S2 Project Badge for Maker Faires, etc.
Phil Pilgrim (PhiPi)
Posts: 23,514
A "Builder" Project for Parallax
A couple weeks ago, I got this phone message from Ken Gracey: "Phil, we've been discussing a brain-dead simple project here, and of course you were the first person that came to mind. ..." (Gee, thanks Ken!
) So, feeling a little brain-dead, I called back to see what it was all about. Basically, what Parallax wanted was a small PCB project that kids could solder at a booth set up for the purpose and could either pin to their shirt or wear on a lanyard. The design objective was to use as few parts as possible, yet end up with something that had a coolness factor and promoted Parallax and their products. It sounded like fun, so I accepted the project.
Ken relayed to me that they had looked at this "smart LED" from RadioShack that has functionality controlled by a push-button switch. He also mentioned that the badge had to be small, which precluded any sort of power bigger than a coin cell. So I went to RadioShack to purchase the LED, a CR2032 battery, and a holder for it to try out. I was concerned about the power requirements, since the LED package spec'd an operating range of 3.2V to 3.5V, and the CR3032 has a nominal output of 3.0V. But it turns out that the actual operating range is 2.5V to 3.5V. It's just that the blue LED segment, which has the highest forward voltage, will start to get dim before the red and green segments as the battery runs down. I was also concerned about battery life, so I let it run overnight. In the morning it was still operating but rather dim. So I turned it off for awhile, using the push-button switch. After the battery had rested for awhile, it snapped back and produced nearly the same intensity as when it was fresh, so I knew the LED/battery combo was a winner. The other thing I like about the LED is that when "off", it enters a low-power mode that consumes only 1 µA. So it's not necessary to remove the battery to keep it from running down.
Since I was given a budget of "under $3" for the whole kit (since Parallax will be giving them away), I had to locate the offshore source of these LEDs. As luck would have it, someone posted that info among RadioShack's user reviews for the LED, saving me a whole lot of research. Parallax ordered samples from Hong Kong, and I verified that they behaved identically to the RadioShack version.
Early on, and I can't remember who broached it, the idea came up that it would be cool if the badge could be made to look like an S2 robot. The idea stuck, and I made it my mission to create a small PCB, no more than 1.5" in diameter, that looks like an S2. It occurred to me that, with wheels attached, a properly-selected tactile switch could be mounted underneath the board to substitute for the S2's idler wheel. That way, the LED's function could be selected by pressing on the top of the PCB near the rear. In order to keep the top as clean as possible for graphics, I decided to mount all the parts on the bottom, with the LED leads being bent in a "U" shape, so it could protrude through a hole in the PCB.
So that was the plan: a badge shaped like an S2 robot with wheels, using the "smart LED", a tactile switch, and a CR2032 lithium coin cell for power. "All" that remained was to design it.
Graphic Design 101
The Shading
After trying to trace the design of the S2 from a photograph and color in some shading for an ultimate halftone to be printed on the PCB, I realized that what I really needed was the original 3D model for the Scribbler/S2. So I contacted Ben Wirz, whose company designed both robots, to see if he would send me some 3D CAD files. He responded with both an IGES model and an STL file, both of which I could import into my Rhino3D CAD program. Here's what the STL model looked like in Rhino with shading:
I used a red background, so I could create a color mask in Corel PhotoPaint that included just the plastic part. This was done after I erased some unnecessary detail, such as the holes in the rear section and for the speaker. Next I called the PCB fab to see how fine a halftone screen they could hold in their process. A 0.006" dot size was the answer that came back. Since the board was to be 1.5" high, that entailed an overall image height of 250 pixels before conversion to halftone. So, after shrinking the image and halftoning it, this is what it looked like. Well, to be fair, this is what it looked like after I modified it much later, per Ken's request, to lighten it a little:
What struck me about this is that it didn't convey much detail -- just a vague ghost of what the S2 looks like. So I needed to come up with something having a little more "snap."
The Detail
Rhino has a feature that converts a 3D model into a multiview drawing that shows all the edges in the design. The nice thing about it is that it does all the hidden line removal and shows only visible edges, if that's what you want. Starting from the IGES model that Ben sent me, I was able to extract the following line drawing:
This, too, had detail in it that I did not want, and the lines were too thin. By importing it into CorelDraw as a vector drawing, I was able erase some lines and thicken the remaining ones, so they could be printed as part of the silkscreen. This is the result:
The only thing that remained was the "S2" logo, which uses the "StarFleet" typeface:
There was a problem with this logo, though. At the scale I needed to use it, the outline and the space inside it were too small for the PCB fab's silkscreen process. So I had to convert the text to curves in CorelDraw to thicken the line and widen the gap. (Because of "dot gain" caused by the next step, I had to come back and widen the gap even more.) After this was complete, I could then sandwich everything together to produce the final silkscreen image:
Conversion to Gerber
Okay, so I've got an image I'm happy with, but how to get it into my PCB CAD program? It doesn't import raster image files. It does import Gerbers, though. There aren't any good "GIF2GBR" programs out there, I discovered, so I decided to write my own in Perl. It turned out that, to get a clean image of the outline and and logo, I could not start with an image downsized to 250 pixels high but needed one six times that big. This entailed converting the CorelDraw vector drawing to an image 1500 pixels high, which automatically upsampled the halftone, so that each dot became a square, six pixels on a side. The program scans the image both horizontally and vertically, replacing each continuous string of dots with a stroke, slightly wider than one pixel, and specified with a circular aperture. Each stroke is written to the Gerber file as a move, followed by a draw. As you can imagine, this produced an extremely large file, but it was sill manageable.
Here's the result, shown on the screen of my PCB CAD program (CADINT):
Here's a "wireframe" detail view without the lines shown in their full width:
Notice the hashing that goes in both directions. This was done to eliminate the bumpiness that would otherwise occur at the ends of the lines. BTW, had I taken the time to program it, I could have traced the outline of each block of pixels and included a Gerber "fill" instruction to do the fill. That would have reduced the file size considerably -- but for the same visual result, and I didn't want to waste the time to do it that way.
There was also some detail for the bottom side of the PCB that got treated the same way, i.e. text that followed a curve. But, due to the small font size, I had to sample the images at 2000 dpi, rather than 1000, to get a smooth rendering from the converted Gerber file.
More to come...
A couple weeks ago, I got this phone message from Ken Gracey: "Phil, we've been discussing a brain-dead simple project here, and of course you were the first person that came to mind. ..." (Gee, thanks Ken!
) So, feeling a little brain-dead, I called back to see what it was all about. Basically, what Parallax wanted was a small PCB project that kids could solder at a booth set up for the purpose and could either pin to their shirt or wear on a lanyard. The design objective was to use as few parts as possible, yet end up with something that had a coolness factor and promoted Parallax and their products. It sounded like fun, so I accepted the project.Ken relayed to me that they had looked at this "smart LED" from RadioShack that has functionality controlled by a push-button switch. He also mentioned that the badge had to be small, which precluded any sort of power bigger than a coin cell. So I went to RadioShack to purchase the LED, a CR2032 battery, and a holder for it to try out. I was concerned about the power requirements, since the LED package spec'd an operating range of 3.2V to 3.5V, and the CR3032 has a nominal output of 3.0V. But it turns out that the actual operating range is 2.5V to 3.5V. It's just that the blue LED segment, which has the highest forward voltage, will start to get dim before the red and green segments as the battery runs down. I was also concerned about battery life, so I let it run overnight. In the morning it was still operating but rather dim. So I turned it off for awhile, using the push-button switch. After the battery had rested for awhile, it snapped back and produced nearly the same intensity as when it was fresh, so I knew the LED/battery combo was a winner. The other thing I like about the LED is that when "off", it enters a low-power mode that consumes only 1 µA. So it's not necessary to remove the battery to keep it from running down.
Since I was given a budget of "under $3" for the whole kit (since Parallax will be giving them away), I had to locate the offshore source of these LEDs. As luck would have it, someone posted that info among RadioShack's user reviews for the LED, saving me a whole lot of research. Parallax ordered samples from Hong Kong, and I verified that they behaved identically to the RadioShack version.
Early on, and I can't remember who broached it, the idea came up that it would be cool if the badge could be made to look like an S2 robot. The idea stuck, and I made it my mission to create a small PCB, no more than 1.5" in diameter, that looks like an S2. It occurred to me that, with wheels attached, a properly-selected tactile switch could be mounted underneath the board to substitute for the S2's idler wheel. That way, the LED's function could be selected by pressing on the top of the PCB near the rear. In order to keep the top as clean as possible for graphics, I decided to mount all the parts on the bottom, with the LED leads being bent in a "U" shape, so it could protrude through a hole in the PCB.
So that was the plan: a badge shaped like an S2 robot with wheels, using the "smart LED", a tactile switch, and a CR2032 lithium coin cell for power. "All" that remained was to design it.
Graphic Design 101
The Shading
After trying to trace the design of the S2 from a photograph and color in some shading for an ultimate halftone to be printed on the PCB, I realized that what I really needed was the original 3D model for the Scribbler/S2. So I contacted Ben Wirz, whose company designed both robots, to see if he would send me some 3D CAD files. He responded with both an IGES model and an STL file, both of which I could import into my Rhino3D CAD program. Here's what the STL model looked like in Rhino with shading:
I used a red background, so I could create a color mask in Corel PhotoPaint that included just the plastic part. This was done after I erased some unnecessary detail, such as the holes in the rear section and for the speaker. Next I called the PCB fab to see how fine a halftone screen they could hold in their process. A 0.006" dot size was the answer that came back. Since the board was to be 1.5" high, that entailed an overall image height of 250 pixels before conversion to halftone. So, after shrinking the image and halftoning it, this is what it looked like. Well, to be fair, this is what it looked like after I modified it much later, per Ken's request, to lighten it a little:
What struck me about this is that it didn't convey much detail -- just a vague ghost of what the S2 looks like. So I needed to come up with something having a little more "snap."
The Detail
Rhino has a feature that converts a 3D model into a multiview drawing that shows all the edges in the design. The nice thing about it is that it does all the hidden line removal and shows only visible edges, if that's what you want. Starting from the IGES model that Ben sent me, I was able to extract the following line drawing:
This, too, had detail in it that I did not want, and the lines were too thin. By importing it into CorelDraw as a vector drawing, I was able erase some lines and thicken the remaining ones, so they could be printed as part of the silkscreen. This is the result:
The only thing that remained was the "S2" logo, which uses the "StarFleet" typeface:
There was a problem with this logo, though. At the scale I needed to use it, the outline and the space inside it were too small for the PCB fab's silkscreen process. So I had to convert the text to curves in CorelDraw to thicken the line and widen the gap. (Because of "dot gain" caused by the next step, I had to come back and widen the gap even more.) After this was complete, I could then sandwich everything together to produce the final silkscreen image:
Conversion to Gerber
Okay, so I've got an image I'm happy with, but how to get it into my PCB CAD program? It doesn't import raster image files. It does import Gerbers, though. There aren't any good "GIF2GBR" programs out there, I discovered, so I decided to write my own in Perl. It turned out that, to get a clean image of the outline and and logo, I could not start with an image downsized to 250 pixels high but needed one six times that big. This entailed converting the CorelDraw vector drawing to an image 1500 pixels high, which automatically upsampled the halftone, so that each dot became a square, six pixels on a side. The program scans the image both horizontally and vertically, replacing each continuous string of dots with a stroke, slightly wider than one pixel, and specified with a circular aperture. Each stroke is written to the Gerber file as a move, followed by a draw. As you can imagine, this produced an extremely large file, but it was sill manageable.
Here's the result, shown on the screen of my PCB CAD program (CADINT):
Here's a "wireframe" detail view without the lines shown in their full width:
Notice the hashing that goes in both directions. This was done to eliminate the bumpiness that would otherwise occur at the ends of the lines. BTW, had I taken the time to program it, I could have traced the outline of each block of pixels and included a Gerber "fill" instruction to do the fill. That would have reduced the file size considerably -- but for the same visual result, and I didn't want to waste the time to do it that way.
There was also some detail for the bottom side of the PCB that got treated the same way, i.e. text that followed a curve. But, due to the small font size, I had to sample the images at 2000 dpi, rather than 1000, to get a smooth rendering from the converted Gerber file.
More to come...
Comments
The next task was to extract a board outline from Ben's 3D CAD file. Again, Rhino to the rescue, where I was able to construct a bounding curve from the 3D model. But I had to simplify it so it could be cut out at the PCB fab place with a router. A call to the fab revealed that they preferred to use a 1/10" diameter bit, but would substitute a 1/16" bit if requested. 1/10" seemed a bit large, so I endeavored to make sure that all inside curves in the design had a 1/32" radius or more. Here's the result:
Whoa! What are those extra appendages for? That brings us to the next section.
Wheels
We wanted to put wheels on the badge, but we did not want to have kids gluing them on and possibly making a mess. The alternative was to snap them on, which entails a material that's somewhat flexible or springy. Since fiberglass PCB material is neither, that means the wheels had to be. Fortunately, there's a material (acetal copolymer -- like Delrin) that's flexible, comes in black, and laser cuts cleanly. It's a bit more expensive than acrylic, but the wheels would be quite small, so a whole bunch can be cut form one sheet. The arrowhead-ended appendages in the above board outline are the axles that the wheels snap onto. The two little plateaus on either side are what the wheels rest against so they don't wobble, but are still spaced away from the S2's body. Here's the design I came up with for the wheels:
The wheel on the left is in its uninstalled, relaxed state. The one on the right has been snapped onto an axle. The slightly expanded ribs apply pressure to the rear of the axle's arrowhead to keep the wheel pressed tight against the two spacers on either side.
Printed Circuit
The printed circuit board had to fulfill the following objectives:
- Easy solderablility.
- Weight toward the rear, so it would rest on the tactile switch when set on a horizontal surface.
- Non-obstructed pathway for the battery into and out of its holder.
Fortunately, the three major components fit in the small space, and routing the connections was easy (the "brain-dead simple" part). Here's a composite view of the board layout sans top silkscreen:Since the board will be fabbed with an ENIG gold finish and the contact area will be a little shallower than the soldermask, I had to make the negative pad for the battery extra large. Even though the negative terminal of a coin cell is slightly convex, I didn't want to take a chance with the battery sitting on the soldermask and bridging over the contact pad without touching.
Pre-prototype Mockup
I wanted to proof the entire concept before committing to fabrication. One reason was that there are no cheap, fast prototyping services which will provide a red soldermask with a black silkscreen and complex contour routing with a small bit. This means that I would have to plan for a more expensive pre-production run for the actual prototypes. So I milled out a board from copper-clad FR4 on my CNC mill, engraving the simple circuit with a V-shaped cutter. Next, I printed and cut a vinyl label for the top side to simulate the soldermask and silkscreen finish. Then the parts were soldered on and tested for functionality. Finally, I cut some wheels on my laser cutter and snapped them on. Here's how it came out:
PCB Fabrication
I sent the Gerbers for the PCB off to the fab for a quote. Naturally, the conversation centered around how these things would be panelized, both to facilitate cutting them out and to maximize yield in order to minimize cost. For maximum cutting efficiency, the parts need to remain joined to the panel after cutting, so they don't flop around or get sucked up into the vacuum. This can be accomplished by three methods:
- Scoring, where flat parts of the board are made collinear with a shallow V-cut that spans the entire panel. The panel is then cut apart with a "pizza cutter" depanelizer to free the individual boards.
- Tab-and-perf, where short perforations are formed from tiny "mouse bites", i.e. closely-spaced tiny holes that join the boards to the panel, forming a weak spot that can later be broken to free the boards. Think of a sheet of postage stamps.
- Double-stick tape, where the entire panel is adhered to the milling base with tape before cutting all the way around each board to keep the boards from coming loose.
The last option was discarded out of hand as too expensive and labor-intensive. Scoring seemed attractive, but the only flat spots available to do it were at the ends of the axles. So I drafted up a sample pattern to see how it played:The board fab didn't like it, because there was too little meat to hold the boards in place. Parallax didn't like it, because the score ran so close to the edge of the fiberglass in several locations that there was a chance for their pizza cutter to jump the track.
This left tab-and-perf. I didn't want to use this method, because it leaves rough edges where the boards are broken off: not a pretty sight. But there was no other choice. Fortunately, the mouse bites could be positioned in places where they would be out of the way and not very noticeable: i.e. inside the S2's wheel wells. So this is the scheme used for the panelization, and I think everyone is happy with the solution:
The final panelization, designed by the board fab, came out as a 5x3 array and has tighter packing than my example above. Tight packing is important for keeping the cost down. So, today, I placed a quick-turn pre-production order for some boards. They should arrive sometime next week. Hopefully, by then, we'll also have final samples of the lanyards that are to snap into the hole at the top of the board.
As always, Ken and the staff at Parallax have been extremely supportive of this "brain-dead simple" project. I've also gotten great cooperation from the sales staff and engineers at Lazer-Tech, my preferred fab house. They do beautiful work, especially with their silkscreening, which is one reason I continue to use them -- especially for a project like this. So, we'll see how the protos come out. Wish me luck!
-Phil
The pre-production boards arrived yesterday. I quickly assembled one and tested it. It worked fine. Here's how they turned out:
They're kinda dark, aren't they? So I've got to make some changes:
The next run will be a full production run, so I've got to get this right! Yikes!
-Phil
Dot gain is caused when the ink dots bleed outside their nominal boundaries. In a halftone, this can cause everything to look darker, but it's a particular problem in dark areas, as the dots tend to merge together completely, producing a solid black. To deal with such a non-linear process, it's necessary to apply a non-linear correction to the grayscale image beofre half-toning. Since I didn't want everything to be lighter -- just the dark areas -- that's what I ended up doing.
Here's an image that compares the original halftoning on the left to the actual silkscreen print. On the right is the new halftoning:
I arrived at it by examining the actual print to discover the "darkest" pattern that the process could hold. Then I made sure that there were no areas in the image that were darker than that. After setting the minimum brightness level, I had to adjust the transfer curve in Corel Photopaint to maintain the subtle shading in the lightest areas without their getting washed out completely.
Adding a Copper Pour
In order to add a copper pour, I needed to know how close to the edge of the board I could run it. You don't want the router bit cutting into copper when the boards are cut out. The bits are designed to cut fiberglass and will produce a ragged edge when cutting copper plating. The board fab told me I could get within 0.015" of the edge. So this is what I ended up with for the copper pour:
The green circle is the nominal outline of the copper pour. Not shown (because they're embedded in the pour) are two dummy SMD pads connected by a trace. This creates a net to which the pour can be assigned. The PCB CAD program also lets me set the standoff from the board edge and remove any areas that are not connected to the net. That keeps everything confined within the board outline. The two crosshatched rectangles are "keep-out" areas that act as additional "fences" that the pour cannot cross. I wanted to keep copper off of the axles, in case the extra thickness were to affect the fit of the wheels.
I sent the new Gerbers to the board fab yesterday, and Parallax ordered a scary-large quantity of boards. I hope they turn out as planned. With a 15-working-day turnaround, I've got three weeks to stew about it. Maybe I should renew my passport, just in case ...
-Phil
How will the serial cable attach? when can we get a schematic? could you pre drill the hole for the hacker port??
ok, ok, I apologize for the "unscheduled" jokes...
-Tommy
Tommy, you must bring some of your fun to the UPEW event. You're right down the coast - will you join us April 13-14th?
Ken Gracey
I might could make it to the event, but erco would have to drive up the 101 instead of the 5
and let me ride in the back seat with the five cases of servo's on my lap,
along with whatever bottled gasses they wouldn't let him ship...
@Phil, that is some peice of work, the kids are gonna love em..:thumb::thumb:
I want to come to UPEW just to build one...
-Tommy
Too bad my hobbies are out of phase. Before I was a forumista, I was a long distance bike racer. I'd bike from SanFran to LA every other weekend. I'd bike right though Pismo, Avila Beach, Oceano. Back then we coulda hooked up and taken over the world, Tommy! But now I'm knee-deep in toddler toys and just trying to keep up with the Joneses...
I will be driving the ford explorer to rocklin, It has all the head and leg room I will need, and it comes with sleeping quarters.
Just have to find the right parking lot.. (WalMart is always nice
You should see how nice highway 1 is now, it has bike lane from Guadalupe all the way to San Luis Obispo now.
It's like a bicycle freeway...
I hear you erco, Spend time with the kids while you can, they grow up quick, and then they are gone.
Funny thing about kids, is that they get old enough to move out, then they get old enough to move back in...
All three of my kids are all growed up and have moved away for good now,
It's nice to know they are able to fend for themselves and their own kids, (Grandkids are a blast!)
@Phil, sorry I killed your thread, it really is a cool mini S2 you have made, I don't know why nobody else likes it.
Maybe the people are still just too stunned to respond, And the beauty has them awestruck and speachless...
-Tommy
No problem, Tommy, and thanks for the compliment! I don't really expect relevant commentary in all my threads. Sometimes I just post stuff for my own benefit. The forum, in many instances, is just a repository where I can store my ideas and methods for later retrieval (lest I forget) and let someone else maintain it for me for free.
Besides, the kinds of threads that connect with most forumistas are those that evoke the response, "Yeah! I could do that. too," in other words, that are enabling. By contrast, this project required some privileged information and specialized tools to complete. So I don't expect it to be as evocative as, say, a Spin program that everyone can try out on their own system or a circuit that can be built with parts from the local RadioShack. It's just a journal of work that I found to be fun and interesting. If someone else finds it interesting, too, it doesn't mean they have to say so, and it doesn't hurt my feelings any if they don't.
-Phil
Definition
Brain-Dead Simple Project - 1. A very intense and difficult project which requires the best tools available and can only be accomplished by a master craftsman, whereas upon completion the brain of the master craftsman has been drained of the mental capacity for further in depth thought. 2. A project of difficulty which simply causes the brain to malfunction and become incapable of further thought.
Instead of a holder, make a cut-out in the PCB where the battery is to be located and solder a strip of metal on either side of the PCB.
As the battery is thicker than the PCB, it should fit pretty snuggly. (May want to use a 2016 as it's thinner, though)
Two pager vibrator motors to turn the wheels
Two tiny mechanical wiskers for sensors to keep it from bumping into things.
Can I solder wires right to the Propeller chip?
Now the Penguin an have its' own little S2 robot.
Anyways nice work PhiPi
Good Luck!
Thanks for documenting this. It's looking good so far; I'm looking forward to updates.
-Phil
Well, maybe I'll have them by Thursday. I was feeling smug about ordering them from Canada, since today is a holiday in the U.S. 'Turns out, it's a holiday* in Canada, too, and not one of the five "working days" in the turn spec.
*Family Day. We celebrate our government leaders; Canada celebrates its people.
-Phil
Bruce
The proposed changes should bring it right into line!
Can the "Parallax Inc" gold text be made to stand out more. The lines making up the text just seem really thin.
I think these are looking great so far!
-Phil
Waiting for that UPS truck like a doggy who sits in the window waiting for his master to come home. Sheesh. We need to figure out how to send physical objects over Skype.
Ken Gracey
I borrowed the logo from the top of the forum for comparison. I think the gold logo would look good somewhere in between the two sizes.
I think the Parallax logo should look lean and clean but not wimpy.
Edit: As I look as some of my Parallax boards I see the logos on those look good. You and Ken will be in a better position to judge with the physical badges in hand. I do wonder if the gold makes the text look thinner?
BTW, These do look really cool!
Observations:
1. Make pads around the pushbutton slightly larger since we don't know what diameter solder and size of soldering tip people will use.
2. Copper layer to brighten up the red on top.
3. Reduction of shading to lighten the board.
4. No change to the Parallax logo - simple is fine.
5. Make all text on back gold.
I'll test it on a 10/12 year old later tonight with the actual instructions. There's nothing difficult about putting this together. By comparison, you should have seen some of the learn-to-solder projects I've seen elsewhere - using SMT parts sometimes! This is very easy by comparison.
This project was clearly more challenging than it appeared on the surface. Much of the project was made interesting by the mechanical design. Choosing the right suppliers really helped, too.
We are building 6,500 kits of these.
Ken Gracey
...'tis time to shine up the ol' Laser Lenses...
-Matt
Those things look so cool!
Run 'er ragged, Laser Lama!
We also need these little LED-bender jigs, too. This is a single-use part, so make the LED-bender from low-cost material, Laser Lama!
Ken Gracey
The gold lettering really looks great. it adds BLING factor..:cool:
I could practice maze solving and line following strategies on just one sheet of paper,
and do tricks, like the kids do with those miniature finger skateboards...
-Tommy
Subject tester is 12 years old with prior soldering experience. Human tester approached project with substantial enthusiasm - supervisor made the following observations:
(a) Studied the LED lead-bending process a bit before undertaking that task.
(b) Didn't know that the half-moon inserts had to be removed from the wheels - documentation didn't say - but he figured it out.
(c) Noticed that the battery holder took a significant amount of heat.
(d) Expressed significant satisfaction upon completion of project - labelled it "really cool, fun, with a smart LED"; he is prepared to impart the skills to other humans. Didn't tell me that every kid would be willing to wear one of these around their neck at school.
We will try this on much younger kids with less experience, too.
Ken Gracey