Godspeed v1.5
It's been a quite a while since I last wrote about Godspeed, so I decided it was time to write a bit about what I've been up to. Originally this was going to be a single log, but as I was writing it I found I wanted to talk about each piece in more detail, so now it's 3 logs - this, the first, being about the board failure and my 1st couple ESC case design.
So, what's new with the board? Oh, it's broken and unusable.
I was really testing it on different elevations, terrains, and slopes. It tackled <30º elevations with ease, and I fixed the issue where the speed controller was cutting out during prolonged accelerations.
So what happened, you say?
On a trial run, I was accelerating down a path, hoping to build up speed in order to clear a coming hill. That worked, but as I cleared the top of the hill, my ESC case fell off the bottom of my board, dragging and dinging as I rushed to brake and keep balance. The path I was speeding on was nice Tarmac but at the top of the hill the road sharply transitioned to gravel. I heard the case hit the floor scraping, dinging, and tumbling for what seemed like an excruciatingly long time (although it could only have been a few seconds) until my motor finally lost power.
Subsequent attempts to re-stick the case to the bottom of the board, power it on and rev the motor resulted in lackluster performance - the speed controller turned on, but could no longer deliver a fraction of the torque it could before, so riding it was a no-go. I guess I kind of deserve that for using Command Strips to affix the esc enclosure to the board, but hey, I wanted to get it out and riding.
ESC woes
So, what's the plan? A new ESC, for one.
New ESC
I ordered the Focbox speed controller from Enertion to serve as the new brain for my board.
My previous ESC is based on Benjamin Vedder's open source design - version 4.12 to be exact. The Focbox is based on Benjamin Vedder's improved yet discontinued VESC 5 design, using a set of surface-mount power transistors for motor control, which results in higher current allowances and better heat dissipation. It also has a smaller, squarer footprint, and a redesigned, sturdier power capacitor junction. On the VESC 4.2, the power capacitor solder joints would experience too much stress because of the wire tension, break, and require re-soldering - even with copious amounts of hot glue taking the strain off.
New Case(s)
Although the Focbox ESC does come with a case of its own, that's not enough to protect it from the environment. Thus, the big thing I'm now working on is a new and improved ESC case and a stronger mounting solution. Since this is a DIY board, I'm designing something I think would be super useful: tool-less removal. I want to be able to access and service the ESC without having to get my screwdriver out. To do that, I'm separating the case into 2 components: brackets, that securely affixes to the board using screws, and a shell that holds the ESC case in place and mounts onto the brackets using a captive bolt and thumbscrews.
I've actually gone through one case design so far, which was meant for the VESC 4.2 speed controller. Here are the drawings for my initial design, called chronos:
Design
I tried to make it a fairly modular design - splitting it up into mounting brackets
I also tried to make it as 3D printer-friendly as I could, trying to reduce the amount of supports I would need to print it.
I then took these drawings and modeled them in OpenSCAD, because 1) I wanted to learn it and 2) I thought I could take advantage of its programmatic-parameterization functionality for a faster iteration cycle. Turns out the parameterization wasn't so helpful in this case, and is a feature available in most other traditional CAD packages now, but it was a good learning experience.
Before I printed out one whole piece, I printed out smaller features of the case like the nut pocket so I could see what the fit was like before printing the whole case out. It's a good thing I did this too, because the first couple tries were either too tight or too loose, and it does take practice and patience to make sure the design and real tolerances are accurate.
This had the added benefit of parallelizing my prints, so I could print multiple features (run multiple experiments) at the same time. Here's how the prints turned out:
If you notice, — issues with the first design.
I'm actually working on a new iteration of the case, Hermes, which I'll be talking about in the next log. It has a reduced form factor I'm using Autodesk's Fusion 360 to model it, an