I understand his reasoning here, but it would be nice if maybe the board was designed to be expandable where those of us who may want I2C can add some more electronics on the board. Of course it would be even better if he sold both versions , but that may be asking too much.
Would a connector mounted on the board be good for this? If so, what kind of connector would you want on it?
Also, Jim had said that he wanted to keep it #1 - cheap and #2 simple.
I don’t think that he wants to create an embeded circuit board for the foot. I2C would be nice, I have to agree, but I think he wants to make a simple design that’s a “bolt on” solution as far as ground contact goes.
This is why I have been hoping to see a smaller single sensor board that could be mounted on a foot. You could then have different boards with different sensors (or combinations of sensors) that could be mounted in the same positions on each foot. One board could be the one with the micro switch and another could have a capacitive touch/force sensor. These would be usable in areas other than foot sensing too, so might have a wider appeal.
The mount could be as simple as a bracket that attaches to an " L" type (not the ASB-06, which would be way too large for this purpose) bracket positioned at the corner of the foot. It should be something that doesn’t require more than two 0.094" diameter holes spaced 0.3280" apart to mount it to the foot or anywhere else. Yes, this is the spacing of the satellite holes from the large hole in the SES patter. Using this spacing would allow a sensor (or sensors) to be positioned to the front and side of the corner of the foot if desired.
As a first go at this, I agree completely with keeping it as simple as possible. I just offer an alternative way of accomplishing what Jim wants to do while extending the audience for the sensor board at the same time. Heck, the foot sized ARF-01 replacement can be exactly that - just having the SES holes at the same place as the ARF-01 and providing mounting places for single or combination sensor boards. The first sensor board could be the one with the micro switch.
I’m not suggesting he do this either, although there is always the possibility of doing a sensor board with on board intelligence later on, depending on the sensor.
There is another very cool thing about my proposal - you can create any number of different sensor boards for the feet and mount them to the same ARF-01 or ARF-02 replacement board.
Perhaps have a simple header that allows future boards with I2C or any embeded design for that matter, to quickly and easily interface the built in pressure switches is all that is needed. Then anyone with a user specific idea has the option of adding anything to the switches via a cable.
The problem with trying to add one soution over another on the board; there is always someone who would prefer somethig else for their user speciffic application.
I think of it like a PING sensor that simply sends a signal when objects are detected and it s up to the programmer as to how to use this data. It does not have all kinds of bells or whistles. Same concept as the tilt sensor, or the ir sensors, the programer makes it all come together.
I don’t think your are understanding what I am proposing. Maybe my idea is just way off base and wouldn’t work.
I am not proposing to add one solution over another. It’s all a single flexible solution to what Jim wants to do. One ARF-01/ARF-02 foot sized mounting board that would allow mounting up to four small single or combination sensor boards to it. The sensor boards could be anything from Jim’s simple micro switch to a capacitive contact/force sensor that could actually measure the force being applied to the ground (or other object).
I think what Jim wants is basically a set of four bumper switches, but for the feet with a different orientation facing downward for the switch triggers. This would be so easy to do with my idea!
I would just like you to understand what I am proposing. I think we have been having a “failure to communicate” somewhere along the path of this project. I’ll bow out, try to keep my mouse away from the quote/post buttons and just watch and see how things progress.
So you want a board that you can add other boards to? Make me understand. Also, where will these other boards mount to? How many other boards would be able to be added to the foot sensor and fit? Don’t forget, there is a servo bracket and a servo in the middle of the foot so the only room you have is in the front and back.
You can already buy a capacitive contact sensor from spark fun, why couldn’t you just buy one or two or four of these and paste them on the ARFs?
I think the origional idea for a simple micro switch solution has been lost somewhere in the discussion. I think some microswitches, popsicle sticks, hot glue, and wire could have a working prototype within a matter of minutes.
Update: Use paperclips in the prototype instead of microswitches to save $$$ and a trip to Radio Shack.
Yes, this is the idea! Imagine, if you will, a 1/8" or so think lexan type board that is shaped exactly like the ARF-01 or ARF-02 feet. This board has the same SES hole location as the respective ARF-01/ARF-02 has.
Yes, I know the space is limited, but I have that all figured out. We use a set of two .094" diameter holes either side of a .315" diameter hole spaced .3820" apart… Yes, that is SES hole spacing. You have four sets of these holes spaced equally around the foot close to the front and back edges. This is where the small sensor boards would mount on short standoffs like those used to mount the EC-02 on the ASB-503 of the BRAT.
Now, we add mounting holes for small sensor boards, at least four, to cover the four corners of the foot. These small boards are what the sensor(s) are actually mounted to. No intelligence whatsoever is necessary for most sensor processing. Just the sensor(s) on the little board that mounts on the ARF-01/ARF-02 foot replacement board. You just have wires routed to the center of the foot where they then wind up through the leg to the master microcontroller.
You can also mount these little sensor boards on other parts of the robot too! They are not usable just on the feet. Even a simple micro switch contact sensor can be use elsewhere, such as for a limit switch.
Yes, you can already buy these, but they are too expensive. I believe they can be done for a better price, and I know Jim knows how to get that sort of thing done. I think the boards could be smaller. I can illustrate this, so hang in with me a little bit longer. I just have to model a modified ARF-01 and I think that will help you understand. I should have done that to start with.
My comments I thought were based on Jim’s original post where he stated he wanted to build a PC board and he mentioned LEDs and the like.
I just wondered if it would make sense to in addition to maybe having a standard 3 pin connector or the like to connect to your microcontroller, you also had unused etch and holes on the circuit board that maybe allowed someone to solder in either a PIC or ATMEL microcontroller (like maybe ATINY24) and maybe some holes for either or SPI or I2C…
But this is probably overkill and I too will drop this.
I did have to adjust the position of the large holes, and the position of some are still not quite right. The center SES holes are right where they are in the original ARF-01.
The PCBs are too large, imo. The simplest way would be to put one of these:
in the middle of the hole with the button facing down. When the BRAT steps down the switch is pressed. The PCB would have minimal requirements. A switch, few pin male header, and a resistor + LED if you want. The max you’d need for this is probably a .5 x .5" PCB… You can panelize a 2x2 board and get a foot (four PCBs) for $2.50 from BatchPCB. Both feet would be only $5!
If you want a more advanced method that can detect how much the bot is pressing down there is another way of doing it. You can put the PCB on the bottom side of the foot. Between the PCB and metal place a small FlexiForce, or similar, sensor. Then to the side of the servo make another PCB which will have a micro that will read these four flexiforce sensors. Then give the micro slave I2C or serial commands so that the Basic Atom and other micros can communicate with it.
For starters, a simple switch will work fine… These switches aren’t large. You’d only need a 2 pin header for the switch, or 4pin for switch + LED.
I wanted to leave enough space for expansion, so the board size did not have to be changed. I have a slightly revised version of the sensor mount, which does not have the larger center hole. I also removed that larger center hole from the modified foot and set the set of four large holes back away from the edge of the foot a little more.
This sensor mount in the revised foot is 1" x 1" so there should be plenty of room now, especially since that large center hole is not there. anymore.
It has to be simple and easy though. I just wanted to allow for enough space to do other stuff on the same sized PCB, and I think I accomplished that. SMD/SMT parts do not take up a lot of space. The switch has to be normally open. I even see a way where you could have both a forward looking sensor as well as the PCB with the downward oriented sensor. All you have to do is attach an ASB-06 with the vertical part towards the center of the foot and attach the PCB to that, through to the foot (or vice versa to allow the holes to be counter sunk like the center SES holes).
There are a lot of possibilities, but for now it must be a simple mechanical contact switch. That allows for a fast reaction time, like a bumper switch would.
I say use empty ravioli cans, and use tin snips to save the environment. But paper clips would be easier. There has got to be some use to toilet paper cores cuz I go through so many of those with a family of four.
“If you feel there is something wrong with my, and/or linuxguy’s, design feel free to speak.”
Don’t pay any attention to the “non producing” peanut gallary ( en.wikipedia.org/wiki/Peanut_gallery ) if they don’t offer any useable input. Anyhow, below is a quick (well, I had to wait for the glue to dry) type of edge detector that could be glued on the top front of a Brat’s foot, giving it a “big toe” to sense when the front of the foot is beyond a drop off. Paperclip, popsicle stick, wire, and glue. Contact between the paperclip and the wire on the popsicle stick is made when there is no surface in front of the foot to lift up the end of the paperclip. Something similar could be made with a microswitch if desired. My multimeter says it works. KISS