high power/high speed BLDC controller build

[dirty_d]

cleaned up the pcb a bit and added pads for hall signals, and pull up resistors for the hall sensor outputs.

 

[dirty_d]

here it is as i have it now, i think im pretty much ready to order tha parts now. anyone notice any last minute mistakes?

 

[wrobinson0413]

.

 

[dirty_d]

im just gonna go straight to the pcb, i wouldnt be able to breadboard soemthing like this up. the pcb is cheap anyway. if somethings wrong i can desolder the stuff off, i got one of those solder suckers. ill add space for another cap on there.

 

[wrobinson0413]

.

 

[dirty_d]

eek, on the stove? i did mine outside, i had fun etching and drilling and putting it all together. the only things im going to use from the old controller are the parts, im going to order some 0.1875" copper rod to use as power rails, ill drill holes in them to mount the fets and solder them in. there will be two rods running parallel to each other with three separate parallel ones inbetween for the phases. ill have the metal tab of the high/low side fets facing each other and an aluminum angle piece will get screwed inbetween all of them, the top of the enclosure will be aluminum flatbar, something like 4" wide and 5" long which will screw onto the top of the angle with thermal grease inbetween, both will be 0.125" thick i guess. ill probably make the sides out of plexiglass painted back or something. im not really sure about the size, the pcb is about 3.2"x3.9" its smaller than the brushed one i made. im not sure how the power rails and stuff will fit inside. the hall sensors i was looking at say to use a 100nF cap as close to the device as possible, i was thinking maybe i could use one of those tiny SMT caps and solder them between the to-92 legs or soemthing, i didnt notice any caps showing in the pics of the crystalyte hall sensors, maybe its just the ones i was looking at.

 

[ZapPat]

Hi dirtyD - great progress!

Good idea using extra copper rod for more current capability on the PCB, and if you want to save money and get something fast, maybe just use heavy-gauge house electrical wire, maybe AWG#10 or so?
As for your hall sensors supply caps, I have done what you suggested (directly soldering cap on to host chip) with some SMD temperature sensor chips. It works well, but have a good small tip handy to do the job! Maybe use temporary glue (crazy glue?) to hold them together before soldering - this might prevent you from blaspheming too much.

Good luck!

 

[dirty_d]

yea i was gonna look for some large solid wire somewhere, but the site im getting the aluminum from has the stuff for like $1 per foot.

 

[nasukaren]

Congrats! I love my little hobby city motor in my moped-EV. It just keeps whirring along like a dream. You've most probably read my whole saga on how I burnt out my first el-cheapo ESC in the first test run and ended up getting a Castle Creations Phoenix.

I'd love for their to be a good home-brew controller for these RC motors -- especially one that can take a Magura throttle or hall-effect throttle directly, without having to mess with a servo tester hack.

Good luck and keep us posted!

Karen

 

[dirty_d]

ok i ordered all the components and metal, so it wont be long! if this doesnt work im gonna cry, lol.

 

[dirty_d]

i found a couple mistakes, i forgot to use pull up resistors on the lm339 outputs, and something else i forgot.

 

[dirty_d]

i fixed the schematic and pcb, i think its all set now. i might try regenerative breaking too, i shouldnt need any additional hardware for it, just another button to switch the throttle from throttle to brake. everything else would be done in software. without additional hardware it wont have torque braking control though as my current sensing deal only works in one direction.

 

[wrobinson0413]

.

 

[dirty_d]

oops, too late for that i think. i already etched and soldered almost everything on, i left out the backemfconditioning stuff since i have hall sensors. after i get it working with halls ill add the backemf stuff and see if it works. it looks kinda funky because the board is transparent.

 

[wrobinson0413]

.

 

[dirty_d]

i wired up the hall sensors and used some small heat shrink tubing on each terminal, then superglued them all together. it fits perfectly inbetween the teeth. they stay put without any glue, but once i get the thing working and im sure they are between the correct poles ill glue them in also. i tested them out before putting them in my waving them by the rotor magnets. they worked fine, switching right inbetween the magnets.

 

[wrobinson0413]

.

 

[wrobinson0413]

.

 

[dirty_d]

not sure now many strands, but all together it looks like 12 or 10 gauge.

 

[dirty_d]

i hooked some leds up to the phase logic level outputs and wrote a quick program for the avr to make sure the hall sensors are working. looks good! i get the right sequence of states as shown byt the leds by slowly turning the motor around. you can imagine how fast this stuff is happening when the motor is at full speed, just spinning it fast by hand makes it look like they are all on at the same time with some flicker, its pretty neat maybe ill take a video.

 

[dirty_d]

hmm, how do you embed youtube videos in posts?

http://www.youtube.com/watch?v=tDkHCi3IU54
[youtube]www.youtube.com/watch?v=tDkHCi3IU54[/youtube]

 

[lepton]

After opening the motor up, do you know if its wound as a Delta or Wye configuration?

 

[dirty_d]

its delta

 

[dirty_d]

hmm, yea the copper bars dont work so good, too much of a pain in the ass. i just used some 12 gauge copper wire, its gonna have to do for now. i gotta make the driver pcb deal now.

 

[dirty_d]

it freakin works! http://www.youtube.com/watch?v=nQtVXJVdxR0. i get about 2A at 19V and 4.5A at 38V, so i guess i have the halls in the right place. the outside gets pretty warm not really sure why since its only 4.5A.