Weber and Coulomb's MX-5

[coulomb]

We're a bit behind with documenting our build, so here are some photos.

We completed the mods for the second Driver Controls Unit (DCU) 2-3 weeks ago. Note the addition of a piezo alarm, so not only 228 BMUs (Battery Management Units) but also 2 DCUs can make a noise now   



The idea was to have two DCUs charging two battery boxes (part half-packs) using two chargers, but by the time the mods were done, there was only one pack left to charge.

[coulomb]

We've slightly changed the way we attempt to keep out rain/mud/dust from the battery boxes. The original idea was to have foam rubber siliconed to the inside of the battery boxes. We still do that, except for the second piece on the ends of the battery boxes. That foam rubber is now siliconed to the end cells. So the ends of the battery boxes are still sealed off by two pieces of foam beside each other, but one of them stays with the cells and the other comes off with the lid.

   

The first photo was taken before the end of the clear tubing covering the battery cables was filled with silicone. Also, the white paper backing in the second photo was removed; it didn't look right. This leaves a sticky area, as the foam rubber is self adhesive. In this case, we don't want the stickiness; it's a nuisance and collects dirt, component pigtails, and small parts. First, we tried dissolving the glue with methylated spirits, but that didn't work so well. What seems to work is talcum powder, but you do have to be careful not to get the powder on surfaces that have to be siliconed to attach the foam rubber to the ends of the cells, and/or to seal the end of the battery box against rain/mud/dust.

[coulomb]

The tailshaft boxes are the most complex in shape, and the most difficult to fit. In retrospect, perhaps they were more trouble than they are worth, but we wanted to get the maximum number of cells in the car, so that the peak battery pack C rate would be as low as possible (design centre: 6C or 240 A).

Here is one of them completed:

   

Note the bright orange conduit between two of the blocks of cells. It gets busy in there, as you will see later. The gap between cell blocks is to accommodate the main chassis rail, which descends at that point. The second image shows the anderson connector resting on top of the battery box; that area gets rather busy too. Did I mention that this MX-5 is a bit cramped?   

Here they are in the car:

   

The second image is looking forward up the transmission tunnel. The crossed black "bars" at the top are actually the handbrake cables.

In fact, the handbrake cables were a special pain; there is more detail in these images.

   

The clips/brackets around the handbrake cables are just a nuisance now; where they used to attach to the chassis is now used as one of the three anchor points for the tailshaft battery boxes. The clips have been removed now.

You can see how busy the area between the tailshaft box battery blocks gets:

   

Even the 35 mm of space above these two battery boxes (there is a fuel-tank box above them, next to be installed) gets busy:



At the right, under the Peter Pan curtains, is the rear fuel tank battery box and convertible top. At the middle left is the gearbox console and handbrake lever, attached at the two red "eyes". You can see two orange conduits that go under the car, roughly along the route the exhaust pipe used to take, to contactor boxes under the bonnet. These are in 25 mm underground-quality conduit, the only type of orange conduit that we could find. This almost but not quite fits two double insulated 16 mm^2 cables, so the conduit had to be flattened somewhat (1.5 turns in a vice) to make them somewhat oval shaped. We then used a citrus based cleaner to lubricate the cables, and drew them in with a steel wire that hooked onto cable lugs. It was quite an effort getting the cables into the conduit. When we realised that we'd have to do the second one (for the other battery half-pack), we were disappointed, and the extra time possibly cost us getting the first half-pack completed yesterday. Still, the second cable took less than half the time for the first one.

The idea of the two pieces of string tied to the handbrake cables is to attempt to have the handbrake cables gently move out of the way when lowering down the front fuel-tank box, which goes down into that space.

[ Edit: added last paragraph; other minor clarifications. ]

[coulomb]

Here is the back of the rear fuel-tank battery box, with a contactor box right next to it.



What's wrong with this picture is that next to every haz voltage cable there should be an optic fibre connection. This one has been deemed "easy enough" to get to; we won't have to remove the battery box to connect the fibres.

However, we were not so lucky with one of the tailshaft battery boxes. It was also missing its fibre, and we had to drop the box some 100 mm to make the connection. It cost us over half an hour (with three of us working on the car). When the box was raised again, the fibre was found to rub against the power cable. This obviously is not good; with vibration, either the fibre will get crushed, or the fibre will damage the insulation of the cable, or both. It cost us another hour or so figuring out how to contort the fibre to make use of a dimple in the car chassis, allowing the fibre and power cable to not pinch against each other.

Speaking of optic fibre, I've just received an email indicating that another 40 metres (!) of fibre will be needed! Each half pack needs two, and each charger needs two, which will run basically the full length of the car. It's a good thing that it doesn't take up much space, and that it won't pick up any electromagnetic interference over all that length.

[Renard]

Looking at your photos, I'm reminded of those diabolical wooden puzzle blocks which are so tricky to take apart and re-assemble.
I've been trying hard, and mostly failing, while doing my conversion, to achieve independent access to components. That is, not to have to take the whole thing apart to get at some small piece. Of course it's true that contemporary automakers also fail at this (unlike fifty years ago). I find that I've been doing a lot of hole-tapping to eliminate nuts that would otherwise be inaccessible.
The task is complicated by the network of wiring which an EV requires. Whereas the ELV wiring can follow the usual pattern of some kind of harness coupled by connectors to individual items, the hazv cables can't be simply unplugged in the same way. And they tie everything up into a bundle.
Alas, if I should need to access a cell in my 24-cell box, I will have to remove -- or hopefully just detach -- the motor controller and the main fuse and contactor box which sit on top.
Your task is so much harder with reduced space and greater complexity, and I admire your intrepidity and salute your courage.

[weber]

A puzzle cube? Yes. That's the MX-ϟ alright!



Independent access is certainly a worthy goal. But yes, we conceded defeat on that, in most places, long ago.

However our HAZV cables can be simply unplugged from every one of our battery boxes. In fact every ELV segment. We use Anderson SBS-50 connectors which are finger-safe (but not waterproof). I hope we don't come to regret it. The fear is the connectors going high resistance and overheating, despite their silver plating.

I admire your intrepidity and salute your courage.

Thanks Renard. I sometimes gasp at our insanity.

"What not to do", at least for a first conversion.

[weber]

About 5 weeks ago I described a possible mod to add current-limiting to the 30 amp DC speed controller kit from the ATA Shop, in this post.

I finally spent the time to get a version working. As suggested by Johny, and Mark Aylott, I gave up on using the MOSFET as the current-sensing element and used the usual shunt in the MOSFET source. Mark recommended two of these Welwyn Open Air Resistors from Jaycar, in parallel, and that's what I used.

Here's a diagram of what worked:



Note that the output of the current-sense comparator also moved from the non-inverting input of the Schmitt-oscillator to its output. i.e. the two comparator outputs are wire-ANDed.

After blowing up two TC4427 driver chips I realised the previous circuit reduced the ON-time for current-limiting rather than increasing the OFF-time, and so passed through a phase where it oscillated at around 200 kHz, very briefly!

[Nevilleh]

Reading both yours and Renard's comments about access, I am moved to tell you that the inter-motor coupling in my BMW failed a little while ago confronting me with a daunting access problem. I originally installed the motor housing first, then the front battery boxes, motor controller, vac pump, PS pump, inverter, contactors and safety cut-outs. Whew! I was faced with a day's work just getting everything out so I crawled underneath to survey the scene from below. Result was I connected the rear motor only and drove down to my friendly local garage and put it on their hoist. The mechanic there is quite interested in the car - he's done work on it before - and we decided that we could unbolt the front cross-member and drop it and the motor housing down as one assembly and then remove the motor housing by itself. Worked a treat and we had the motor housing sitting on the floor after only an hour's work. Left all the other complicated stuff in place.
I have mentioned this as an example that there may be more than one way to skin a cat! If you want to know more about the failed coupling, look in my BMW thread.

[weber]

I was very pleased to have Coulomb and Newton (Jeff Owen) here again today to work on the MX-5. Jeff and I ran the third 25 mm orange conduit down the transmission tunnel. This one contains the hazardous voltage cables that go between the two chargers in the boot and the contactor boxes under the Tritium Wavesculptor up the front. These contactor boxes contain fuses for these cables, and contactors to connect them to the two half-packs when required.

Coulomb completed those contactor boxes and screwed them in place.

Bladecar also turned up to give us a hand.

Jeff brought the 25 mm orange conduit we needed. We found that the Key Plastics brand of orange corrugated conduit is far more usable for our purposes than the Clipsal brand which is too rigid and brittle. [Edit: See later correction viewtopic.php?title=weber-and-coulombs- ... 667#p40667]

We had a big distraction this afternoon because we were invited by Ross Pink to Electronic Innovations Pty Ltd (EI) for a late lunch of gourmet pizzas and to witness presentations of the the results of "project week". This is an annual EI tradition where staff get to work on a project of their own choosing. This year they only had 3 days in which to complete their various interesting electro-mechanical devices. I will mention only the most ambitious of the 5 projects -- a segway-type scooter by Bruce Frazer. It wasn't yet operating stably, but a remarkable effort for 3 days. I hope they will put information about all of the projects online soon.

Jeff and Bladecar returned the favour by showing their electric vehicles to the EI staff. Ross Pink enjoyed driving both of them. We also reported on the progress of the MX-5 project. That one's taking a little longer than 3 days.

On our return to the MX-5, we installed the largest battery box so far installed, containing nearly 1/4 of the cells. 54 x 40 Ah cells. See the photos below.

         



[Edit: Punctuation and "usable" -> "usable for our purposes"]

[Renard]

Congratulations!

That must be very satisfying to get those boxes in and bedded down.
Fingers crossed that they don't have to come out again.

I can see that your construction methods require you to put the boxes in assembled rather than to populate them in situ.

The orange conduit which I used responded well to being dipped in very hot water and curved to fit.

[weber]

Thanks Renard.

It turns out I owe Clipsal an apology. Apparently, the only reason the Clipsal corro conduits were brittle was because in order to fit two 10 mm diameter cables in a 25 mm conduit (which is 19 mm inside dia) we squashed the conduit slightly in a vise, along its whole length, to make it slightly elliptical. It seems to work-harden and embrittle the conduit. I expect exactly the same would happen to the Key Plastics 25 mm conduit. But we didn't need to squash it because we only put two 8 mm diameter cables inside it. It also seems that heating the squashed conduit, after getting the cable into it, anneals it, and it is no longer brittle and can again be cold-bent.

[Edit: Spelling]

[weber]

Today Mark Aylott helped me eliminate the conduit/cable crossings (orange) in the 35 mm gap between the diff battery boxes and the front fuel-tank battery box above them. Then we ran a bundle of 8 optic fibres (in 10 mm black loom tube) down the transmission tunnel and fanned it out within the same 35 mm gap to 3 different destinations. It looks like a railway switchyard. Then at last we lowered in the front fuel-tank battery box. Woohoo! Photos follow.

       

       

    

[weber]

We think this is an interesting route we found for the conduit carrying the DC cables from the chargers in the boot. This is in the left rear wheel arch. At first it appeared to be threatened by the upper wishbone above it and the half-axle below it. But in the photos the wishbone is already at the lowest point of its travel, and the half-axle doesn't go that high. The parts the conduit is zip-tied to don't move and the holes in the sub-frame were already there. But we did need the hair-dryer to soften it, to get it into place.

   

[srs]

Congrats on the build guys, great read

cheer Scott

[PlanB]

Let Them Eat Cake

Sitting in a cafe reading about all the hard work to shoe horn in 415 x 80 = 33.2kwhr = 119520 kJ of batteries, I couldn't help notice the caramel slice in the display case, 2060kJ per 100g slice. 119520/2060 = 58 slices = 5.8kg. What a shame caramel slice is not rechargeable.

[weber]

Thanks Scott.

Interesting calculations PlanB. But don't forget the well known fact that Caramel Combustion Engines (CCEs) only have an average efficiency of 20%. Electric motors are more like 80%, so you'll need 4 times as many caramel slices as you thought. 232 slices = 23.2 kg. Hmm. OK. Still way ahead of our 340 kg of batteries.

[zeva]

At the risk of further increasing the signal-to-noise ratio of the thread, I feel I must point out that a 2060kJ caramel slice at say $3 per unit works out at $5.24/kWh, which is 19x more expensive than electricity.

Factoring in the efficiency difference weber points out, it'd be 76x more expensive to run a vehicle on caramel slices. That kind of cost vs weight savings ratio wouldn't be justifiable outside the drag racing community.

[klaasdc]

Those brake discs look like they could use some start-stop traffic

[Johny]

At the risk of further increasing the signal-to-noise ratio of the thread, I feel I must point out that a 2060kJ caramel slice at say $3 per unit works out at $5.24/kWh, which is 19x more expensive than electricity.

Also the vehicle would slowly gain weight thus reducing the power to weight ratio as it got fatter.

[PlanB]

Will there be caramel slice at the official MX launch up there in Brisi or will the red terror be skulking over the Sydney harbour bridge & around the opera house forecourt under cover of darkness like the EI scooter?

[weber]

Jeff Owen came over again this afternoon and worked towards completing the under-boot battery box while I got things ready to install the roll-bar and its attached battery box. Jeff helped me carry the rollbar & box to the crane just before he left and I installed it after dinner.

Here you can see the method we used to complete the sealing out of road dust and gasses from the cabin where cables come up from below -- some upholstery-foam offcuts from Clark Rubber stuffed into the ducts.

   

In the following, I note that the webbing strap is carefully centred and is wrapped one and a half times around the rollbar so it won't slide. The soft-top needs to be propped part-way up like that to allow the rollbar feet to tuck under it. When the rollbar is in place and the crane removed, the soft-top has to be put all the way up to allow the feet to be bolted down.

   

The following photo of the rollbar in place (complete with mozzies) was taken by opening a gap between the soft-top and the top of the windscreen just wide enough to take the photo through -- hence the wrinkly rear window.

Anything in this photo that isn't already black will eventually be covered by black carpet (except the mozzies). Below the rollbar, a curved sheetmetal cover (original equipment but cut around the rollbar battery box) will first be bolted in place to hide the red battery box and all those cables and connectors and grey contactor boxes.

[weber]

Thanks zeva, Johny and PlanB for that important reasoning and calculation. It's a relief to find that we don't have to rip out all those batteries and start from scratch with caramel slice. But celebrating the launch with some, sounds like a good idea.

Hi KlaasDC, those rusty disks just show the benefit of the regen you get with AC. You never have to use the friction brakes.

For EV-day yesterday we had Jeff Owen, bladecar, coulomb and myself. Most of the day was spent finishing off the under-boot battery box. This is the largest (60 x 40 Ah cells). It was the first to be designed and built (nearly 4 years ago) but is the last to go in.

The photos show the exciting part at the end of the day, where we finally installed it. We still have lots of wiring and contactor boxes to do, and we need to weigh the MX-5 now so we can order some new springs.

    

    

    

    

[jonescg]

^^
Ma'am, your garter straps are showing

[Renard]

Two of my -- fortunately much smaller -- under-the-rear-seat boxes also have to go in from underneath the car. So I'm curious to know -- what I can't work out from the photos -- whether you raised the box or lowered the car or both?

[weber]

Mark and I just weighed the MX-5. To see our method go to
viewtopic.php?title=weber-and-coulombs- ... 436#p10436

We got tares or kerb weights (no spare tyre, chargers in boot)
Front 656 kg, Rear 706 kg, Total 1362 kg.
The original tares or kerb weights were (with full petrol tank and spare tyre)
Front 517 kg, Rear 466 kg, Total 983 kg.
Percentage increases
Front 27%, Rear 51%, Total 38%

That rear increase is a bit scary. Any suggestions?

[Edit: Changed "tares" to "tares or kerb weights" since they are the same thing, at least in this case]