Plan A fails, onto Plan B

Hey Electrified Miata team, here's the latest update in the never ending quest to
go faster. First, I'd like to acknowledge the amazing number  of friends and family that have already helped get this project off the ground.   Second, I'd like to share a bit of the details that go into going faster. Number one on the acknowledgment list is my wife, Teresa, who listens to my dreams, helps me edit these blogs, listens to me work through my problems, and is there to celebrate the many small victories along the way. The friends and family who have contributed this week are in no particular order: Sawyer Bitner, Doug Kimber, Steven Quick, Bruce Bitner, Joe Kroesche, Ed Bailen. Also, thanks to my TKD family who always bring me to the present moment instead of dreaming about the future.

Going faster starts with assembling battery packs, reassembling the old go-kart, constructing the Battery Management System (BMS), developing CAN software, and separating the wire looms on the  Leaf Hardware. A group of us met last week and we popped the top and bottom off a Leaf charger, discussed the various tasks to make progress towards the electric Miata. Instead of tasks getting completed, two MORE tasks were added to the list! Most of the other tasks have partial progress. 

Read more to learn about the journey. Warning, long and detailed descriptions below.

In reference to last weeks tasks, here is the progress.

1. Assemble two battery packs in a 20S20P configuration.


  One thousand batteries have arrived. Thank you very much  to the 18650BatteryStore. They know how to treat their customers right. The day I was going to order my first set of batteries, Dec 23d, a Deal of the Day was announced.  Suddenly, I'm doing research on LG batteries, as for a short time, they were on sale. I find the batteries with the right specifications, and whip out the credit card (Merry Christmas to me). I get a reply email that day (although I didn't read it till the 26th) saying they had sold out of that particular battery but would it be okay to send me a better battery for the same price. Heck yeah, doing the happy dance! I got upgraded by 300 extra amp/hours. The batteries arrived very quickly (Dec 30th) via fedEx. Thanks to Bruce Bitner, I now know that each pack of 100 batteries weighs about 11 pounds. Now I need to test them and assemble them. Here is  a picture of the pre-assembly state.

 After measuring the battery holders on the go-kart, it seems that a 20S10P cell is all I have room for on a single layer. That means I'll have 4 physical packs, combining two each make 20S20P packs.  Here is one of the packs partially assembled.

I also started measuring each battery for state of charge, a mind numbing task.   As I will be connecting them to directly to each other (parallel), they had better have the same relative state of charge.  So far I've gotten 204 cells measured, with 200 cells measuring 3.597 +/- 5 millivolts and 4 cells measuring 3.581.   The difference isn't that great, but I'll put them aside for a test/practice battery assembly.

​2. The go-kart needs to be re-assembled and the Alltrax 7245 controller needs re-programmed to 72 volts.
   The go-kart is out of the shed and the motor out of the attic. Thanks to Sawyer for the heavy lifting.

3.  The Battery Management System BMS will need some assembly.
    Joe Kroesche (of B2600 fame ) has volunteered to prototype one of the boards so we can see how it operates. In fact, he's already ordered some parts to make sure everything works before we need to assemble about 100 of these little boards.  Thank you Joe!   This is a great reminder that it's so much easier to do things as a team.  Both Joe and Doug Kimber (a friend from Bell Labs and my circle track racing days) are familiar with the process of building Surface Mount circuit boards.   I'm learning so much about types of files and assembly procedures.

4.  Separate the wire looms on the Leaf hardware to have just the inverter and the required motor resolver interface.
   No progress yet.

5.  Develop hardware and software needed to communicate via CAN with the motor inverter.
    My 'A' plan was to use the beagle-bones.

 The beagle-bones have two embedded CAN controllers, so they would be ideal. I still have to have a external chip to both get the right signal levels and isolate the BeagleBones from external events (things that would harm the BeagleBones). I found a tiny board from Waveshare that had the chip I needed on it. Unfortunately, I couldn't make this setup work.  I did hook the CAN_L to the CAN_H (which is wrong) and it felt like one of the waveshare boards got kind of warm.  After that I ordered some replacements boards, but still couldn't get it to work. Not sure if the software was too old, or what the problem was. I had a backup plan in mind, so on to plan B, which involved using raspberry Pi's

 Since the Pi didn't have a CAN controller embedded, the little add on board was a little more complicated and cost more at 18.00 per board.   Of course the OS on my Pi was badly out of date, so my first task was to get the latest software loaded. Thanks to the internet, updating isn't hard.  So much faster than the days of modems.. After getting the software up to date, I found I had no jumpers to connect the Pi to the CAN boards.  I was headed to Fry's and asked Bruce if he wanted to come.   It turns out he had some jumpers, thanks Bruce, and I finally started making progress. I could get the CAN link up, but still no communications. I found a grounding problem and everything started to work.   After that, (which took most of the week) I finally got to the software portion of the project.   I found code on github (thank you internet) which allowed me to use Python to communicate with the CAN bus.  Whew!   I thought that would be much easier than it was.    And then two more tasks got added to the list..  And yes, they should have been on the list in the first place.

6.   A new task, Charge the batteries.
 Lithium Ion batteries need a special charging procedure, else they are likely to blow up. It won't be long before we have battery packs assembled. It will definitely be longer before we have full battery monitoring capability but will need to charge the batteries to get the full power on the go-kart.  Chargers can be expensive, so I"m hoping to use the charger that came with the Leaf motors. Bruce, Ed, Joe and I took apart the case from the charger which was much harder than it looked. That gray gasket goo Nissan used was really sticky

It will require some reverse engineering on our part.   It is possible according to this  article.

7.   A new task, Obtain a coupler that fits the the Leaf motor shaft.

  Soon after the battery packs are assembled, I'll be trying the spin the motor. Right after that, I'll want to put a motor on the go-kart to do a thorough test of its capabilities. We'll need a way to put a gear on the end of that motor shaft.   It's not a stock part from Grainger, and the guys at EvWest.com didn't have any suggestions as to where to find one. I may have a guy here in Texas willing to make them (for a price).   To do that, I need to make a CAD drawing of what I want, which is a new tool to me.   Luckily, my friend Doug Kimber turned me on to ONSHAPE, a free cad tool.  Hope the learning curve isn't too sharp. I also talked with Steven Quick as he works with a machine shop.   Need to get that CAD drawing done. Anyone out there know how to do this? 

My head is still spinning with all the details and bouncing between all these mini projects. Overall, this past week has been a lot of fun and truly enjoyable.   Again, a huge thanks to friends and family that make this all possible. I'm paraphrasing a quote from the movie Ford Vs Ferria where Carroll Shelby is giving a speech.  "When a person knows exactly what he/she wants to do, they'll never work a day in their life".   While it's not 100% fun (checking 1000 batteries is rather boring), it is mostly play. Thanks so much for reading. It's not too late to join us in the fun - contact me (using this link if you don't have my number)