Nevilleh's BMW

[Nevilleh]

I first thought about an EV conversion quite a few years ago, but only this year gathered the energy to actually do it!
I wanted a fairly light car, with rear wheel drive so as to eliminate the gearbox and one that wasn't too expensive. A 1988 BMW 318 for $500 fitted the bill nicely. In pretty good nick except for no WOF and no Rego, plus needing 4 new tyres, but mechanically sound with new brakes and shock absorbers.
Early on I decided PbA was a waste of time, Li is the only way to go! Also, I planned to go with ac but eventually decided it was a bit hard and couldn't find an affordable controller of a reasonable size, so went with dc - for now, anyway.
The transmission tunnel of the BMW isn't large enough for a big motor, so I decided to use two smaller ones coupled together. The Advanced L91 is my motor of choice.
Here's the car ready for "extraction"!



It took a while to remove that lump of metal, but it eventually came out to leave a nice, gaping hole. The motor and (auto) gearbox together, plus the exhaust system, fuel tank and radiator weighed in at 236 kg, All that dead weight - gone!



Now to figure out how to put something back in there.



Quite exciting when the motors arrived, here's what one looks like. It weighs 37 kg:



I built a dummy frame out of MDF to get the fitting and dimensions worked out. The cast Al bracket on the end is the BMW gearbox mount crossmember:



Coupling the two motors together puzzled me for a little while, as a commercial flexible coupling was very expensive and also quite large in size. But the BMW uses a rubber doughnut to couple the gearbox to the front drive shaft, so I thought one of those could be modified to do the job. A trip to the local wrecker gave me a rubber doughnut and a couple of front drive shaft sections that I could cut the 3-prong "spider" from, and my mate with a machine shop turned them up, drilled them out and cut keyways to make a very satisfactory coupler. Here it is fitted to the motors in the final Al mounting frame:





The motor frame itself is a work of art! Made from 10mm Al plate, cut and curved and welded, an engineering masterpiece made by a true craftsman! Not me, a fellow I have come to know quite well who has an engineering business just down the road. He hasn't sent in the bill yet, so we are still able to afford to eat!



Fitting the motors and the coupling was quite an intricate task and took me most of a day as I had to get the motor shafts aligned to a high degree of accuracy. The flexible coupling can handle angular displacements, but I thought it would struggle with co-axial offsets. Anyway, the removable mounting plates are designed to take small shims to be able to correct any mis-alignment and I managed to get it to better than .2mm which I thought would do quite well.



And another view:



You can see the reversing and second motor isolating contactors, the LogiSystems controller and the way the two motors fit together. The long channel sections are the front mounts which go onto the original BMW rubber engine mounts.
The "second motor isolator" allows me turn off the power to one motor. I thought this might make a useful "Power/Economy" selector and also will turn off one motor when in reverse. I don't need 100 kph going backwards!
Here's the "gearbox" mounting end of the frame assembly showing the part of the coupling that will attach to the BMW drive shaft:





Yesterday, I finished wiring the motors and contactors, all with 70 sq mm cable, and applied a 12V battery to it to see what would happen! The motors rotated beautifully, with no vibration that I could feel, proving that my motor alignment is pretty good. Big sigh of relief!
These motors are capable of around 47 kW of power and 216 NM of torque - each! The LogiSystems controller can deliver up to 1000A which is 150 kW from my 150V battery pack (ignoring the fact that it would sag mightily at that current) so I expect pretty good performance from this thing.

The Sky Energy cells are here and I have 135 of them. They are 40 AH cells and I am going to connect them as 45 series blocks of 3 in parallel. ie effectively 45 120 AH cells. I didn't use 120 AH cells as they are too tall to fit where I plan to put them.

I tested some of these at 120-odd amps discharge and measured a capacity of typically 46 AH, so I expect my battery pack to store 3x46x45x3.2 = 19.87 kW-hrs. That means I have 16 units of electricity available if I discharge them 80%. Might take me 100 kms!

Next job is to give the BMW engine bay a good clean and then fit the motor frame. Watch this space!

[Johny]

Great photo journal and nice looking motor mount. Watching with interest.

[EV2Go]

That’s the same controller NetGain are selling isn’t it? Keen to hear how it goes.

[Nevilleh]

I have now wired the motors - at least the interconnect stuff and the forward/reverse contactor, together with the second motor isolating contactor and installed the whole assembly into the car. It sits on the original rubber engine and gearbox mounts and couples straight to the original rubber driveshaft coupling - all with new rubbers and bolts, of course.
This is what it looks like from the front:



I had to unbolt the LogiSystems controller to get the motor assembly in, but there is now enough room to re-fit it easily.

The rear motor is down where the gearbox used to be, you can't see it all from the front.

The grey painted bracket on the right is to mount the Toyota electric power steering pump that replaces the old belt driven one. I welded that in place after fitting the motor assembly, so I hope it doesn't have to come off if I ever want to take the motors out!

A couple of side views of the engine (should that now be motor?) bay:





The 2 big, black hoses are the original power steering pipes.

At the front of the car you will see the air con condenser. I intend to fit the a/c compressor and set it up for reverse cycle capability so as to provide heating and cooling. The compressor will mount on top of the motor assembly.
The big open space between the a/c condenser and the motor assembly will be filled with SE Li cells. Here's what a few of them look like:



They are 40AH cells connected in parallel in threes and I plan on fitting 27 such packs in the front area and the remaining 18 under the boot floor, replacing the spare wheel well. That will give me 122 kg in front and 81 kg at the rear, which together with the 90 or so kg of the motor assembly, should give me a weight distribution not too dissimilar to the original - a little more weight to the rear, in fact. Probably need it to stop the wheels from breaking traction!

Here's a view from under the car showing the drive shaft coupling:



I did connect a 12V battery to the rear motor terminals and - wonder of wonders! - the wheels went round and round. That's a very gratifying thing. It sucked about 40A, but don't know how fast it was going as I don't have any of the 12V electrics (speedometer and tacho) working yet.

And that's it so far. Back to the blasted BMS software!

[Johny]

Nice going. Good pics too!

[Taffy]

I love the engine mounts! That is a work of art.

Great pictures, i will be very interested to see how you get on.

[Thalass]

Nice job! Look forward to seeing some video of it in action.

What BMS are you using? Your own?

[Nevilleh]

Yes, I will be using the BMS I've been developing. I hope! If I can ever get the software finished.......

[HeadsUp]

i dont intend to criticise anything , but just a couple of design aspects to contemplate for troublefree motoring.

the shaft coupling doesnt take advantage of the full length of keyed shaft available , check with torque tables for keyway length

( other than that .. love ya work , youre doing some beautiful things there )

[Nevilleh]

Thanks for those suggestions, I will check them out.
The coupling does use the full length of the keyway when it is properly assembled and the couplings have a bolt into the centre of the shaft to secure them.

[Nevilleh]

Here's a couple more photos showing not a lot of progress, but it IS happening!

In this one you can see the first battery pack being installed. A second pack will sit on top of it and a third one goes in the boot.
Also, I had to shift the LogiSystems controller to the firewall as I didn't allow enough space for the electric power steering pump with my original mounting arrangement. The steel bracket on the right of the mounting frame is for the power steering fluid reservoir. The black cylinder to the right again is the power steering pump - ex Toyota MR2. (That's a story in itself as it needs 8V to run)
You can see the throttle pot. box in the top right corner - throttle cable not yet connected.



This is the battery layout. There are twelve 120AH 3-packs all connected up with the copper straps supplied although I had to make up some extra long ones to link the triple-packs together. (They are the square-ended ones)
The coloured wires are for the BMS.
The curves in the box sides are parallax errors from the wide-angle lens!

[HeadsUp]

Thanks for those suggestions, I will check them out.
The coupling does use the full length of the keyway when it is properly assembled and the couplings have a bolt into the centre of the shaft to secure them.

no worries

there are different types of fit for shaft couplings

interference fit is when you have to bang it on with a hammer

extreme interference fit is when you have to heat up the female part first ( but any married man knows that )and then push it on with a press so as not to damage the bearings with hammering.

and sometimes even shrink the male part in liquid nitrogen or dry ice
( nobody wants to go there but sometimes the job demands it )

[CroDriver]

Hi.

Cool project.

And just a suggestion: Connect the cells in series using the terminal of the middle cells in the 3 cell parallel sub-packs

The way you have connected them now will stress the cells to which the bars are connected the most.

The best solution would be to have a bar on each terminal, so every cell would get the same load

[woody]

An intermediate solution would be zig - zag connections, i.e. Connect the positive of one end cell and the negative of the other end. This gives equal paths of 2 links for each battery.

But crodriver is right - more copper links would be best and also protect against some wiring failures...

[Nevilleh]

Yes, I originally planned to do the series links from the middle cells, but the bolts are not long enough to go through three Cu bars! And I haven't been to the shop to buy some longer ones.
Might try it like it is and see if I can notice any changes in the cells.

[HeadsUp]

what size screws are holding the battery box onto the chassis rails ?

the last battery box i drew up for a BMW is pictured below

you could probably use 40 x 40 x 4 angle coz you "only" have 55 kg of batteries but you have to consider the effect of vibration and twisting of the chassis under average driving on average roads.

its got to be engineered to stay crack free for years ... not weeks or months

cheers and best wishes

[HeadsUp]

ps

if you cant get inside a chassis rail to put nuts onto a bolt , there are beautiful things called a nutsert or rivnut

( google it )

great for putting nuts inside a blind compartment , you may be able to use them without buying the special tool , i have air powered ones which are worth about $ 900 each , but you can hire them too

[Nevilleh]

Thanks for all those comments and suggestions, much appreciated.
The front, lower battery box is held in place by 6 x 6mm bolts.
I have used rivnuts lots of times and generally successfully riveted them with a centre punch and a big hammer!
"if it's not working, hit it harder".

[Squiggles]

ps

if you cant get inside a chassis rail to put nuts onto a bolt , there are beautiful things called a nutsert or rivnut

( google it )

great for putting nuts inside a blind compartment , you may be able to use them without buying the special tool , i have air powered ones which are worth about $ 900 each , but you can hire them too

Is it acceptable to attach to one side of the chassis rail without adding reinforcement?

[weber]

Great work Neville.

You only need to move 4 battery links to solve both the mechanical strain and current sharing problems raised by Crodriver and Woody. As shown in the following edited photo.

[HeadsUp]

ps

if you cant get inside a chassis rail to put nuts onto a bolt , there are beautiful things called a nutsert or rivnut

( google it )

great for putting nuts inside a blind compartment , you may be able to use them without buying the special tool , i have air powered ones which are worth about $ 900 each , but you can hire them too

Is it acceptable to attach to one side of the chassis rail without adding reinforcement?

depends where you are fastening and what is being attached there

has to be a case by case analysis

generally speaking , engineers dont like chassis being drilled without seeking prior approval

its harder to undrill a hole than it is to check

general rule is you dont drill near the edge or near the corner of a chassis rail

if you have to drill , better drilling nearer to the center of it , or being at least 30 % of width away from the edge , but check with engineers in your state

also , there is chassis harmonics to consider , we prefer not to drill at a flex node point in the chassis , if you looked at the chassis from side on when under acceleration or braking , the flex node is like the point where a sine wave crosses the zero line , stresses are higher in that location.

chassis are designed to flex , in fact they look like a big violin string vibrating , or like a stretched out sine wave .

anywhere you have to attach something to a chassis , its better if there is something sitting on top of the chassis rail to exert loads that way plus having a bolt , rather than a bolt taking all the load in shear , see the battery box drawing above where an angle iron is sitting on top of the chassis with bolts through that , even if one bolt fell out the battery box will still sit there

the main flex node in a utility chassis is usually about half a metre behind the cab , the part of the chassis forward from there is relatively rigid , although there could be 2 or 3 minor and major flex points in a chassis depending on how many coffees the design engineer had.
at the rear there might be another minor stress node depending on if the ute tray is loaded and how many brackets there are under the ute tray and where they are. chassis rails will often be thicker at the points where stress nodes are , and thicker or deeper to support structural fittings , such as suspension pivot points.
a well designed chassis can have varied thicknesses and height of chassis rail throughout , but if you had strain gauges on the chassis during use , you would find a consistant load per square centimetre at any point you measured.

got to also keep in mind that a chassis rail flexes and twists under use so anything rigid in nature is best attached with a flexible mount between it and chassis , sometimes a 3-bolt mount will allow enough movement rather than a 4 bolt mount causing something to crack
see this post on mounting
nuther' post on chassis mounts

there ..... you all had ya sermon , now go and do 7 hail marys and dont try to get a perve down the teachers blouse again.......

[Nevilleh]

Been a while since I added anything here, but there has been some progress.
Both front battery boxes are installed and connected and they hold 27 of my "3-pack" cells. The battery box in the boot has been installed on Al rails bolted to the floor - it can slide back and forth so as to provide access when required and not take up much boot space. There are 18 3-pack cells in this box (making up the total of 45) so its quite heavy. It slides right back for battery access, then right forward against the back of the rear seat panel when in use.
The 70 sq mm cables running from the front of the car to the rear have been measured, cut and terminated, ready to install under the car. I plan on holding them in place with wide plastic cable ties POP-rivetted to the underside, couldn't think of anything better.

Once these cables are installed, all the power wiring is in place and I expect to have the thing moving under its own power next week. My aim has always been to have it running by Xmas, but it still has to be certified and there's a bit of work to do yet!

Here's a picture of the Big Red Button, fuse, main contactor (all thanks to Ian at ZEVA) and Hall-effect current sensor. You can also see the pre-charge resistor at the top right (not connected yet):



The current sensor is a bit hard to make out, but its the black "box" on the right side of the picture with a big, silvery terminal sticking out, waiting for the main battery cable to be connected. The 3 pins sticking up are for the connector to the electronics.

[antiscab]

looks very nice.

what is the part number of the hall sensor out of curisoity?

cheers,
Matt

[Nevilleh]

I used a Tamura sensor, the 500A one from this:



I haven't built any electronics to drive anything with as yet. All I did was connect it to a 5v supply and measure the o/p with a DMM. I pushed 120A through it with my dummy load and checked it against my clamp ammeter. I think from memory, its output changes about 3mV per amp and it was close enough to the clamp meter to not cause any concern. Somewhere around I have a 500A shunt resistor and I might use that to calibrate it as the clamp thing is not terribly accurate.

One thing that springs to mind is that it will be hard to get a very accurate reading of the charger current, so maybe a different sensor for monitoring that might be good. I am planning on feeding its output to an ADC channel on my mega48 powered controller. The resolution per bit is only 4.88mV though, which is why I say it won't be very accurate for the charger current.

[Nevilleh]

Took the car out on the road yesterday, the first time I have taken it out of the carpark. Quite interesting! I haven't attempted to alter the LogiSystems controller settings yet, so it retains the factory ones which seem fairly gentle. It accelerates to 50 kph very quickly and then more slowly - as you would expect - and I took it up to 80 without any difficulty. Didn't go any faster as I was in a 50 zone! And it is still not certified or registered.
A number of people who have used the LogiSystems controller complained of a very jerky takeoff, but I found it to be quite smooth although the accelerator needs to be treated quite gently.
I need to get a passenger along to monitor all the instruments and stuff, so changing the settings might have to wait a little while.
I have the two L91 motors wired in parallel and I found no difference in performance when I switched one of them off. I guess this verifies the fact that the performance is pretty much entirely due to the controller - at least while the motor(s) are capable of handling whatever it puts out.