Thanks Jonny!
The design is working out really well. However I did want a shorter 'bot, but because of the necessity of mounting the floor-sensors out front quite a bit (so the 'bot wouldn’t fall off of the table), it made it a bit long. I realized I had enough room to slide in another R/C servo for the jaws, so that’s what I’m working on now.
Alan KM6VV
Congrats Alan! 
Very well done and made!
Thanks Kåre!
That means a lot!
Alan KM6VV
yes iv seen your design for a gripper. nice design. would be very cool to see that on the AS13b! what are your plans regarding the material to use.?
I’m currently making the gripper out of aluminum sheet metal. I did temporarily use two pieces of PCB stock for links to the R/C servo horn, but I’ve received my bell-crank links now, and will replace my test material.
I’ve been experimenting with using latex (?) tubing over the jaws, but it gets in the way of the microswitches I want to use.
The LPA gears as seen in the CAD drawing above would allow me parallel motion of the gripper jaws, but really not that much range.
Idea is that acquiring the object (target) closes one or more (low pressure) microswitches on the back of the gripper, then the jaws close on the object until the (higher pressure) microswitches on the jaws close). At that point, the gripper R/C servo stops closing and just holds. At least, that’s the theory. The “Close Jaws” behavior is initiated when the object is detected in the gripper, and when finished, the behavior shifts to the “Search for goal” behavior. Should the object be lost, then before the behavior returns to “find object”, an “open jaws” behavior is executed. All the while, a “Don’t fall off the table” behavior is running (actually has the highest priority). Should the edge of the table be encountered, the behavior “Back up and reverse course” takes over until the new direction is achieved. Then we’re back to a base line behavior of moving forward for a while, and then doing a first left, then right sweep in the forward direction. A 180 spin could also be way to first find the object (target). The highest behavior, “find the goal” is currently the same as the “find target” behavior, we just use sensor data from the higher-mounted Sharp IR rangers.
All of this is accomplished with FSMs, (Finite State Machines) or AFSMs (Augmented FSM). Sensors are checked about 10 times a second via a task flag set by the system timer implemented with an interrupt. Sensors trigger behavior shifts, and the tasks are all controlled by flags. there is no “blocking” or waiting for sensors or any other status.
With the addition of the new R/C servo for the gripper, and the extra “jaws closed” microswitches, I’ve used all of the I/O and analog pins on the Arduino UNO!
With the form factor being the same, the new Basic Micro “Mad Hatter” board should be a “drop in”. Only tricky part might be getting a system timer interrupt set up under Basic.
Alan KM6VV
Alan KM6VV
Yes the copper one. Nice arrangement of sensors.
Some more jaw pix for the ASB-13 'Bot:
http://www.MarconettEngineering.com/Jaws1_bottomViewB.jpg
Bottom View, jaws with servo
http://www.MarconettEngineering.com/ServoBracket_bottomViewB.jpg
new servo bracket
http://www.MarconettEngineering.com/Jaws3_bottomViewB.jpg
Bottom View, jaws with servo
Got the jaws working with BLS-01 ball links and 2-56 threaded rod (TROD-01).
Slots and holes in jaws are for microswitches, can be seen on earlier pix.
http://www.MarconettEngineering.com/ASB-13_bot_jaws_side1.jpg
Side view showing jaws servo mounted. Don’t mind the make-do PCB links, they’ve been replaced by ball links and threaded rod.
Code for jaws written and working!
I’m still thinking of using the LPA gears and a variation of the servo bracket above to implement the “parallel jaw” version (CAD drawing in previous post).
Alan KM6VV
P.S. Just pulled out one of the ASB-13 brushed brackets I’d received last week or so, and ALL the holes are there now! Thanks, I needed that!
A little Status,
From another thread:
*We’ve kicked-off the MicroMouse challenge for our HBRobotics club, tomorrow I’ll drag in my sections of maze, and try to get my hastily-built ASB13-MM 'bot (take-off of my ASB13 “MicroMoose”) bot to do simple wall-following in the maze. As of last night it wasn’t running too well. I plugged in one of my infamous “blue” jumpers from a servo pin to +5 V, instead of a servo data pin (Arduino UNO). Probably why the program “crashed” the first two times I threw the servo power switch to on. I probably fried the pin. More later, I’ll probably take it over to the ASB13 'bot thread.
*
It seemed like it would be a quick project.
RUN MICROMOUSE AS Wall Follower in MicroMouse Maze
HSR-1425CR (57 oz.in.) Continuous Rotation Standard Servo 8uS deadband 400uS ± control band?
ARW-01 Aluminum Wheels on HMSH-02 Hitec hubs
ASB13 chassis w/plastic disk
Front Castor
Double IR sensor Bracket
MicroMouse Sensor bar
Sharp GP2D120 4-30cm angled left/right side looking, and forward looking IR sensors
5-Cell Battery 7.2V
Arduino UNO board (could use new BotBoarduino Jr. board)
Arduino rev 2 prototype board
SLD-01 Single Line Sensor
Schmart Board jumpers and .025" wrap-posts
Code is that of the MicroMoose, with a new behavior added: Wall following (Left, Right, Both).
Pictures when I get a chance. It looks like a very early MicroMoose, with three Sharp IR sensors on the front.
Alan KM6VV
Update.
I need to add odometry for both table-top or Micromouse projects.
Trying to get SLD-01 Single Line Sensors to work with 36 stripe (18 black, 18 aluminum) wheel rim encoders.
Black stripes created by marking off wheels,then painting every other one black with enamel model paint.
Current distance from sensor to wheel is less then 1/16", wouldn’t work at all at 1/8".
New metal Hitec servo horns on HSR-1425CR continuous rotation servos. Probably .040-.050 runout (hub does not fit tight on servo shaft).
Observed: wheel run-out is often enough to cause a black (LED on) dropout. Can’t adjust either pot or distance to get a useable square wave (dropouts) out of sensor.
Wanted to use digital input, but might have to go to analog and a simple QRD1134 instead.
Program operation:
outputs of two SLD-01 line sensors are fed into digital input pins of uP. Pins 2 & 3 are set up to interrupt on change (both edges). Interrupt code clears count variable for encoder on each call after copying the accumulated count. Main timer routine interrupts at 1mS rate, incrementing the counter. 10 time a second, the accumulated count is displayed. Should be a accurate display of the speeds of the two wheels. That’s not happening.
Alan KM6VV
I don’t think the speed of these sensors has ever been tested. You could be running out of bandwidth.
The OPB745 sensor or the LM311? I’m a little more familiar with QRB1134 sensors (initial comparison of data sheets didn’t show much difference, but I’ll read more thoroughly).
Other observations, rim is only about .165" tall (radius), sensor area is perhaps .140", and width of stripe about .120". Sensors were initially slightly low, which could have resulted in the loss of a good signal.
I’m not through with this sensor/setup just yet. I hit upon the idea of cutting a thin fiberglass disk, and making the strip taller (radius), and with fewer stripes, 32 instead of 36 total. I’ve also been comparing with a QRB1134 sensor, which I know has the speed.
8uS for the QRB1134 (phototransistor)
---- for the OPB745 (photo Darlington)
I was initially hoping to use the QRB1134 parts, but ran out of analog inputs on the uP. Also, as I realized later, analog sensors take more work (processor time), and are trickier to handle (code). the comparator is quite useful in this case. I’ll probably build up a comparator for use with the QRB1134 sensors, and A-B them (along with the revised encoder disk).
Also on the agenda are the newly updated Wheel Watchers,
nubotics.com/products/ww11/index.html
Although they don’t work with the nice aluminum wheels.
Alan KM6VV
In making an encoder disk to bolt up to the LM aluminum wheels I ran into a problem:
The wheel measures about .656" between bolt centers, as does the corresponding holes of the metal servo horn hole pattern on a universal bracket.
The metal servo horn hole pattern measures about .670" between centers. They bolt up to the aluminum wheel OK (clearance holes).
What doesn’t work is drilling a .062" aluminum disk with .0656" holes, and expecting the stack to bolt up properly. Ya wouldn’t think it would make that much difference!
Alan KM6VV
New Gripper pix:
http://www.marconettengineering.com/4-barGripper01B.jpg
Bottom view of jaws
http://www.marconettengineering.com/4-barGripper03B.jpg
Top view of jaws
http://www.marconettengineering.com/4-barGripper05B.jpg
Close-up of rear jaw microswitch
http://www.marconettengineering.com/4-barGripper07B.jpg
Side-view of jaws
http://www.marconettengineering.com/4-barGripper09B.jpg
Bottom view close up of LPA gears
http://www.marconettengineering.com/4-barGripper11B.jpg
Front view close up of LPA gears and bearing
http://www.marconettengineering.com/4-barGripper13B.jpg
Top view of jaws partially open
http://www.marconettengineering.com/4-barGripper15B.jpg
Top view of jaws closed
http://www.marconettengineering.com/4-barGripper17B.jpg
Jaws gripping on short can
http://www.marconettengineering.com/4-barGripper19B.jpg
Jaws gripping on short can
http://www.marconettengineering.com/4-barGripper21B.jpg
Bottom view of jaws closed
These jaws (grippers) will be added to the front of the ASB-13 Table Top 'bot.
8 gram microswitches
HS 5485HB servo.
LPA gear set
Ball Bearing Hub Set
Jaws, links and frame are 1/16" aluminum
(Ignore the aluminum block used only as a support).
Jim,
I did have to add a pair of 2-56 tapped holes into the drive gear, I suggest four (4) 2-56 tapped holes (std. servo horn pattern) be added to the drive gear as well. There is very little room under the mounted gear, otherwise the same clearance (drilled) holes could be used.
Alan KM6VV
Interesting gripper design. Needs more surface area to contact the can in order to grasp. Pretty cool.
Thanks.
I’m machining a short pair of jaws (probably out of 1/4" Delrin) to mount on top of the microswitches, that’ll give the jaw faces more width. Could even bend up a piece to mount on top of the microswitches.
It does pick up a can just as it is!
Can you get holes drilled and tapped in the second (drive) gear sometime in the future?
I’m also considering adding another R/C servo on one of the ends of the yoke (U-shaped frame) to allow the jaws to be tilted up. There would be a bearing on the other end. I realized after I made the parts that this might be possible. But nothing to secure the “ears” of the 2nd servo to so far. Maybe someone has some ideas?
Alan KM6VV
Is there a specific reason for the non symmetry of the parallelograms? No wait, trapezoidograms.
The links have two lengths, which cause the jaws to work closer to the body of the 'bot.
And, I’m still experimenting! Do you have some experience?
I got a simulation program AFTER I cut these parts.
Alan KM6VV
Looking good Alan!
Having the SES holes on the servo gears too was a good suggestion. Could be useful in some cases. But it isn’t that hard to drill and tapp the holes by hand though. 
I like it…
Some CNC work… !
Hi Kåre,
Thanks!
Yeah, not that hard to do by hand, but I like to use a drill! ;.)
After I’d had the gears mounted for a while, I noticed that I couldn’t align the two links mounted on them to be parallel. Oh no! I thought; I must have drilled for on-tooth rather then off-tooth (when gears mesh, one is on-tooth, the other is off-tooth. Which would mean I’d have to remove the gear, locate and similarly drill and tap the other two holes. Turns out, I just had to rotate the gear a tooth. Next time it’s off, 'tho, it gets more tapped holes!
Alan KM6VV
