The drill file is missing I am working on recreating it.
I think that is more of a problem of people mounting something that needs cooling inside of an enclosure.
Mount the controller with the heatsink outside of the box and it will run way cooler.
Ideally I get more feedback from a larger sample group before jumping to v1.0. Unfortunately we are at the mercy of the community for this.
Sure!
You’ll see a divide in opinions here. A lot of people will say using regular soldering iron to solder is less scary than hot air reflow…and people are not as likely to have a reflow station. The TO-220 FETs can be soldered by common soldering irons.
This is what I have in mind as well but keeping to through hole caps.
So directFETs are going to be for a more premium controller. You leave the “cheap” and easily sourcable realm when you leave TO-220s (common) to a brand-specific package like directFET.
6 Likes
Exactly. My #1 pet peeve about esk8 designs is that it usually just has the controller tossed in a plastic box with no airflow. A heatsink for the controller needs to be incorporated into the esk8 enclosure so that the benefits of airflow can be had while still protecting the batteries, controller, bluetooth, ect.
5 Likes
Exactly thermal saturation can kill your performance.
2 Likes
The logic hear was to sort of emulate the effect of spring clips instead of screws. There is documentation out there showing how spring clips apply a more even pressure to the FETs and improve thermal coupling.
No it can’t. Look at the design in KiCAD and see how many things would be in the way of this.
Have you found where the other components are getting excessively hot? The DRV would be the next highest priority thing to keep cool besides the FETs but I’ve found it is coupled to enough copper in the PCB to be fine.
I know it’s common but try to not do this. power electronics are happiest when exposed to airflow.
1 Like
Exactly the problem I am going to solve with this bad boy:
ps sorry for the potato video quality, took it from a messenger chat
4 Likes
So don’t get me wrong. I’ve though about at least using SMD FETs but not specifcally direct FETs.
I most likely won’t go with this exact design but this is the direction I would go for an open-source FOCer with SMD FETs. However I’ve already got a different design that makes it about half the footprint…
Don’t get too hung up on this one. This design won’t see much attention until after the CFOCv1.0 release.
14 Likes
I really like the current design. It will be easy to repair too.
I appreciate all the work you put into the design.
2 Likes
Thanks! Repairability was something I considered with the CFOC2 design.
6 Likes
My vote for 1.0 is the following.
-IMU that JLCPCB can assemble.
-Bigger pads for the SMT caps.
-Standard servo header for the PWM input.
-Switch to the non-standard orientation standard so that JLCPCB can’t fuck it up.
-Change the logo on the back to Mr. Pink since he was cheap.
4 Likes
@shaman I managed to burn something in my cfocer, I think I put a wrong resistor on the pull down pads for the ADC, probably a resistor too close to be short, I really don’t know. But the cfocer doesn’t have any light.
Can you point out on which component could be damaged if that’s the case of shorted pulldown resistor?
1 Like
I can help the best I can. Can we take it to PM?
2 Likes
Hi @shaman ,
I don’t have the controller yet, but from my experience with power circuits is that components will eventually heat up even the passive ones. SMD stuff on a pcb with flat bottom mounted to a big heatsink will ensure an optimum even cooling for every component.
Btw, I have question.
I read somewhere that some SMD resistors are not populated in the ADC section. Are these necessary to run an eBike throttle through ADC input ?
1 Like
These are the optional pullup and pulldown resistors. They’re not necessary for the throttle. I’ve run into situations where it’s nice to have them for hall effect based brake signals though.
Hey long time lurker here. The Shunts are currently out of stock, would 1W 0.001R 2512 be sufficient ? The only other availible one has 0.0015 ohms which seems as bad as the first one
1 Like
Unfortunately 1W resistors aren’t enough…
Calculating the dissipated heat for 70A motor current and .0005ohm total shunt resistance puts you at 2.45W. 2x 1W .001ohm resistors would only handle 2W at best. You could try to order some of these from LCSC and solder them yourself.
Otherwise, we’ll just have to wait for JLCPCB to restock.
EDIT: actually these instead
3 Likes
Thanks a lot for your fast and detailed answer, handsoldering it is.
1 Like
Potential Issue with SMT Assembly of the DRV8301 from JLCPCB
A few have reported blown up D1 diode arrays upon initial powering on of the CFOC2. The issue has been discovered to be inadequate soldering of the DRV which leads to an unregulated 5V of which could actually be 7V or greater. This doesn’t happen with every unit and seems to be intermittent.
It is recommended that you reflow the DRV and/or touch up the soldering of both the legs and the large pad underneath before initial power on.
The potential cause of this is the small hole in the DRV’s ground pad that was meant to allow soldering of the DRV with even a normal iron (like the CFOC1). This hole may be interfering with the SMT assembly process. This will be removed in the CFOC2 v1.0 release and the pad returned to a normal SMD pad.
12 Likes