esk8.news esk8 calc donate now

Cheap FOCer! (VESC compatible 4.12 redesign)

Cheap FOCer
(original thread here)


Top

Bottom

Advantages

  • Lower build and BOM cost
  • TO-220 FETs allow for big heat sink attachment for better thermal performance
  • SMD components no smaller than 0805 make hand assembly practical
  • Added ON/OFF capability to turn off control circuitry when controller is not in use. A simple mechanical switch will achieve this when connected to the “ON/OFF” 2-pin header.
  • Can fit in typical “350W” ebike enclosures
  • Can fit in Hammond 1590b enclosures
  • Runs cooler than other VESCs

Disadvantages (that I know of)

  • Larger than original VESC. Cheap VESC is 45mm x 92mm
  • Higher profile with TO-220 package FETs
  • Additional assembly steps to beef up high-current traces. 2 layers of 1oz copper can’t handle the current flowing through during operation. Assembler will have to apply wire/solder wick/bus bar to the exposed Power, Ground, and Phase traces that you can see in the “bottom” image.

All this wouldn’t be possible without Benjamin Vedder’s hard work to build on. Please consider supporting him for his efforts through the link below.

https://vesc-project.com/donations

Building off of Benjamin’s shoulders isn’t free either. It takes time, money, and lot’s of analysis in the lab to do it right. Anything you can toss my way helps me a ton in ordering materials for R&D.

GitHub repository for the gerbers, schematics, and BOM

18 Likes

I’m creating this new thread to continue off of the old one from esk8 builders. If you are new to this project, read the summary below for the gist of what’s going on. You can also click the link in the original post that will take you to the original thread.

Summary:
I did my own take on the VESC 4.12 hardware by designing a 2-layer PCB for lower cost and using TO-220 MOSFETs for better thermal management. I made an initial prototype by hand, did some analyisis on it in the lab, found some problems, fixed the problems (mostly), had beta versions made, and the beta testers are currently providing feedback. So far so good with the beta testing and I’ll be posting some of those results here.

The initial problems were mainly due to bad layout of the main power and ground traces. I learned that bad layout of high-current traces causes a lot of parasitic inductance. Parasitic inductance causes voltage spikes that can kill your entire controller. I fixed this by running the high-current traces directly over/under each other to minimize loop area thus minimizing parasitic inductance. I also added RC snubbers to the switch nodes to further dampen oscillations and spikes. All of this has allowed it to operate under 12s with no major issues so far.

6 Likes

It would be nice to include few of the development posts. Like a just of what problems you had and how you solved it. So that it’s like the original thread but more efficient.

Edit: I see you are doing just that.

2 Likes

how much copper did you add on the lower side? Also would you recommend a stencil? or is it unnecessary?

1 Like

I populated the prototype without a stencil and applying solder paste with a syringe. It took a while but it’s doable.

As per suggestion of @b264 There is now a through-hole pad underneath the DRV in order to easily solder the pad on the bottom side of the IC.

9 Likes

This is amazing :smiley:

3 Likes

I’ve got a silly question… I don’t think I recall seeing rectangular holes in PCB’s before. Can the cheap PCB fab guys make those?

Thanks.

3 Likes

That’s a very good question. My guess is that they’ll route it out with a normal round bit which will cause the corners to be a little rounded. It’s probably something I should confirm before someone else finds out the hard way. Anyone else have any insight on this?

1 Like

would it even matter if it is not rectengular?

1 Like

Should be ok? But is will be rounded to 0.8mm for pcbway at least

https://www.pcbway.com/pcb_prototype/PCB_Milling.html

why are they not using lasercutter? Much cleaner isnt it?

mmmm I could not tell you, this is not my field. accuracy, copper traces or finish perhaps

Laser “cuts” by burning away materials. So the laser would burne the pcb. pcb charres when it burnes and could become conductive in unwanted places.
my guess why they don’t use laser cutters

4 Likes

that makes sense.

We were not permitted to cut fiberglass on the lasers at the TechShop because the glass doesn’t cut (although you can etch glass) and the epoxy produces bad fumes.

1 Like

I found this PCB factory video. There is a lot more steps than I thought there was. :lab_coat: :nerd_face: Quite interesting.

9 Likes

wow, indeed quite alot more steps than i thought.

1 Like

fascinating stuff

1 Like

The through hole for the solder pad under the drv…simply amazing. It would allow some skilled soldering iron changing by those of us without a air station.

I’m in

3 Likes