Let me refrase that: The “Cheap Focer 3” thread was a fake, and nothing else.
And the cheap focer 2 has been a great inspiration for a lot of people. Though, the really awsome work on open source vesc hw is done by MxLemming and badgineer - the MP2 and MESC. Think I have mentioned these guys 10 times before. Go there, if you can’t wait for this thread.
1 Like
The “Cheap Focer 3” thread was a fake
It’s still there. Perhaps it will be stalled forever. But I can’t imagine he set out to start it and not finish.
Presumably, there must be some shortcomings to this design in your mind since you’ve decided to make your own design. Would you mind sharing? I apologize if you’ve already done that.
you were saying? have some faith, it’ll make your life much less stressful 
Congratulations to you, go and buy one…
Here’s a short demo of the previous design (not the latest):
2 Likes
Not really any shortcomings of the MP2, except maybe the price and physical size. It hasn’t got an IMU though. Finally it seems that there are at least two cheap VESC alternatives available, though still only 1 is Open Source. This is good news for the community.
I will continue the development of FOC-king, because in my daytime job as an engineer I work with robotics and I’m responsible for DC motor control development, so it fits quite nicely.
The FOC-king is modular and will stay like this. The good thing here is that it makes everything so simple, and hopefully this will inspire others to start developing their own FOC boards, when they realize that FOC hardware is SO SIMPLE.
The latest updated PCBs should arrive on monday, and I’m excided.
3 Likes
It looks like the price per board (all included except wires and shipping, in a batch of 5) will be:
$35 or €31.
All components assembled by jlcpcb. Spec: 75V, 12 Fets, CAN, IMU.
DIY, you need to solder the Mosfets, and rearrange the positions of the MCU board and Power board.
NO release yet. Latest updated board needs testing (arrives in a couple of days)
5 Likes
Me too! I’m eager to try and build a couple of these. I’m even thinking I may design my own little daughter board for the driver components to make it slightly more modular.
Eagerly awaiting the news.
Only problem with more modules is that “pins” are a source of noise pickup.
So if you separate mosfets, then shunt resistors and operational amplifiers should be on the seperation board as well - I would say.
1 Like
I’m only a hobbyist level EE, so any tips are greatly appreciated. My plan was to leave the MOSFETs and shunt resistors on the main board and make a daughter board that carries just the discreet components that drive the MOSFETs. Does that work?
A couple more questions…
- When you mention op-amps, are you referring to the IRS21867S gate drivers?
- I notice that MOSFET #1 is different than 2-6. Why is that?
Thanks again.
Pins are also a possible source of mechanical failure from heavy vibration.
1 Like
Understood. I may secure the daughter boards mechanically (i.e. screws) in addition to using pin headers that go into socket headers. Is that a reasonable solution?
You must have an old version. I recently doubled the number of mosfets, without actually having to make the board larger. A cheap way to get as much power as possible. You need to get the latest from github. The mosfets should all be the same.
2 Likes
But hey, tell me some news instead. This thread is about Open Source hw available on github, and not proprietary products for sale at some site at $180 + shipping. What happened to cheap?
1 Like
You’re just moving too fast for me! Not only did I have an outdated version. Turns out I had two different outdated versions. So I’ve deleted both and downloaded the latest ones just now. Yup - all MOSFETs are the same.
But this brings me to another newby question… I remember someone once asked Shaman about doubling up the FETs. I don’t remember his explanation, but he said it’s not that simple. Any ideas what that’s about?
Directly paralleling them? That can cause all sorts of problems.
They won’t turn on at the same time or at the same voltage, causing more stress in some vs the others. They also won’t share current equally. Just adding FETs to an existing setup increases turn-on and turn-off times and that increases the switching heating and could stress the FET drivers. Lastly, without isolating their gates a bit with individual series resistors you risk oscillations and destruction of the FET(s).
Depending on the specific situation you might be able to get away with it, permanently or just for a while. But it is definitely not something to design a product for IMO.
2 Likes
I feel like i’ve looked at blown vescs and they all have multiple parallel fets per phase no?
what extra details do they have to do to avoid all those timing and other concerns?
They may well exist. There’s certainly a variety of versions out there. But none of the VESCs I’ve actually had in my hand used parallel FETs.
Basically, avoid the problems I listed.
- Drive each FET separately or, at least, have a gate resistor for each paralleled FET.
- Use good FET gate drivers to ensure quick turn on/turnoff to minimize time spent switching.
- Use FETs from the same batch with careful attention to their pcb layouts to ensure that the current dividing was as even as possible.
- Overspec the number FETs used to ensure that uneven current dividing and timing differences never result in the excessive stressing of any one FET.
- Use a pcb layout that minimizes circuit inductance but this is a best practice for any FET switching circuit.
Or, considering how many ESCs pop, they might do nothing at all to avoid potential problems with directly paralleling fets.
“Hardening” the design of an ESC costs money. If having some fail doesn’t affect sales more than some limit they set then they can avoid the extra parts costs needed to reduce the number of failures. Why add parts to the design if the parts cost more than replacing blown units? 
Especially if their reputation and sales doesn’t take a big hit from selling units that aren’t very reliable. If their ESC is cheap enough, or widely available enough in a market starving for options, then they can get away with a pretty crappy design.
I should add that this is definitely not the way I approach circuit design! In my opinion it’s just an all-too-common (very unacceptable IMO) approach to business for some companies. Particularly ones in a certain country.
8 Likes