We would like to remind you of this upcoming event.

Monday Night Chat

Date: Monday, July 1, 2002
Time: 7:00PM - 9:00PM PDT (GMT-07:00)

Monday Night Chat happens every Monday at the SRS. Come ask
questions or offer advice live.

Directions:

The chat requires that you use an Internet Relay Chat (IRC)
program. Details on finding one are below.

For you experts, the chat is held on us.chatjunkies.org in the
#SeattleRobotics group.

For Windows users:

See the SRS website at

http://www.seattlerobotics.org/contact.html#Monday

Robin specifically said she wanted to include all SRS members; she knows how
big our mailing list is.

Doug

FThompson9@... writes:
>Does it count if I don't live in Seattle?
>
>Pherd
>Somers, CT
>
>
>In a message dated 06/27/2002 1:29:10 PM Eastern Daylight Time,
>doug.kelley@... writes:
>
>
>> Members, you can take advantage of Robin's generous offer by placing your
>> name
>> and address in the following table:

I had simply sent an email off to N&V when the very first
notice came out and they very happily put me on the August
"sample" list which is the intended distribution. I did
explain I was on the SRS email list but lived in Houston so
I couldn't go to the meeting. N&V indicated that for now it
will simply be a supplement to the regular N&V but that they
hoped that enough interest would be generated to make it a
separate publication.

Mike

> Robin specifically said she wanted to include all SRS
> members; she knows how big our mailing list is.
>
> Doug
>
> FThompson9@... writes:
> >Does it count if I don't live in Seattle?
> >
> >Pherd
> >Somers, CT
> >
> >
> >In a message dated 06/27/2002 1:29:10 PM Eastern Daylight
> Time,  >doug.kelley@... writes:
> >
> >
> >> Members, you can take advantage of Robin's generous
> offer by placing your  >> name
> >> and address in the following table:
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
> To unsubscribe from this group, send an email to:
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
> Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/
>

According to Nuts and Volts web page the special offer will be included with
those of us who already are subscribers.  So is there any need to have Nuts and
Volts subscribers put out name in this database?

Doug Leppard
Doug.Leppard@...


-----Original Message-----
From: Doug Kelley [mailto:doug.kelley@...]
Sent: Monday, July 01, 2002 10:23 AM
To: FThompson9@...; seattlerobotics@yahoogroups.com
Subject: Re: [SeattleRobotics] Re: RE: Special Mini-Magazine
DevotedExclusively to th.


Robin specifically said she wanted to include all SRS members; she knows how
big our mailing list is.

Doug

FThompson9@... writes:
>Does it count if I don't live in Seattle?
>
>Pherd
>Somers, CT
>
>
>In a message dated 06/27/2002 1:29:10 PM Eastern Daylight Time,
>doug.kelley@... writes:
>
>
>> Members, you can take advantage of Robin's generous offer by placing your
>> name
>> and address in the following table:


Visit the SRS Website at http://www.seattlerobotics.org

To unsubscribe from this group, send an email to:
SeattleRobotics-unsubscribe@yahoogroups.com


Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

Only people who are not subscribers who would not be getting the August
issue should put their name in the database.

(Does that make sense?)

Normal subscribers to Nuts & Volts will receive the supplement with their
August issue.

(I think the more I try to explain, the worse it gets ... must be Monday!)
>:)
>
Robin



>> -----Original Message-----
>> From: Doug Kelley [mailto:doug.kelley@...]
>> Sent: Monday, July 01, 2002 9:00 AM
>> To: robin@...
>> Subject: Fwd: RE: [SeattleRobotics] Re: RE: Special Mini-Magazine
>> DevotedExclusively to th
>>
>>
>> Hi Robin,
>>
>> Could you answer this for me?
>>
>> Doug
>>
>> According to Nuts and Volts web page the special offer will
>> be included with
>> those of us who already are subscribers.  So is there any
>> need to have Nuts and
>> Volts subscribers put out name in this database?
>>
>> Doug Leppard
>> Doug.Leppard@...
>>
>

You will need 3 10k resistors and 2 transistors.  The first is a
2N3906 PNP type (or similar).  The second an NPN type that can
drive the solenoid.

Tie the emitter of the PNP to the +5V power supply.  The base is
connected through 10k (4.7k if needed) resistor to the 68HC11
output pin.

Tie the emitter of the NPN to the -12V supply.  Connect the
collector to one side of the solenoid.  Connect the solenoid's
other wire to the +12V supply.

Connect 2 10k resistors the the base of the NPN transistor.  The
first connects to the PNP transistor's collector while the second
is connected to the -12V supply.

This circuit will work in reverse.  When the output pin is a 1
the solenoid will be off.  When the pin is a 0 the solenoid will
be on.

You can use an N-channel MOSFET for the solenoid driver.  Connect
the source to -12V, the drain to the solenoid, and the gate to
the resistor junction.  Replace 10k resistor from the PNP
collector with a 2.2k to bring the voltage at the gate to a
higher level.

I hope you can make sense of the circuit description I can't send
a picture from here.


  > dickenssrs wrote:
> OK you power-circuit folks, please help a programmer out.
>
> I want to drive a bunch of 24V DC solenoids from a 5V
> 68HC11 microcontroller. The power supply I'm using has
> 5V and +12/-12V. I need a simple (cheap) driver circuit
> to have the 5V HC-device energize a 24V solenoid. The system
> I'm looking at building will have dozens of solenoids, each
> driven from a 5V HC device, so I need a simple and cheap
> circuit so I can build dozens of them.
>
> I was thinking a MOSFET or transistor as the 24V "switch",
> in the proper configuration and current-limiting control,
> but I'm real rusty with these circuits. I could use a 5V-coil
> relay, but I'm sure there are much better methods to use.
> Also, I could go with a 24V supply if needed, but I'm currently
> using a +12/-12 supply and would like to find a good circuit
> using it.
>
> Thanks for any suggestions or circuits you might want to share.
> I'll post a summary of what solutions I find for this.
>
> tom@...
>
>
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
> To unsubscribe from this group, send an email to:
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>

Fellow Roboteers,
Southern Oregon Robotics club is attempting to meet the requests of
hundreds of Amateur Robotisists across america, less than one year ago
we asked the question "What would it take to get you to travel from
far and near to attend a robotic competition?" You responded and we
began to build RoboMaxx around your ideas. The event is less than 4
months away, that's just enough time to get your latest prject done
and claim some of the cash prizes that will be awarded for Open Class,
Mini Sumo, Maze solving, Lego Mindstorm Creations, and a brand new
competition we call Mini Bbots.
Watch this link for further details --> www.1sorc.com/robomaxx the
site is still under construction, but it's taking form fast.
I hope to see you all at RoboMaxx!

Send questions to info@...
Respectfully,
Current President of
Southern Oregon Robotics Club
Dan Gates

Hi All,

Sorry about the delay getting this stuff posted.  Please access it in
the "Files" area of Yahoo! Groups.  There are a couple image files
attached that won't be there if I post this directly to the
maillist.  Also, I'd rather you viewed the formatted HTML instead of
the plain text.  It's much prettier.  As always, if any obvious
errors are spotted, let me know and I'll fix the archived version.

Cheers,

Steve Kaehler, SRS Secretary

http://groups.yahoo.com/group/SeattleRobotics/files/Monthly%20Meeting%
20Minutes/06-15-02_SRS_Meeting.html

We would like to remind you of this upcoming event.

Monday Night Chat

Date: Monday, July 1, 2002
Time: 7:00PM - 9:00PM PDT (GMT-07:00)

Monday Night Chat happens every Monday at the SRS. Come ask
questions or offer advice live.

Directions:

The chat requires that you use an Internet Relay Chat (IRC)
program. Details on finding one are below.

For you experts, the chat is held on us.chatjunkies.org in the
#SeattleRobotics group.

For Windows users:

See the SRS website at

http://www.seattlerobotics.org/contact.html#Monday

I am trying to help my son with an amplifier assignment in class.  I hated amps
in school.  Can you point us to some web resources that would help build the
following amp using Bipolar Junction Transistors:

Small signal voltage gain of 50 or higher
Input resistance of 40K ohms or higher
Output resistance of less than 100 ohms
Undistorted maximum output swing of 2 V peak to peak
Power supply voltage of 12 V
Lowest possible power dissipation
Highest possible bandwidth
All specifications of 20% tolerance

Thanks for the input.

Doug Leppard
Doug.Leppard@...

Can any one out there help me?  I have a Compaq Pentium running
Windows 98 2nd Ed.  I had to reformat the hard drive to correct a
malfunction.  Now after reinstalling the OS the CD-ROM and CD-RW
drives won't work right.

You can read the directory but can't access any of the files on
either drive.  I can't get the CD software to acknowledge the the
second drive is a CD-RW.

How does one configure Windows 98 to use the CD-ROMs?  I've tried
Microsoft's web site and downloaded several tech sheets to no
avail.  I can't figure out the difference between real and
protected mode drivers etc.

Real mode drivers are the type to use when the desktop is not showing, when
you have only a command prompt. In C:\AUTOEXEC.BAT the Windows installer
should have REM'd out the "MSCDEX" line and it's matching device driver in
C:\CONFIG.SYS (those are real-mode, the Windows desktop runs in Protect
mode).

Try unchecking the boxes "Sync negociation" and "DMA" in the device manager
properties page of the two CD drives.
	 -Bryan R. Minugh

-----Original Message-----
From: Randy Carter [mailto:rwcarter@...]
Sent: Monday, July 01, 2002 8:34 PM
To: SRS
Subject: [SeattleRobotics] OT: Win98 and CD-ROM drives.


Can any one out there help me?  I have a Compaq Pentium running
Windows 98 2nd Ed.  I had to reformat the hard drive to correct a
malfunction.  Now after reinstalling the OS the CD-ROM and CD-RW
drives won't work right.

You can read the directory but can't access any of the files on
either drive.  I can't get the CD software to acknowledge the the
second drive is a CD-RW.

How does one configure Windows 98 to use the CD-ROMs?  I've tried
Microsoft's web site and downloaded several tech sheets to no
avail.  I can't figure out the difference between real and
protected mode drivers etc.

Hi all,
I am building a Ni-Cd battery charger and it uses a 5600uf single ended
electrolitic cap in it. I purchased such a cap today (a 5600uf 40v cap) and
I notice it has a third leg. The third leg is shorter than the +ve and -ve
legs. Am I correct in guessing this is some sort of ground and I can safely
snip it off and only use the two longer leads?

Also, the battery charger requires a 24v 1.5A power supply. I have a Sony
switching power supply that I recovered from an old printer. It has a number
of different voltage outputs two of which are 24v outputs. How can I tell if
this power supply is capable of supplying the needed 1.5Amps? I have googled
and also searched the sony site, but no luck...

Thanks for any and all help,
Regards,
James Fitzsimons

--- In SeattleRobotics@y..., "Doug Leppard" <DLeppard@c...> wrote:
> I am trying to help my son with an amplifier assignment in class.
I hated amps in school.  Can you point us to some web resources that
would help build the following amp using Bipolar Junction Transistors:
...........


Doug, there are a number of sites on the web with electronics
tutorials and schematics for various circuits. My favorite is:

http://www.epanorama.net/

I also have a bunch of links to others here:

http://www.oricomtech.com/misc/teklink1.htm#Design2

hope this helps,
- dan michaels
=========================

A couple of good resources are Electronics101 and Electronics_101
on Yahoogroups. Check out the files and bookmarks pages there.

http://groups.yahoo.com/group/Electronics101
http://groups.yahoo.com/group/Electronics_101

Steve Greenfield

--- Doug Leppard <DLeppard@...> wrote:
> I am trying to help my son with an amplifier assignment in class.
>  I hated amps in school.  Can you point us to some web resources
> that would help build the following amp using Bipolar Junction
> Transistors:
>
> Small signal voltage gain of 50 or higher
> Input resistance of 40K ohms or higher
> Output resistance of less than 100 ohms
> Undistorted maximum output swing of 2 V peak to peak
> Power supply voltage of 12 V
> Lowest possible power dissipation
> Highest possible bandwidth
> All specifications of 20% tolerance
>
> Thanks for the input.
>
> Doug Leppard
> Doug.Leppard@...


__________________________________________________
Do You Yahoo!?
Sign up for SBC Yahoo! Dial - First Month Free
http://sbc.yahoo.com

A couple of days ago there was a discusion about how to use
accelerameters to help a 2 wheeled robot keep its balance.
The hard part seemed to be getting rid of the acceleration
caused by the robot's motion so that you could have a good
reference to a true vertical.

Just a moment ago a small LED (I'm not confident enough for
it to be a lightbulb ;)  ) lit up over my head!

What if you mount 2 accelerameters equidistant from the
center of rotation but on opposite sides of the line of
rotation (like counter weights) such that the axises being
measured were parallel and pointing horizontally directly
ahead when the robot is in a vertical position (or pointing
in a vertical position, either way should work). Any
acceleration induced by the robot's forward motion would be
equal and in the same direction for both sensors but any
rotational acceleration would be equal but in opposite
directions. It should then be rather simple to subtract one
component with the result being either a clean measurement
of the robot's forward motion or rotation. If you have no
rotation and your wheels are not turning then you are in a
vertical, stationary position. If there is no rotation and
your wheels are turning then you are in a stable, constant
motion forward or aft. If thereis rotation use the wheel
encoders to determine if you are at the speed you desire and
adjust the rotation accordingly to maintain balance.

When the bot is first starting up it should strive for zero
rotation (no falling) and zero forward motion and use that
as a vertical reference. Everytime the bot is instructed to
stop it should recalibrate this by achieving no rotation and
no wheel rotation.

If the sensors are 3 axis sensors and are mounted on a rod
passing through the center of rotation (wheel axial's line)
and horizontal (that puts one sensor in the front and one in
the back) with the sensing axises measuring pitch, yaw and
forward motion, you now have a system that gives you
balance, forward acceleration, and turning rate. the key is
to have the two sensors on opposite side of the axial.

OK, now that I've voiced ... err... scribed the idea, what
holes can we punch in it (just be nice!)?

mike

I am helping to start a robotics class at our local high school.  I am working
with two sharp teachers, a husband wife team.  The wife is being inducted into
the Teachers Hall of Fame this summer.

We have bought several MarkIII mini-sumo kits for the class.  We will build
several of the kits for the kids to program for line following and mini-sumo
contests.

Several other of the kits would be built into experimental stations.  They would
be mounted on a board with experiments programmed into the MCU.  The various
pieces of the robot would be mounted on the board for experiments.

Some of the experiments would be:

Voltage testing on:
line sensor
distance sensor
voltage regulator
motor input (PWM on scope)

test MCU reading from:
line sensor
distance sensor
PWM

I figured with an experimental station like that they could learn about the
indivual parts before trying to program the whole thing together.

Do you have any input?  Any ideas for the experimental stations?


Doug Leppard
Doug.Leppard@...



[Non-text portions of this message have been removed]

How about testing how hard the robots can push? Based on battery voltage,
say? Give them an idea how the motor's push power is connected to battery
power. Maybe try different servos, too, and different wheel sizes to show
how friction relates too. Different weights of robots affecting push power,
so on and so on.

Mark Medonis


>
> I am helping to start a robotics class at our local high school.  I am working
> with two sharp teachers, a husband wife team.  The wife is being inducted into
> the Teachers Hall of Fame this summer.
>
> We have bought several MarkIII mini-sumo kits for the class.  We will build
> several of the kits for the kids to program for line following and mini-sumo
> contests.
>
> Several other of the kits would be built into experimental stations.  They
> would be mounted on a board with experiments programmed into the MCU.  The
> various pieces of the robot would be mounted on the board for experiments.
>
> Some of the experiments would be:
>
> Voltage testing on:
> line sensor
> distance sensor
> voltage regulator
> motor input (PWM on scope)
>
> test MCU reading from:
> line sensor
> distance sensor
> PWM
>
> I figured with an experimental station like that they could learn about the
> indivual parts before trying to program the whole thing together.
>
> Do you have any input?  Any ideas for the experimental stations?
>
>
> Doug Leppard
> Doug.Leppard@...

--- In SeattleRobotics@y..., "Doug Leppard" <DLeppard@c...> wrote:
> I am helping to start a robotics class at our local high school.  I
am working with two sharp teachers, a husband wife team.  The wife is
being inducted into the Teachers Hall of Fame this summer.
...............
> Do you have any input?  Any ideas for the experimental stations?
>


Doug, I don't know if you are a member, but just down the road at
PARTS is a guy named Edwin Pilobello who has been teaching robotics
classes to high-school teenagers for several years. You might track
him down and get some first-hand experience info.

http://groups.yahoo.com/group/PARTS/


- dan michaels
www.oricomtech.com
=========================

Along those lines, there are several experiments in "Mobile Robots" that are
either fleshed out or strongly hinted at that might be appropriate.  Take a
cruise through it and see if anything blows your knickers up.

Dan

-----Original Message-----
From: Mark Medonis [mailto:mark@...]
Sent: Tuesday, July 02, 2002 8:27 PM
To: MiniSumoMarkIII@yahoogroups.com; SeattleRobotics@yahoogroups.com
Subject: Re: [SeattleRobotics] High School robotics class


How about testing how hard the robots can push? Based on battery voltage,
say? Give them an idea how the motor's push power is connected to battery
power. Maybe try different servos, too, and different wheel sizes to show
how friction relates too. Different weights of robots affecting push power,
so on and so on.

Mark Medonis


>
> I am helping to start a robotics class at our local high school.  I am
working
> with two sharp teachers, a husband wife team.  The wife is being inducted
into
> the Teachers Hall of Fame this summer.
>
> We have bought several MarkIII mini-sumo kits for the class.  We will
build
> several of the kits for the kids to program for line following and
mini-sumo
> contests.
>
> Several other of the kits would be built into experimental stations.  They
> would be mounted on a board with experiments programmed into the MCU.  The
> various pieces of the robot would be mounted on the board for experiments.
>
> Some of the experiments would be:
>
> Voltage testing on:
> line sensor
> distance sensor
> voltage regulator
> motor input (PWM on scope)
>
> test MCU reading from:
> line sensor
> distance sensor
> PWM
>
> I figured with an experimental station like that they could learn about
the
> indivual parts before trying to program the whole thing together.
>
> Do you have any input?  Any ideas for the experimental stations?
>
>
> Doug Leppard
> Doug.Leppard@...


Visit the SRS Website at http://www.seattlerobotics.org

To unsubscribe from this group, send an email to:
SeattleRobotics-unsubscribe@yahoogroups.com


Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

The problem is that if the OAKCDROM.SYS isn't in the CONFIG.SYS
and the MSCDEX.EXE isn't in the AUTOEXEC.BAT both of the drives
disappear.  The most annoying thing is I CAN read the directory
but none of the files.

How do I get the other kind of drivers to load?  What kind of
settings should I see in the SYSTEM.INI, CONFIG.SYS,
AUTOEXEC.BAT, etc.?

I have the CD-ROM for the CD-RW but it reports that none of my
drives are the recordable type.  Currently the CD-RW removed
until I can get first 40x drive to work right.  I tried to
download the drivers from the Memorex web site still no go.

I know all of my problems are Caused by bad settings in Windows
98 but, I have no idea what is causing them.  Every thing I tried
on the Microsoft web site hasn't worked.  One curious thing is
one of the reports states that I could have an old MSCDEX.EXE
file that is incompatible with Windows 98 2nd edition.  When
replaced it with a virgin copy from the same CD-ROM that the OS
came on, it reports a copyright date of 1995.



  > Bryan Minugh wrote:
> Real mode drivers are the type to use when the desktop is not showing, when
> you have only a command prompt. In C:\AUTOEXEC.BAT the Windows installer
> should have REM'd out the "MSCDEX" line and it's matching device driver in
> C:\CONFIG.SYS (those are real-mode, the Windows desktop runs in Protect
> mode).
>
> Try unchecking the boxes "Sync negociation" and "DMA" in the device manager
> properties page of the two CD drives.
>  -Bryan R. Minugh
>
> -----Original Message-----
> From: Randy Carter [mailto:rwcarter@...]
> Sent: Monday, July 01, 2002 8:34 PM
> To: SRS
> Subject: [SeattleRobotics] OT: Win98 and CD-ROM drives.
>
>
> Can any one out there help me?  I have a Compaq Pentium running
> Windows 98 2nd Ed.  I had to reformat the hard drive to correct a
> malfunction.  Now after reinstalling the OS the CD-ROM and CD-RW
> drives won't work right.
>
> You can read the directory but can't access any of the files on
> either drive.  I can't get the CD software to acknowledge the the
> second drive is a CD-RW.
>
> How does one configure Windows 98 to use the CD-ROMs?  I've tried
> Microsoft's web site and downloaded several tech sheets to no
> avail.  I can't figure out the difference between real and
> protected mode drivers etc.
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
> To unsubscribe from this group, send an email to:
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>

Howdy

Mike Jones wrote:

> A couple of days ago there was a discusion about how to use
> accelerameters to help a 2 wheeled robot keep its balance.
> The hard part seemed to be getting rid of the acceleration
> caused by the robot's motion so that you could have a good
> reference to a true vertical.
>
> Just a moment ago a small LED (I'm not confident enough for
> it to be a lightbulb ;) ) lit up over my head!
>
> What if you mount 2 accelerameters equidistant from the
> center of rotation but on opposite sides of the line of
> rotation (like counter weights) such that the axises being
> measured were parallel and pointing horizontally directly
> ahead when the robot is in a vertical position (or pointing
> in a vertical position, either way should work).

I'm not sure I understand.  The idea is to use a pair of
accelerometers to cancel each other's platform accelerations
and leave the tilt angle as a residual, yes?


            robot side view

              ___________
              |         |
              |         |
              |         |
              |    _    |
ADXL202 A -> O_ /   \_ O <- ADXL202 B
                 \ _ /

                 wheel

Like this?


The accelerometers, when stationary, respond to the "acceleration"
of gravity between 0, edge-on to gravity, and 1, face-on to gravity.
For the robot to balance we need a lot of precision around 0, maybe
plus and minus 3 degrees.  This small signal is easily swamped by the
accelerations of the platform once it begins moving.

Since we know the accelerations of the platform from the wheel encoders,
we could in principle use those readings to correct the accelerometers
by subtracting out the difference.  In practice we have not be able
to make that work.

> acceleration induced by the robot's forward motion would be
> equal and in the same direction for both sensors but any
> rotational acceleration would be equal but in opposite
> directions.

I don't think that's right.  Both rotations will be in the same
direction, even though one is moving up and one is moving down,
they are both rotating from edge-on to face-on in the same direction,
clockwise or counter-clockwise.

> It should then be rather simple to subtract one
> component with the result being either a clean measurement
> of the robot's forward motion or rotation. If you have no
> rotation and your wheels are not turning then you are in a
> vertical, stationary position. If there is no rotation and
> your wheels are turning then you are in a stable, constant
> motion forward or aft. If thereis rotation use the wheel
> encoders to determine if you are at the speed you desire and
> adjust the rotation accordingly to maintain balance.
>
> When the bot is first starting up it should strive for zero
> rotation (no falling) and zero forward motion and use that
> as a vertical reference. Everytime the bot is instructed to
> stop it should recalibrate this by achieving no rotation and
> no wheel rotation.

When the robot is stopped, the tilt angle can be read directly
from the accelerometers.

>
> If the sensors are 3 axis sensors and are mounted on a rod
> passing through the center of rotation (wheel axial's line)
> and horizontal (that puts one sensor in the front and one in
> the back) with the sensing axises measuring pitch, yaw and
> forward motion, you now have a system that gives you
> balance, forward acceleration, and turning rate. the key is
> to have the two sensors on opposite side of the axial.
>
> OK, now that I've voiced ... err.. scribed the idea, what
> holes can we punch in it (just be nice!)?
>

The industrial inertial measurement units we have looked at
seem in general to combine heavily low-pass filtered accelerometers
for a long term gravity reference, combined with the integrated
output of a rate gyro (hi-pass filtered to remove bias drift) for
rapid hi-frequency response.  The IMU's based on accelerometers
alone seem only accurate when static.  The sourceforge "autopilot"
project is currently developing a (hopefully) lowcost 3-axis IMU
for their autonomous helicopter:

	 http://autopilot.sourceforge.net/

But the ADXL's are so cheap, especially compared to industrial rate
gyros, that an all-accelerometer solution would be pretty cool.
Analog Devices will sample you a couple of ADXL202s for experimentation,
it you are so inclined.  You might bread-board it together and see what
transpires!  Remember that the platform experiences other accelerations
besides lateral and rotational including bumps, irregularities in the floor,
edges of carpet, etc, indoors, and very irregular and tilted surfaces outdoors.

As an alternate, a two-wheel balancing robot can be constructed rather
simply if you accept the limitation that it can only run on a (more or less)
level floor or table.  For most of the robots in the SRS and DPRG, that
is already a given.  With that caveat, the tilt sensor can simply be
a potentiometer attached to the bottom of the robot with a "feeler"
dragging on the floor.  Here's a picture:

	 http://www.geology.smu.edu/~dpa-www/robo/nbot/02-00000.jpg

The pot is connected as a voltage divider feeding an A/D channel.  In some
ways this is the most robust angle sensor, low noise, no resonances, nothing
to adjust, and allows the robot builder to focus on the balance algorithm
rather than oddities in the IMU output.

Of course it's not nearly as sexy as a non-contacting IMU...
but much easier to make work, and a good starting place for either.

best regards,
dpa

--------------------------------------------------
SR04 and The two-wheel Balancing Bot IMU:
http://www.geology.smu.edu/~dpa-www/myrobots.html
--------------------------------------------------

Hi

thought I might try the ascii art again, and use
periods to keep yahoo from removing the spacing:

...........robot side view............
......................................
.............___________..............
.............|         |..............
.............|         |..............
.............|         |..............
.............|    _    |..............
ADXL202 A -> O_ /   \_ O <- ADXL202 B
................\ _ /.................
......................................
................wheel.................

Mike, is this the idea?

regards,
dpa

ok I give up.

dpa

> Howdy
>
> Mike Jones wrote:
>
> > A couple of days ago there was a discusion about how to
> > use accelerameters to help a 2 wheeled robot keep its
> > balance. The hard part seemed to be getting rid of the
> > acceleration caused by the robot's motion so that you
> > could have a good reference to a true vertical.
> >
> > Just a moment ago a small LED (I'm not confident enough
> > for it to be a lightbulb ;) ) lit up over my head!
> >
> > What if you mount 2 accelerameters equidistant from the
> > center of rotation but on opposite sides of the line of
> > rotation (like counter weights) such that the axises
> > being measured were parallel and pointing horizontally
> > directly ahead when the robot is in a vertical position
> > (or pointing in a vertical position, either way should
> work).
>
> I'm not sure I understand.  The idea is to use a pair of
> accelerometers to cancel each other's platform
> accelerations and leave the tilt angle as a residual, yes?
>
>
>            robot side view
>
>              ___________
>              |         |
>              |         |
>              |         |
>              |    _    |
> ADXL202 A -> O_ /   \_ O <- ADXL202 B
>                 \ _ /
>
>                 wheel
>
> Like this?
>
>
> The accelerometers, when stationary, respond to the
> "acceleration" of gravity between 0, edge-on to gravity,
> and 1, face-on to gravity. For the robot to balance we
> need a lot of precision around 0, maybe plus and minus 3
> degrees.  This small signal is easily swamped by the
> accelerations of the platform once it begins moving.

Or rotate it 90 degrees so that the more important extreme
is around 1 rather than 0, which ever way works best.

> Since we know the accelerations of the platform from the
> wheel encoders, we could in principle use those readings
> to correct the accelerometers by subtracting out the
> difference.  In practice we have not be able to make that
> work.

The wheel encoders, at least for the balancing act, are most
useful to determine if the platform is in motion and a tilt
is normal or the bot is stationary and there should be no
tilt.

Actually, I was saying that the accelerameters are used to
measure both the platform's motion and the rotation around
the axial. The acceleration due to platform motion would be
the same for both sensors. If there is rotation then one
sensor will have a value added to it because it is moving
forward slightly faster than the platform and the other
sensor will have the same value subtracted from that of the
platform. To visualize it better lets say the platform is
accelerating at 100 units and the platform rotation is
contributing 2 units in the forward direction. This would
give a sum of 102 for the rear sensor and 98 for the forward
sensor. Now subtract the one from the other and the
difference is 4 units in the forward direction, divided by 2
(because 2 sensors), gives acceleration of 2 in the forward
direction and (102+98)/2=100 for the platform component. You
can double check (or increase reliability) by doing the same
thing for the vertical plane since the gravitational
acceleration is constant and the delta would be simply
caused by rotation. They key is that since each sensor is on
opposite sides of the axial they will respond destructively
for rotation and constructively for platform forward or
reverse motion.

Note that for sensitivity reasons, having a vertical
orientation of the line between the sensors may be more
desirable and the bigger the distance the more precise the
sensors will be, but this would be limited to a bit less
than the diameter of the wheel so experimentation would be
in order.

> > acceleration induced by the robot's forward motion would
> > be equal and in the same direction for both sensors but
> > any rotational acceleration would be equal but in
> > opposite directions.
>
> I don't think that's right.  Both rotations will be in the
> same direction, even though one is moving up and one is
> moving down, they are both rotating from edge-on to
> face-on in the same direction, clockwise or
> counter-clockwise.

But acceleration is a vector so direction is important.
Depending on how each sensor is mounted relative to the
other sensor you may have to insert a minus sign in front of
the value. I have not used accelerameters yet but will be
getting a few to play with very soon, but don't they give
both a positive and negative reading in the same line of
action, depending on the actual direction the acceleration
is in? Or do they simply give a scalar value?


> > It should then be rather simple to subtract one
> > component with the result being either a clean
> > measurement of the robot's forward motion or rotation.
> > If you have no rotation and your wheels are not turning
> > then you are in a vertical, stationary position. If
> > there is no rotation and your wheels are turning then
> > you are in a stable, constant motion forward or aft. If
> > thereis rotation use the wheel encoders to determine if
> > you are at the speed you desire and adjust the rotation
> > accordingly to maintain balance.
> > When the bot is first starting up it should strive for
> > zero rotation (no falling) and zero forward motion and
> > use that as a vertical reference. Everytime the bot is
> > instructed to stop it should recalibrate this by
> > achieving no rotation and no wheel rotation.
>
> When the robot is stopped, the tilt angle can be read
> directly from the accelerometers.
>
> >
> > If the sensors are 3 axis sensors and are mounted on a
> > rod passing through the center of rotation (wheel
> > axial's line) and horizontal (that puts one sensor in
> > the front and one in the back) with the sensing axises
> > measuring pitch, yaw and forward motion, you now have a
> > system that gives you balance, forward acceleration, and
> > turning rate. the key is to have the two sensors on
> > opposite side of the axial.
> > OK, now that I've voiced ... err.. scribed the idea,
> > what holes can we punch in it (just be nice!)?
> >
>
> The industrial inertial measurement units we have looked
> at seem in general to combine heavily low-pass filtered
> accelerometers for a long term gravity reference, combined
> with the integrated output of a rate gyro (hi-pass
> filtered to remove bias drift) for rapid hi-frequency
> response.  The IMU's based on accelerometers alone seem
> only accurate when static.  The sourceforge "autopilot"
> project is currently developing a (hopefully) lowcost
> 3-axis IMU for their autonomous helicopter:
>
>     http://autopilot.sourceforge.net/
>
> But the ADXL's are so cheap, especially compared to
> industrial rate gyros, that an all-accelerometer solution
> would be pretty cool.   Analog Devices will sample you a
> couple of ADXL202s for experimentation, it you are so
> inclined.  You might bread-board it together and see what
> transpires!  Remember that the platform experiences other
> accelerations besides lateral and rotational including
> bumps, irregularities in the floor,  edges of carpet, etc,
> indoors, and very irregular and tilted surfaces outdoors.
> As an alternate, a two-wheel balancing robot can be
> constructed rather simply if you accept the limitation
> that it can only run on a (more or less) level floor or
> table.  For most of the robots in the SRS and DPRG, that
> is already a given.  With that caveat, the tilt sensor can
> simply be a potentiometer attached to the bottom of the
> robot with a "feeler"  dragging on the floor.  Here's a
> picture:
>
> http://www.geology.smu.edu/~dpa-www/robo/nbot/02-00000.jpg
> The pot is connected as a voltage divider feeding an A/D
> channel.  In some ways this is the most robust angle
> sensor, low noise, no resonances, nothing to adjust, and
> allows the robot builder to focus on the balance algorithm
> rather than oddities in the IMU output.
> Of course it's not nearly as sexy as a non-contacting
> IMU... but much easier to make work, and a good starting
> place for either.
> best regards,
> dpa
>
> --------------------------------------------------
> SR04 and The two-wheel Balancing Bot IMU:
> http://www.geology.smu.edu/~dpa-www/myrobots.html
> --------------------------------------------------
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
> To unsubscribe from this group, send an email to:
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
> Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/
>

Hi Mike

If this works it will be very useful,
keep us informed of your progress.

best regards,
dpa

I came across an interesting chip. It is a "rotational accelerometer" made
by ST called the LIS1R02.

Does anyone know how I can buy or get any samples from ST? They seem to
point to distributors on their website but I searched for this chip name at
several listed and no one seems to have it.

Mike

I have had luck calling Mouser and asking them to get me samples from ST on
items they do not carry.
http://www.mouser.com  That is how I got the ST TD340.

Kip

----- Original Message -----
From: "Mike Jones" <michaelj@...>
To: <seattlerobotics@yahoogroups.com>
Sent: Thursday, July 04, 2002 6:48 PM
Subject: [SeattleRobotics] Re: 2-wheel balancing


> I came across an interesting chip. It is a "rotational accelerometer" made
> by ST called the LIS1R02.
>
> Does anyone know how I can buy or get any samples from ST? They seem to
> point to distributors on their website but I searched for this chip name
at
> several listed and no one seems to have it.
>
> Mike
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
> To unsubscribe from this group, send an email to:
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>

I have most of the 68hc11/12 GNU C compiler installed and running.  I still
have a couple of problems that I need some help on.

Can someone who has gone through this pass me a crt0.s file (C startup
program) that creates the needed soft registers (_.frame, _.xy, etc.)?
Hopefully, this same file includes the variable initialization code, so I
don't have to bother writing that, either.  :-)

Attempts to link library object modules, such as memcpy(), currently fail.
Partly, this is because I don't yet have the above soft registers correctly
defined.  But it is also because the linker appears to be trying to load
the entire object library into the final executable.  Perhaps this problem
will sort itself out when I get the correct crt0.s file.  But if this is
really an unrelated problem, better to get feedback now.

					 Karl

Karl Lunt
email: karll@...
web:  www.seanet.com/~karllunt

Has the Encoder been phased out of existence?
    Kevin.