Paul & Sabrina's cheap DIY 144v motor controller

[motor_control]

Rule of tumb ! Do not rise current of MOSFET more than 25% from which is in datashett for 100 grd C !

[MPaulHolmes]

Originally Posted by motor_control

I think you need harware overcurrent protection.

Hardware overcurrent protection will definitely make this a more robust controller. I agree. It works fine without hardware overcurrent protection, but it is an important extra safety feature, so you don't have to just depend on the microcontroller. It will be implemented. Fran Sabolich on the EVTech list has helped me with this subject. He drew up a schematic as well which I plan on using in the final draft.
Fran's Schematic

I'll check out your schematic as well. Thank you!!!

Spaceeba! Voui Harasho gavaritia pa angliski!

[motor_control]

Originally Posted by MPaulHolmes

Fran Sabolich on the EVTech list has helped me with this subject. He drew up a schematic as well which I plan on using in the final draft.
Fran's Schematic

I've moded it slitely

Originally Posted by MPaulHolmes

I'll check out your schematic as well. Thank you!!!

In my circuit mosfet itself is current sensor. And then hoter mosfet is, then higer Rds-on and then lower protection treshold is.

IR2127 starts masurament after 750 nS after gate ON and if voltage from "D" throw resistors network is higer then CS treshold voltage - IR2127 turns off mosfet and FAULT pin is LOW (for MCU inform) FAULT will be osolute after "IN" lowering by control sirc.

www.irf.com/technical-info/appnotes/an-1014.pdf

IR2127 can work like low side driver too.

Originally Posted by MPaulHolmes

Spaceeba! Voui Harasho gavaritia pa angliski!

Russia, Moscow.

[bennelson]

Originally Posted by motor_control

Russia, Moscow.

Cool! This project is going totally global!

It's a good thing that electronics schematics are a language of their own!

[jimini]

Originally Posted by MPaulHolmes

I think you could let the battery sit for a while, check the resting voltage, from that guess at it's state of charge. Run it at a pretty low current (so the Peukert effect doesn't come in to play?) for a specific amount of time, let it rest for a few hours, check the state of charge and figure it out from that. That sounds really annoying.

I tried it with the wax paper. It needs fine tuning. I pressed to hard, and it smooshed the ink all over the place!

You might try this trick...
Get an old copy of National Geographic (apparently it has a high percentage of clay), Print your circuit on a sheet, Iron it onto your clean copper(400 wet or dry sandpaper should work), for a couple of minutes. Then etch as you normally would.
Worked really well for me.

[MPaulHolmes]

Originally Posted by jimini

You might try this trick...
Get an old copy of National Geographic (apparently it has a high percentage of clay), Print your circuit on a sheet, Iron it onto your clean copper(400 wet or dry sandpaper should work), for a couple of minutes. Then etch as you normally would.
Worked really well for me.

Thank you! I never would have thought of that! I have a whole bunch of them on my book shelf.

[motor_control]

I've use glossy photo paper for inkjet printing (Lomond A4 170 g m2 one side is glossy and ather is matt ). Glossy layer sealed microvoids on the toner routs on the copper.

I make routs from 0.12 mm ! I do not move hot iron. I assamble all together with load and turn on iron (therm 130 C) after 20 min I turn off iron and let them cool down 20 min. Then lay pcb to water for 15 min, after that I friction suaface of paper by figer (NOT deattach paper from pcb) and paper rolls in tyni scraps.

Photo of my pcb on my page Sorry, in Russian only.

My Procces



and pic icd2 programmer clone bcp - routes from 0.15 мм

[MPaulHolmes]

That etching job was beautiful! Thank you motor_control! I appreciate your help. This gives me a whole new way to think about how to do it.

P.S., I might end up switching to the ATMega16 too. I'm starting to run out of pins on the ATMega8 I'm using.

[motor_control]

Same as mine current protection is used in Curtis controller (maked out on LM339)

http://www.evdl.org/docs/curtis_1221b_schematic.pdf

http://cafeelectric.com/curtis/Curtis_manual.pdf

Some notes about MOSFET sizing:

Curtis has 35x IRF640 0.15 ohm 11A (100 C) 200V

Current protection 400 A

and rating: 400 A for 2 min
250 A for 5 min and
150 A 1 hour

Originally Posted by MPaulHolmes

P.S., I might end up switching to the ATMega16 too.

I sugest ATmega32 (same as ATMega16 but 32 kb flash)

[MazdaMatt]

If you are making any major changes, please consider a "plug-in" cpu board. I have a few "scrap" boards here that have very nice 32-bit 80mhz mcus on them (coldfire) and they would be an excellent processor for control... if i ever get around to building something