Hey I missed this at the time but that looks very risky IMO! The original design is already quite questionable because it uses the legs of the mosfets as mechanical support; the mosfets are bolted to the heatsink but the board is only half supported in there so the weight of the board kind of hangs on the legs of the mosfets. This one is even more so, because the board is completely unsupported. I know the board isn’t incredibly heavy, but those legs are not at all designed to be load bearing and especially not in a very high vibration environment where they flex back and forward at every vibration
For reference, serious designers would be very upset if an individual capacitor of smaller size than those ones was allowed to freely stand vertically without attaching it with silicone to something stable like a heatsink that’s screwed down or the chassis. Allowing those mosfets to free stand wouldn’t fly at all normally even with zero weight on them, never mind an entire board with big caps
i get what you are saying but there really isn’t much difference, the only real differance is mine is not in a case, its bolted up to the homemade HS in the same way. if i remember correctly when i opened the case it was kind of just suspended there too with no support.
so far the trike has about 500 niles on it and nothing has shifted or failed, thats not to say it wont happen 2 mins in on my next ride but so far so good.
yeah the main difference is that the board in the original enclosure is sitting in slots in the aluminium, and capped at both ends. There’s still a small amount of play in there and it’s not bolted down (in mine I 3D printed braces for the capacitors and the PCB and siliconed it in place to remove this) whereas in the open air situation the whole board can swing from the mosfets and the capacitors can swing from the board. I know yours hasn’t failed yet but the difference between bad support and none at all really is fairly big
yeah I was trying to think of what to do about it. You could dremel out the original case so that your current connection to the heatsink is still there but the rest of it still exists, and silicone it down? The alternative is trying to make another enclosure with 3d printing or CNC and that’s a much much bigger job
I thought I had posted full pictures somewhere but this is all I can find at the moment
this is bordering on out of my knowledge, but silicone is very useful for damping and as a light duty adhesive; if you have two components standing side by side that can support their own weight and you’re just trying to stop them vibrating or wobbling then silicone is ideal, as in A below (side to side arrow is vibrational force). But in B it’s not enough. B.1 is arguably almost ok, the side to side vibration of the capacitor relative to the board can probably be managed but the gravity pulling the capacitor down (blue arrow down) is a lot. In B.2 it’s not enough, the weight of the board is a big lever pulling down from that point to bend the mosfet legs and strain the connection to the heatsink, and the only thing resisting that is the silicone
Agreed, any vibration or movement just results in metal fatigue and/or breaking of seals or the fracturing of the black epoxy case of a component. You want the component locked to the pcb (silicone in the right place and quantity can definitely do this) and then to minimize the vibration/impact the entire PCB experiences.
Set the ESC to 100 battery, 100 motor amps, but never got over 77 amps on the motor. I think the phase wire are too long and small (12-13 gauge or 3mm2) to get the full 100 amps. Would like to switch to 10awg silicone at least.
The bike had a fairly heavy load. He was pulling a trailer with his 6 year old in it, had several batteries in the tail bag for spares for other bikes on the ride, and is a fairly big guy. Batteries were only charged to 80%. We swapped batteries at the mid point.
Overall performance was pretty good. Unloaded it accelerates a bit slower than my skateboard (12s4p 30Q single 6374 motor with a Trampa V6 on 90mm thanes with tall 16/36 gearing set to 80A battery, and 80A motor) but has nice power delivery and keeps pulling well to top speed.
The ESC touched 80C a few times, but no noticeable thermal throttling. Field weakening was not active. Was about 60 degrees out. We are thinking of moving the ESC to the front of the bike to get some air flow over it and cool it down.
Curious to see if it will keep working through the season. The temperature sensor just sits on the aluminum bar that the mosfets are attached to, so I think the mosfet temp reading would be somewhat delayed. I guess as long as the temperature doesn’t rise too quickly, it should be able to thermal throttle in time to prevent a failure. I still think cooling is the most important problem to solve with this ESC.
I agree…hugely delayed IMHO and significantly below the actual MOSFET temp. Especially if a thermistor is being used. They already have a delay (response time) of about a minute and their shape makes it hard to thermally couple them well to anything flat, leading to low temp readings.
Hey. I am looking for a v6.0 beta (latest from GitHub) binary for Flipsky 75100V202.
I am required to use the Lisp feature from the v6.0, but after compiling 75_300 (with phase filtering disabled and adjusted maximum settings) I always ran into issues. Most of them noticed when using Lisp with M365 Dashboard script.
For example, issues like:
“dashboard light is turning on and off when accelerating motor”, which is being read by gpio-read 'pin-rx and somehow starting motor is causing it to read something from RX.
and motor doing weird noises and jittering making it impossible to properly drive.
So, there are likely a few hardware differences between the 75100V202 and the default 75_300 firmware. Can someone help me out with this?
Yes, you cannot use the 75_300 firmware configuration. If you are already at the compiling stage, just use my .c and .h files from the “Source & archive files” section on the main post. Then adjust the conf_general.h to use those files.
Thank you, I have now compiled the firmware from GitHub with your .c and .h files. It works good and seems fine for now. I will report back if I notice any other issues, but so far all issues seem to be gone.
I also noticed, temperature reading was wrong on 75_300 as well, it always showed like 30-31, now it shows 21 (like on v5.2 and 5.3).
A lot of new info has come out in the past 5 months since I first created the original post.
I updated the first post to detail out Flipsky’s shady business practices including the story behind the original creator of the unit, and why there are now two versions.
Hi, i bought one of these controllers, and i would like to thank you a lot for the work you did, i installed it on my scooter and managed to configure the motor for now, i also updated the firmware and it went flawlessly.
Many thanks.
is this version compatible with new flat aluminium PCB 75100? did anyone try?
I have disassembled the ESC, but it doesn’t seem to have version written anywhere on the PCB on the front side - I don’t want to disassemble heatsink side, as they used thermal paste to from MOSFETs to back plate. Also, it uses 2 PCBs now - top one holds chipset with connectors and bottom side hosts all power components.
