Location: Brisbane, Strathpine
Re: Luke's 2001 NB MX-5
They will look even better with a new saw.
Re: Luke's 2001 NB MX-5
Speaking of looking better. I got a chance recently to make an attempt at fabricating my battery cell heat sinks out of 1.6mm aluminium. The sheets were water jet cut again and this time I asked for the finest cut and to have them run through the linisher to remove the burrs so I didn't have to file them myself.
This is what I received
They were worse than the first batch. Suffice to say I wasn't happy and it's unlikely I'll use the waterjet again. But with lockdown and the lead time to have them cut yet again I decided to just get on with it and worked them down to nice rounded edges and smooth faces. Very time consuming but they came up great.
Last time I folded these sheets I marked fold lines on each sheet and tried to line them up for each and every fold. This lead to inconsistent results so I clamped a stopper on the folding tool to ensure the sheets folded at the same point each time. The problem was, the damn sheet kept moving but it did at least lead to better results.
This is what I received
They were worse than the first batch. Suffice to say I wasn't happy and it's unlikely I'll use the waterjet again. But with lockdown and the lead time to have them cut yet again I decided to just get on with it and worked them down to nice rounded edges and smooth faces. Very time consuming but they came up great.
Last time I folded these sheets I marked fold lines on each sheet and tried to line them up for each and every fold. This lead to inconsistent results so I clamped a stopper on the folding tool to ensure the sheets folded at the same point each time. The problem was, the damn sheet kept moving but it did at least lead to better results.
Re: Luke's 2001 NB MX-5
I didn't want to have to realign my stopper sheet in the folding tool every other sheet so I decided to get serious and make a jig. I'm rather glad I've invested in some decent tools for this project, it makes this kind of thing much easier.
This time around the heatsinks came out near on identical to one another. It's not quite machine quality but for a garage job I'm pretty happy with the result.
Now it's time to tape, glue and clamp cells into these damn things and get some working battery boxes.
And back on the workbench with a stopper that's not going to move at all...
This time around the heatsinks came out near on identical to one another. It's not quite machine quality but for a garage job I'm pretty happy with the result.
Now it's time to tape, glue and clamp cells into these damn things and get some working battery boxes.
-
- Senior Member
- Posts: 826
- Joined: Mon, 17 Sep 2018, 16:30
- Real Name: Francisco
- Location: Brisbane, Strathpine
Re: Luke's 2001 NB MX-5
Why didn't you get the sheets cut on a guillotine?
That would have been quicker and cheaper than the water jet and would have given a very nice edge.
That would have been quicker and cheaper than the water jet and would have given a very nice edge.
Re: Luke's 2001 NB MX-5
I had my first run of 3mm copper busbars guillotine cut but the edges we're deformed a bit (partially bent and torn) and to be honest I didn't think of using a guillotine for these. Given their softness (grade 5005) and thickness they probably would have come out a lot better.
TL;DR, inexperience is the reason
TL;DR, inexperience is the reason
Re: Luke's 2001 NB MX-5
So I attempted another battery pack using the new folding jig and a slightly different technique with the thermal paste. Instead of beading the paste into the top and bottom corner of the heatsink and sliding one edge of the cell in after the other which caused some of the paste to get between the cells I instead decided to tape the cells together and apply the paste to the sides.
Once applied I then slid the heatsink over the top which pushes any excess paste out of the cell. It's a wee bit messy and I need to work out the right amount of paste to use but it's much easier and effective than beading.
A bit of side compression followed by placing 100kg of weight on top and I left it to dry overnight.
So findings from this method was overall it's way easier, cleaner and quicker. I do need to let the paste thicken little first before applying it so it doesn't run down the sides on it's own though.
I unfortunately also discovered an issue with the final cell though. It was only ~0.5mm thinner than the first batch (but far more consistent due to no paste between cells) at around 49mm which is still 1mm too much. If you take a look at the finished cell you can see there is a gap between the side/base of the bottom cell and the heat sink.
The folding tool is stretching the sides of the heat sink when they're folded and adding ~1.5mm to the overall width which throws everything out. Again, this puts me back to the drawing board and I'm not sure I'll be able to use this cell pack unless I can fill that gap somehow. I most certainly don't want any hot spots.
I've finally caved on this and am having heatsinks laser cut and precision folded at a sheet metal workshop with very exact post fabrication measurement specifications.After getting a quote it's actually only about $50 more than the water jet cutting (which was pretty cheap already). I pick these up tomorrow and will tape and prepare yet another pack. Third time lucky!
Once applied I then slid the heatsink over the top which pushes any excess paste out of the cell. It's a wee bit messy and I need to work out the right amount of paste to use but it's much easier and effective than beading.
A bit of side compression followed by placing 100kg of weight on top and I left it to dry overnight.
So findings from this method was overall it's way easier, cleaner and quicker. I do need to let the paste thicken little first before applying it so it doesn't run down the sides on it's own though.
I unfortunately also discovered an issue with the final cell though. It was only ~0.5mm thinner than the first batch (but far more consistent due to no paste between cells) at around 49mm which is still 1mm too much. If you take a look at the finished cell you can see there is a gap between the side/base of the bottom cell and the heat sink.
The folding tool is stretching the sides of the heat sink when they're folded and adding ~1.5mm to the overall width which throws everything out. Again, this puts me back to the drawing board and I'm not sure I'll be able to use this cell pack unless I can fill that gap somehow. I most certainly don't want any hot spots.
I've finally caved on this and am having heatsinks laser cut and precision folded at a sheet metal workshop with very exact post fabrication measurement specifications.After getting a quote it's actually only about $50 more than the water jet cutting (which was pretty cheap already). I pick these up tomorrow and will tape and prepare yet another pack. Third time lucky!
Re: Luke's 2001 NB MX-5
A little progress update.
I got my chill plates back from the engineer who welded on all of the blocks I fabricated.
Now that they're welded on I can finally begin assembly of the battery boxes. First up is sealing up the channels on the underside of the chill plate with 3mm thick nitrile cork gaskets held in place with 5mm thick polypropylene using 16mm zinc plated torx head screws. I chose not to use an alloy cover for two reasons, weight and soak. As one box will sit above the motor and the other two will be further back in the car exposed to tarmac heat I thought this solved two problems in one.
And finally the delrin end plates (that act as clamps) can be properly attached. I still need to get the m10 stainless torx screws nickle plated to avoid galvanic corrosion with the aluminium chill plate but I'll do that after assembly when I send off all of the copper busbars and conduits to also be tin plated (once I've fabricated all the tricky connections between the cells and other components in the boxes).
I got my chill plates back from the engineer who welded on all of the blocks I fabricated.
Now that they're welded on I can finally begin assembly of the battery boxes. First up is sealing up the channels on the underside of the chill plate with 3mm thick nitrile cork gaskets held in place with 5mm thick polypropylene using 16mm zinc plated torx head screws. I chose not to use an alloy cover for two reasons, weight and soak. As one box will sit above the motor and the other two will be further back in the car exposed to tarmac heat I thought this solved two problems in one.
And finally the delrin end plates (that act as clamps) can be properly attached. I still need to get the m10 stainless torx screws nickle plated to avoid galvanic corrosion with the aluminium chill plate but I'll do that after assembly when I send off all of the copper busbars and conduits to also be tin plated (once I've fabricated all the tricky connections between the cells and other components in the boxes).
Re: Luke's 2001 NB MX-5
I also got back my laser cut and machine folded heatsinks a couple of weeks ago, the accuracy difference between the home made and machine made ones are pretty noticeable (handmade right, machine made left).
Compared to thickness 47.7-48.3mm and height 124.5mm:
They're rather time consuming at:
- 20 minutes filing edges round and sanding off minor scratches so as not to puncture the cells.
- 20 minutes to tape with polyimide (unbranded Kapton) tape.
- 50 minutes to tape together, glue, insert into heatsink, clean excess glue and compress overnight.
Compared to thickness 47.7-48.3mm and height 124.5mm:
Re: Luke's 2001 NB MX-5
And here's all of the completed packs so far, 19 with one setting. Good for at least 21kWh. 12 more to go!
The original packs can be seen on the two right most stacks, hence they're slightly higher whereas the first three stacks are nearly identical in height.
The original packs can be seen on the two right most stacks, hence they're slightly higher whereas the first three stacks are nearly identical in height.
Re: Luke's 2001 NB MX-5
I've now assembled 31 battery cell packs in total. With 2 more to go with the cells I have left and then I'll have to try and disassemble the first 7 packs that I made as their dimensions don't work with the third and current attempt at cell packs.
These packs as seen above are super consistent in size and after preparing the surface of my largest chill plate I trial fit them. I was able to fully tighten sides/clamps by hand which was a bit relief as the first set of packs would never have fit horizontally (they were too thick and inconsistent).
After fitting these I began putting packs into groups of 4 and glueing them using the same thermal epoxy I hold the cells in place with. I added an extra layer of polyimide as a bit of cheap insurance. If I ever need to replace a pack all I need to do is remove the layer of tape on each pack that is glued to the other pack. And the benefit of glueing all of the packs together is that they then form a structural bond to the end plates which are held in with 10 M10 screws. The packs will easily survive 20g of force.
After glueing 12 cell packs into 4 groups (I'll do the final glueing together and to the end plates once I've got everything fit properly) I quickly placed them into the box and clamped them together with a 1.mm aluminium sheet at the end (to make space for the final L shape heat sink on the exposed top cell in the end pack). With the added epoxy and extra 1.mm sheet there was a bit of a bulge in the end plate. Not entirely unexpected but it did create a time sensitive issue...
A little bit of garage ingenuity fixed this problem though. I sat the box on top of two 20kg bumper plates and got out one of my serious clamps.
After compressing the bulge in ~1-2mm for a couple of minutes I released the clump so I could get the center screw in and noticed that the end plate sat perfectly on it's own, no clamp, no screw. The screw went in by hand without any resistance. Awesome!
These packs as seen above are super consistent in size and after preparing the surface of my largest chill plate I trial fit them. I was able to fully tighten sides/clamps by hand which was a bit relief as the first set of packs would never have fit horizontally (they were too thick and inconsistent).
After fitting these I began putting packs into groups of 4 and glueing them using the same thermal epoxy I hold the cells in place with. I added an extra layer of polyimide as a bit of cheap insurance. If I ever need to replace a pack all I need to do is remove the layer of tape on each pack that is glued to the other pack. And the benefit of glueing all of the packs together is that they then form a structural bond to the end plates which are held in with 10 M10 screws. The packs will easily survive 20g of force.
After glueing 12 cell packs into 4 groups (I'll do the final glueing together and to the end plates once I've got everything fit properly) I quickly placed them into the box and clamped them together with a 1.mm aluminium sheet at the end (to make space for the final L shape heat sink on the exposed top cell in the end pack). With the added epoxy and extra 1.mm sheet there was a bit of a bulge in the end plate. Not entirely unexpected but it did create a time sensitive issue...
A little bit of garage ingenuity fixed this problem though. I sat the box on top of two 20kg bumper plates and got out one of my serious clamps.
After compressing the bulge in ~1-2mm for a couple of minutes I released the clump so I could get the center screw in and noticed that the end plate sat perfectly on it's own, no clamp, no screw. The screw went in by hand without any resistance. Awesome!
Re: Luke's 2001 NB MX-5
After getting the cell packs to fit I removed them, again, to fit the PCB's that serve as the mounting points for the meaty busbars (3mm x 45mm = 135mm2 good for ~500A continuous). Keeping the packs in groups of 4 makes them pretty easy to work with at ~16kg each.
A trial fit of the busbars to one of my packs.
And all of the packs back into the box, again, for some more fit testing.
The end busbar links don't quite fit and contact the top clamp rod. These busbars will need to be bent around the rod and a layer of polyimide wouldn't go astray either. Now I need to have a bit of a think on how I'll get a nice curve bent into these.
A trial fit of the busbars to one of my packs.
And all of the packs back into the box, again, for some more fit testing.
The end busbar links don't quite fit and contact the top clamp rod. These busbars will need to be bent around the rod and a layer of polyimide wouldn't go astray either. Now I need to have a bit of a think on how I'll get a nice curve bent into these.
- jonescg
- Senior Member
- Posts: 4721
- Joined: Thu, 21 Jan 2010, 23:05
- Real Name: Chris Jones
- Location: Perth, WA.
- Contact:
Re: Luke's 2001 NB MX-5
The proximity of those busbars to the aluminium gives me the heebijeebies...
AEVA National President, WA branch director.
Re: Luke's 2001 NB MX-5
Haha, they're going to be bolted together by way of a acetal block so there won't be any way for them to come into contact with one another. I'll also give both the top of the busbar and the aluminium angle a nice wrap of polyimide to ensure no fun sparks when assembling a fully connected battery. There was only that single busbar fitted in this photo.
I gave a spare busbar a bend test in a vice and it's pliant enough to bend by hand. So getting those ends to bend around the black covered rod should be trivial too.
Re: Luke's 2001 NB MX-5
I picked up my battery box mounts this week from a laser cutting fabrication shop. They're a custom design using 2.5mm mild steel sheet with 10mm rivnuts pressed into them. There are right and left side versions of each and they're about 15cm long. I'll have to cut them to size for each position but the same basic design will be used throughout.
These face up in the engine bay as the battery box will be lowered down onto them. But for the rear two battery boxes they will both be loaded from underneath so the mounts will face down instead.
And in the rear.
These face up in the engine bay as the battery box will be lowered down onto them. But for the rear two battery boxes they will both be loaded from underneath so the mounts will face down instead.
And in the rear.
Re: Luke's 2001 NB MX-5
So I could have gone very basic and cheap with these mounts. Originally I considered using 2.5mm or 3mm square tube cut to size with bolts and nuts. But as with all decisions with this build, weight must be taken into account.
I then considered taking some of the material out of the square tube (bottom and perhaps an angle cut on the end). This would have turned into more work or if I was to get someone else to do this manually it would have been pretty expensive.
So after running through some stress simulations at 200g (yes, 200g not 20g as per NCOP14) I sent over my design the laser cutter and at a tad under $20 per mount including powder coating it was a "pretty" and easy choice. The final design is ~120g per mount vs 300g+ for square tube and at about 85% of the strength of square tube.
Here are the details of the simulations I ran. First up 3.0mm square tube:
And then the open mount design at 3.0mm:
Now using 2.5mm and a gusset inside the mount:
I then considered taking some of the material out of the square tube (bottom and perhaps an angle cut on the end). This would have turned into more work or if I was to get someone else to do this manually it would have been pretty expensive.
So after running through some stress simulations at 200g (yes, 200g not 20g as per NCOP14) I sent over my design the laser cutter and at a tad under $20 per mount including powder coating it was a "pretty" and easy choice. The final design is ~120g per mount vs 300g+ for square tube and at about 85% of the strength of square tube.
Here are the details of the simulations I ran. First up 3.0mm square tube:
And then the open mount design at 3.0mm:
Now using 2.5mm and a gusset inside the mount:
Re: Luke's 2001 NB MX-5
And finally 2.5mm with no gusset.
Now that the mounts are sorted I need to learn how to weld them in. Actually I'm practising this weekend making a jig to put a nice half circle kink into the end busbars to clear the top rod. I should have some photos of the jig which is just 20mm round bar, sheet metal and some angle in a clamp. Suuuper simple but perfect for the application.
Now that the mounts are sorted I need to learn how to weld them in. Actually I'm practising this weekend making a jig to put a nice half circle kink into the end busbars to clear the top rod. I should have some photos of the jig which is just 20mm round bar, sheet metal and some angle in a clamp. Suuuper simple but perfect for the application.
Re: Luke's 2001 NB MX-5
I dropped off all of my busbars and a bunch of stainless steel screws to be plated today.. And I thought I'd better update how I went with getting a nice bend in the end busbars around the clamping rods.
I spent a couple of hours watching YouTube tutorials and practising with my mig welder. Let's just say I've got plenty of practise to go but it was easy enough to make a jig, iterate on the design and actually get a decent kink bent into the busbar.
To start off with I used some 20mm round steel bar and some 6mm plate and marked up where things would sit for the convex side of the jig. I then cut up some pieces of 10mm plate to sit either side of the bar to essentially turn it into a semi-circle and give the busbar something to bend back to straight on.
On the opposite side I used some equal angle steel (which bent open more than 90° after a couple of tests and required a little fixing with some 45° cut square). All put together the jig looks like this from the side.
The kink ended up being a bit too big so I used some scrap busbar material and raised the plates a further 6mm. Good ol' bear tape holding those plates in place, real professional! The kink however was pretty good.
After some further fiddling I bent my first actual busbar the wrong way and had to hammer it back flat with some nice heavy steel plating. But after a little (um, a lot) of swearing things were back on track and I'm really happy with the results!
I spent a couple of hours watching YouTube tutorials and practising with my mig welder. Let's just say I've got plenty of practise to go but it was easy enough to make a jig, iterate on the design and actually get a decent kink bent into the busbar.
To start off with I used some 20mm round steel bar and some 6mm plate and marked up where things would sit for the convex side of the jig. I then cut up some pieces of 10mm plate to sit either side of the bar to essentially turn it into a semi-circle and give the busbar something to bend back to straight on.
On the opposite side I used some equal angle steel (which bent open more than 90° after a couple of tests and required a little fixing with some 45° cut square). All put together the jig looks like this from the side.
The kink ended up being a bit too big so I used some scrap busbar material and raised the plates a further 6mm. Good ol' bear tape holding those plates in place, real professional! The kink however was pretty good.
After some further fiddling I bent my first actual busbar the wrong way and had to hammer it back flat with some nice heavy steel plating. But after a little (um, a lot) of swearing things were back on track and I'm really happy with the results!