Opel Corsa Project

From RC-Autopilot Wiki

Revision as of 18:30, 25 June 2010 by Ruqaya (Talk | contribs)
(diff) ← Older revision | Current revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Contents

Opel Corsa Project

In September 2008, I decided to try converting a conventional car to electric drive, something lots of people all over the world have been doing for years. Having thought about electric cars for years, I had always tried to avoid this method because I always looked at the high weight of standard cars (ranging from 800kg to 1500kg) and it was always my ambition to persue the direction of light and efficient cars. The reasons being efficiency (how fast/far can you drive with a given energy amount) and comfort (how much acceleration can you get with a given motor power). It's simple, a 500kg car with a 15kW motor accelerates twice as fast as a 1000kg car, and thats a big difference. Sure you can compensate with bigger and stronger motors, which is exactly what the auto industry has done over the past 100 years (try to find a car with less than 33kW motor power! Have fun looking! There are some, but rare!) but in the case of electric drive, the more power you want, the bigger and heavier your battery gets, which in turn weighs the car down again and you need still more power. For example, the REVA car which is really a pretty good vehicle, weighs 400kg empty, plus 260kg in (lead acid) batteries, amounting to over 1/3 of the total vehicle weight!

But conventional cars do have the big advantage that they are widely available and cheap (I got my Opel Corsa test car for 400 EUR) and already have everything in them that you need. And (initial) cost is still a major issue that is affecting the spreading of EV's. So, I'll try to build a cheap electric car, even if not as efficient and innovative as I might like to.

Used Components

The Plan

The plan is to strip down the car (removing only the motor, exhaust, tank, radiator, and all the other gasoline-related parts) and hopefully get the weight down below 600kg. Then add the drive components, watching the weight. My controller weighs 15kg, my motor 38kg, which will hopefully amount to about 650kg without battery. The battery weight will depend on which type I choose, I'll probably start with lead-acid simply because it's comparitively cheap (1500 EUR for 10kWh) but heavy (260kg), along with the poor Peukert factor, severely reducing the effective energy that can be drawn with high currents. Plan B will use a Lithium battery, either Li-Ions (maybe Thundersky, LCP or LFP) or Li-Polymer (i.e. Kokam). The Lithium battery will weigh only 100kg with the same energy capacity, deliver much more of it's energy even at high currents, but also cost a whopping 5000 EUR, more than 3 times as much as lead.

After I was already well into his project, and had already bought motor+controller, I found out that another guy in Australia has already successfully done a similar conversion: http://evalbum.com/1149 The main differences are, I will be retaining the gearbox (if only as an easy connection point for the motor), I'll be using Lithium batteries (much lighter, only 82kg for 11kWh) and the car is somewhat smaller & lighter, and more aerodynamic. But my motor+controller are smaller too.

Lightweight Donor Cars

  • Opel Corsa: 780kg
  • Citroen AX: 690kg
  • Fiat Panda
  • Suzuki Carry (Small Van): 775kg
  • Daihatsu / Piaggio

Controller

The controller is meant for 3x400V or 3x500V AC input, but also has connectors to the intermediate circuit (after the bridge rectifier) which has an operational range of 484...778V. So this is the range my battery voltage must always stay in, under heavy load and during regen charging, to avoid the controller switching off because it detects a fault. The middle (average) is 631V, which should be the nominal battery voltage.

Battery

Using 3.7V lipo batteries, 631V means 170 cells in series (16Ah each for 10kWh). Or using 12V lead batteries, it means 53 units (also 16Ah each) in series. This high voltage means all switches and circuit breakers must be made for 1000VDC.

A glance on Ebay reveals a best price of 33,61 EUR (plus VAT) for the cheapest lead-acid 12V/16Ah battery (a motorcycle starter battery, not best suited for cyclic use), amounting to 1781,50 EUR for the total pack.

I found a manufacturer for LiFePO4 batteries with quite inexpensive prices, and I'll probably order a pack to see how they work out.

  • 1C discharge: 56 mins to 3V, 63 mins to 2V
  • 2C discharge: 26 mins to 3V, 31 mins to 2V
  • 3C discharge: 20 mins to 2V

Motor

The motor is a 400V 3-phase industrial motor rated at nominal 7,5kW at 3000rpm. The nominal torque is rated at 50Nm (0-3000rpm). Like all electric motors, it can be overdriven for short periods of time. The max torque is rated at 2.2 times the nominal torque, which amounts to 110Nm. At 3000rpm this would result in 16,5kW of power.

Talking about power and torque... Notice this power diagram from a 2008 Opel Corsa (44kW motor) brochure. You will notice that the 44kW motor rating is only reached at about 5500rpm. How often do people really drive with that high rpm? Only during high acceleration (like, turning onto a highway) or going maximum speed. I certainly know that my girlfriend (who also drives a Corsa, just like the one I'm converting) rarely goes over 2500rpm and starts complaining if I go over 3000rpm. So she effectively only uses the 15-25kW range of her motor. And thats assuming she also had 44kW to start with... but her motor is the 40kW version (at 5600rpm).

Now, you'll also notice that the torque (red curve) varies between 70 and 88 Nm depending on rpm. Below 1000rpm, the motor has no torque because it will stall. (Thats why ICE engines need a clutch, while electric motors don't). I expect to get over 100Nm when accelerating, which means my electric conversion should actually perform better (at least in some situations) than the original.

Progress Report

Sept 5, 2008

I'm delighted with the design of the Corsa, no wonder it's so popular. Without ever having seen or worked on one, I was able to remove the motor (incl. generator and starter) and radiator in only 3 hours with only a handful of tools (screwdrivers, wrenches and socket wrench sizes 10-19). Nothing else! (Except of course for a hoist to lift out the motor) - Anyway, the old motor is now out and I stood the new motor inside to see if it will fit. It will :-)

Sept 6, 2008

Layed a new 16A/3-phase cable to my garage and connected up the controller, motor and notebook, and built a small control panel with on/off, direction switches and a speed pot. Without an incremental sensor I can't use the CFC mode, so I reconfigured the controller to use VFC mode, speed controlled. The motor runs, but the controller turns off with a fault when I turn down the speed again, because it wants braking resistors to "burn" the excess current when braking the motor. Also pressure washed the car under the hood.

Sept 7, 2008

Today I removed the exhaust pipe, gas tank and heat shields from the car, and found how to turn off braking in the controller config. Now I can control the motor speed from 0-3000rpm without any faults. Time to figure out how to mount the motor in the car! Once thats done, I can "drive" as far as the power cable reaches (about 30m) and if that works out well, I can start thinking about what battery to get. Obviously the front end rides much higher with hardly any weight in front.

November 2008

Update: I now have a load of 168x 20Ah LiFePo batteries waiting to be tested/installed. At the moment I can only work outside, and I have another fulltime project, so I probably wont be continuing work until April 2009.

May 4, 2009

Now I have more time, and the weather is nice, and I can even work indoors. Perfect for continuing my work! Today I started assembling the battery cells into handy packs of 15 cells in series. I had to experiment with different methods of making the connection. Initially I was going to solder them, but that proved to be more difficult and unreliable than I thought. Then I tried clamping with screws, but the torque applied to the foil contacts when tightening the screws is not good (danger of tearing). I'll probably go with rivets and washers, similar to the last picture. The holes in the foil cant be drilled either, they must be punched, a 4mm paper puncher seems to work (lets see if it still works after 500 holes), otherwise i'll need to use a holepunching iron.

May 9, 2009

Today I mounted the motor. Thank God I had a forklift handy, as I had to lift the motor in and out about 10 times to make different changes to the adapter plate.

May 10, 2009

Today I hooked up the controller to the motor and ran a few tests. The flywheel is a little off center (about 1mm) so I'll have to take it back to the machine shop and have it centered. I pulled the handbrake, put in the 1st gear, and gave it a little power, and to my delight it easily pulled the car over my workshop floor, with locked rear axle. And not only forwards, but backwards too.

So now I need to mount the controller in the car somehow, wire up a pot to the accelerator pedal (I've ordered one here: http://www.fourthgen.net/potentiometercart.shtml) and get the batteries wired up. Also of course there is then the question of a DC-DC-converter to get 12VDC and a charger that can charge upto 600V...

May 13, 2009

I found a good solution for the battery contacts. I had a strapping set for strapping large parcels with 1/2" nylon ribbon, which includes a crimp tool for 1/2" metal crimps. After successfully crimping a few cells together with crimps, I found that no clamp was needed, so now I just roll the battery contact foils together (folding them about 3-4 times) and then just crimp them directly. Sure saves time and effort, and weight. Also the contacts can be seperated again in case I need to replace a cell. I still need almost an hour to make one 48V block (3 done, 8 to go).

Also I found that the 4mm aluminium sheet I used to mount the motor with, is not strong enough even for testing. So I'll get a 10mm Al sheet and start fitting that in today.

May 15, 2009

Today I successfully installed the motor using a 10mm aluminium plate. This time the results were much better, the motor runs nice and quietly and hardly off balance. This way I ran it at 2000rpm without being afraid it would break something. Still experimenting with controller parameters, it seems in some situations the motor tends to "run away" when the load is suddenly changed (i.e. when pushing the clutch). The SEW hotline support is very helpful and made some suggestions on what parameters to tweak.

May 17, 2009

Today I finished assembling all 11 battery packets and started installing them into the car.

May 18, 2009

Finished installing the battery pack (well, in a makeshift bread crate) and took the car for a first test drive up and down the driveway. Now I can turn on the regen function again and take advantage of regen braking (which works).

Ordered 2 13.8V/25A power supplies and will hook them to the battery across a center tap, to generate 13.8V/50A to charge the stock battery and operate the 12V system. Looking for a big 12V operated relay for the main battery switch. And when that and the potbox is installed, I can take it to the TÜV and see if they will allow me to register the car, or if I need to make any changes. (which I'm sure I will... I'll just have them make a list). Also there I can finally weigh the car and find out if it's any lighter than the original.

May 19, 2009

Still need the following components:

  • DC-DC-Converter for the 12V supply (ordered something to try, see above)
  • Potbox to control the throttle with the gas pedal (on its way)
  • Charger (need to rig something up, most solutions are too expensive)
  • Battery Meter (or at least Current & Voltage meters), found some cheap ones from China on Ebay: http://shop.ebay.de/merchant/sureelectronics?_nkw=dc+amp+meter (better of course would be a capacity display, which required summing up the current & voltage over time, the ones I've found so far cost 200-500 EUR)
  • Inverter (sine wave generator to provide a 230V AC outlet in the car) - not necessary, but nice to have
  • Contactor for 600V/50A with a 12V coil... not easy to find :-(

May 20, 2009

Today the 12V PSU's arrived, will try hooking them up to 250VDC and in parallel tomorrow.

Regarding the charger, I modified a PC power supply to provide 700VDC (by voltage doubling). Now I need to find a way to regulate the current (preferable µC-controlled)

I also ordered a contactor, and 2 cheap digital volt/amp meters from China, to serve as long as I dont have a battery meter yet.

May 24, 2009

After spending the last 2 days with my family, I'll continue working today. Unfortunately I cant do very much because I'm still waiting for the potbox, contactor, and volt/amp meters to arrive from India, China, and Germany. But I can re-install the stock 12V battery and wire up the DC-DC-converters and see if they do a good job of supplying the car's appliances (lights, fan, windshield wiper, etc.) with power. I think I'll mount them under the passenger seat for now. I had a closer look at the controller, wondering if there was any way I could reduce the necessary space, since half the unit seems to be just cooling fins and fan. Ideally I'd like to disassemble it and screw the power unit (IGBT's) directly (maybe via an aluminium plate) to the vehicle chassis for cooling. But that looks like too much trouble and not really worth the space saving for now. Maybe I'll look for a different type of controller next time.

May 25, 2009

Turns out the weather was just too nice to do anything yesterday :-) but today the Potbox came, so I'll probably install that. (Although the weather is even more tempting today... 30°C sunshine and my pool needs some work ;-)

First things I notice when I unpack the potbox:

  • The pot cable only has 2 connections... You need 3 for a reliable signal, and the better controllers have a failsafe circuit that wont work without all 3 connections
  • There's a lever to hook the gas cable to, but no cable pinch and no counterholder part... so I have to rig something up mechanically. Thats not what I had in mind, or I would have built the whole thing from scratch. So basically the box is just a pot with a spring-loaded lever and an end-switch on it, and I even have to open the sealed box to get at the third pot contact. Thank god there is one... they sanded off the pot specs but luckily they didnt cut off the third lead. I have to admit I'm disappointed, I would have expected more for 65$. It's even a bitch to open, screws overlap in the wrong order to be tightened easily.

May 27, 2009

The potbox is installed. I connected all 3 wires to the pot and now I can finally drive the car with my foot instead of with a knob :-) In the process I found that the pot is really a 20k pot that only rotates about 20% of the possible range. So the resulting voltage swing is only 0-2V instead of 0-10V which is what we want. Luckily we can reconfigure the controller to accept 2V as the max throttle setting.

Car still drives very jerkily. I'm still experiementing with different modes and parameters in the controller. But I sort of wanted to be able to monitor the battery a little better (which I still can't charge yet) before I do too much.

May 28, 2009

Today the volt+amp meters came. But unfortunately the seller didnt tell me that the inputs and the supply voltage apparently aren't isolated, so a) I can't use 3-4 battery cells for the supply voltage as I was planning, and b) I cant even use the same 12V source for both the volt+amp meters. Great. Do they really expect me to use 2 seperate batteries??? I'll order a couple of isolating DC-DC-converters to solve the problem. Not a huge problem, but further unnecessary delay and added costs (another 12 EUR per meter).

June 02, 2009

Today the contactor finally came, so now I can turn the inverter on+off with the ignition switch. Since regen also goes through it when braking, it's important to have a utility contact to tell the inverter to "immediately stop" all actions, or the switching transistors might be damaged.

June 04, 2009

The last few days I was mainly working on a charger circuit. Since I couldnt find one for a reasonable price, I decided to build my own 3000W switched mode power supply (SMPS). Today I finished the circuit and ordered the parts, hopefully I can start assembling and testing next week.

Also today the DIO card came, I ordered one for the inverter to supply a second analog input, which will enable current regulation mode (instead of voltage to frequency). I hope to get smoother driving that way. If that doesnt woirk, the last step is then to get an encoder and switch to CFC mode, thats the surest way it'll work, but unfortunately also the most expensive route.

August 28, 2009

File:23chargers.jpg The last 3 months I was pretty busy with other stuff and I didnt have much time to work on the Corsa. Also, after the DIO card didnt work as I expected, I decided to look for an encoder, which took until July to find a suitable one. I was also a little demotivated with the charger diffuculties and so the Corsa had a break. Now I have an encoder (Heidenhain ROD436, 1024 steps) and I'll try fitting it and hope for better results. Also I thought of a makeshift way to charge the battery by daisychaining 23 (!) 24V power supplies together (1A regulated, switchable voltages 9...24V) and charge the battery that way. If I closely monitor the voltage & current, I can adjust the total voltage by switching down a few of the PSU's a few steps and then slowly increasing it again by 4V at a time when the current falls below 0.5A. That solves another one of my worries, running down the battery too much while experimenting, before I can find a way to recharge them.

August 29, 2009

Today I mounted a new stronger motor mounting bracket, replacing the flimsy temporary one. Also started mounting the encoder on the rear shaft. The battery is now fully charged (about 530V floating voltage). After mounting the encoder and the amp/volt meters on the dashboard, I can start some test driving and experiment with controller programming. Next week I'll have little time for working on he Corsa, but after September 7th I should make more progress. I really wish I had some kind of logger for battery voltage/current, wouldnt even be very complicated (a simple microcontroller would do) but i dont have time to build one just now.

September 11, 2009

Now things are rolling :-) finally I had time to spend a full day on the Corsa. I finally wired up the main contactor to the 12V ignition switch, mounted voltage+current meters on the dashboard (or rather, in the glove compartment because they aren't very bright), fine tuned the controller ramps so driving is more natural, picked up some temporary license plates and for the first time took the car on the road.

Acceleration in 3rd gear is satisfactory, even 4th gear is ok for cruising, I set the RPMmax to 3000rpm to start with. Starting up from standstill without violent jerking of the motor is a little tricky in 3rd and almost impossible in 4th gear, so still need some work here. One major issue I encountered is, the only parameter for controlling the braking intensity is the down ramp. It always goes from 100% to 0% in n seconds, at the moment its set for 20 seconds for a natural driving feeling (feels like ICE motor braking when I let off the gas). But if I slow down faster than the ramp is set (like from 3000 to 1500rpm in less than 10 seconds) by using the footbrake, the car actually tries to speed up again (against the brake) to regain it's ramp. This behaviour is an absolutely no-go for street driving, so I'll have to find a way to fix that. (The only way at the moment would be to turn off regen braking alltogether, but I really want regen braking).

Also I finally found an affordable 12V electric vacuum pump for the footbrake amplifier, a part from an Audi S6, for 85 EUR. Although I had the impression that the "emergency" functionality of the footbrake (without vacuum) should be fine for driving (brakes just fine as far as I can see, just without amplification), after talking to the TÜV engineer, I learned that it would not be accepted for road safety without either a) a footbrake installation that wasnt ever meant for vacuum operation or b) an electric vacuum pump. Also this week I finally got a suitable charger from a small German company that can provide 560VDC from 230VAC at 2-3A output. I tried charging the battery after yesterday's drive and it seemed to work great.

Mounting the encoder is still a little tricky without a professional machine shop, hopefully I can finish that in another day. That should solve a lot of the controller behavious problems.

Oh yes, while driving the battery current went upto 30A while accelerating and upto -20A while braking. The voltage dropped to about 455V under max load. (As far as I could see while driving - I still need to hook up a laptop for data logging to log & graph these values).

September 14, 2009

Today I mounted the encoder, configured it on the controller, and... what an improvement! It was even better than I hoped, acceleration now smooth as silk, no unwanted acceleration anymore, behaviour just as you'd expect a "gas pedal" to behave. It was well worth the touble and cost. Maybe other controllers (Danfoss?) can be used without an encoder but the SEW really needs it. Also took the car for the 6km drive to the TÜV and applied for registration, accompanied by a camera team from the German TV channel WDR, who are reporting on my progress. I didnt expect it to fully pass immediately, and it didnt, but the to-do list is sensible and quite doable and I'm confident I can fix at least most of it in the next 2 weeks.

September 28, 2009

Today the camera team and I took the car to VDE, the German regulatory institute for EMI compliance, and had the car tested for EMI and hazardous area protection according to IPXXB (which tests if a "standard human finger" can touch any dangerous parts (moving, hot, high voltage). It passed both tests after minor modifications on the fly, which I was prepared for. This (mainly the EMI) was my main concern regarding TÜV compliance, since if the controller or motor had generated too much EMI which couldnt be fixed, that would have been a major hurdle. Unfortunately it was also very expensive (over 1000 EUR), although the VDE was extremely cooperative and gave me a very special price. But since thats passed now, the rest of the requirements are "peanuts" (things like a fuel gauge, lighting, blower-heater, buzzer, somewhat better fastening of the battery box, weighing the car, vacuum pump for brake assistance, etc.)

September 30, 2009

After some more extensive driving tests, I've discovered that the 4-pole motor was not the best choice, since its pull is rather poor over 2500rpm which means the max speed in 4th gear is 80km/h and in 5th gear its around 100km/h. But the torque is a dream!!! I can drive it under normal conditions ALWAYS IN 5TH GEAR! (Even acceleration from a stoplight in city traffic is sufficient!) As a result its always nice and quiet (low RPM in gearbox) and as long as 80km/h is ok, thats a pretty good setup. The clutch is definitely not necessary, I had left it in but I never use it, changing gears (normally only done between reverse and 5th when standing still) works fine without clutching at all. So one of my next mods will be to remove the clutch and connect the motor shaft directly to the gearbox. Also weighed the car, it now has a curb weight of 860kg, which means it got only 25kg heavier. Although I doubt the car really only weighed 835kg before, since I only added a 40kg motor, 80kg batteries, 16kg controller and maybe 10kg of cables, metal frames and nuts+bolts. But I removed lots of stuff (even the old motor controller, the indoor hotwater heating element and the spare tire) so I'd be surprised if if got any heavier. But I saved every screw that I removed so I'll toss all that onto a trailer and weigh that seperately :-)

October 01, 2009

Today's the big day - after completing all the required modifications, the camera team and I will take the car to the TÜV again for the final inspection. If everything goes well, I'll have a certificate this afternoon which will allow me to register the Corsa as a normal road vehicle.

And hey presto, it worked :-) now I have the TÜV certificate and can go register the car with normal plates.

October 27, 2009

Today a TV program featured my (and another) conversion: www.wdr.de/tv/servicezeit/mobil

Retrieved from "http://rc-autopilot.de/wiki/index.php/Opel_Corsa_Project"