Stickers!

I am spending Christmas in Wyoming with family, so no work on the car lately. I have been debugging the BMS code, and designing a sticker for the car:
The equation is the differential form of Faraday's law, which is one of Maxwell's equations, which describes induction. If this seems like something that belongs on your electric car, I can send the full size image.

Battery charger and other stuff

It has been over a month since I last posted to this journal. I have been very busy at work, where we just launched our first product- a miniature XRD/XRF instrument for identifying minerals in the field. Even so, I have made some progress on the car, mostly buying parts and working on the next version of the BMS. I am now the proud owner of a PFC-20 battery charger from Manzanita Micro. This charger is nice because it can output the 408 volts that my pack will require. One drawback is that it is not isolated, which means that the negative side of the battery pack is connected to one leg of the mains. This should not be a safety issue, as all the other equipment I am using is fully isolated.

Oh yeah, I said next version of the BMS. Although I'm sure the voltage regulator style BMS that I already developed would perform just fine, I have gone ahead with the Eierlegende Wollmilchsau (literally, an egg-laying woolly milk pig-- a German expression that can better be translated as "Swiss army knife") version. This version has a small microcontroller (an Atmel ATtiny25) on each cell that monitors the cell voltage and communicates with a tiny linux computer over an optoisolated i2c bus. Each microcontroller can bypass current, either on its own or on command from the computer.

The advantage of this system is that it gives individual information on the state of each cell. Also, if a cell controller fails, the computer would immediately notice, and alert the driver that the pack needs service.

The other positive aspect is that each cell controller will cost less. Microcontrollers are truly amazing. For under a dollar, you can get a chip that is much more powerful than, say, the Tandy Color Computer 2 that was my first computer.

Mr Fusion

I've started buying the safety equipment needed to deal with a 300 Volt battery pack with something like 15 kilowatthours of energy (roughly enough to run our house for three days.) In addition to some fancy insulated screw drivers to keep from shorting things out while I'm working, I bought this monster fuse. It will go between the battery pack and everything else. Lens cap for scale.

British racing green

Well, I've finally fallen into the biggest trap for people who have electric car websites-- I've been working on my website instead of my car. At least reading my journal (I'm allergic to the word "blog") won't make your eyes hurt so much anymore. Plus, it's now authentic British racing green.

BMS, round two

I received and assembled the first three samples of the BMS boards. Everything looks good. I did some thermal tests, and each unit should be able to bypass one amp with a heatsink installed on the main transistor, and about twice that with a fan in addition. If I use the BMS to cut back on charging current when any cell goes over voltage, the bypass current should only be a few tenths of an amp, and I'll be able to dispense with the heatsink. That would be nice, as the heatsinks are a fair amount of work to install. Here's the top of the board, with heat sink installed:And here's the back. The cool thing about making your own layout is that you can write whatever you want on the board:


Only 93 to go...

Ol' sparky

I have finished figuring out where the batteries will sit. The spare tire well is almost perfectly sized for the pack. I will probably relocate the spare underneath the car in place of the gas tank.

While I was looking things over, I learned why British cars are infamous for their dodgy wiring.

Nothing like unsupported wires rubbing against the gas spout... and is that a twisted-together "connector"? I hope that was an aftermarket accessory. The factory connectors look pretty smashing as well:


I'm hoping these practices won't rub off on me when it's time to do the high-voltage wiring

BMS round one

I received my first prototype board from the fab house this weekend, and soldered it together.
The circuit worked the way it was supposed to, so I have ordered 3 more "final prototypes" that will have a few small changes, as well as a more polished look, with solder mask and silk screen and a slick oval shape. Once those arrive, I will probably borrow a few batteries and test the whole system.

TPS Report

Throttle Position Sensor, that is. I need a way to communicate to the motor drive that I want the car to go. Traditionally, electric car builders have used something called a "Curtis potbox", which is a variable resistor in a box, controllable by a cable hooked to the accelerator pedal. Trouble is, this device is made for controlling forklifts, and so isn't really suited to the abuse of thousands of stop-and-go cycles in a car.

Luckily, (newer) cars already have something very similar built in. The throttle position sensor tells the engine management computer to adjust spark timing and fuel mixture according to how much go the driver wants. It should be easy enough to adapt it to my purposes. The best part is, I got this one on ebay for $5. I think it came from a Ford. I 'm gonna keep all the useless-but-cool gasoline-related accessories on the module.

Batteries...

With the bulk of the mechanical work done, it's time to start thinking about the electrics. I've chosen the Lithium Iron Phosphate batteries from Thunder Sky, which don't catch fire like some lithium batteries do when overcharged. Nevertheless, the batteries can be damaged if they are overcharged or overdischarged. In a long string of cells wired in series (I'm using 96) some cells are inevitably stronger than others, and will end up being overcharged. The string voltage remains constant, so this means that some cells get severely discharged (even negatively charged). Both of these destroy the affected cell, so I need a way to keep the voltage of each cell between 2.5 volts and 4.25 volts. There are several ways of doing this. Charging each cell individually, then monitoring all the cells for undervoltage during discharge is the best way, but constructing a charging system to individually charge 96 cells, then connect them all in series to the drive circuitry would be a difficult design challenge, as well as a wiring nightmare.

Instead, I'm taking a simpler approach. Each cell will have a shunt regulator across its terminals. A shunt regulator is a device that conducts no current until the voltage across it reaches a set value. When the voltage rises above the set value, the device begins to conduct just enough current to maintain the voltage at the set point. This takes care of the overvoltage protection. Undervoltage is a bit trickier. I think I've come up with a clever way to signal the controller to shut off when the voltage of any cell drops below the setpoint using a single circuit (instead of 94), but I need to develop it a bit first.

The idea of building 96 circuit boards is a bit daunting though. Each board will have ~10-15 components and cost about $5 for parts. It only adds about 5% to the cost of the battery system, so is well worth it, but spending my evenings hunched over a soldering table doesn't sound like fun. Maybe I will send it out for assembly... with 100 pieces, the prices should be pretty good. I just need to prototype one or two to get everything right.

Back on track

I finally finished machining the new hub, tearing apart the clutch to get the spline, welding the two together, and putting the motor back in. I fired up the motor, and everything worked fine. I enjoyed sitting in the car listening to the motor & transmission turning for a while...

Setback

I finished the subframe and got the motor properly aligned with the transmission (not so easy) and bolted everything down for a live test (without the driveshaft installed-- no driving through the shop wall). I rigged up a 300-volt DC power supply to run the motor drive, just to make sure it wouldn't complain about running off of batteries instead of its normal diet of 3-phase 240 Volt AC. It worked great. On to the powered test of the motor/transmission assembly. I ramped up the speed of the motor slowly, and everything was fine. I took things up a notch by quickly accelerating the motor, to simulate an actual load. Suddenly, the adaptor hub that I had spent a few evenings turning on the lathe twisted right in two. Oops. I had followed the design guidelines given by Lovejoy, the shaft locker supplier in machining the hub. The lock holding the transmission to the adaptor hub works by compressing a thin tubular section of the hub onto the shaft. I had my doubts that such a thin tube would be strong enough to carry the 150-220 ft-lbs (200-300 Nm) of torque Lovejoy advertised. It wasn't. Although, strictly, the shaft lock device was still securely in place, so I guess the book value was right. That's a little bit like building a ship in a bathtub, though...

Nevertheless, I shall continue! The next plan is to snag a splined hub from an old clutch disc, and attach that to the hub that is still locked on to the motor shaft, using bolts or perhaps my new bowhunting TIG welding skills.

It's a motor!

After quite a bit of sweating and head scratching, I got the motor into place.
Don't worry-- the C-clamps are temporary. The engine bay looks cavernous, but soon enough it will be crammed with batteries.

This style of design and fabrication is quite a change from my job, where everything gets designed using computer tools and precisely manufactured. This is more the "Orange County Choppers" school of metal work, only with a lot less swearing and throwing things. I sometimes wonder about those chopper shows on TV, whether they have a whole sweatshop full of 3D CAD guys sitting in a poorly lit room somewhere, swilling mountain dew and churning out chopper designs.

Subframe

To make the motor, transmission, and car more or less one unit, I need to build a subframe. This will also serve as a platform for the front battery boxes. It was difficult to measure everything outside the car, so I decided to install the transmission, block it up to where it should be, and build the frame to it. Here's the transmission in place with the first cross member:
This is my first stab at TIG welding, so the welds aren't going to win any beauty contests. I'm happy with the overall result, though. Next, the top half of the subframe:

Mr Motor, meet Mr Transmission

I have been out of town for the last couple of weeks, so I was excited to work on the car all weekend. The shaft locks arrived while I was gone, so I set to work attaching the motor to the transmission. I had one small setback. The motor is a metric 112M size, which means that it is 112 mm from the base to the center of the shaft. It also should mean that it has a 28mm diameter shaft. For some reason, though, this one had an 1 1/8" shaft. This was a problem because the shaft lock I bought is 28mm. No worries, though. 28mm is smaller than 1 1/8", so I just turned the shaft down on the lathe.
This was good, as it gave me an excuse to take the motor apart.

Next, I test fit everything, and ran the motor for the first time attached to the transmission


Everything ran smoothly, so time to start putting it all in the car!

Shaft locking

Since I finished pulling the engine, I have been working to source parts to couple the electric motor output shaft to the transmission input shaft. Instead of using a splined or keyed shaft, I will be using keyless shaft locks from Lovejoy, combined with a hub I will machine myself. The keyless shaft locks work by drawing two concentric wedge profile rings together with jack screws. As the rings come closer, they wedge against the outside of the shaft and the inside of the hub, to form a very tight fit. I will post some pictures of this once the parts arrive.

Engine pull day two

Here is the sequel to last week's movie, where the engine finally comes out. It took a bit of work to get the tail of the transmission over and out, but there were no mishaps.

Engine pull day one

Today I started pulling the engine. I nearly finished, but the hoist I scrounged up was not strong enough to lift the engine and transmission. I am doing the work at TechShop, down the street from my house. It's nice to have a place with enough space to work and all the tools I need. The big engine hoist they have there needed some repairs, so I will finish the job later this week. Meanwhile, check out the time-lapse video from today:

Thoughts about energy

Since I've been thinking about building an electric car, I've gotten a lot of questions about them. I'm far from being an expert on the subject, but I have picked up a few answers as well. I think the most frequent question is something like "Can you put some solar panels on the roof and make it run that way?" The answer is yes, but the car would be slow even by classic British standards. For comparison, an average-sized solar panel makes about as much power on a bright day as a healthy adult in a hurry on a bicycle. You could maybe fit three of these panels on a car without looking ridiculous or getting in trouble with the law. That doesn't mean it is a stupid idea. Solar car races (I guess "rayces" if you're into that sort of thing) are pretty cool. And then there's this guy if you prefer the Mad Max look. The best thing is to have a solar powered carport to park your car under while you're at work, or just a roof full of panels at home. Even if you get all your electricity from the power company, though, electric cars are an improvement over the gas variety from an environmental view. The movie "Who killed the electric car?" presents arguments from this side quite well.

Engine sold

I listed the gas engine on Ebay last week, and there was a fair amount of interest. Some in the motor itself, but also quite a bit in the idea of an electric conversion. I'm glad I started this journal, so I could point to it as a place to begin learning about electric cars. The auction ended last night, at a little over two hundred dollars. I'm mostly happy to have the engine out of the way so I can begin the conversion, but I don't hate two hundred dollars. That's almost two batteries!

Bench test

I tested the motor + drive system last night, and overall it was a success. I ran the motor up to 7500 rpm with no problems. The cooling fan sounds like a jet taking off at that speed, though. It's nice to have a lot of cooling air, though. Unloaded, the motor consumes about 750 watts at top speed, which seems reasonable. I can run the motor down to about 30 rpm, where it uses about 50 watts. It would have been nice to get an idea of torque output, but I don't have an easy way of loading the motor.

The motor controller has two modes: V/f and sensorless vector. At its simplest, V/f mode makes the voltage to the motor linear with frequency, so for example, to run the motor at 30 rpm, the drive would supply 10 Volts at 1 Hz, and at 7500 rpm it would supply 230 Volts at 250 Hz. This is fine for constant load applications, but if the load increases too much, the motor can stall. In sensorless vector mode, the drive uses a mathematical model of the motor to calculate the torque applied at any time. It can then change the voltage and frequency to match a requested torque or speed. This is much better for an electric car, because motorists are used to an accelerator pedal that requests more torque from the motor. Because the motor's rated capacity is much smaller than the drive's (5 HP vs 30 HP) the drive was unable to model the motor, so sensorless vector mode is not working. I will see if I can input parameters specific to the motor and get it working. Otherwise, I will probably need a bigger motor eventually. In the meantime, I can set the acceleration and deceleration times for the drive to be long enough so that everything works. This just means that the drive will take the car's inertia into account a little better, and won't ask the motor to turn at 7000 rpm when the car's speed is telling the motor it must turn at 500 rpm.

Now to get rid of that pesky gas engine so I can start putting things in the car...

The motor drive is here!

I ordered a motor drive last week from the internet. In the past, these were the kind of things you would have the purchasing representative from you company call the regional sales rep about, a price would be quoted, and a lot of paperwork generated. Now, there are a couple of places to get industrial equipment on line. I bought this one from factorymation.com, as they had brands that I had heard of before. Whatever that's worth.



The drive weighs about 30 kg.

The business end: This is where the cables from the batteries and to the motor will run. The fans dissipate waste heat. The unit is about 90% efficient, but that's still nearly 2kW of waste heat at full power.

DC or AC?

Most electric cars built by hobbyists (see http://www.austinev.org/evalbum/ for a nice collection of home-built and other electric vehicles) run on DC - direct current. It's pretty simple - get a bunch (4-15) of batteries, hook them together, and run a motor with it. Speed/torque control is done by limiting the current through the motor by an electronic chopper (Pulse Width Modulation). AC systems, on the other hand, need higher voltage (300 Volts DC for a 240 Volt motor, or around 600 Volts DC for a 480 Volt motor) This means a string of 26 batteries to run a 240 Volt motor. Also, instead of a single switch to control current, as in the DC case, AC motor drives need to have 6 switching transistors and a bunch of digital signal processing logic to produce the three sine waves needed to run a 3 phase motor. AC Propulsion and Siemens and a few others make automotive-type AC drives. See http://acpropulsion.com/ and http://metricmind.com/for examples. These go for about $20k, so they're quite out of most hobbyists' reach. So why bother with an AC conversion?

• Pure snobbery
• Simpler, more efficient motors
• Regenerative braking
• Higher voltage means lower current, which means lighter wires and other components

An interesting industrial-component AC system from Australia is here:
http://www.austinev.org/evalbum/1149

Here's the new motor. It's a 5 horsepower (3.75 kW) 3 phase, 240 volt induction motor (framesize 112M -- ebay $150).
My approach with this conversion is to use off-the-shelf industrial components as much as possible. Industrial AC motor drives are coming down in price significantly these days, and AC induction motors have always been cheap compared to their DC counterparts. This motor was manufactured in 1980 and apparently sat in a warehouse for the last 27 years, as it was brand new in the original crate.

The old motor

Here's the old gas motor, which is for sale, by the way. Abingdon's finest.

Begin at the beginning

Here's the car. Not exactly Kermit-green, but kinda looks like a frog. I'm a big fan of cars with names. As far as I know, this car doesn't have a name yet. So. Little green MG, I dub thee Kermit The Car.

Why?

I've started this journal as a way to keep track of my progress converting a 1970 MGB GT to run on electricity. Why a pushing-40 British car? Why rip out a perfectly good gas engine and replace it with something that belongs on an air compressor? I don't know. I probably won't know when I'm all done.