Zoomcat, that’s the best concept I have seen so far! I love the spring operated contact which allows the foot to be completely flat.
The micro switches woud have to be inverted for this operation, but it would work very well.
I think the simplest idea is to use tactile switches on the bottom of the foot board, but incase the buttons require two much force, a design like I have below could be used.
http://img404.imageshack.us/img404/8035/footswitchnt0.png
Atvatages:
-
No need to have board fabed
-
Parts already available on LM shef except switches - ballbearings ( lynxmotion.com/Product.aspx? … tegoryID=6 ) can be used for the fulcrum.
-
Simple aluminum brackets are needed for fulcrum and micro switches
-
The exsisting ARFs would only require small screw holes added for the hardware upgrade components.
Hi Mike,
I see you’re using a lever-actuated microswitch, presumably to allow overrun of the foot pad. Couldn’t you also just turn the switch over, and use a push-pin inside the foot perimeter to actuate the lever of the switch? No fulcrum or bearing.
Alan KM6VV
To keep things simple and minimize added weight to the foot, I’d go with an inverted microswitch. Using hot glue (which is quick and generally removable when desired), I’d glue a small thin piece of wood (I use the cheap three layer hobby plywood from Lowes) vertically on the foot top, and then glue the inverted microswitch to the side of the piece of wood. Then glue a small touch probe to the microswitch lever, which would extend out beyond the front edge of the foot. The touch probe could be made from something simple like a bamboo skewer. I think the general idea is to get something simple working to see if further development is worth while. Developing software to make the sensor actually useful maybe the difficult part.
There are several ways to do something like this. Adding a push pin would still require a bracket for the microswitch to hold it inverted . The push pin wound need a shoulder to keep it from passing through the foot and held in place from the tenssion of the microswitch lever. This is a good idea actually. I will whip up a drawing and post it here.
KM6VV,
How’s this design? A handfull of stainless pins would need to be turned for this application and ther might need to be a matching insert to allow free movement of the pin. Aluminum can be a little sticky.
One thing about this design is the stiffness of the button. Without a lever like my previous design, you have the potential of 4 switches combined requiring a lot of force to depress all 4 at the same time. This is something yet to be tested.
Hi Mike,
That’s what I had in mind. I’d press on the end of the lever, 'tho, rather directly on the microswitch button. I’ve been looking for my collection of microswitches! I want to adapt one to my 1/8" aluminum Tibia.
True, the forces could add up. Are 4 switches really needed for a single foot? Really light (force) switches are probably needed. Some of the microswitches I have seen have a nice long lever arm; I’m thinking that the end could be bent at right angles to extend through the foot, and would actuate the switch just before contacting the foot, thus limiting the travel and further protecting the switch. No push-pins needed! Now to adapt this to a tibia!
Alan KM6VV
Sure, I could move the microswitch to have the pin directly under the end of the lever. Some microswiches have a roller brearing on the tip of the arm as well which would be nice for the pin to push on.
I just did a rough drawing to plot down the basic idea.
Hi guys,
been a while since I’ve been around here, new job and all that…
Anyway I’ve been reading up on the posts here and found this discussion, thought i might give my two cents, dunno if they’d be usefull though.
We use tiny detector switches at work which seem to be useful in the application of a foot sensor switch, they’re 3.9x3.3x1 mm in size and need only 0.3N to be operated.
here’s
a link to a datasheet of the lil buggers.
/edit: this is the site of the manufacturer, there’s other series in the same detector switch line that might make integrating them into the feet simpler.
In fact, just try “detector switch” in google. Lots of manufacturers have the same sort of thing.
I’m also trying to get some silicone rubber somewhat conductive so i can try to make something like a rubber “sock” that would give an ohmic representation of the pressure on the foot, I’ll keep you updated if i make any progress on that.
I really like the idea Jim has for this and some of the different variations that have been suggested, especially the simple ones. Yes I would love to have a more advanced interface to it, but…
When I first built my Brat, one of the things on my list to do someday was to add support to keep it from walking off of the table. My original idea was to do it in much the same way that the iRobot keeps from falling down the stairs. What I was planning was to mount a Sharp GP2D12 sensor on both feet, probably pointing a little forward. If while it was moving, I would detect the output voltages and if it dropped below a certain threshold, I would trigger the oops condition…
Never completed it, but that was my earlier thoughts and who knows maybe I will play with it now. But first I need to take my Brat out of surgery…
http://img239.imageshack.us/img239/6766/bratgk1.jpg
Again I like the ideas up here and depending on what finally gets built, I may upgrade my brat
Kurt
Hi Guys,
Yeah, the roller could work! And the SMK “sensor switch” could no doubt be pressed into use as well. Haven’t figured out what Kurt’s pix is!
I’m wondering if users of this foot (I’m not) would want to detect which portion of the foot was seeing the ground? Or would any closure do? Maybe obstacle detection could be incorporated? The 'bot could sense if it “kicked” something? Maybe a “toe” sensor is needed. But it doesn’t seem like a lot of foot action is used, more like a shuffling action.
I’m more interested in hexapod foot/leg sensors at this time. Maybe there’s some commonality.
Alan KM6VV
It’s ok, I think I can take your eagle files into my full version and modify the size. My full version is older though… I made some progress on my schematic I will try to post it today. It’s simple, only take 1 Analog input per foot and will be very inexpensive.
I have read this entire thread. I like lots of the ideas. Still to early to know the final outcome. Thanks for the input…
My free version of eagle cad will let me draw the full size of the foot, it just won’t let me place the components past the size limits. I can do most of the work in the free version, I just will not be able to lay it out.
How about a Capacitive Sensor? The QT113-D touch sensor looks like it could do the job. A sheet of thin PCB could be made practically the same size as the current foot.
I haven’t worked with this sensor yet. It would appear that a short insulating tube extension to a hexapod Tibia could have a sensor plate (inside surface?) added, and would thus be able to detect objects near it.
Perhaps a small disk at the END of the Tibia would allow sensing just a contact with the ground surface.
Obsolete, replace with QT100 DigiKey 427-1120-1-ND
The QT1081 DigiKey 427-1127-1-ND is 8 channels.
Alan KM6VV
Has anyone tried the optical fibre tactical sensor approach?
A simple loop consisting of a led/laser shining into a looped fibre, received by a photodetector.
Apparently touching (thus bending) the fiber gives a fairly decent light variation through the fiber because of more / less reflections which should be detected by the photodetector.
Haven’t tried it myself yet though, but maybe it could serve some use.
I like the capacitive idea too though… hmm maybe i should start testing some things instead of constantly thinking about them 
Well last week i drew up some schematics, but have aparently missplaced them. Man I gotta get organized. 
So anyway my first idea was to use a passive resistor ladder and some switches to generate a unique analog output for each of the possible closed switch conditions.
I also wanted to use a normally open single pole single throw switch and have LED’s light up when each switch is closed.
Well the problem is an R2R ladder requires either a high or a low on each input, which is not possible using only a NO SPST switch. LED’s are an obvious problem as well. I’m not willing to leave off the LED’s so I am ok with adding some active components. Gots ter haven der lightsenblinkin!
So the easiest way to get highs and lows from the switch is with a simple hex inverter IC. Here is my sad excuse for a schematic. lol…
Forest Mimms? is that you? 
Question:
Are those three resistors with the dotted lines going to 5v?
I still have some of those books! 
The dashed lines go to additional switch, inverter, led circuits, four in total. Sorry I left that info out.
Ah ha! I understand, you only drew one example for the first resistor that can be duplicated for the remaining three. Got it! 8)
somewhere over the southern horizon I just saw a big flash as the light bulb went on… 