Arduino Controlled Pneumatic Servo

For an upcoming project I needed a pneumatic ram with a closed loop control system so I could position it accurately. Didn't have the budget for an off the shelf solution, so I bodged one together with an ardunio, a couple air solenoid valves, and a pair of potentiometers.

How it works is one potentiometer is the target while the other is mounted to the ram. The arduino code compares the two, figures out the direction that the ram needs to move to match them up, it then cycles the solenoid valves on and off accordingly. Once the values match, it turns both valves on, more or less locking the ram in place. The target pot can also be replaced with any analog input.

For the rest of the hardware I am using a pair of clippard mouse valves and an arduino UNO. The valves work well, 10ms response time and very low power consumption you need a transistor and diode to hook them up to the Arduino, look up the solenoid tutorial. The ram is a double acting one. Also It works best if you have an adjustable inline flow restrictor on the incoming air stream. It allows you to fine tune the rate of fill so the servo is less twitchy.

The big secret on this was to mount a wheel onto the knob of one potentiometer, this wheel has a string wrapped around it that is tied to the end of the ram. When the ram extends, it pulls the string, spins the pot and tada you now have the rams position. You need a second string with a spring attached to roll the wheel back and keep everything taut. To work out the diameter of the wheel just take your rams stroke and multiply by PI to get your radius. Keep in mind that your pot might not turn 360 degrees, mine was 270 so factor that into your calculations. Best to make it a bit oversize anyway, worst case is that you lose some resolution if its too large, if its too small the ram will destroy the pot or break the string when it runs out of room. In this case the wheel is made of birch plywood, with a groove turned into it on the lathe to guide the string.

Circuit diagram and breadboard:

I drew these up with Fritzing, which is an awesome open source electrical diagram program. (I have no connections to the project, just a happy downloader)

Pneumatic servo Arduino code:

nt targetPin = A2;// select the input pin for distance sensor or potentiometer
int actualPin = A0;// select the input pin for the potentiometer attached to the ram
int outSol = 12; //digital pin with the solenoid that forces the ram out
int inSol = 11; //digital pin with the solenoid that forces the ram in
int actualRange = 838; // range of the potentiometer attached to the ram -test it with serial out)
int targetRange = 650;
int pollDelay = 60; //how fast to cycle, figgure out how fast your valves will cycle, I am using 5-10ms valves
int fuzzyFactor = 100; //how close the two values need to match

int targetValue = 0; // variable to store the value coming from the sensor
int targetValueLow = 0;
int targetValueHigh = 0;
int actualValue = 0;

void setup() { // declare the solenoid pins as outputs
pinMode(outSol, OUTPUT);
pinMode(inSol, OUTPUT);
Serial.begin(9600);}

void loop() {
pollDelay = random(80,120); //generate a random delay between sensor checks

// read the values from the sensors:
targetValue = analogRead(targetPin);
actualValue = analogRead(actualPin);

actualValue = map(actualValue, 0, actualRange, 0, 1023); //remap to whole range
targetValue = map(targetValue, 0, targetRange, 1023, 0); //remap to whole range and invert

Serial.println(actualValue); //write stuff to the serial for debugging
Serial.println(targetValue);

//Set up the value range thats acceptable
targetValueLow = targetValue - fuzzyFactor;
targetValueHigh = targetValue + fuzzyFactor;

if (targetValue < fuzzyFactor) //if the goal is full retract lock the ram there { digitalWrite (outSol, LOW); digitalWrite (inSol, HIGH); Serial.println ("Case 1"); } else if (targetValue > (1023 - fuzzyFactor)) {//if goal is full extend lock the ram there
digitalWrite (outSol, HIGH);
digitalWrite (inSol, LOW);
Serial.println ("Case 2"); }

else if (actualValue > targetValueLow && actualValue < targetValueHigh){ //if the ram is close enough to the goal apply air to both sides digitalWrite (outSol, HIGH); digitalWrite (inSol, HIGH); Serial.println ("Case 3"); } else if (actualValue < targetValue){ //if its less than the target extend digitalWrite (outSol, HIGH); digitalWrite (inSol, LOW); Serial.println ("Case 4"); } else { //if more than target retract digitalWrite (outSol, LOW); digitalWrite (inSol, HIGH); Serial.println ("Case 5");} delay(pollDelay); }

Curved Slot Cam Plate

Prototype of a mechanism for an upcoming project. I think I am going to use it as a locker mechanism for a wooden box.

Steam Engine

First big metal project on the lathe, a small single cylinder single action wobbler engine, turned from solid brass. I wasn't working from plans, more just winged it with the materials I had in my shop. It stands about 4" high. Runs well on about 15psi of air, starts to shake above 40psi, and turns into walking device at about 80psi. Not the most balanced of engine designs. Fun to make though. See below for a video and construction photos.

Turning the flat on the piston. The heat shrink tubing you see is packing material, behind the brass piece are two lathe tool bits used to provide support and space the piece being turned away from the back of the chuck. I had to flatten one side so it would clear the crank at the bottom of the engine.

Finished piston next to the cylinder/column/crank assembly.

That little pin is what attaches the piston to the crank wheel.

Its press fit into place using a 2 ton arbor press. The trick is to turn the pin slightly oversize so it doesn't quite fit in the hole, then use the press to persuade it. I did the same thing with the crank shaft and flywheel.

Turning the column. I heard that turning a square into round on the lathe is supposed to be hard on the tooling but I found with a piece this small I just set the lathe to its slowest speed, and took lots of light cuts. I did manage to jam the bit once, but fortunately my lathe is belt driven so the belt slipped and didn't cause any damage.

Finished cylinder and column. Those two large extra holes on the side are where I miss drilled the column, they are now lubrication holes. The smaller hole is where the air supply/exhaust hooks up.

I added this chunk of brass to the bottom of the engine to stabilize it. It is attached with a single screw and a pair of alignment pins. Also gives me a good place to tap a couple holes for the mounting screws.

When I tapped the holes in the base I used the wrong tap. So after some swearing I realized that I could just make my own brass screws. here is one being parted off the lathe. I used a die to cut the threads. Once parted the slot at the top is cut with a hacksaw.

Desktop Periscope

This is a little desk toy I made for my office cube. When cranked to its full height of five feet it affords a view over the cubicle walls. It can also be dropped to about three feet so as to not raise suspicion when its not in use. The top mirror rotates 360 degrees so I can see behind me as well. Both mirrors are mounted on pivots to adjust their angle.

I used red oak for most of the wood parts, the winch drum is turned from solid maple. The metal bits are aluminum, the pulleys, mirror mounts and rope guides were all turned on a small lathe.

The mechanism is pretty simple, the mirror support rods are constructed in two telescoping sections, there is a series of pulleys on both that allow a winch drum to pull the top section up, a ratchet and pawl keep the winch from winding back when the handle is released.

I will put together a video once I get it installed in the office.

Umbella Spotlight

A project I made to keep the darkness off.



Youtube link: http://www.youtube.com/watch?v=gRaC10y8_eg

It has an articulated baltic birch, maple and bamboo frame with 6 Luxeon Rebel 1000ma LEDs each putting out about 310 lumens, or 1,860lm for the set. For comparison, your average LED flashlight kicks out about 100lm or so. This thing is rather bright so the fedora is a safety feature as well as a fashion statement.

The wooden frame articulates, allowing the LEDS to pivot between fully inward facing to fully outward simply by raising and lowering the umbrella slider. This allows the umbrella to cast everything from a tight spot, to a wide doughnut of light.

Right now I am powering it off an 18v cordless drill battery, at full brightness I can get about 3 hours before I have to recharge it. The ultimate plan is to run a few cylindrical rechargeable cells inside a hollow metal handle.

Oh, and in real life it doesn't make that choir noise.... yet.

Tarn - Lightweight Canoe Plans

Finally got around to drawing up the plans for my lightweight canoe, click the image for larger. Or even larger here.

As a quick disclaimer, this boat was built with absolutely perfect stock, no run out, no knots, so I was able to pare down the dimensions of the parts to the minimum. This means that if you have any flaws in your materials it would be advisable to upscale the parts accordingly. The same goes for the rest of the design, it has a minimum of ribs and a minimum of redundancy. Paddle it accordingly.

Usual restrictions apply, Commercial usage of this design is forbidden. Personal use is allowed, up to two boats, but no reproduction of the boat or this design in exchange for money or goods is allowed. Also, if you do build one, send me an email and/or some photos: dbynoe@gmail.com

Machine for Preserving the Wind show video.

Here is a video showing my latest installation work, Machine for Preserving the Wind. Its up till June 30th, 2011 atTruck Gallery in Calgary, Alberta, Canada.

There is also a review of the show up here at FFWD Weekly.

Oh, and Adam Savage tweeted a link to my sculpture, my inner geek is unspeakably happy.



Each of the 40 poles that make up the piece are mounted on a string pivot, allowing them to sway freely. They have a cast concrete counterweight at the bottom that balances the wooden pole so they can sway with the slightest movement. These are then connected with a series of strings that allows them to be pulled back, and released to sway freely on their own. In this way the poles are indirectly coupled to the drive mechanism, they can be pulled back, but then they are free to sway and swing on their own, resulting a much more organic and random movement than if they were directly coupled.

To drive this mechanism the poles are connected to a pair of large wooden cams. A cam is basically a wheel with an irregular shape that causes a follower on the wheel to move back and forth in a fashion reflecting these bumps. In this case each bump outward corresponds to a gust of wind, and pulls the poles backward. The two wheels turn at slightly different speeds, and the output of each is mixed and averaged, resulting in a program that takes over 14 minutes to repeat.

The cam wheels themselves are driven by a multi stage reduction drive. An electric motor, of the type usually used in a household furnace, is hooked up to a worm drive speed reducer, and then to a two stage chain drive, this setup brings the speed of the motor down from 1725rpm at the input to half of an rpm at the output, a reduction of over 3000:1, this process increases the torque from half a foot pound, to over 400 foot pounds after friction losses. This was fun to design.

Upcoming Show - Opening June 3

Just a heads up that I will be having an opening for my new show "Machine for preserving the wind" on June 3rd, 2011, at 8pm, at Truck Gallery. It will be a kinetic installation that will look like a field of wheat blowing in the wind, only without the actual wind. Instead it will be driven by a series of ropes hooked up to two large cams. Basically I am making a mechanical system to mimic the movement of the wind poles outside.

Wind Poles - North Glenmore Park

Here's a small video showing those wind poles in action at North Glenmore Park. They move remarkably easily once you get the counterweight mass balanced properly.




Now I just need to make a system of ropes and cams to replicate that movement indoors.

Canoe Sculpture

The latest project to come out of my shop is an cedar canoe frame for a nice family who have a place in Canmore. This boat is similar to my ultralight canoe, but shortened to 8' so it can fit better on a wall or over a pool table.

It's built mostly out of a nice unique piece of western red cedar. This tree produced lighter colour wood for about 20 years before going back to the normal dark red cedar colour. This left the board with a stripe running down the middle that I was able to incorporate into the inwales, floorboards and lamps. The rest of the boat is the usual hodge podge of yellow cedar, mountain ash, yellow cedar and fir.

Wind Field prototype and production.

Started working on a new large sculpture, or more accurately a large series of small ones. Basically a set of wind feelers. They consist of a 6' wooden rod that is attached via a pivot to a base. They are held upright by a molded concrete counterweight. The tricky part is that I want to build at least 50 of them.

Scroll to the bottom of this post for a video of it blowing in the wind. 

This is the plasticine sculpt for the counterweight. I calculated the rough volume and dimensions in Rhino 3D, it had to be fairly accurate as the volume will affect the final weight and thus the performance of the sculpture.

The plasticine was built around a wooden core, once the sculpt was done this was removed leaving a hole. In this way I can avoid drilling the concrete after its cast.

To make my production casts I started by making a mold box out of 1/4" plywood, the tapered shape was used to reduce the amount of casting rubber needed for my mold.

I then installed nails in my carefully smoothed model, and suspended it a half inch above the bottom of the mold.The balls form registration pins for the two halves. Once the first half was poured, the nails were removed, the holes filled, and the upper half was poured in.

This is the finished mold made from poured urethane casting rubber. The core in the center has a removable steel rod to keep it aligned.

The finished counterweight. I found out that if I use Quickset concrete I can demold the object in about an hour and half. Once pulled out of the mold I let it set up overnight in a water bucket.

Lots more finished counterweights.

Detail of the sculpture base, The arm assembly pivots off of a piece of string.

Lots of bases, I have made a series of jigs to simplify the construction, there are about 20 of them here, only another 30 to go. 

For extra credit I designed the bases so the legs can pivot inward and essentially flat pack. 

World Skills Wings - Construction Video

Heres a video showing the wings I built for the City of Calgary Public Art at World Skills 2009.  I also included a bunch of footage showing how they were made.

Canoe Finished

Got my new canoe done. Still need to name her but details details. The shot to the right was taken up at Consolation Lakes, about 3kms into the backcountry from Moraine Lake, close to Lake Louise. Stay tuned for a video showing the trip.

Final weight came out around 15lbs for the boat, a little more than I wanted but still about 30lbs lighter than your average rec boat. I ended up adding a couple extra ribs in the middle to make it a bit stronger, probably could have gotten away without them but now its pretty much bombproof.

By the time I get my drysuit, lifejacket, throwbag, paddle and bailer loaded up my pack ends up being around 30lbs total, including the boat, so its pretty easy to carry in for any reasonable distance.  

Some construction details:

To skin the hull you basically just drape the fabric over, clamp it to the gunwales and sew the fabric together at the ends. For this boat I sewed the skin on wet and a little slacker than normal, I was afraid of the tension distorting the light frame.

Detail of stems. The stitch is basically just a saddle stich with two needles for the first pass. The fabric is then trimmed close, and the ends are folded under with a double whip stitch.

To attach the skin to the gunwales I had to get creative. Normally one would staple the skin to the outer gunwale, then screw in a rub strip to cover up the staples. This method is fast, but I didn't like the look, or the extra weight.Instead I drilled holes along the inner gunwale. These were drilled at about a 45 degree angle so they came out at the bottom corner.

The skin is pulled taught over the gunwales (while wet) and sewn through the holes. The top edge gets cut and I run another row of stitches to lock it back. Unfortunately this method means that I have to sew the entire length 4 times. I think it looks cool though.

Whenever I got to a rib I just skipped over it. There is enough tension on either side to keep it taught.

Same deal with the thwarts.

I ended up coating it with that same goop from the skin boat school that I used on my other kayaks. Its a two part polyurethane and goes on pretty quickly and painlessly.  You can see how translucent it is, that blue stuff is the waterline.

Broken Islands Tidal Life

Was digging through my harddrive and came across a couple videos I shot of marine life in the Broken Islands. Now that youtube lets you upload in HD I thought I would share.

Hermit crabs and a bunch of other critters in a tide pool:



Life in the lagoon. Taken between Jaques and Jarvis Islands in the Broken Group, west coast of Vancouver Island, right about here. Lots of sea stars, lots of shells, and a number of red sea cucumbers.

New Lightweight Canoe

Its been a busy last few months, been working at Street Characters Inc. building mascots. Spent my evenings working on a wooden drawing machine (should upload some photos of that...) and yet another boat for my fleet. This ones a canoe, and I am working to keep it under 15 pounds so I can carry it into backcountry mountain lakes. The complete frame is to the right, it came out at around 10 pounds (4.5 kg). Which is not bad for an 11' boat

Its been a fun boat to build, to keep the weight down I had to use the clearest straight red cedar I could find. The stems and centre thwart are red cedar, and the ribs are local ash wood I got from Drew Beatie, who runs a portable bandsaw-mill.

For those wondering how you build something like this, read on.

This is all the cedar stock, I start by running it through my bandsaw to rip it down to size. I dont have a jointer or a thickness planer so I hand plane all the pieces down smooth. I also dont have a carpenters bench so I use a 2x6 stuck in a workmate and resting on a saw horse for a planing bench. It works great.

This is the creation of the stem pieces (the curved bits at the bow and stern of a canoe) They need to be really strong as they are what you tend to run into things with. I took a piece of yellow cedar, roughly 2.5x3cm and ripped it into thirds. The pieces were then steamed, and bent over a form together.

They hold their shape pretty well once steamed.

I then glued them together with PL Premium construction adhesive. Use a lot of clamps.

The end then gets scarfed into the keel stringer. Most sane people use a table saw to do this. I don't trust mine, so I use a sliding miter saw. Cut a piece of scrap at 45 degrees, and clamp your piece to that. With a bit of futzing on the set up you can get a really accurate 8:1 scarf cut. Just be advised the the offcuts will launch, fortunately away from you.

Installing the stems onto the keel stringer. It would probably be wise to drill a couple dowels to keep the two parts from sliding while you glue them. Learning is fun.

These pieces are the breasthooks for the bow and stern. They attach the ends of the gunwales to the stems. To make them use a sliding bevel gauge to eyeball the angle that the gunwales meet the stem at, average all your measurements and take your best guess and set the miter saw to that.

I tape them up so I dont have to clean off the glue, and then just glue and peg them in place.

This is how you rip long stock in a very small and messy shop. I set up my bandsaw in the middle of the garage so then I can use my miter saw stand as an out-feed. It works better than those roller stands as you cant tip a mitersaw stand over easily. Here I am ripping thin stock off a big hunk of green ash for the ribs.

Once ripped, its back to planing, by hand. Someone finally told me to stick a block of wood at the end of the bench for the piece being planed to rest against. Previously I was clamping the wood down. The block of wood makes life so much easier.

Here is the hull blocked up with temporary ribs. These were left over from a previous experiment at making gluelam ribs. I learned that if you overclamp epoxied joints you squeeze all the glue out. Resulting in a useless rib for a functioning boat, but it works great as a form.

The ribs were steamed and shoved in next to the temporary frame. This was my first experience steaming green wood, and hallelujah I have seen the light. After 15 minutes in the steamer you can pull these out and bend them into a 3 inch radius with no problems at all. And this is 3/8" thick ash wood.

The hardest part of building this boat was working out all the little details, so here is a few shots showing some of the more tricky bits and how I worked them out.
First, this is where the stringers meet the stems. I cut them off at an angle to match the surface. Then tied a couple constrictor knots around the stringers. These keep the lashing from sliding off the front, I then wrapped around both stringers, made a couple frapping turns to secure it, then did a couple loops around the stems.

This is the mini thwart at the bow and stern. I added this piece in because I felt the boat needed a little more wood to keep the gunwales from spreading. Its a simple Y lashing through a hole drilled in the thwart. You can see how the inwale and outwale are also secured with a lashing to hold them together.

The centre thwart is lashed in much the same way. I added in a block of ash between the thwart and gunwale to help spread the load, red cedar is really soft wood. This joint isn't pegged or glued, the lashing holds it in place just fine.

Shot of the final stem piece. The breasthook is glued to the gunwales, but only lashed to the stem. I wasnt sure what to do with the inwales so I just cut them off and lashed them to the outwale.

Stem from a profile view, you can see the scarf joint on the keel stringer. I should probably put a couple lashings on it to secure it.

The complete frame.

For you detail nuts, the dimensions of all the pieces are as follows: (all units cm)
Gunwales: inwale: 1.1 X 4 outwale:1.3 x 4
Stringers: 1.5 x 1.7
Keel stringer: 2.3 x 2
Floorboards: 4 x 0.9
Ribs: 2.6 x 0.9
Stems:2.4 x 2.2
Centre thwart: 3.3 x 2.2 in the centre tapering to 3 x 1.7 at the ends.
End thwarts: 1.1 x 1.77

The overall proportions of the boat are 11' long and 28" wide, its 13" deep amidships and is slightly wider aft of centre.