Bryan and Scott's Porsche Work

Over the past few days Bryan and I have been working on the wiring project for the Porsche. The wiring under the car and in the mid-car section was a complete mess, and we worked on cleaning it up to make it look better and make it easier to work with in the future.

Bryan holding a bundle of wires from the middle of the car

To start out, we wired up a terminal box in the front and back of the car that we'll be using to rewire the mess underneath and inside the car.

After we finished wiring up the terminal boxes, we started redoing the wiring inside and under the car. First we pulled out the DC/DC converter and decided to mount it underneath the terminal box in the back of the car. I designed and built a mount for it using sheet metal. I also made room on the mount to attach a relay so that when the car is plugged into an AC outlet, the DC loop is disconnected. The mounting device was also designed so that one cannot easily stick their fingers in any electrical components and get electrocuted.

This is a picture of the wiring on the DC/DC converter.

While I was working on this, Bryan and Dan first had to analyze where each of the wires in the bundle from the DMOC were actually going. After this was documented, they removed the unnecessary bundle and started rewiring.

They are currently working on wiring everything through the terminal box, including regenerative braking and accelerator wires. This will not only make everything look a lot cleaner and more streamlined, but make access easier for future work.

More pictures to come soon!

Over the past few weeks I have been in the process of designing the drive train for the elEVen vehicle. During the process I explored several different options including using a transaxle or a speed reducer in conjunction with a differential. I eventually settled on a direct drive system using a Ford 9” differential with 7.3:1 gears. This reduction ratio was chosen to give the car a top speed of 100 mph while minimizing the 0-60 time. By using a direct drive system we will be able to minimize the complexity and weight of the drive train.

The Ford 9” was chosen due to it’s the large aftermarket community and wide range of reduction ratios. Because the Ford 9” is a solid rear axle differential several modifications will have to be made to install it in the front independent suspension. These include mounting the differential upside down and using a independent rear suspension carrier.

Since deciding on the Ford 9” we went ahead and ordered a rebuilt 9” Trac-Loc limited-slip differential assembly with 7.3:1 gears. We also ordered an independent rear suspension carrier made by Dutchman Motorsports. Since the IRS carrier came with Porsche 930 CV stub axles we went ahead and ordered matching CV joints for our half shafts. Below you can see the assembly, carrier and stub axles assembled.

As I mentioned before, the elEVen is going to have direct-drive system with no gearing between the motor and differential. The coupling will be done using a u-joint and a rubber flex-plate. The u-joint will allow us more freedom when mounting the motor, while the flex-plate will dampen out oscillations caused by any misalignments in the drive train.

The u-joint will be assembled out of a u-joint style slip yoke that will slide over our motor spline shaft and a custom adapter yoke, which I will be machining next week. The adapter yoke will have a 1310 size u-joint yoke on one side and a flat flange that connects to the flex-plate on opposite side. On the other side of the flex-plate will be another custom coupler with two flanges, one to connect to the flex-plate and one to bolt to the differential pinion. This coupler will also have to be machined by hand next week. Today we placed the order for the 4140 steel round stock for the pars so hopefully I will be able to get into the machine shop Tuesday or Wednesday to start making them. So far I have made some rought solid models of both couplers, which I will be refining as I discuss my plans with shop guys. Below on the left you can see the model of the adapter yoke and on the right the differential coupler.

Here are both parts attached to a model of the flex-plate.

Monday July 6, 2009:
Charger:
In the afternoon, I also looked into what charger we want to get for the elEVen car. Chargers cost between 2000 and 5000 dollars for a high powered good one so this will be a pretty significant purchase for the team. There are about a half dozen companies that make electric vehicle chargers. Right now, we are looking at getting a Manzinita micro charger. They are very robust and versatile and give the user a lot of flexibility.
In the evening, we had our weekly meeting and then Dan presented the results of his modeling work. According to the drive cycle and the modeling code he used, the Mercury Milan will have a 0-60 time of about 8 seconds and will average about 460 watts/mile. This mileage estimate seemed high to us so we are going to look into it more and make sure all the parameters in the model are correct.

Tuesday July 8, 2009:
Motorcycle Battery Pack:

Today I worked on the CAD model for the motorcycle battery pack. I am going to try to push to have another prototype of the enclosure done by the end of this week. I finished most of the CADing and just have to put on the finishing touches. Here are a few pictures. The black strips are rubber tap for vibration isolation, there are two barbed hose fittings to attach cooling air hoses to.

Wednesday July 1, 2009:
Battery Testing:
I finished testing another 1500 cells or so. That brings our total number of good cells to 2700. At this rate, it will take another 4 days or so to finish testing all of the batteries.

Motorcycle Pack Design:
Mike and I worked on the design of the motorcycle pack some more today. We plan to seal the four modules to keep water and other harmful elements out. In order to cool and vent the cells, we plan to have two barb fittings, one on each side of the enclosure that we will attach hoses to. These hoses will in turn be attached to a central manifold assembly with a high powered blower that will pull air through the pack. We think that by centralizing the point where air is exchanged with the outside, we can better control the airflow and prevent condensation or excess water or particles from passing through the pack and causing a short. We are going to revise our CAD model in the next couple of days and post a picture of the new design when we finish it.

Tuesday 4-30-09: Battery Testing:
The long tedious testing of the individual batteries has begun. After talking to our contacts at A123 and other EV teams, we decided that the two most important indicators of a battery's health are voltage and impedance. We plan to first measure the voltage of all the cells that A123 donated to the team and throw out any cell that are not between 3.296 and 3.340 volts. The cells have been sitting for 3 years so if the voltage is within this range, it is a good indicator that the cell is sealed well and that there is very little current leakage through the cell. This range was given to us in the A123 battery pack design manual that is available to the public. This design manual has a lot of very useful information about building a pack and particularly about building a pack with lithium iron phosphate batteries. Today I checked over a 1000 cells. My goal is to get all the cells checked before the car arrives next week. Today, Mike and I also went to the electrical engineering lab and used their impedance tester to check a box full of good cells (cells that passed the voltage test). The average impedance of the cells was 7 milliohms with a very small standard deviation. It seems to me, and the other team members agree, that we only need to measure the voltage of the cells in order to tell if they are good or not. In other words, if a cell passes the voltage test, it should pass the impedance test as well. If we decide that this is a reasonable assumption to make, we will be able to save a lot of time by just testing the voltages of all the cells.

Cell Level Battery Pack Fuses:
Today I ran a couple of tests to determine if the individual cell fuses are sized correctly. Our nickel sheet is .010" thick and the tab fuses are 5.5mm wide. A123 sizes their fuses to be 3.6 mm wide and .007" thick nickel sheet. They rate there fuses to blow at 2100amps in .1 seconds. Our fuses are currently about twice the cross sectional area of A123's. We wanted to investigate this further so we ran a few tests using an A123 26650 cell. We estimated that an A123 cell can output 300 amps in a short. We cut two test strips of nickel. One that was the size of the fuses that we are currently using on the motorcycle packs and one that was the size of the A123 fuses. Both blew when we shorted the battery across them. This was a good sign because the purpose of the fuses is to protect against a short. After talking it over, we decided to stick with the 5.5mm width rather than reducing it to 3.6mm because we do not want to risk the nickel heating up when we rapid recharge these packs. I went ahead and ordered enough nickel to finish the motorcycle packs and also build one of the car modules. We bought our nickel from National Electronic Alloys Inc. They have a very good selection and they also have good prices.

BMS Updates

After several meetings over the past week, we have decided to use a LTC6802 BMS IC as the main BMS circuitry for the slave boards. We will interface them to a main controller via an optically coupled interface. The main controller will be designed around an MSP430. The data sheet for the LTC6802 can be found here:

http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1037,C1134,P86662

More info to come soon...

-Eric

The Battery Pack

The past week or so has been spent designing a prototype battery pack for Lennon's electric motorcycle, eMoto (http://www.electricmotion.org). We'll be replacing his 240 pounds of old-school lead acid batteries with a 30 pound pack of A123 LiFe cells. Of course, this pack will only have about 75% the capacity of the lead acid stack, but the power and weight improvements will be significant.

This pack isn't a favor for Lennon though. It'll serve as a test platform for future pack design for the elEVen, as well as demonstrate rapid-recharge capabilities on a smaller scale. Consequently, the pack has a very similar architecture to that of the proposed elEVen pack. We've run into a lot of problems during the design process which we never anticipated. It's very likely we'll run into the same problems on the full sized elEVen pack, so hopefully we'll be able to get a lot of bugs out with this small scale version.

Here's an overview of the basic design, both for the elEVen and the eMoto:

Modularity: It is rather difficult to make an electric vehicle without using high voltage, which can be very dangerous to work with if proper precautions are not taken (just like gasoline in ICE cars. It's just a property of high energy systems). The obvious solution for dealing with the dangers of high voltage is to eliminate the high voltage itself. We are accomplishing this by building several independent low voltage modules, which will be wired in series to create the high voltage required only at the very end.
The eMoto pack will consist of two 19.8 volt (6 cells in series) modules and two 16.5 volt (5 cells in series) modules wired in series to reach the nominal voltage of roughly 72 volts (required for the motor controller). Each module will have 10 cells in parallel to achieve a capacity of 1.6 kWh for the entire pack. We originally intended to use three 24 volt modules, since the elEVen will be using 24 volt modules, however, we decided to use the existing battery mounting on the eMoto. This made arranging seven of the ten cell batteries in series very difficult. Thus, we arrived at the much simpler pack design we currently have.
We anticipate that the elEVen pack will consist of 24 volt modules with 72 cells in series. However, the issues we ran into with the geometric constraints of the motorcycle could arise when designing around the existing chassis of the elEVen. Hopefully the experience with the eMoto will help us keep our future designs a little more flexible. Additionally, the eMoto will have the four modules wired directly to each other. We plan to connect the elEVen modules with contactors, allowing the packs to disconnect and drop the maximum voltage to 24 volts. This will provide a safer working environment as well as making the vehicle safer if it is ever involved in an accident. Finally, each module will have a built in Battery Management System (BMS), which will then communicate to a master module to balance all of the modules together.

Inter-cell Connections:The individual cells of the modules have to be connected some how. We took a lot of advise from the likes of Bob Simpson (http://www.evdrive.com) and Bill Dube (http://www.killacycle.com/) and have decided to use nickel plates spot welded to all of the cell terminals as the interconnects. These nickel plates will be bent to help arrange the cells in each module. They also have tabs for individual cells with integrated fuses to protect the pack should a cell internally short.

Cooling:As we rapidly recharge the cells and continuously discharge them, they will heat up. Though warm cells have a lower internal impedance and are therefore more efficient, running the cells too hot will damage the internal structure and reduce the life of the pack. We'll be air-cooling the cells across the electrical contacts, since they are both electrically and thermally coupled to the inside of the cell. We don't anticipate that a lot of cooling will be required (see the section on battery data which will be posted soon), but we want our design to be flexible in case more cooling is required. We accomplished this by spacing the cells vertically to allow airflow. However, we have not finalized the method of spacing (the leading idea can be seen below). We're also looking into various methods of cooling, including passive cooling and using various arrangements of computer cooling fans for active cooling. We'll be testing the effects of various cooling systems once we finish the prototype eMoto pack.

Cell enclosure: This has surprisingly been one of the more difficult aspects of the pack design. The cells need to be restrained somehow, and in the case of the eMoto, the enclosure also needs to be able to take a compressive load. We were originally looking at using FR4 as the primary cell enclosure for its insulating and fire retardant properties, and an aluminum secondary enclosure to provide rigidity. Since then, we've looked at a single FR4 enclosure design, an FR4 primary and polycarbonate secondary, polycarbonate primary and secondary, and a single polycarbonate enclosure. Currently, we've decided on a 1/8” polycarbonate primary which will likely be encased in a polycarbonate secondary with possible rubber shock-mounts in between the two layers. The shock-mounting will reduce the effects of vibration on the cells, hopefully preventing any of the connecting welds from breaking.
After figuring out the enclosure materials, we had to design a good way of creating a box. This sounds trivial, but there is a lot more that goes into it than meets the eye. In our case, we were limited to making the box out of sheets of polycarbonate. This was the only cost effective way of obtaining all of the necessary features, such as cooling slots and cell dividers. Machining a solid piece of polycarbonate into the shape we wanted would take too much time, cost too much money, and waste too much material. Therefore, we had to come up with a good way of joining the polycarbonate sheets together into a box. We looked into drilling and tapping screw holes into the thin face of the polycarbonate, but in order to do so, we would need to use #0-80 machine screws. That method of fastening seemed impractical and labor intensive while possibly being insufficiently strong. We also looked into chemically bonding the polycarbonate together. We've decided to do so on the prototype eMoto pack, with the addition of interlocking tabs for added mechanical reinforcement. We are considering using polycarbonate angle stock to reinforce the corners. As of now, the box seems pretty rigid.



As of now, we've assembled the primary enclosure with dead cells, and we are waiting on the battery welder. We've received the weld samples, so the welder should come in sometime this week. Once we have that, we will begin matching cells and electrically assembling the prototype pack. To prepare for this, Paul and I have been testing boxes of cells to weed out the obvious bad ones. We're looking to manufacture the secondary enclosure this week, as well as experimenting with various cooling designs. Additionally, Shane and Lennon have been working on our first rapid-recharge charger for the motorcycle. If we're lucky, we'll be rapid-recharging our first pack by the end of the week.



Look for test data and more images soon.

-Mike