400 series brushless motor deck conversion

[mfortuna]

Here's my latest project. I was able to find a brushless motor that is very close in dimension to the OEM 385 style motor. I had to trim back the transmission since the brushless motor is just a little larger in diameter. The yellow bushing end is a tight fit and I ended up gluing it on with silicone adhesive. The speed control is intended for RC airplanes and is 4S rated, meaning it can take up to 4 LIPO cells or 12 NICAD/NIMH cells.The PCB is a propeller proto board, the same one as used in my budget ConnectR project.

The ESC is self arming meaning you have to go through a set process to get it ready to control an airplane motor. The ESC requires the transmitter to be set to full throttle before turning on the ESC. Once you turn on the ESC you need to perfrom certain throttle settings for a certain amount of times.

Since the ESC supplies power to the receiver I'm hoping the ESC firmware allows for a delay in seeing the full throttle signal. I have messed around with prop code and have not yet got the ESC properly armed. Next up I will measure the servo pulse from a RC system and duplicate the settings in prop code.

I may need to modify the project a bit if I can't get the ESC to arm because of the delay in seeing a full throttle signal. I will have to add a MOSFET which is controlled by the propeller. It would delay turn-on of the ESC until full throttle has been set. I would also have to power the propeller through a regulator instead of the ESC (which has a built-in regulator).

[Gordon]

...I was able to find a brushless motor that is very close in dimension to the OEM 385 style motor.

He's just soooooo cute!!

... The yellow bushing end is a tight fit and I ...

All of those that I have handled are made with bores too small for the motor case! iRbt just does not care.

...The ESC is ...

"ESC" = Electronic Servo Control? Or what?

... Next up I will measure the servo pulse from a RC system and duplicate the settings in prop code.

If you are referring to measuring the PWM wavetrain from a 4XXX-Roomba's hand-held Remote Control unit, ...
{EDIT: Ooops, you are not; so I have deleted my non-applicable Roomba data (just to reduce clutter).}

In looking at your pix, a physical handicap, so far as the auto-load accommodation adjustment working as planned, came to mind. Perhaps you have addressed it in your own thinking, but those fat conductors have got to be thinned as much as possible to allow wind-up as the motor case revolves. Seems like this Roomba application of that BM will not be taxing anywhere near its full current draw, so smaller wire-size should be OK for it. Comments?

[mfortuna]

ESC = Electronic Speed Controller.

The remote controller's pulse train is not the same as a radio control pulse train. The RC is more simple. It is just a variable frequency pulse. I used similar code on my Security Bot. I think one issue is I may have the code backwards. I think I was starting with the lowest frequency, not sure if that corresponds to brake or throttle.

The motor leads look thick but they are very flexible. Hopefully they will not be an issue.

I'm hoping this experiment will lead to longer run times and longer life. The three major components plus the pinion gear add up to about $90 so this is far from a economical hack. But I can always find a use for the motor and ESC on an RC project.

[mfortuna]

I forgot to add is if I get this all working I plan on adding a couple switches for up/down speed control. This will let me tune the brush deck for cleaning ability/battery life. Once I find an optimal speed I will hardwire it in the code.

On a side note I have gone back to using brushed motors in my RC race vehicles. I prefer the smooth throttle response plus the high brushless power is not good for a racer of >50 years old.

[mwkeefer]

So your doing a BLDC motor on a roomba... 2 or 3 phase? I mean, willl you use a small hobby RC controller for speed, what is the purpose... high rpm + reduction to improve the speed and I assume scraping power + suction power (this is the technique used in the really expensive ball type vaccume, 10K+ RPM reduced for cyclone vortex... that uses a BLDC Inrunner.

Need any help or ideas, drop me a PM

-Mike

[mfortuna]

It was mainly to get around the heat and wear on the brushed motor. I wasn't looking for an improvement in cleaning power but maybe an inprovement with run time. This project has been on the back burner, I need to get back to it before the weather gets warm.

[mfortuna]

Here it is almost 2 years later and I finally made some progress (although not all good). My main issue was I could not get the ESC to arm. I used my USB scope to measure the servo pulse from a RC transmitter and receiver. What I found was the pulse width for full brake was about 0.9msec and the pulse width for full throttle was about 2.0 msec.

I had been using a full range of settings which are about 0.5 msec to 2.25 msec.

By using the same 0.9 msec to 2.0 msec range on the prop board I get the ESC to arm 100% of the time. That is the good news.

There is bad news.

At about 1/3 throttle and using an 11.1V LIPO battery the motor gets extremely hot after about 10 seconds of no load running. I could not hold it my hand which is bad!

Also my glued on yellow bushing/endcap fell off which means I would need to mechanically attach it. I would probably need to get a small drill tap and make some screw holes.

So at this point this project is on the back-burner again!

[mfortuna]

I have some other brushless motors kicking around. I'm going to try them (even if they wouldn't fit in the roomba) and see if they still get as hot.

[mfortuna]

I finally tested some other brushless motors and they work fine. They run very cool and show no signs of the heat produced by the original motor. I am now re-thinking this project a bit to see if I can use different motor.

The prop board does have a useful function now. It can be used as a brushless motor tester!