Smart Roadster (with Prius Gen3 Inverter and Outlander Rear Motor)

[Pete9008]

Hi, I'm new to this site, and new to the whole EV bit but have been reading though everything here and have been inspired to start an EV project of my own.

Over the years I seem to have moved from small fun cars to larger comfortable ones, and in the process lost a lot of the enjoyment in driving, so the plan is to get back into something small and light that is fun to drive. The search for a suitable car led me to a Smart Roadster, not something I had ever considered before but a but of research convinced me it could work. The main problem with the roadster seems to be the terrible gear change and weak engines which led me to think about an engine swap. A bit more research, finding this site and realising just how far DIY inverters and control software have come has convinced me to go electric.

So far I've got the car, it's a but tired and needs some work for the MOT but should form a good basis.

In terms of other components:

Inverter - I have a Prius Gen3 inverter in pieces on the bench and am looking at the options to replace the logic board. I think it is going to be a custom board based on the openinverter hardware and software as this will allow me to integrate a few other features. The biggest problem with this at the moment is getting components! So far I've bought a few STM eval boards from which I'll rob the processors and I'm hoping to reuse components and connectors from the Prius logic board where possible. The plan is to use MG2 to drive the main motor, MG1 to drive an early generation Prius AC compressor, the boost/buck converter for HV battery charging and the DC/DC for 12V battery charging. If this works it should make for a very neat and compact install.

Motor - It's looking like a Mitsubishi Outlander rear motor and diff at the moment. The packaging is not ideal as it will push the motor out behind the rear axle while it would be better in front but once I have one I'll have a look at the possibility of rotating or flipping it. Worst case it mounts in the standard orientation in which case some of the suspension arms may need relocating.

Batteries - Probably a BMW 330e pack with the modules rearranged into a new housing. I'm hoping that this will fit in the area currently occupied by the engine and fuel tank but it will be tight. This may all change once I can get a better idea of the available space. I'm in two minds what to do about battery heating and cooling. The BMW battery packs have a very nice AC based system integrated into them which it would be nice to keep but is this just an unnecessary complexity for UK weather with slow charging? Overall I'd like to get 12-16kWhr of batteries into the car. Based on other conversions 160Whr/mile seems a reasonable target for the Roadster which should give a 50-70mile range which is more than enough for day to day use.

BMS - Yet to be decided, if the BMW packs are used then probably try to reuse the BMW BMS boards.

There is going to be a fairly slow start to this project as I'm still in the end stages of a house refurbishment/move and that has to take priority. The plan for now is to collect all the bits and sort out the new logic board ready. While doing this I'm also going to get the car back on the road in standard form and try to do loads of CANBUS logs as I'd like the new powertrain to be as closely integrated into the existing car systems as possible. In particular I'd like to try and keep the existing ABS and stability control systems working (which includes engine torque control). This may be a bit ambitious but worth a try. I'm planning on using GVRET and SavvyCAN for this unless anyone can suggest something better?

My background is in electronic design and embedded software but have never done any high power electronics or motor control so all that side is very new to me. I have built a couple of kit cars so am fairly comfortable with the mechanical side of the conversion.

Got to say I'm more than a bit nervous about the high voltages involved in this!

[Ctwidle]

Hi Pete, welcome to the madhouse. I am working on a 74 beetle with a Yaris inverter and an Outlander motor. Using the Y61 diff provided difficult due to lack of driveshafts. They didn’t come with the unit and were difficult to obtain either new or s/h. Its easy to detach the diff from the motor and you are left with a compact 40kg motor that might be easier to splice into the confined space of the roadster drivetrain with a suitable coupling and adaptor plate.
Chris

[seanyt]

Very nice build. I also like the appeal of the roadsters almost
Lotus Elise like dimensions.

I also have an outlander rear motor and outlander inverter.
Have you both considered using the outlander inverter and feeding it canbus messages alone to drive it.

Without the use of a v3 board. I also have a v3 board.
But plan to go the canbus route first.

[Peter]

Hi Pete9008. I have an SDU in my Smart 450, go the whole nine yards and upset many very expensive cars

[Pete9008]

Hi guys, thanks for the warm welcome!

Hi Pete, welcome to the madhouse. I am working on a 74 beetle with a Yaris inverter and an Outlander motor. Using the Y61 diff provided difficult due to lack of driveshafts. They didn’t come with the unit and were difficult to obtain either new or s/h. Its easy to detach the diff from the motor and you are left with a compact 40kg motor that might be easier to splice into the confined space of the roadster drivetrain with a suitable coupling and adaptor plate.
Chris

I did think about that but would like to keep the weight of the conversion down and the smart gearbox is quite a bit heavier then the Outlander one. Bit worried about you comments on the driveshafts, I have previously looked online and they seemed available second hand (not cheap but available). I'll have another look but if I can't get them may need to reconsider. How did you go about making up a coupler for the gearbox and motor splines?

Very nice build. I also like the appeal of the roadsters almost
Lotus Elise like dimensions.

I also have an outlander rear motor and outlander inverter.
Have you both considered using the outlander inverter and feeding it canbus messages alone to drive it.

Without the use of a v3 board. I also have a v3 board.
But plan to go the canbus route first.

I did consider the Outlander inverter but I'm hoping to extract a little more power than standard out of the motor and I think the Prius inverter provides a few more options here. Do you know whether the 70kW limit is fixed within the Outlander inverter?

Hi Pete9008. I have an SDU in my Smart 450, go the whole nine yards and upset many very expensive cars

I'd love to but the Tesla part costs are a bit too much for my pockets!
Do like the sound of yours though, I imagine the 0-60 time is limited mostly by having to keep the front wheels on the ground!

[Peter]

Hi Pete9008. Shame your budget isnt higher. You are correct, limited wheelies with 0-60 in 4 secs
However 70Kw will be very quick so I imagine you will be a happy chap when its up and running.
Where are you based?
Good luck with the build.

[Ctwidle]

Local dealership quoted nearly AU$ 1000 each but when they checked there was no stock even in Japan. Bought a pair from a wrecker that turned out to be from a different model and the inner splines were correct but nearly 10mm too long. Might have been possible to shorten and cut a new circlip groove but I don’t have the facilities. That was last year, supply may have changed but i had the beetle gearbox rebuilt with just 2nd and 4th gears instead.

Check out the “Outlander rear motor and inverter” thread for a few options on couplers. I believe there is also a Suzuki clutch spline that can be used if you don’t want to extract the one from the original diff. Tom aka bobby_come_lately has just done a youtube on making his coupler.

Without the necessary programming skills or, at post 70 the capacity to develop them to a safe level, I am waiting for the brains around here to come up with a plug and play solution. In the mean time it is fun just to watch projects coming together.

In following Outlander threads I have seen a reference to the 50, 60 and 70kW versions of Outlanders using the Y61 being dependant on battery performance rather than differences in the motor and others have suggested it may be capable of quite a bit more but no evidence was presented.

[midway]

Why is no one using a gearbox from a Peugeot 3008 hybrid. Power declared 80kw built-in gearbox with differential

[4markowen]

Also, why not a Mini Countryman / BMW Series 2 hybrid rear diff? 100kW motor, built in diff.

There's a few electric rear diffs that offer a nice little package for front or rear drive /axle replacement conversions IMO. Unlike the Toyota/Lexus MGR which was focused on a partial 4wd design, these offer full electric driving from the rear.

I suppose as none of us have gotten around to buying one and hooking it up
Mainly as outlander/prius components are at the cheaper end of the scale in the breakers.

[seanyt]

Hi Pete

the evolution miev used the stock motor and inverter upto 100kw and much higher torque levels.
the outlander using 80s batterys will do 70kw i plan to run 96s and 85-90kw.

From other peope experimenting the inverter is a dumb unit and just gives torque output when requested. Its up to another ecu to cap the overall power.

Outlander phev motors and gearbox have been out longer than 3008 and 2 series so this i assume is why most have been considering them . Its been know that outlanded uses canbus communication for power and regen. So this is far easier than setting up an inverter . This is my goal anyway.

I have my outlander motor and inverter fitted to the rear of a hatchback car with the front engine still in place also acting as a hybrid system .

[seanyt]

Why is no one using a gearbox from a Peugeot 3008 hybrid. Power declared 80kw built-in gearbox with differential
_20210307_152418.JPG_20210307_162149.JPG_20210307_143937.JPG_20210307_144541.JPG

Dont confuse old 3008 and new 3008
Old smg180 motor is 27kw power the one you have in the pictures.

I dont have details of the new one right now. But when i considered it. it was axle rpm limited.
as in 10:1 ratio as an exampl and a top speed of 140-160km/h using a 19-20" wheel same as used in 508 pse where it then decouples.
Also 166nm torque i believe outlander can command 300nm for short periods.

So when you work out using smaller wheels the motor and gearbox was rpm limited. also the price of 3008 hybrid motors and 2 series are far more expensive along with also considering xc90 t8 rear axles or v60 d8.

[midway]

I also used the outlander reducer in my project

[Pete9008]

Where are you based?

Lincolnshire now, used to be near Warrington which would have been nearer you but moved over here a few years ago.

Ctwidle, thanks for the details. For them moment the plan remains to use the whole unit but it's nice to know there are other options if it doesn't work out.

Midway & 4markowen - the other options do look good but I keep coming back to the Outlander unit for cost and availability reasons. At the price they are I'm happy to abuse it (overvoltage and overcurrent) just to see what it can do, knowing that replacements should be easy and cheap to get.

the evolution miev used the stock motor and inverter upto 100kw and much higher torque levels.
the outlander using 80s batterys will do 70kw i plan to run 96s and 85-90kw.

From other peope experimenting the inverter is a dumb unit and just gives torque output when requested. Its up to another ecu to cap the overall power.

Outlander phev motors and gearbox have been out longer than 3008 and 2 series so this i assume is why most have been considering them . Its been know that outlanded uses canbus communication for power and regen. So this is far easier than setting up an inverter . This is my goal anyway.

I have my outlander motor and inverter fitted to the rear of a hatchback car with the front engine still in place also acting as a hybrid system .

100kW would do nicely . My rough planning calcs to date give 93kW and 907kg giving a power to weight of ~140bhp/ton which is in the right ballpark for a fun car that can use all its power. As mentioned above with the relatively low cost of the motor and inverter I'm happy to push them a bit further than this to see what happens.

I'm planning the same as you on the voltage, maybe even a little higher. A higher battery voltage seems to win in a number of ways, more capacity and more power for a given cell current. I am struggling to understand the effect battery voltage has on regen braking though but I'm going to ask that (potentially stupid) question in another thread.

The simplicity of the Outlander inverter is attractive but I think I'll stick with the Prius one for now. If it works the way I hope it should be pretty much a one box solution for all the HV stuff.

Have you had yours running yet? If so how does it performs?

[Peter]

An ex-Warrington resident, never mind. If you are ever 'Up North' again you are welcome to scare the pants off yourself with my Smart

[Pete9008]

The bit I was in wasn't that bad!

Thanks, still work part time for a company over near Chester so may well be over your way at some point...

[Peter]

Good, will have a fully charged battery and disposable underwear ready

[Pete9008]

[Pete9008]

Have been working through all the HV components and come up with a new architecture, curious whether anyone can see any issues with it.

While originally I was planning to use MG1 to drive a 3phase feed A/C compressor (mainly because it seemed a neat solution) but have now realised that it would probably be better to completely forget about treating MG1 as a 3-phase driver and instead think of it as being the switching components for three separate buck converters. This will obviously need external inductors and output capacitors (which would be added to a new enclosure welded or bolted to the side of the existing inverter housing) but would then provide three individually programmable HV outputs. These would be used for battery heating, cabin heating and to supply an A/C pump with a built in controller.

The battery would connect to the inverter bus bar connections and would use the standard contactors within the BMW pack.

As before the existing buck/boost converter would either be used to buck the battery voltage down to run the DC/DC converter when driving, or in boost mode for battery charging (if the battery voltage is always higher, if not then an additional diode on the input should allow it to run in boost/buck mode for charging). I think enough PFC, certainly for lower current charging, would be possible within the Prius buck/boost block.

If this works then almost all the HV connections could be made within the Prius inverter box. It would need a custom logic board but I was planning to do that anyway. The only additional HV box would be for an input filter and single phase rectifier for the charger input connection.

The battery pack would start as a standard 6 module (96cell) pack. Later I would like to increase this to 8 modules though (128cell). This would be within the voltage rating of the inverter's HV bus but is it too much for any other reason?

I plan to use Catphish's battery monitor interface to monitor the cells and control balancing so should have plenty of flexibility on pack configuration. Bit wary of removing cells from modules but if it needs doing then I'm willing to do it.

The contactors within the later BMW batteries are only rated for 500V. With 128cells I would have to terminate charge at 3.9v to stay below this. Is this too low? If so would 125cells charging to 4V be better? I only ever plan to discharge/charge between 20% and 90% to protect the cells. I'd prefer not to be modifying battery modules and would prefer to use full modules if possible but I'm willing to remove cells if necessary.

Would it be better to drop the battery voltage further to give a margin on the contactor voltage?

Comments and thoughts appreciated.

[Pete9008]

No progress on the car itself as the garage is full of another car that I need to fix, MOT and sell before I can move the Roadster in. I have spent a fair bit of time thinking about the implementation though and I'm fairly decided on the above architecture - essentially using the inverter as the HV distribution/junction box, to switch the HV supplies to loads and as a charger (the contactors and BMS will be within the battery box). This should avoid a lot of HV wiring and give a fairly neat install.

I still haven't decided on how many battery modules to use. The standard 6 BMW modules would be easiest but is a little limited in capacity. Increasing to 8 modules would get the range to where I want to be and improve performance/reduce battery current but is pushing the voltage limits of a few components (the contactors in particular). It also puts a lot of weight in the back of the car. I'm starting to think that a better compromise would be to start with a 6 module battery pack in the back where the engine used to be. The car can then be used like this to see how well it works. I'll also include an extra set of parallel battery connections so that a second pack could be fitted in the front in the future. The second pack would have a separate set of contactors within it so that they could be used/charged either in parallel or independently under control of a board I've been working on.

Below is the board:

The original plan was to go with Damien's dual controller board but I decided that I'd like to include a few custom features and tweak a few things (with hindsight I should have stuck with Damien's - I'd heavily underestimated how long it would take!). The main changes are:
1. It has a dedicated resolver to digital IC (recovered from the Toyota logic board) to read the motor position. I know this is probably unnecessary, but I'm more comfortable with it in terms of accuracy and noise immunity. I'm just hoping I've got the circuit right as otherwise I'm going to feel a but stupid (as the board doesn't have support for the ADC based resolver solution)
2. I've tweaked a couple of the timer lines on the F103 and I'm hoping to use the spare channel on the main PWM timer to trigger simultaneous bursts of injected ADC conversions for the two motor phase currents and also trigger the sampling of the resolver data. This should give synchronised data for the motor current and motor position.
3. It uses the 100pin version of the STM32F103 to drive MG2, partly because I found one left over from a previous project but also because it allows a faster parallel interface to the resolver IC and adds an extra ADC which gives a little more flexibility in software. The board should also support the STM32F405 chip if needed in the future.
4. MG1, the DC/DC, charger interface and all comms are handled by a STM32L496. The STM32L series are very low power devices and the intention is for this section of the board to be permanently powered. This chip is connected to a 10Mbit RS485 port and a SD card which will be used for high speed data logging - ideally logging all the motor data on every pass though the control loop (so that if anything misbehaves it should be possible to figure out exactly what's going on).
5. The F103 (MG2 motor control) is held in reset if the ignition isn't on.
5. The L496 has no resolver or encoder input so won't be able to control a motor. Instead it will use the 3 HV bridges to supply HV loads (AC compressor, cabin heater and possibly a battery heater). This processor also has an input for an isolated current input (charger input current) and an isolated 4-channel ADC (for measuring the MG1 HV output voltages).
6. There are two CAN ports. One goes to both the L496 and the F103 while the second only goes to the L496. The port on the L496 will be permanently powered and the intention is to dedicate it to the BMS.
7. Moved to an ESP12F WiFi module to save a bit of space and improve robustness. It's shared between the two processors and has an on board programming port.
8. It still uses the Toyota connector. Two reasons, I liked the separate dedicated sub-connector is has for the resolver connections and it has more pins than the ampseal alternative. The ampseal would have 34 pins, as standard the Toyota has 40 pins and I'm hoping that I can increase this to 50 by adding extra ways to the unpopulated positions. Should still be OK if I can't but the extra ways would provide extra grounds and a few extra spare inputs/outputs which would be nice.

The schematic still needs a thorough check over and the board a full design rule check but hopefully it will be going out to manufacture sometime in the next week or two. If anyone fancies having a look though the schematics for problems let me know and I'll put them up.

[Pete9008]

Look what arrived this morning

Hadn't been planning on buying the motor yet but one came up at a reasonable price complete with the mounts and the HV and LV cable tails. The place selling it also had the driveshafts so was able to do a deal on combined delivery.

Probably a bit late to be asking this question given that I've now bought the driveshafts but - is there any real benefit to keeping the original manual gearbox and mating the Outlander motor to it?

The Plan had been to use the Outlander reduction box but now looking at it using the Smart box would put the motor lower down and in front of the rear axle - exactly where I would like it positioned from a weight balance and handling point of view. It would also allow the gear ratio to be tweeked to find the best compromise (or compromises, sport and economy modes?) Downside is the added complexity of the motor to gearbox adaptor and the extra weight of the gearbox (say an extra 20kg). I have a reasonably well kitted out workshop so can handle the engineering side of either option.

What I don't want to do is complete the conversion and then find the car feels a bit sluggish at some speeds, is this likely to be an issue or am I overthinking it? I do have the option of running the motor at a higher voltage than standard which might help here?

What I'm really after here is views from anyone that has a relatively light car running the Outlander rear motor/box on how the package performs and what kind of acceleration is seen, particularly at say 40-60mph. Also interested in thoughts from people with gearboxes on their cars, do you ever change gear or does it always get left in the same one?

[Pete9008]

Hmmm, just done a few sums:

Standard car - peak of 0.56G in 1st to 33mph, 0.4G in 2nd to 45mph, 0.31G in 3rd to 59mph. Not great but not bad

EV with Outlander motor and Smart box - peak of 0.86G in 1st to 50mph, 0.63G in 2nd to 69mph, 0.48G in 3rd to 90mph. Much better!

EV with Outlander motor and box - peak of 0.47G to 94mph

None of the above take account of torque reduction in field weakening and all use standard Smart and Mitsubishi peak torque figures (I'd expect to be 20-30% better on the Mitsibishi but don't want to include it without confirmation from someone who has tied it).

I'm guessing field weakening would start around 3krpm at standard voltage, maybe 5k running at higher voltage and so would start to be felt around 45-50mph with the outlander box, not ideal but really no worse than an ICE in higher gear, 0.47G is also pretty much at the limit of what standard road tyres can cope with so would much more make any difference in the real world? If the Outlander motor can do 90kW, as previously mentioned by seanyt, this would be nearer 0.7G which would spin the tyres at 50mph!

Still very interested to hear other views but I'm swinging back towards the outlander box given it would appear to offer enough performance and is the simpler conversion.

[arber333]

Still very interested to hear other views but I'm swinging back towards the outlander box given it would appear to offer enough performance and is the simpler conversion.

Dont forget the Outlander sums were made with OEM battery which is really the limiting factor for the motor-inverter combo. I suspect you will be able to get more torque out of the motor than OEM.
Now that i tried the leaf gearbox with my Pug i will never go the transmission route again. Transmission is noisy, it is worn out or has other issues.

By far the only factor when choosing transmission over OEM gearbox are the driveshafts. Some people are afraid to cut/splice the driveshafts and with good reason.
But if one has a good machine shop available than this is not a problem.

[Pete9008]

I've not done driveshafts before and modding them does worry me so I'm likely to outsource them. On the other hand I've outsourced a prop shaft in the past and not been happy with what came back! I can handle all the bits involved in a gearbox adaptor myself but it would take more time.

Still considering battery options. Really need to corner weight the car to see how much weight can be added on each axle before making any decisions on pack sizes. I'm also wondering how high it is safe to go with battery voltage. I'd like to go to a 128cell pack, but would only charge to 90%, giving 500V. Is that too high? Either way I should be able to get to 90kW.

How does an electric motor feel when it starts to go into field weakening, how noticeable is it?

[arber333]

I've not done driveshafts before and modding them does worry me so I'm likely to outsource them. On the other hand I've outsourced a prop shaft in the past and not been happy with what came back! I can handle all the bits involved in a gearbox adaptor myself but it would take more time.

Still considering battery options. Really need to corner weight the car to see how much weight can be added on each axle before making any decisions on pack sizes. I'm also wondering how high it is safe to go with battery voltage. I'd like to go to a 128cell pack, but would only charge to 90%, giving 500V. Is that too high? Either way I should be able to get to 90kW.

How does an electric motor feel when it starts to go into field weakening, how noticeable is it?

Not noticable with pull. Later on in high rpm sure. Noise does change. Maybe you can hear it if you open the window and drive along a wall.

I would not advise to go over 104S. I did and i am at the end of my inverter comfort zone. And i only charge to 4v per cell. Most inverters just shut down higher than 400V. I know outlander does...

[Pete9008]

I would not advise to go over 104S. I did and i am at the end of my inverter comfort zone. And i only charge to 4v per cell. Most inverters just shut down higher than 400V. I know outlander does...

The inverter is the Prius Gen3 which I thought could work with bus voltages of up to 600V? Are there any other components that can cause problems at higher voltages?