 |
 | Mazda RX7 Conversion |  | Ian Hooper, 2012 | |
 |
Specifications in brief
- Donor vehicle: 2000 Mazda RX7 RS
- Power: ~300hp
- Torque: 700Nm from 0rpm
- Range per charge: 120km
- Performance: 0-100 in approx 5s
- Weight: 1440kg (+12% on OEM)
- Balance: 51.5:48.5
Components
|
Getting Started
Introduction
Welcome to the construction journal for my Mazda RX7 electric conversion. There's quite a lot of content so you can skip to chapters of interest using the links above, and feel free to leave a message in the comments sections if you have any questions!
It had been a long-time dream of mine to build an electric car with performance enough to compete head-to-head on a race track with petrol vehicles. I believe such an EV can make great strides in dispelling common misconceptions about EV performance, and actually make people want an electric car. I was inspired by US EV racing pioneers like John Wayland (White Zombie) and the Pro EV team (Electric Imp) and wanted to bring some of that excitement to Australia.
Back in 2007 I started my first EV conversion, based on a 1990 Mazda MX5. The conversion went pretty well and it was a nice car to drive, but due to inexperience and budget constraints, its performance wasn't really worthy of competition use. So in 2011 I decided to sell the MX5 in order to help fund a new EV conversion. With greater budget in mind, there was only one choice of donor vehicle for me: a Series 8 Mazda RX7 Type RS, one of my all-time favourite cars and arguably one of the finest sports cars ever made. |
|
After a few months of unsuccessful searching for the perfect donor vehicle here in Australia (they're not very common, and many have been abused or have accident history), I instead ended up importing directly from Japan. In the end it took about six months from the sale of the MX5 to having an RX7 sitting in my workshop! But by late March 2012, the electric conversion officially began.
Engine Removal
For the MX5 conversion I got a couple of friends to help me with the arduous task of engine removal. It was a pretty awful job, and I could feel myself procrastinating about having to do it again with the RX7. So this time I took the easy option, and enlisted local RX7 specialist Rotomotion to do the job for me. Having a workshop remove the engine for you is not cheap, but it is a big job so saved me a lot of time, plus they're helping me find a buyer for the old engine.
Rotomotion also removed the gearbox, radiator, fuel tank, fuel pump and exhaust. I weighed the car before and after removal using load cells borrowed from EV Works. This is invaluable data for knowing how much weight allowance you have to play with, and where it should be in the car in order to maintain good weight balance. Apart from for performance reasons, the Department of Transport prefer to see vehicles with a converted weight and balance as close as possible to the original, so they know that the driving characteristics will not be compromised.
It came out at a total of 385kg removed, with 252kg front and 133kg rear.
Initial component choices
Controllers: Seduced by the idea of higher efficiencies and regenerative braking, I had hoped it might be feasible to use an AC drive system for this conversion. But after looking long and hard at the available options, I was unable to find anything which could remotely compete with a high-end DC system in terms bang-for-buck. (My budget is not unlimited!)
The best candidates I could find were the Tritium WaveSculptor 200, and either a rewound industrial AC induction motor, or something more exotic like an Ultramotive motor. The Tritium looks like a great controller, and (for AC) not bad value at AU$6600 for 165kVA peak power ($40/kVA). But compared with a CafeElectric Zilla2K EHV with 680kVA for about AU$5500 (AU$8/kVA), there's still a massive price difference between AC and DC controllers - all the more pronounced in very high performance vehicles (I would need four WaveSculptor 200s to equal one Zilla 2K EHV). So, with a touch of regret, I rationalised that DC was still the way to go.
At the time I was deciding, the most powerful controller on the market was the CafeElectric Zilla 2K EHV. These have been around for many years and have been used in most high performance EV conversions around the world, so seemed the logical choice. (EVnetics, who make great controllers, recently announced their ridiculously powerful Shiva controller, offering 1200kVA peak! But I won't be going that extreme; as well as being almost twice the price of a Z2K, I don't think I could fit in either motors or batteries big enough to justify a Shiva.) |
|
Motors: For safety and traditional reasons, most DC motors for road-going car conversions are designed to work around the 120-150V mark. So if using the full voltage capabilities of the high voltage Zilla2K, the best way to go is dual motors. When driven by a large controller, there is so much torque that a gearbox becomes unnecessary, and the motors can be coupled directly to the tailshaft - known as direct drive. (In fact a single 9" or larger DC motor with a 1000A+ controller is sufficient for direct drive.) Dual motors also opens up the possibility of series/parallel switching, which gives the electrical equivalent of two gears - though I don't expect to need this. Brand wise, I believe NetGain make the best DC motors for high performance applications, using very large windings and brushes to allow high peak current handling.
Netgain Impulse 9 motor |
Based on the specs and NetGain's performance graphs, I identified either the Warp 9s or the Impulse 9s as the best candidates. The Warp 9 has approximately 30% more torque per amp than the Impulse 9, but consequently about 30% higher back-EMF. In practice this means the Warp 9 would offer better low-speed acceleration, but the Impulse 9 would offer more consistent acceleration to higher speeds.
Some back-of-envelope calculations suggested dual Warp 9s would yield approximately 1.1Gs acceleration, and the Impulse 9s around 0.85Gs acceleration. Being skeptical that street tires would be able to hold on at over 1G acceleration, the Impulse 9 therefore seemed the better choice, and would offer more consistent performance over the whole rev range. Using the slightly smaller Impulses would also save around 25kg. |
Batteries: In all previous conversions I've done personally or been involved with, either ThunderSky (now Winston) or CALB batteries have been used. For many years these two brands have been the best value lithium solution for EV conversions, and in my opinion are still the way to go for normal road-going commuter EVs. However with only a ~5C real-world peak discharge rate, their power is a little low for race cars. So for this conversion, I had to find something better.
Manufacturer Headway have been around for a while, and their medium-sized (10-16Ah) cylindrical cells have been demonstrated in several EVs around the world with good results. So I arranged for some sample cells to test, and confirmed that their internal resistance of ~10mohm in a 10Ah cell is about (proportionately) half that of ThunderSky cells - suggesting approximately twice the power density. In a typical EV-sized pack, that yields a peak power of around 200kW, which should be plenty for a high performance EV (especially considering the superior torque and power band characteristics of electric drives). |
Some of the Headway cells for the RX7 |
Other components: Once the main items above had been selected, including settling on a nominal battery voltage of ~280V (88 LiFePO4 cells in series), the rest of the parts were easy enough to decide on:
- DC/DC converter: IOTA DLS-55 HV (200-320VDC version, well proven and economical)
- Charger: Zivan NG3 288V (not the cheapest brand but reliable / well proven)
- Power steering pump: Mocen 12V 500W (16 July 2012: actually I'm looking into a different option)
- Instrumentation: One of my ZEVA Fuel Gauge Driver Plus devices for battery SoC and current monitoring
- Vacuum pump: MES 70/6E (nice compact unit with inbuilt vac switch, Swiss-made automotive grade construction)
- Contactors: 2x Kilovac EV-200 (one per battery box - economical hermetically sealed contactors)
- Main fuse: Bussmann 500V 600A
The BMS was not so easy.. Being unfamiliar with any Battery Management Systems (BMSs) designed for Headway cells, and partly for interest's sake, I decided to design and build my own BMS for the car. More information about the BMS design can be found here. I'll also be using one of my EVMS units as the master unit.
With a little over 200kg of batteries, 120kg of motors, and perhaps 50kg of other equipment on board, I am estimating (hoping) the converted weight will be extremely close to the original factory weight.
Almost all of the components are coming from local Australian supplier EV Works (who I used to work for). The traction circuit will be quite similar to the "More Complete Wiring Diagram" I drew up a few years ago.
Design Stage: Motors
With the all the engine-related components removed and initial component decisions made, it was time to start designing how it would all fit into the car. The engine bay now had a vacant space around 700mm wide, 900mm long, and perhaps 500mm of total height - with many things to squeeze in!
The first components to fit are the motors, since there's little choice about where they can go (must connect to the tailshaft). The best place for them is as far into the transmission tunnel as possible, to leave maximum space in the engine bay for batteries etc. In the MX5 I had a fair bit of trouble getting the correct alignment from the motor to driveshaft to differential. If this is slightly off, it can introduce vibrations into the drivetrain.
Final section of the gearbox (upside down)
|
As such although I will be running direct drive (no gearbox), this time I'll be reusing just the final section of the gearbox, which has the existing driveshaft and torque tube attachments, to ensure correct alignments. (The torque tube is a C-channel frame which rigidly connects the gearbox to the differential.) It also has the speedometer drive sensor in this section which will save me having to re-engineer a speedo sensor mount.
As a general rule, the more of the existing car you can reuse, the easier the conversion will be! So the motors would be mounted in-line, resting on the original engine mounts, and with the rear motor face mounted to this gearbox section. |
To verify fit in the tunnel, I mocked-up an adaptor place which would attach the gearbox section to a steel frame holding the two electric motors, which would rest on the old engine mounts. (I do a lot of prototyping with Medium Density Fibreboard / MDF, because it's easy to work with and I believe it's invaluable to actually build and test-fit 1:1 models before spending money on steel/aluminium! Prototyping with wood first often identifies issues or improvements early.)
Once I was sure the motors would fit where I wanted them to go, it was time to design the frame which held the motors in place. What I came up with basically uses two 75x75x6mm equal-angle steel "rails" running the length of the motors, and four perpendicular 6mm plates holding both faces of both motors in-line. Then there would be legs protruding sideways out to the original engine mounts, through large rubber vibration mounts.
This steel frame alone will have an approximate weight of 15kg, which is not insignificant, but it does need to be pretty tough so I don't want to skimp.
To join the motors mechanically, after consulting with local transmission experts Transeals, I've ordered a Rotex RR38 spider coupler. Ideally the motor shafts will be accurately aligned, but the spider coupler allows for slight misalignment if the framework flexes under load. |
|
Design Stage: Batteries
For a comfortable 100km+ range in a normal car (my minimum range requirement), you need around a 20kWh battery pack (200wh/km). With a ~280V pack, this meant a 70Ah capacity. As such a reasonable pack configuration would be 88S7P. The Headway 38120 cells are 330g each, giving a total battery weight of just over 200kg.
Empty rear end where fuel tank and
muffler used to be |
The obvious places for batteries to be installed in the RX7 was to have some occupying much of the remaining space in the engine bay, and some where the fuel tank was (just behind the rear wheels, beneath the boot area).
There is a pretty large potential space in the rear for a battery pack - around 800W x 600L x 300H mm. As it turns out, weight is more the limiting factor. The rear pack sits behind rear wheels, so actually contributes more than its own weight to the rear wheel loading. Also the motors are a little behind the front wheels, so would also be contributing a little to rear wheel loading too. As such, if I wanted to maintain comparable rear axle loading (Δ133kg), it would be necessary to restrict rear pack weight to under 100kg (approximately). As such I decided - as an initial target - to have a 40S7P pack in the rear, in series with a 48S7P pack up front. |
Due to the space occupied by the motors and other ancillary equipment planned for the engine bay such as the aircon compressor and power steering pump, there proved to be very little height left in the engine bay for the front battery box - about 250mm, which is less than the height of a stack of 7 cells. So after much headscratching and considering different battery configurations, I came up with an interesting way to fit 48S9P in the front. The image right shows the side view of a 12S9P arrangement, and there will be four of these across the width of the engine bay giving 48S9P. |
|
Mock-up of front battery box in engine bay |
The resulting 740L x 610W x 240H mm battery box would fit above the motor frame, sloping down slightly towards the front of the engine bay to follow the bonnet contour, leaving just enough space behind it to mount the Zilla controller. I haven't purchased enough cells for running a 90Ah pack so intend to just start with 70Ah, but will design the boxes for 90Ah because it's nice to have the option of upgrading to 90Ah in future (once I can afford some more batteries!). The image left shows the test-fitting of a prototype box of correct dimensions in the engine bay, including the motor frame underneath it.
The exciting thing about potentially fitting in a 88S9P pack is having over 25kWh in a small aerodynamic car like this may yield 150km+ range, as well as having ~30% more peak power than my initial expectations. |
For the rear battery box, as I mentioned before, there is an excess of space in the rear to fit the required 40S9P pack. For consistency the cell layout is a similar hex packing to the front box, but only 8 rows wide, and 5 rows deep. This gives a battery box size of 760W x 500L x 240H mm.
I'll be cutting out the floor of the boot including the spare wheel well and mounting the battery box flush with boot, with a polycarbonate cover for display purposes. |
|
Structurally, the battery boxes will be made from 3mm aluminium panels TIG welded into boxes, and lined with 1.6mm polycarbonate (which is both electrically insulating and flame retardant). There will also be 1.6mm polycarbonate panels separating each of the battery arrays.
Continue reading --> Drivetrain
Comments| | Steve Gates on 10th Apr 2012
Hi Ian Matt passed on your site - looking good - you don't lose any time getting this going! One thing you may like to consider is whether the motor frame has the torsional stiffness - maybe close the section to form a box/tube or use tubulars lenghtwise like my Silvia. Happy to come by sometime if you would like any input on structure. (I have my scooter going now too).
Cheers... | | |  | | | Ian Hooper on 15th Apr 2012
Thanks Steve, after giving your comment some thought I've added an extra brace to the motor frame design in the upper part of the gap between the motors - should add a lot of torsional stiffness in that section! |
|
| | Matthew Clifton on 15th Apr 2012
Impressive blog Ian. Without being too impatient but rather curious, in view on your rapid progress, do you have an estimate of project duration - to on-road and then to speed tested and proved? | | | | | | Ian Hooper on 15th Apr 2012
Hi Matt, most likely mid year before it's finished. Things are going well but there's certainly lots of work ahead yet! |
|
| | Connie C. Khan on 16th Apr 2012
Hello can I reference some of the material here in this entry if I reference you with a link back to your site? | | | | | | Ian Hooper on 16th Apr 2012
Hi Connie, yes feel free to do so (thanks for asking!) |
|
| | David Grieve on 8th May 2012
I came across a EV builder using a gear vendors overdrive transmission. This gives 0.78 and they come as a tail housing.
I thought of your conversion and maybe it's something useful for you.
http://www.diyelectriccar.com/forums/showthread.php/good-ohmn-here-we-go-23492p52.html |
| | Tyler Watts on 28th May 2012
Hi Ian. Very interested in the project. When do you think you'll have some updates please sir? Any chance you could note when the page is updated to allow your followers to keep tabs on it please?
Very impressed with yoru work, adn been dreaming of this conversion for a long time myslef, you're an inspiration. I might be plagerising your designs or encouraging a kit out of you sir!
Regards
Tyler | | | | | | Ian Hooper on 29th May 2012
Thanks for the message. Probably a month or so before the next significant update - currently building the motor cradle and battery boxes etc. Feel free to send me your email via the contact page and I'll start an update notification list. I expect to upload all the CAD files etc once the build is complete and proven, in case others want to build their own! |
|
| | Andrew M on 24th Jun 2012
Please can we have an update. Waiting to see the next step! | | | | | | Ian Hooper on 26th Jun 2012
Hi Andrew, see link(s) at the top for updates! I was kind of waiting to finish major parts before writing them up, but it might be best if I just add to this conversion journal incrementally as I go. So there will be updates more regularly in future! |
|
| | Glen T on 28th Aug 2012
Hi Ian, how much is this conversion going to cost you? I'd like to do a similar conversion on my Toyota Supra, also from Perth. Lookin' forward to your next update. | | | | | | Ian Hooper on 12th Sep 2012
Hi Glen, apologies for the delay - I was out of the country for a while. Ballpark cost for all the conversion parts is approaching $30K. (One can of course do conversions much cheaper than this, but performance-wise you get what you pay for..) |
|
| | Kvh1 on 6th Nov 2012
Great read. Nice choice of vehicle!
I'm keen to see the end result in regards to performance and total build cost. Being a import I don't think it will ever be road registered?
But on the race track, no problem.
Keep the dream alive! | | | | | | Ian Hooper on 6th Nov 2012
HI KVH, it will actually be road registered! I imported the car under the Specialist & Enthusiast Vehicle Scheme. Unfortunately bringing a car in under SEVS makes them much more expensive than bringing one in for race use only on a CAMS license, but I'm planning to use the car as my daily driver as well as a track toy. It got its first test drive yesterday so I'll start putting up some videos in the Testing section soon. |
|
| | Kim C on 12th Dec 2012
Hi Ian, have you got it road registered yet? Regards, Kim | | | | | | Ian Hooper on 12th Dec 2012
Hi Kim, nope it's not road registered yet! I'm at that tricky stage of trying to integrate and re-enable some of the original systems, like the heater and A/C. Also the odometer isn't working without the ECU so I'll have to figure out a way around that - could cause some headaches! |
|
| | David on 28th Dec 2012
whats the weight??? what a masterpiece!!!
P.S> I remember there is a Japanese house that makes CF complete doors for this model and a CF hatch with sun louvre and lexan on the inside. The CF doors are 9kg-12kg each. | | | | | | Ian Hooper on 28th Dec 2012
Hi David, I'm yet to weigh it after conversion (I will have to do so before I can license it) but I believe it'll come out pretty close to original weight. Usually the licensing department are OK with ±5% weight vs original, any greater variation and they want braking tests done etc, which I would rather avoid! |
|
| | Conception Trahin on 21st Jan 2013
There is definately a lot to know about this subject. I really like all the points you made. |
| | Tristan on 29th Jan 2013
It's funny that you're required to get the weight as close to original as possible to lessen any headaches with registration.
Completely counter intuitive to minimising the weight of an EV.
Awesome project it is an amazing car! | | | | | | Ian Hooper on 29th Jan 2013
Hi Tristan, well if you manage to come in *under* weight it's never a problem, but unfortunately due to the weight of batteries required for a decent driving range, conversions usually end up being overweight, which of course can be a safety concern as far as braking performance, chassis integrity etc. |
|
| | Wayneus on 14th May 2013
Hi Ian,
Nice build! Very impressed with what you have done! One awesome RX!! How long did it take you to complete from start to finish (it looks like around 7-86 months)?
Have you got any figures on 0-100km times or quarter mile times? | | | | | | Ian Hooper on 14th May 2013
Hi Wayne, I think it was about 9 months from start to first drive, then a few more months before getting it licensed (mainly due to problems with motor controller and drive coupling). In fact the drive coupling is still giving me grief (keeps coming apart) so I haven't had the chance to do performance testing - hopefully in a month or two once I'm confident everything is working reliably! |
|
| | eVO on 17th Aug 2013
awesome project! what is the total cost of it? | | | | | | Ian Hooper on 17th Aug 2013
Hi eVO, I think total cost of the project was around the AU$55K mark including the vehicle itself (which was about half the total; the S8 RX7s are an expensive starting point!) |
|
| | eVO on 26th Aug 2013
IM INSPIRED BY YOU AND PLANNING TO START RX7 EV PROJECT MYSELF. |
| | Ron Johnson,Vancouver, Canada on 3rd Oct 2013
Hi Ian,
I enjoyed your Web site and explanations re your XR-7.
My car is a 1987 XR-7 5 speed. I am thinking of using a warp 9 motor. Considering what you know now, would you advise me to use a warp 9 or a Ac 50 9"?
Thanks, I will be watching for your reply.
Ron | | | | | | Ian Hooper on 3rd Oct 2013
Hi Ron, the short answer is that the Warp 9 (with say a Soliton 1 or Zilla 1K controller) will give higher performance, while the AC50 will give higher efficiency.
The Warp 9s are a great DC motor, good for up to 192V systems and coupled with a 1000A controller will give in the region of 250Nm torque with a peak power output around 150kW.
The AC50 kit is similar money and will probably give you 10-20% more range from the same battery pack. Half of that comes from higher motor efficiency, and half from the benefit of regenerative braking. However, the two controllers commonly used with the AC50 motor (Curtis 1238-7601 or 1239-8501) are much lower power than DC offerings; you end up with about half the torque and half the power vs Warp9 + 1KA controller, which of course makes a huge difference to how the car feels to drive. It's like the difference between a sports car and a commuter. So if I were doing your car, I'd go the Warp 9.
(Nobody denies that AC is technically superior and is the way of the future, but it is still a long way from DC's "bang for your buck"!) |
|
| | Ron Johnson, Vancouver, Canada on 3rd Oct 2013
Hi Ian,
Thanks for the clear advice. I was leaning the same way for the same reasons, but still wonder why so many pros are using the AC 50.
I have decided to use a clutchless adapter and the original complete transmission.
Kindest regards,
Ron |
| | Mike on 28th Dec 2013
Hi Ian,
When I start building my EV I'll definitely use this blog as a reference, excellent job.
I was wondering for power and weight and space would it be practical to fit a chain drive to the back axle with both ends of the motor driving 2 sets of chains and the motor mounted in the centre next to the axle.
I know they used chain drive in trucks way back in the 1920-30's.
Chain works well on motorcycles lol.
Cheers,
Mike | | | | | | Ian Hooper on 28th Dec 2013
Hi Mike, a chain driver certainly could work, but would be somewhat more difficult to achieve. A common "rule of thumb" for EV conversions is the less re-engineering of the original vehicle you do, the easier it will be to both convert and re-license. (E.g since the car wasn't designed for chain drive, it may be difficult to fit the motor in the available space near the rear axles, and you would need the new drive system evaluated by a recognised engineer, which could be expensive). Also you would need to either retain the differential somehow, or run dual motors - one for each rear wheel. Otherwise you'd get a lot of resistance and tire scrub going around corners! |
|
| | Terence Miskimmin on 6th Mar 2014
Hi Ian, thank you for sharing. Do you know any-one who builds conversions, I would like to convert a4 x4 Nissan Patrol |
| | WalterKi on 15th Mar 2014
I love your blog.. very nice colors & theme. Did you create this website yourself or did you hire someone to do it for you? Plz respond as I'm looking to create my own blog and would like to know where u got this from. thank you | | | | | | Ian Hooper on 15th Mar 2014
Hi Walter, I did it myself - nothing fancy used, just old-school PHP+CSS :) |
|
| | Mike H on 14th Aug 2014
Hi Ian,
With regards to your evaluation of the two 9" motors, and this might sound stupid, but would you gain anything by using a Warp9 and an Impulse9 in series?
Would this tend to broaden the overall torque curve or just cause problems and inefficiencies?
My original thoughts were due to wanting more torque than dual Impulse9s, but not wanting the length of two Warp9s.
Thanks,
Mike | | | | | | Ian Hooper on 24th Aug 2014
Hi Mike, apologies for the late reply - was out of the country. A Warp 9 and an Impulse 9 should work OK in series, and efficiency should be unaffected, but I'm not sure what the overall torque curve would look like! I suspect you'd end up with behaviour that's basically the average of the two. (The Warp 9 has higher torque per amp but higher back-EMF, which I think would result in the Warp always using more of the power than the Impulse.) |
|
|
