Solenoidbot
Nat Taylor and Basar Gulcu


© Society of Robots
     A solenoid is a coil of wire that creates a uniform magnetic field in the direction of the axis of the coils by running a current through the coil. A pull type solenoid actuator has a magnetic rod inside this magnetic field, which is pulled by magnetism. A spring pushes the rod back to its original position when the current stops. In this project we're going to use solenoids as actuators for our quadruped, "SolenoidBot," to achieve fast motion in a straight line on a flat surface. One of the problems with servo-bots is speed. With solenoids we can move faster than servo’s. We can always have three points of touch while staying still. We chose legs over wheels to make the robot more versatile in terms of what surfaces it can operate on. While being statically stable, the quadruped will rely on dynamic stability when moving at faster gaits. The legs are very simple swing legs with two degrees of freedom. They swing laterally, and change length. Motion is achieved by the swinging of four legs. Instead of joints like human legs, the robot has retractable feet. When the leg is swinging, the foot raises 1" so that it is off the ground for the full swing.


News

May

5/9/2007

     Triumph! CATASTROPHIC FAILURE!!!!
Progress is slow, but all the robots legs move and the sonar sensor works.
UPDATE: Unfortunately the Solenoidbot is fatally flawed. There are severall problems with the design. I will try to explain them below.

  • We used pull type solenoids. That means the robot sits on springs. After the added weight of the 8 relays, the robot is too heavy for its own springs. We might be able to reduce the weight, but that is a limiting factor at the moment.
  • Weight plays another role. The solenoid plungers have to retract into the solenoid itself. There are three forces the pull type solenoid is acting against. The first is gravity, which is essentially negligible when it comes to the weight of the plunger itself (not the robot.) Second is the force of the spring, which is what the solenoid is built to overcome. Third is friction, which is causing CATASTROPHIC problems. The weight of the robot is causing the plungers to become out of alignment with the tight solenoid entry. Therefore it is taking too much power, or in some cases making it impossible for the solenoid to retract. This renders the robot functionless.
  • Third is the inability to pulse the solenoids at a frequency that could keep them partially plunged, which is what our plan to have the robot turn around relied on. Granted at the moment this is really the last of our concerns, it is still a hurtle we would have to overcome.
  • Another problem is the power consumption. The weight of the batteries required to power the robot is astronomical, which is why we kept it tethered. Even so, it seems to be drawing about 150 watts, nearly all of which is lost to heat, which explains why the things get so damned hot. The warning on the website seems to be a drastic understatement.
  • Finally is the counstruction materials, which would just have to be more rigid for the robot to work as intended. The mylar plastic flexes too much in some cases.

5/8/2007

Finally got the danged sonar sensor to work. Note to self: When mounting an electronic device, make sure metal contacts are not in contact with any other metal. This may seem fairly elementary, but it was a very easy thing to over look and very difficult thing to trouble shoot. Anyway, I finally got that work and am very happy.

5/7/2007

Continued work on the program. Used Pf. Parker’s trick to make the chip output to multiple pins at one time. The robot mimics a quadruped gate while it rests on its stand. In other words, its flying! Ha.

5/6/2007

Got all of the relays and solenoids hooked up, and proceeded to test each one individually. One of the relays was malfunctioning--it was determined to be irreparable and its solenoid was attached to the neighboring relay which will risk burning out another relay by putting too much power through it. Oh well.

April

4/30/2007

Acquired connectors at JANs. Made circuit diagram. Soldered half of the relays.

4/29/2007

Practiced soldering. Checked availability of connectors are RadioShack.

4/27/2007

Relays mounted to board. Informed by US Customs that other set of relays is being held. Refund requested.

4/26/2007

Confirmed functionality of relays

4/25/2007

Relays finally arrive!

4/23/2007

Relays arrive at JANs. Tested them back at the lab. Wrong relays! Reordered.

4/17/2007

Researched relays. Checked stock at JANs. Ordered relays.
Relays are electrically operated mechanical switches. They come in many shapes and sizes, but we will need a SPDT relay, which has 5 terminals. 2 to control the relay, one common, one always open circuit and one always closed circuit. The stamp will be hooked up to the control terminals and the solenoid to the always open terminal. When the stamp outputs to a given relay, it will flip its switch and activate the relay.

4/16/2007

Today we decided that the robot will walk and when it gets close to something turn around and then walk again. Hopefully it will be very quick. To accomplish this we will use a sonar sensor while will give distance information to the robot.
We have also discovered that a basic stamp outputs about 5V which will not be enough to power the solenoids. We will have to use relays.

March

3/28/2007

The robot is now fully assembled and ready to have the basic stamp and accompanying wires soldered. See photos.

3/7/2007

After several failed attempts that resulted from exporting problems, we finally have the robot cut out.

3/1/2007

Officially designed the robot. We wanted to maximize the speed of the robot, which would be achieved by covering a large distance for every step. Design criteria were the following:

  • Maximum a foot can lift is 1"
  • Maximum swing of the upper leg is 1"
Next came some geometry. By moving the pivot location and varying the leg length different swings for the foot can be achieved. However, the foot can only raise 1", so the maximum length along a bisecting radii for the swing between the arc and the chord was 1". This length is a function of swing size and leg length as well. We used a spreadsheet to optimize these lengths, while allowing for some noise.
Then we also added very conservative "emergency" points in case we have overestimated the solenoids or some other problem.

February

2/26/2007

We have officially hooked up a solenoid and test it out. Not surprisingly the plunger has to be all the way inside the solenoid for it to engage. The unfortunate result of the test is that a single solenoid is drawing about 30 watts. While is workable, the robot will probably have to be tethered.

2/22/2007

Today we created a foam board mock up to mimic the motion of the legs and also explore the relative sizes of it all.

2/14/2007

We have made the decision to switch to a quadruped. The goal of our project is to explore the use of solenoids as actuators, and we can test them better when used as a quadruped because it eliminates the need for balancing and ankle joints

2/8/2007

Today we decided to try to replicate biped locomotion with solenoids. Some of the challenges we anticipate are:

  • Balance
  • Weight
  • Mobility of ankle joints
  • Power supply