Machining the Sprite Adapter

Warning!!! Innocent carbide tooling was harmed in the making of this adapter plate. Viewer discretion is advised.

We turn our attention to the Sprite, and the most challenging part of the conversion: machining the adapter plate. See this post for more information on adapter plate function and purpose.

This is a 12" X12" X 1.5" (300 mm x 300 mm x 38 mm) chunk of 6061 aluminum. 

This is the Remy 250 HVH that will power the Sprite.

There are two important features that we need to design into the adapter plate: a bolt circle of tapped holes with diameter 8.46" (215 mm) and a raised lip with a diameter of 7.09" (180 mm). These features are concentric with the motor shaft, where tolerance in positioning is critical.  Any error will cause misalignment between the motor shaft and the transmission shaft, with possible vibration, excessive wear, or malfunction.

There are other features that do not need to be located with high accuracy: A 4 inch (100 mm) clearance hole for the motor shaft and hub, and a nice round perimeter is nice aesthetically and sheds some weight.

The first step is to rough out the shape.  This was first attempted on the band saw.  Based on the rate of cut we decided to speed up the process and used a table saw with a carbide tipped, high tooth count blade.  This faster method only took a couple of hours.  Probably none too soon for Paul, who was getting hit in the arms with hot aluminum chips, while feeding a hot and sharp edged part into the blade.

 Only 8 more cuts to go, Paul!

Here is a 16 sided polygon, in the rough shape of our desired circle.

We had two options to finish the part: milling machine or lathe.  The lathe can clean up the outer perimeter of the adapter, and cut the groove for the motor face lip in short order.  With a rotary table on the mill, we can easily locate six evenly spaced bolt through holes on an 8.46" (215 mm) diameter.  The mill can clean up the outer edge and cut the groove, but not as quickly or nearly as cleanly as the lathe.  Because the features need to be concentrically located, being able to do all three operations with the same set up made the mill the clear choice for us.

We made and installed a collar on the center axis of our part.

 Now the part is aligned to the center of rotation of the rotary table.

Taking 0.050" (1.27 mm) depth of cut with each rotation of the table, we were able to make an attractive disk in a mere 30 passes.  Only 10 more to go, Paul.

Next we turned our attention to the groove.  This feature needed to be about 0.125" (3.2 mm) wide and deep.

 

Disaster struck as our fairly expensive carbide end mill suffered an unfortunate fate.  And we thought we were taking "light" passes.  We shall never speak of this incident again.

This step was really satisfying.  This was the first time I have had the luxury to use a rotary table to lay-out a bolt circle, and I must say it is far superior (faster and more accurate) to using a divider, center punch, and drill press.  All you do is move the X axis of the mill to the correct radius for the bolt circle, as displayed on the DRO (did I mention this mill also had a digital read out), and rotate the rotary table in 60 degree steps for each hole.  Here we are "spotting" the drill locations with a stout center drill.  This maintains the correct location of each hole, because twist drills tend to walk on you, if you try to start them on a flat surface.

The final drilling operation was uneventful.  

We are left with a pile of aluminum chips to clean up, and a part that we can be proud to say we made ourselves.

Battery Racks, Part 1

It is time to cut some metal. We can fit 48 of the Thundersky cells under the Doka’s rear bench seat.

The rack will be made with a structural steel called angle iron. The name was popularized when wrought iron was the material of choice. The name stuck.

The process is straightforward: Cut the material to the correct length.

Tack weld. This is a short burst of the welder, located in a single point along the joint line.  At this point, the material can still be positioned a small amount, and if an error is discovered, it is easy to break the tack and make a correction.

Check that everything is square. See the square on the table?

If everything looks good, fully weld up the joints.   We are committed now.  Any change requires a lot of grinding.

A welded joint results in a raised bead over what used to be two pieces of steel.  The bead of a skilled welder is  like a work of art.  However, sometimes the bead is in the way, so we grind the welds in areas that need to be flat (like the bottom of the rack).

Check out the sparks in SloMo.

Doka Battery Pack and Some Math

For the Doka, we will be using 100 Thundersky lithium iron phosphate (lifepo) cells wired in series for a pack voltage of about 320 volts.

Lifepo chemistry has a long cycle life of 2,000 charge and discharge events. Pack capacity will linearly decay, and at the 2,000th cycle, capacity will be about 80% of the rated capacity. Other lithium chemistries get between 500 and 100 cycles. Ever wonder why your smart phone starts running out of charge before the end of the day after a couple years? The trade-off that the lifepo cells have for long cycle life is charge density – they weigh more and take up more volume than the other lithium ion chemistries.
The cells we are using have  100 Ah (amp-hour) capacity. They could deliver 100 amps for 1 hour, 1 amp for 100 hours, or any “reasonable” combination of current and time that multiply to 100 Ahr. Combining the volts and amp hours, we get the energy storage of the battery. 320 volts * 100 Ah = 32,000 watt hours.

So what do these numbers tell us about what we can expect in range? Just a bit more math is needed. We need to know how fast we are consuming the energy from the pack as the vehicle drives down the road.

A really rough estimate for energy consumption is to take the mass of the vehicle in pounds and divide by 10. The result is the amount of watt hours consumed to drive 1 mile. For metric calculations, take the mass of the vehicle in kg, and divide by 7.3 to get watt hours consumed per kilometer.

For the Doka, which weighs about 5,000 pounds (2400 kg) the energy consumption estimate comes out to about 500 watt hours per mile (325 watt hours per kilometer). The energy in the pack divided by the energy consumption gives the expected range of the vehicle per charge. 32,000 watt hours / (500 watt hours/mile) = 64 miles of range per charge. In metric, the range estimate is 32,000 watt hours/ (325 watt hours/kilometer) = 98 kilometers.

These are all just guesses.  It will be a lot more fun to build the Doka and find out what it can do. 

Joining the Motor and Transmission, Part 2

This is the business end of an AC 55 motor.  It is big, mean, heavy, and not much use in its current form.  We need to add some hardware.

First up is the hub. The hub is a chunk of steel that slides over the splines of the motor shaft, and grips them securely with a taper lock bushing. There are tapped holes and a shoulder feature that will prove useful for mounting the flywheel.  Just behind the hub is an aluminum sleeve that I machined on my lathe.  Its only purpose in life is to hold the hub in the correct position, while tightening the taper lock bushing.

The next part to be bolted to the motor is the adapter plate.  This plate does two things.  It serves as a spacer, so the thickness is critical for locating the clutch depth into the transmission bell housing.  This plate must also be structural, and provide a strong connection between the transmission and the motor.

The next plate is 0.75" (19 mm) thick, and takes its form from the shape of the transmission bell housing.

Paul is making sure all these bolts are well torqued.

Seeing Paul sitting on the motor reminded me of another image.  Anyone know the movie?

The flywheel is pressed into the hub, but not bolted in yet.  The next step will be to re-thread some metric M8 bolts over to imperial 5/16".  I'm joking.  We have metric/imperial confusion when mixing both standards on the same project.  But in this case we have a good excuse.   There is only 2.5 thousandths of an inch (63 microns) difference between 8 mm and 5/16 of an inch.

Joining the Motor and Transmission

If you are designing an electric car from the ground up, you may find that a fixed gear or two gear transmission is the most efficient solution. If you are converting an existing gas vehicle to electric, re-using the existing transmission is a very attractive option. It allows you to reuse the existing transmission mounts and the mechanical drive system downstream of the transmission, like axles and drive shafts.

The gearing on a gasoline engine will do a decent job of covering the gears we want for the electric motor. The electric motor will have a lot more torque available at zero RMP, so 1st gear will pretty much be useless. An electric car will be quicker off the line in 2nd gear than the gas motor starting in first. Electric motors are happy turning higher RPMs than gas motors, so 3rd gear is usually good for highway speeds.

We need to come up with an adapter to join the electric motor to the transmission. This will be two adapters, actually. The first is a chunk of aluminum to physically join the body of the motor to the bell housing of the transmission. The second is a hub that mounts to the output shaft of the electric motor, and bolts up to the flywheel. This hub is a two piece system.  By drawing the hub together, the taper squeezes the motor shaft with quite a bit of force.

The VW Doka is a rear engine, rear wheel drive that uses a trans-axle, where the differential function is included in the gear box. The Sprite is a traditional front engine, rear wheel drive configuration. Both will use a similar arrangement to connect the electric motor to the transmission.

The goals of a good adapter: keep the electric and transmission shafts rigidly aligned, and locate the flywheel into the transmission bell housing the same depth as it was with the gas motor, so that the clutch fork will correctly engage the pressure plate to release the clutch disk.

That's the theory.  Now to put it into practice...