DPRG
DPRG List  



Fwd: [DPRG] RE: Re: Motor Control and parts shopping

Subject: Fwd: [DPRG] RE: Re: Motor Control and parts shopping
From: Gary Burton mechgeek2000 at gmail.com
Date: Mon Aug 31 19:47:38 CDT 2009

Alas, I should check my work at least twice before hitting the send button

---------- Forwarded message ----------
From: Gary Burton <mechgeek2000 at gmail.com>
Date: Mon, Aug 31, 2009 at 7:46 PM
Subject: Re: [DPRG] RE: Re: Motor Control and parts shopping
To: Paul Bouchier <bouchier at classicnet.net>


Mr Bouchier and other DPRGrs,

Have you built a tracked bot? Was it a success or failure?

I've been investigating several different types of drives for a bit and
still investigating.  I'm doing this for fun, I don't care if it can cruise
all over the place and end up at the same place it started at, or within a
few millimeters of that point. An ability like that is awesome if the bot is
in an environment of known constants, but mine isn't going to be. It's going
to have deal with carpet, linoleum, going between the two.

I picked the tracked drive for several reasons, it's free, it was already
built and if it doesn't work out, I can try out something else. Besides, I
didn't figure on using dead reckoning for it's sole means of navigation.
Humans and animals don't use it solely for their own navigation purposes,
basicly we use whatever works for us, whether it's using reference points,
scents, just stumbling upon it or asking for directions.  As for bots, they
are small enough that anything can knock them off of their path, such as a
human, an animal, a slick spot of the floor, a bump in the carpet or a new
obstacle.  This is going to happen whether it's tracked, wheeled, hovers or
walks. It happens to humans as well, how many of us have gotten turned
around in the dark and bumped into something that was not supposed to be
'there', or stepped on something in the dark?

The tracked body is as good as any other in a domestic environment without
stairs or steps. The tracks are encoded to keep them running at the  same
general speed, to know when  they are stuck, and something else for the MCU
to monitor and use as part of it's nav routines.  Yes there will be slippage
on carpet, obstacles, wet spots to slide through, all parts of it's life
that through programming and experimentation it will overcome most of them
and learn how to avoid the rest.

Your wisdom and knowledge is heeded and appreciated, but this is for fun
after all, is it not?  Nobody is under contract to produce a robot that can
do something. I'll be glad when mine is able to bump into something
*chuckle*.

I hope everybody had a wonderful weekend




On Sun, Aug 30, 2009 at 4:07 PM, Paul Bouchier <bouchier at classicnet.net>wrote:

>  Gary - think hard before you go the tracked route. You can't use encoders
> to tell how far you've turned - you virtually have to use a compass (which
> brings magnetic susceptibility) or an IMU or more exotic techniques. Even
> straight-line dead-reckoning is harder, because you don't know how many
> degrees the robot turned when the left track ran ahead 100 counts, or
> whether it will turn the same amount back when you run the right track to
> catch up. Speed isn't the issue - it's dead reckoning accuracy.
>
> regards
>
> Paul
>
>  -----Original Message-----
> *From:* dprglist-bounces at dprg.org [mailto:dprglist-bounces at dprg.org] *On
> Behalf Of *Gary Burton
> *Sent:* Saturday, August 29, 2009 7:47 PM
> *To:* dprglist at dprg.org
>  *Subject:* Re: [DPRG] RE: Re: Motor Control and parts shopping
>
>   I'm still going to go to with the L293D for right now. It's tried,
> tested and has a minimum of parts. The voltage needed for the motors is well
> within it's range. There is an upgrade path for both chip and motors. Plus
> the fact the programming for it is quite simple. I'd prefer to use steppers
> or gearheads, instead of geared down cheap variable DC motors. I'd also like
> to get it moving and move on to encoding the drive shafts and working on the
> sensors. This is my first bot, so vast improvement may seem built in and I'm
> taking notes on the next one which should be a vast improvement.
>
> So far Leo will have tracked drive driven by geared down DC motors from a
> L293D, bumper sensors and IR sensors for object finding at range all
> controlled by an AtMega328.  Wheels might be faster, steppers would probably
> be more efficient with the proper board to drive them and either a few ARM
> modules or a few networked At328 chips would be preferred. Baby steps is a
> good thing and makes debugging for somebody who's not done this
> succcessfully before, much easier.
>
> Hopefully, soon I'm using the stuff like you guys probably are, but till
> then bear with my simplistic and naive decisions.
>
>
> On Fri, Aug 28, 2009 at 3:14 PM, Jose I Quinones <jiq at avayan.com> wrote:
>
>>  You would think it is like a Servo controller, but that is in fact a
>> miss conception. I am writing a tutorial depicting the differences between
>> RC PWM and Speed Control PWM.
>>
>>
>>
>> But basically, RC PWM encodes the position information on  a pulse that is
>> 1.0 ms to 2.0 ms wide with 1.5 ms encoding dead center (if properly
>> calibrated).
>>
>>
>>
>> With Speed Control PWM, there is no information encoded. What you are
>> doing is segmenting the available energy into percentages. Per example, if
>> you have a 12V battery, a PWM on the ENABLE or the PHASE pin translate into
>> how much time the motor is ON and how much it is OFF, which is equivalent to
>> a fraction of the 12V. Hence, a 50% duty cycle PWM with a 12V battery is
>> analogous to a 6V battery. That is why you can use this to control motor
>> speed. Because speed is directionally proportional to applied voltage. You
>> can see this better on a video I released last week. I expect to present
>> more videos in the upcoming weeks, in case you are interested in motion
>> control basics.
>>
>>
>>
>> http://robot.avayanex.com/
>>
>>
>>
>> JIQ
>>
>>
>>
>> *From:* Jose I Quinones [mailto:jiq at avayan.com]
>> *Sent:* Friday, August 28, 2009 1:27 PM
>> *To:* 'dprglist at dprg.org'
>> *Subject:* Re: Motor Control and parts shopping
>>
>>
>>
>> Hi guys!
>>
>>
>>
>> Allow me to recommend something totally different and to my understanding
>> considerably easier to use, better deal, and better performance. Maybe it is
>> “less fun” but I believe it can still be an enjoyable experience.
>>
>>
>>
>> TI has a new device called the DRV8800PWP. It is a 2.8A H Bridge. The cool
>> thing about is that all you need to supply is ENABLE and PHASE. So you do
>> not have to worry about all the IN1/IN2 combinations, decay modes and stuff.
>> Just to add to the coolness, it is protected against output shorts, thermal
>> overrun and demonic possessions. This thing is practically immortal! The
>> only thing that it will not survive is voltage inversion or over voltage but
>> I have hardly heard of an H Bridge that contains those features. I believe
>> we are supposed to be minded enough to not let something as simple to
>> happen.
>>
>>
>>
>> That being said, you can PWM either the ENABLE or the PHASE and you get
>> speed control on both situations. You may be saying “WHAT THE HECK??? PWM
>> THE PHASE?” If you have never heard of this, it allows you to control both
>> direction and speed with a single PWM timer. Per example, if PWM duty cycle
>> (DC%) is less than 50%, then the motor moves Counterclockwise (CCW). The
>> closer the DC% is to 0%, the faster the motor moves. But if DC% is above 50%
>> then the motor moves clockwise (CW). In this case, the closer the DC% is to
>> 100%, the faster the motor moves. This is of course only possible when
>> ENABLE is asserted.
>>
>>
>>
>> I have a design on my Open Source Design web page around this chip and am
>> working some new versions. Feel free to take a look and copy whatever you
>> want. It is as free as free gets. I also have some bare boards to sell in
>> case you want to build your own. They are sold at whatever cost me, so in
>> other words as cheap as I can afford.
>> http://www.avayanelectronics.com/Products/AE-DC1/ae-dc1.html
>>
>>
>>
>> BTW if you need superbly high currents (which does not seem to be the case
>> for those considering to use a 600 mA device), then there is another larger
>> design here:
>> http://www.avayanelectronics.com/Products/MDL-7960/mdl-7960.html
>>
>>
>>
>> Thanks and best regards,
>>
>>
>>
>> JIQ
>>
>>
>>
>> Feel free to visit some of my pages at:
>>
>> http://www.avayan.com
>>
>> http://avayanings.blogspot.com
>>
>> http://www.comicups.com
>>
>>
>>
>> _______________________________________________
>> DPRGlist mailing list
>> DPRGlist at dprg.org
>> http://list.dprg.org/mailman/listinfo/dprglist
>>
>>
>
>
> --
> Find my blog at:
>
> http://mechgeek2000.blogspot.com/
>
>
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org
> http://list.dprg.org/mailman/listinfo/dprglist
>
>


-- 
Find my blog at:

http://mechgeek2000.blogspot.com/



-- 
Find my blog at:

http://mechgeek2000.blogspot.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.dprg.org/pipermail/dprglist/attachments/20090831/c54515e8/attachment.html

More information about the DPRG mailing list