Paul & Sabrina's cheap DIY 144v motor controller

[ChedMpls]

finally got my cougar completed and I am bench testing. i'm having a problem with power control - as i try to throttle up i get no power and then reach a point where i get full power - nothing in between.
I removed my pot and it checks out ok - linear, 0-5k.
my test setup uses a 10v rcCar battery and a 6v salvaged drill motor.
rtd explorer shows nice throttle response graph but shows PWM going from zero to 510 all at once.

Any ideas? Can't wait to get this sucker tweaked so i can start building!

ched

[flores]

I guess the current is to small to measure.. then the controller will automatically increase the duty cycle

[jackbauer]

Its a vw/audi coolant pump. No idea about flow rate. If it can cool and engine it can cool my igbts

[MPaulHolmes]

Originally Posted by ChedMpls

finally got my cougar completed and I am bench testing. i'm having a problem with power control - as i try to throttle up i get no power and then reach a point where i get full power - nothing in between.
I removed my pot and it checks out ok - linear, 0-5k.
my test setup uses a 10v rcCar battery and a 6v salvaged drill motor.
rtd explorer shows nice throttle response graph but shows PWM going from zero to 510 all at once.

Any ideas? Can't wait to get this sucker tweaked so i can start building!

ched

I think flores is right. There is a dead zone with the throttle, and as soon as you are out of the dead zone, the microcontroller tries to make current match the throttle position. So, 5% throttle means the microcontroller does the only thing it can to make the current be 5% of 500 (25 amps). Namely, it quickly ramps up the pwm in the futile attempt to receive 25 amps of feedback from the current sensor. It looks weird on the bench, but feels nice in the car.

Oh, I finished (mostly) all 4 control boards for the 1000amp controllers. I can't feel my hands though, because it's freezing in the garage. I should have soldered my hands a little to warm them up. I can't even type aahhhh.

[DJBecker]

Originally Posted by flores

I guess the current is to small to measure.. then the controller will automatically increase the duty cycle

You might consider using one of the current sensors from Device Tech. They sell the small current sensors on FleaBay for under $10, shipping included.

They have a round 1/4" aperture, which is appropriate for their measurement rating. If you have a low current motor and need more measurement sensitivity you can loop the wire through the toroid a few times.

They run on 4.5-6.0V with the same centered ratiometric output of the LEM sensor. Ratiometric means that the output scales with the supply voltage: with a 5.00V supply the no-current output is 2.50V, with a 4.80V supply the output is 2.40V. This exactly what you want if the reference voltage for the A/D converter is the same as the supply voltage.

One aspect that is nicer than the LEM sensors is their lower power draw, with a disable input that puts the sensor in a micro-power off mode.

[flores]

I'm a bit confused on how to place the busbars on the 500A controller.

In pictures I see sometimes the busbar ends all on one side and sometimes the middle busbar is on the reverse side of the case.

Is there a special reason for this difference? Or is it just personal preference? Or a difference between rev D and C?

thanks

[adamj12b]

Originally Posted by flores

I'm a bit confused on how to place the busbars on the 500A controller.

In pictures I see sometimes the busbar ends all on one side and sometimes the middle busbar is on the reverse side of the case.

Is there a special reason for this difference? Or is it just personal preference? Or a difference between rev D and C?

thanks

Which end they stick out is your decision. Some layouts of the car work better 1 way instead of the other. It does not matter.

If you want, you can use a longer B+ bar so you have the B+ and B- on one side, then B+ and M- on the other. I dont think ive seen one like this yet though....

-Adam

[thargrav]

Originally Posted by DJBecker

A bunch of interesting parts are available only in surface mount.

Small boards like this are way easier than you expect. The critical element is a solder mask layer on the PCB.
I've been amazed at how well it works.

I use solder paste (not "soldering paste" e.g. flux) and a hot air gun.
Specifically the inexpensive "Mechanics" brand bought from DealExtreme and a $10 hot air gun from Harbor Freight.

Dab a tiny amount of paste on each pad, place the part, wiggle a bit and remove it to make certain that the ink-like paste has spread evenly. Reversing the part helps even out the paste. Smearing outside the pads is no problem, it actually helps see melting progress later.

Once everything is placed, put the board on a level heatproof surface. With low air speed heat the board from far enough away that you don't blow the parts away. Nothing will happen for the first 30 seconds or so, then the solder will suddenly "flash". All of the paste smeared on the solder mask will just instantly disappear, sucked into the joints or left as almost invisible grains that dust away later. Keep heating for another 10 seconds or so. Now the hardest part: don't touch the board for five minutes. I've picked up boards after seemingly minutes (OK, maybe only 20 seconds) only to watch parts shift or slide off, molten solder trailing.

This doesn't work as well on bigger boards will different/bigger size parts. There you have to gently preheat everything to a controlled 100C or 125C, and be careful to use the hot air evenly during the re-flow stage. The preheat keeps the big parts from absorbing so much heat during reflow, and it needs to be gentle to drive off moisture so that you don't get micro steam explosions.

A trick for reflowing SMT solder is to use a toaster oven. I have used one for years and they work great. You place the populated board in the toaster oven, turn it on broil and set a 15 minute timer. Once timer goes off you turn the oven off and wait 15 more minutes.

Tom

[flores]

Originally Posted by thargrav

A trick for reflowing SMT solder is to use a toaster oven. I have used one for years and they work great. You place the populated board in the toaster oven, turn it on broil and set a 15 minute timer. Once timer goes off you turn the oven off and wait 15 more minutes.

Tom

Tom,

Can you share the model/make of this oven? And at what temperature do you set it? Maybe a picture? Thanks.

[thargrav]

Originally Posted by flores

Tom,

Can you share the model/make of this oven? And at what temperature do you set it? Maybe a picture? Thanks.

It's a cheap one from Wal-Mart, no particular model but it needs to have a top heating element and a broil setting, which is basically all the way on.