FIXED! Repairing a Flipsky 75100 Pro, Wont connect

A-DamW

1 kW
Joined
Dec 19, 2019
Messages
415
Location
Idaho county Idaho
Hi everyone,

I recently saw @DanGT86 posting about some connection issues on a Flipsky 75100 Pro. (VESC compatible)

Subsequently, he posted the controller for free+shipping(Thanks Dan!), and the condition of the buyer attempting repair, and posting the results.

I took the bait:p

I am going to approach this repair procedurally, roughly in the order of:

1. Disassembly, thorough visual inspection for loose solder bits or other foreign matter, solder bridging, PCB touching case,
PCB defects, components touching case or each other, sharp metal edges abrading wires, sufficient thermal paste, etc.

Note, I do the above to all controllers I buy. I have 3 other Flipsky controllers, all three had QC issues, loose solder blob, PCB touching case, component(inductor,capacitor) touching case, disaster averted, all three work to this day.

2. While disassembled, apply low voltage ~30v EDIT: try 18v first, from a current limited supply,
check 5v and 3.3v rails, look for hot spots, attempt vesc-tool connection via usb and BT.

2a. If connection succeeds, attempt firmware upgrade to current 6.05 release, research if bluetooth firmware can be, or even needs to be, upgraded.

2b. If connection fails, apply 3lb sledge hammer vigorously.
Or, connect a vesc-express(esp32-c3) over canbus, cross fingers, attempt connection.

3. If the simple connection methods fail, use a spare vesc controller as an SWD(Serial Wire Debug, god mode) port programmer(yes, vesc can do this natively, ask me why I love the vesc platform so much!)

3a. Use ST-Link-v2 and STM32CubeProgamer and SWD port, if the above is unsuccessful.

4. This starts to get into uncharted territory for me, as I have used all the above methods before, so I am somewhat familiar.
If we make this far, I may have to get the 'ol Tektronix o-sillyscope running, and ping some of the big brains on ES for help.
Possibly use AI for debugging tips.
Some other things in this category(that I do have a little experience with):
a. Hotplate+hotair reflow the logic and/or power PCB's.
b. Hotplate+hotair remove the BT module, if it has failed and is tying up the com-port connection.
c. Hotplate+hotair replace the stmf405rgt6 cpu, I have couple spares around.
d. Etc.


That's it for now, will post some pics later, esp. if I find something juicy.

Adam
 
Last edited:
Other than to use a deadblow hammer for step 2b to minimize the noise and component-spray, I can't think of any changes I'd make to the steps above. ;)

For reference, these are some of DanGT86's posts about the controller / problem, in case they contain any useful info on prior troubleshooting/testing that will help those following this new thread by the new owner of the unit ;) :
https://endless-sphere.com/sphere/t...iny-controller-vesc-based.113445/post-1822768
https://endless-sphere.com/sphere/t...iny-controller-vesc-based.113445/post-1833045
https://endless-sphere.com/sphere/t...iny-controller-vesc-based.113445/post-1833326
 
Thanks for the links amberwolf.

Looks fairly clean inside.
1st & 2nd pic:
Some notable improvements over the 1st gen 75100 alupcb:
Milled pocket above logic pcb for component clearance to case.
Silicone on electrolytic capacitors to prevent vibration fatigue failure.
Adhesive on logic pcb to power pcb connectors.
Battery wires are soldered directly to the bussbars, no pcb trace gap.
Someone must be getting paid by the amount of thermal paste used,
holy cow, there is enough for 5 controllers, better than no paste I guess. (thermal paste is generally thinner=better)

3rd pic:
Well what do have here, I see 4 bare pads marked vcc, gnd, dio, clk, pretty sure that is our bluetooth SWD port,
handy if we need to break out the big guns.
Marking on the bluetooth ic: "N51822" probably an nrf51822 clone.

4th pic:
This end of the power pcb is pressed down by two long legs milled into the top shell.
I am a little concerned about time and vibration wearing through the silkscreen and shorting the DC bus.
When I reassemble, I think a couple layers of kapton tape will suffice.
Far end of the aluminum power pcb is held with screws and metal tabs.

Next up:
Keeping the magic smoke inside.
 

Attachments

  • IMG_20250114_145641344_AE.jpg
    IMG_20250114_145641344_AE.jpg
    5.6 MB · Views: 12
  • IMG_20250114_150023522_AE.jpg
    IMG_20250114_150023522_AE.jpg
    916.8 KB · Views: 12
  • IMG_20250114_145709984_AE.jpg
    IMG_20250114_145709984_AE.jpg
    1.3 MB · Views: 11
  • IMG_20250114_150848584_AE.jpg
    IMG_20250114_150848584_AE.jpg
    5.7 MB · Views: 11
Last edited:
Test rig setup:

3d printed Milwaukee 18v battery adapter.
270 ohm power resistor, 'cuz that's what I had laying around.
5 amp fuse.

IMG_20250114_162244952_AE.jpg
 
Ok, powered up with the test rig, blue status light was flashing dim and rapidly.

So I removed the 270 ohm resistor, that resulted in a semi-stable boot.

Connected to usb on a linux system, looking at dmesg kernel messages in a terminal, shows the connection being reset every couple seconds.
Same thing for vesc-tool on linux.

So, I tried bluetooth on android mobile app, Whoohoo! I can read the controller config!

In the moble app, terminal->Print Faults shows a FAULT_CODE_BOOTING FROM_WATCHDOG_RESET
so something is going on.

At this point I'm a little torn between trying to update firmware over bluetooth(slow, may brick it),
or just try to disable bluetooth and hope it un-blocks the usb connection.

Ether way, we still have SWD god mode to update the firmware.
 
Last edited:
YEAH BOiiii!! it livesss! (on the bench at least)

The saving grace here was bluetooth allowing a single setting change.

@DanGT86 , all I changed was "APP to Use", from "ADC and UART" to "No APP".

Stable green led light, stable usb and bluetooth connection.

I'm guessing this may be due to Flipsky being lazy and just using the 75_300_R2 firmware(incorrect),
and somehow conflicting "ADC and UART" and the bluetooth uart.

Relevant backstory:
How to update firmware on the Flipsky 75100 & 75200 FOC ESC

So now that we made it this far, I want to find out if there is a proper hwconfig/firmware, and
see if "ADC and UART" can be used without bluetooth uart conflicts.
 
Last edited:
This end of the power pcb is pressed down by two long legs milled into the top shell.
I am a little concerned about time and vibration wearing through the silkscreen and shorting the DC bus.
When I reassemble, I think a couple layers of kapton tape will suffice.
Kapton tape can be mechanically damaged if between moving parts; if you add teflon tape with a cloth backer to the top of the kapton, the kapton can do the electrical insulation and the teflon can do the mechanical protection.
 
all I changed was "APP to Use", from "ADC and UART" to "No APP".

Where did you make this change? Was it made in the app before you got the controller to connect?

Or in other words would I have been able to do this with the regular windows VESCtool with the connection resetting intermittently?

Glad to see it's not dead. It will be interesting to see if this problem reappears for you every few rides.

If it turns out the Bluetooth module is bad and was occupying the serial port this would have me scared off of buying future VESCs with integrated bluetooth. I would like to know if you can successfully disable the bluetooth such that an external BT module could be used in case the onboard one fails.

Good stuff. Thanks for keeping us updated.
 
Where did you make this change? Was it made in the app before you got the controller to connect?
Android mobile vesc-tool connected over bluetooth, APP setting tab.
No, I had to connect over bluetooth first, then change, (was def kinda buggy, slow, 3-2 random disconnects, but worked in the end)

Or in other words would I have been able to do this with the regular windows VESCtool with the connection resetting intermittently?
I would guess no, I could not connect over usb at all, I did not try bluetooth on PC, just android mobile.
Bluetooth and usb now appear to work fine, tested for about an hour each.

If it turns out the Bluetooth module is bad and was occupying the serial port this would have me scared off of buying future VESCs with integrated bluetooth. I would like to know if you can successfully disable the bluetooth such that an external BT module could be used in case the onboard one fails.
I think the stm32f405 has 3 uart ports, I'm guessing the hwconfig/firmware was causing "ADC and UART" and bluetooth to fight over the the same uart.
I do think you can disable bluetooth on the PC vesc-tool, not mobile though.
I definitely prefer external modules for maximum configurability/repairability.
 
Back
Top