suspect bad FET driving engineering in Chinese controllers

[emilx]

Hello everyone,
I'm new on this forum and I'd like to use this opportunity to say hi to everyone!
I happen to be an ebike fan like everyone else, but my last 3 diy bicycles got stolen, so so lately I'm doing more of an documentation work rather than a practical one.
At the address
http://www.avdweb.nl/Article_files/Solarbike/Motor-controller/China-BLDC-motor-controller-36V-250W.pdf
I found a pretty accurate schematics of my old ku63 which was very helpful in order do fix it after my friend managed to fry it :lol: .
But doing some further analysis I learn that they basically used the same FET to switch both high-side and low side. After copying parts of the schematics in Orcad/Pspice I'm pretty convinced that there's no way the controller works the way it should, as the gate voltage if the high-side FETs drops under 5V soon after the motor gains a decent speed. I believe is that the RdsOn in High side Mosfets increases, the efficiency drops and the max speed is reduced drastically.

So Id like to ask all of you, am I missing something? Is it ok to use n-channel MOSFET to switch both LOW and HIGH side?
If the Chinese manufacturer got it wrong, how difficult would it be to implement a FC8201 based controller paired it with a kickass arduino control module with a screen everything?
I guess that after a year of developing and testing I could come up with a valid controller, with infinite customizable options at a pretty competitive price, and I could also release it as an open-source project, so everyone could modify it the way they need. I think it could be even manufactured at an industrial-level with a modest expense
What do you think about this project? Do you think is it worth?
Thank you!

 

[bearing]

Is it ok to use n-channel MOSFET to switch both LOW and HIGH side?

It's not only OK, it's the preferred way. N-ch are far superior as power MOSFETs compare to P-ch. P-ch only offers an advantage when Vdd is less than max gate voltage (and some other special cases).

The gate of the high side is often driven by a boot strap circuit. Thats is also the case in your schematic. C14 is the bootstrap capacitor, which will have about 14V. It gets charged every time the low side FET is conducting.

 

[emilx]

thank you very much! I guess I still have to learn a lot. Now everything makes sense. I was about to take out the mosfets and switch the high side mosfets with the p-channel counterparts and to modify the circuit accordingly only to see what happens.
Probably I'll make a couple more simulations and maybe have the Vgate checked with the oscillo, only to make sure that the boot strap circuit is working.

 

[amberwolf]

There are also several controller development threads here on ES that you will want to find and read; you'll probably find a lot of useful information in them.

 

[HighHopes]

thanks for sharing that schematic.. it was really interesting to see how to do an ultra cheap - very low reliability - design. my guess is that the PCB + all the parts for that design would be at most $100 .. and that's one-off pricing.

 

[HighHopes]

oh, to answer your question .. if you were to stay at 36V and 250W as the china controller you posted .. yes i think your project could be worth it. but .. what is your application? 250W is fairly low power even for a bicycle. maybe for a kid's push scooter? maybe for a kick-ass remote control truck? but a bi-cycle for an adult.. pretty low power.

if you went to higher power and/or higher voltage i suspect your plan would not work.. at least as your plan exists today.