thanks a lot for this! i’m using it on a ebike on 16s with the flipsky 75100 pro.
i had great experience with flipsky vescs on ebikes before (regularly pushing the mini 4.20 to 60bA and 100pA on lipo 12s) but this time, nothing worked… the first flipsky 7550 didn’t work at all, then seeing it wasn’t sold anymore, i suspected some hardware issue and switched over to the 75100 i bought in the meantime… but was having similar problems… then tried for a few days playing around with settings, but nothing worked…
finally found this, seeing it wasn’t built specifically for the 75100 pro, i was hesitant and was finishing everything else first… but now did it, and after some testing, found that the 75_100_V2_no_limits was the correct file for the flipsky 75100 pro! everything seems to work fine, including bluetooth.
my bike is put together now, but i think i jumped the gun by switching firmware too quickly. im on casainhos firmware i believe. from his zip folder im using the bin from path \bldc-fix_noisy_currents-firmware_6.02\package\75_100_V2\75_100_V2_no_limits. bike pulls strong except for on takeoff. if i accelerate too hard from a stop/low speed it will jerk and sometimes kick the abs overcurrent fault. from googling it seems like it could be motor saturation. could also be because the motor wiring is wound for high speed. sounds similar to casainhos description of his first motor. i dont know if phase filtering will help anything but figure i might as well try it since this controller supports it. is there firmware here that has phase filtering still enabled? or is there an easy way to compile casainhos firmware version with that one option changed? i dont know anything about programming have just been using the bin files.
dont know if i should post in this thread or not. im sharing in case it helps someone else. tried jaykups firmware with the flipsky 75100 pro v2. tried all kinds of settings i could find online but still seemed to stutter seemingly around the point it switches over to closed loop when launching hard from a stop. when i raised sensorless erpm to high value ~6000rpm seemed to work fine and fix the issue.but eventually while testing further started to cog and would not even take off under light throttle without jerking/cogging. noticed it would clear up when jiggling the hall sensor connector alittle and holding throttle. graphed rotor position in realtime data and found hall signal absent unless putting alittle pressure on hall sensor connector to one side or the other. from googling this looks like a jst gh connector.Common JST Connector Types – Matt's Tech Pages. i thought it might have a bad crimp so i extracted and replaced some pins from the spare 4 pin connector i had but the results were the same. connector feels fine and clicks into place when i plug it in, but does not seem to make good contact on either side. from the pics you can see it should touch the pins on either side of the socket but doesnt do that unless i push on one side. ive ordered another connector and hopefully that fixes the issue and is not a poor socket connection to the pcb as i dont think ill be able to repair that.
Is there any test on VESCTool for the hall sensors? maybe that can help validate hall sensors…
On VESCTool more recent, from the git main branch, it have an option to see the difference from the hall sensors to the sensorless calculate angles…
Here a screenshot I took yesterday using the Flipsky 75200 V2 - so in my case, at that specific speed, is about 6 degrees that difference - maybe this can help to validate the hall sensors:
icant remember exactly where in realtime data but i had just been turning the wheel slowly and it looked like it was showing the steps from the hall sensors. also i checked voltage at hall sensor connector power and ground at motor and was only getting .8 volts. also hall sensor detection was showing failed result. in the end i think the socket at the pcb was too loose to squeeze the pins enough. i soldered the wires directly to the pins on the board and secured it with epoxy, and hall sensor issue is fixed and runs fine now. but still has jerking under heavy throttle/low speed unless i set sensorless erpm to a high value, basically running on hall sensors for longer. casainho is the firmware version you shared for the flipsky 75100 pro v2 as well? ive tried the bin in your zip folder under \75_100_V2\75_100_V2_no_limit, but am wondering if the one in FSESC_75_200_ALU\fsesc_75_200_alu_no_limits would be better for the flipsky?
Pretty common for them to get noisy at high current levels, they operate based on magnetism, so does the motor but at massively higher levels, then theres the part where the wires for the sensors are bundled up with and run parallel to 3 wires with huge amounts of power running through them in comparison to the hall signals.
Low rpm stutter i have found is solved by raising sensorless erpm to 5000-6000, raising “hall extra samples” to 4-6, lowering “gain” under foc>general, and if needed, lowering motor current . Its hard to get it perfect since the hub motors operate at such low rpm and the hall sensors only read every 60 or 120 degrees (of erpm, not actual rpm). For more precise low rpm positioning, encoders or resolvers are far better than hall sensors, but not found on typical hub motors.
Also sometimes playing with saturation compensation, or trying a different observer can help, sometimes mx_lemmings works better, sometimes ortega works better, every motor/controller/tune seem to respond different, so play around with them and see what works for you.
I just tested / use with the Makerbase 75200_v2. I build the firmware in the hope it will work for the other models.
So, if running with hall sensors are working well, then the issue is with the sensorless. On that case, I would say the issue is with the detected parameters like the inductance, etc.
When you do the FOC detection, do you see detected parameters always changing? if so, most probably the parameters are not well detected. With this firmware, the parameters detection are always almost the same - the seems good to me!!
thank you for the input guys. the detections were pretty consistent with resistance, inductance, flux linkage etc. at +or -5 values. i tried different current values and time constants for detection, saturationcompensation, extra hall samples, zero vector frequency, vadicus method, high current sampling mode, different observers and everything else i could find online. but it doesnt matter now since the hall sensor issue starting occuring again and when i tried to solder it again and bend the wires to fit out of the cover hole some of the traces pulled off the pcb. ive ordered a mkb 75200 and will try again when it arrives.
Only the 75100 Pro V2 and 75200 Pro V2 from Flipsky supports phase filters, NO makerbase one supports phase filters except for the 84_100
keep in mind you should NOT use Jaykups FW on the 75100 Pro V2 and 75200 Pro V2 due to that and other differences, they’re actually fully compatible with official VESC firmware
they added the option to remote turnon/off though.
i actually had messaged makerbase and they sent me a bin file to use and said phase filter should work on the 75200 version 2, since the one i received was version 2, but wasnt updated and firmware in vesctool was showing version 5. instead of 6.02. 75200-FW-6.02.zip (683.3 KB)
lines 54-56 in hw_mksesc_75_200_v2_core.h also seem to say it should support it? Request to merge my hardware into hwconf-Makerbase by makerbase-motor · Pull Request #672 · vedderb/bldc · GitHub
unfortunately ive become too good at breaking these things and managed to damage something while trying to mod it for my voltmeter/key switch. now i have the flipsky 75200 pro v2. i finally have it running well enough that im satisfied, with these settings:
when i tried to get it working well with phase filters on, it didnt seem to make much difference, so i tried using the bin from casainhos github [fsesc_75_200_alu_no_limits]. and thats what im still using, but if as you say the stock firmware is better on the pro v2, i may flash that and try it again.
Hello, can You post more photos of this revision, from both sides. I saw that 75100v203 have changes in cooling. Is there aluminum PCB with soldered MOSFETS, for better heat dissipation? Can You share more detailed photos which will prove that?
