Cheap FOCer 2 (Open-source, Low-cost, VESC 6 based ESC) (v0.9 Release. Beta testing ongoing)

zhangzq71 · January 12, 2021, 1:48am

Sure, it works with the 5065 motor that in blue closure that you also see in this picture. It works quite good just without the IMU.

ziploc · January 12, 2021, 2:24am

I have no doubt it works. I was just concerned those solder joints might not hold up to the vibration. But you’re there and I’m not. If it looks good to you, it’s probably fine.

Are you adding an IMU? Doing a DIY OneWheel?

zhangzq71 · January 12, 2021, 3:10pm

I tried to assemble the IMU today but failed, I will buy another new IMU to try. I want to make a longboard.

ziploc · January 12, 2021, 5:32pm

You’re installing the SMD IMU on the PCB - right? Not using an external IMU. But also, why do you need an IMU on a longboard?

I just installed an IMU a couple days ago. Happy to offer tips if you like.

zhangzq71 · January 12, 2021, 11:26pm

Yes, that is the IMU on the PCB not theexternal one, because I never have a VESC had IMU I just want to try. My problem is there was no any solder in the pin of the IMU after hot air flow for a while. I’ve soldered many many this kind of footprint, the pin of the IMU I had must had been oxidized.
I want to know what is your tips to solder the IMU?

ziploc · January 12, 2021, 11:36pm

If you’ve done this many times, you’re almost certainly better than me at it. But I’ll tell you my process anyway…

I spread a bit of solder paste across all the contacts on the pcb. I use a jeweler’s screwdriver to make sure I have a pretty thin coat so it doesn’t ball up and create a short between pins. I also try and make sure I don’t get too much under the chip itself (except for the contacts).

I then place the IC with a pair of tweezers and I look at it through some relatively powerful magnifying glasses while I center it up on the footprint using the tip of an exacto knife. Of course I also check and make sure I have it in the right orientation (i.e. pin 1 on pad 1). Pretty easy to get that wrong.

I run my hot air station at about 450F and lowish air (around 3 out of 10). I approach the IC from probably 5" to get things warmed up, moving the tip in small circles, and approach the IC gradually until I’m probably an inch or two from it. I watch for the solder to flow at the base of the IC or the nearby smd resistors, and then remove the air.

Finally, I test it in VESC_Tool looking at real-time IMU data.

Hopefully the problem is what you suspect. I wasn’t able to get an IMU IC recently, so I bought one from Amazon for $5 and removed it from the board it was on. Presumably that wouldn’t give it a chance to oxidize.

Good luck.

zhangzq71 · January 13, 2021, 1:34am

Thanks for your tips. Actually what I did was almost exactly the same as you, I never had this circumstance before, maybe I bought the fake IMU, it costed only $0.5. I am ordering another one to try.

ziploc · January 13, 2021, 1:37am

Wow - where’d you get one for $0.50? That’d be quite a bargain if it worked. Somewhat less so otherwise.

zhangzq71 · January 13, 2021, 1:40am

Yes, I got from TAOBAO.

it is RMB3.7, about $0.5. and I will try another one, it is about $1.

nhedlund · January 13, 2021, 7:13pm

Hi all!
Biggest thank you to shaman making this open source, and also for the superb ordering and assembly guide!! I just finished soldering up my first FOCer 2 out of five, all went well and it’s running fine when uploading pre built firmware in the “latest” folder of the repo. This is done from my win10 computer with a ST-LINK V2.

I was going to try to make an custom app, thus i need to build from source code.
When i run “make upload” from the master branch firmware from shamans repo i get some issues. The FW compiles and uploads without problems, but if i change any motor or app config settings and power cycle the FOCer, it’s not recognized as a usb device anymore. I also tried the newest FW from Vedders repo aswell, just changed the HW_SOURCE and HW_HEADER to V2, and tried different HW60_IS_MKx but this never booted ok its just blinking all leds in irregular pattern and wont show as a usb device.
The build and uploading i do with my Linux laptop with the same ST-LINK V2.

So it kinda works, but for some reason when a setting is changed, the fw breaks and it starts to blink all leds in a irregular pattern… I see that there is an issue on github for this, but without any solution, I’ve tried searching but havent had any luck… Any help as to what needs to be changed in the source code or where i can find more info about this is greatly appriciated!

Charclo · January 14, 2021, 1:47pm

On my v0.9 hardware I can compile and upload the firmware after changing HW_SOURCE and HW_HEADER to V2. It boots fine after changing settings. I uploaded the firmware with vesc tool, haven’t tried with a ST-LINK yet

nhedlund · January 14, 2021, 2:59pm

Wait, i might have missed something very basic here… Please bare with me… I was under the assumption that I had to use a ST-LINK when programming it the first time since there wouldnt be a bootloader in there yet… But you only change to HW_SOURCE and HW_HEADER to V2. in conf_general.h from the latest shaman repo then run make, afterwords you upload the compiled file from build_all folder with the newest VESC Tool and everything works? You do this from a linux computer? Big thanks for your reply!

Charclo · January 14, 2021, 6:33pm

I had already uploaded the bootloader before, this part you have to do with an ST-LINK. I just tried compiling and uploading the firmware from the repo to see if that could be the problem. Can you get you’re cheap focer running on the prebuilt firmware again? Have you tried reuploading the bootloader?

nhedlund · January 14, 2021, 7:28pm

I actually never uploaded the bootloader the first time around, i just ran make upload and after that i got connection via usb and could use vesc tool to configure input and run motor detection etc…

I tried to upload just the bootloader.hex from repo in folder bootloader with ST-LINK, then pulled the repo again just to be sure and just change HW_SOURCE and HW_HEADER to V2 in conf_general.h.
And the FOCer works and can connect via usb, if i dont change settings and power cycle, it connects again. But as soon as i change anything, and power cycle the leds start to blink in a different pattern and its not even recognized by the computer… Very strange! I will try to solder up another one and see if i get the same results.

I can get it to function as it should if i flash it with ST-LINK and prebuilt FW from the latest folder. I also got it up and running as it should with prebuilt FW from the variESC project. I could run make upload on the code in variESC also, but the controller was not recognized when plugged into usb.

I wonder what I am missing…

nhedlund · January 15, 2021, 2:10pm

Alright, today i soldered up a new FOCer (this name “FOCer”, I LOVE IT! )
After initial power up as per the instructions i uploaded the bootloader.hex file with ST-LINK, it does not even take a second, which is a bit odd since the guide linked in the repo says it might take several minutes?. After that, i could not connect nor did it appear as a USB device, Then i tried to upload the prebuilt FW from the “latest” folder. That also uploaded okay. I changed som settings and power cycled and it was fine as the previous one.

I moved over to my linux machine and ran “make upload”. It built and uploaded fine, i switched back to win10 computer and the FOCer booted up and connected via USB. I tried to change a motor setting and power cycle, and that actually worked! I then tried to change a app setting, and then a power cycle, and after that it was faulty and would not connect.

If someone else could test this i would really appreciate it, as i really would love to be able to build from source and upload the code.
By the way do I dare to run “full chip erase” in STM32 Utility so that I know im working with a “clean” chip?

nhedlund · January 21, 2021, 9:34pm

Hi, so i think I found the root of the problem I have not yet tested it fully but seems to be working now. I had the newest gcc-arm compiler installed, and apparently something is not built correctly if you dont use a older version. I solved the issue by installing GNU Arm Embedded Toolchain: 7-2018-q2-update June 27, 2018.

This is where i read about the compiler issue:

I followed this guide to install a specific version:

doomy · January 29, 2021, 1:26pm

I just compiled the new firmware (5.02) with the source from vedder and i noticed that i hit the amp limit (70A) easily with my onewheel. This was not the case with the old firmware so i checked the hwconfig and i found that CURRENT_AMP_GAIN is set at 10 in the new firmware.

Charclo · January 29, 2021, 1:55pm

Are you using v0.9 hardware or v1.0? It seems there are some small differences
https://github.com/shamansystems/Cheap-FOCer-2/tree/Developer-Branch/guides/upgrade
The firmware in vedders repository is for v1.0 hardware

shaman · January 29, 2021, 2:14pm

The firmware from the vesc project is for hardware v1.0.

doomy · January 29, 2021, 2:17pm

I just did the hardware mod:

github.com

shamansystems/Cheap-FOCer-2/blob/Developer-Branch/guides/upgrade/readme.md

# Upgrade hardware from 0.9 to be 1.0 compatible
The changes from beta 0.9 to 1.0 are mostly hardware only, however there is one hardware change that required a software modification. 
This makes 1.0 firmware incompatible with 0.9 hardware. 

Thankfully this change can be done to existing 0.9 hardware boards. 
This will not make your board into a 1.0 board, however it will make your board 1.0 compatible. 

1.0 compatible boards have the following advantages:
1. You can use 1.0 firmware which is available in the main bldc repo and VESC Tool. 
1. It will fix the DRV error locking issue where DRV faults require a power cycle (you can still get a single DRV fault, but it wont require a power cycle).  

## How to upgrade!

Replace resistors R64 & R65 with 10k ohm resistors. Easy peasy!
- ***Resistory Details:*** 10kΩ ±1% 0.1W ±100ppm/℃ 0603 Chip Resistor
- ***LCSC Part #:*** [C25804](https://lcsc.com/product-detail/Chip-Resistor-Surface-Mount_UNI-ROYAL-Uniroyal-Elec-0603WAF1002T5E_C25804.html)

![](./upgrade_guide.jpeg)

But now motor detection settings are way off.

UPDATE:

I have now flashed the precompiled bin file instead of my own. The motor setting seem ok now.
But it still goes into limp mode when a DRV error happens and only a power cycle resets it. This should have been fixed with the hardware update no?