Pneumatic Test Stand

Our team is looking to set up a pneumatic “test stand”.

Objectives:

  • Gain familiarity with pneumatic components and function
  • Rapidly lash up pneumatic gizmos and test them out (grippers, lifters, etc.)
  • Measure performance and validate model/analysis, confirm margin
  • Optimize and validate designs BEFORE adopting for the robot design

We need to have some capability to measure pressure, force, position, etc. We will also need some data logging/analysis capability.

Has anyone done something like this that they could share?

Is anyone interested in collaborating on this project?

We have a small pneumatics board that we use for testing pneumatic mechanisms. It doesn’t have any testing/analysis capabilities though. You can calculate the force using the air pressure in and bore size, unless you just want to test if the cylinders are functioning correctly.

Update 20 May:
I have completed the basic "Air production, storage and delivery modules (see attached photo):

  • Power supply from an old computer
  • Three air tanks: you can shut off two with a valve
  • Three double solenoids (one 24 v, two 12 v) to power up to six actuators
  • Manual push buttons to actuate each solenoid
  • Quick disconnect for additional loads

Can you guess what my control panel is made from?

Still to do:

  • I built a “Lash up board” to quickly try out actuators: Horizontal and vertical peg boards
  • I need some sort of “tinker toy” parts quick to assemble arms, gears, cams, grippers, levers, etc.
  • I need to have alternate inputs to allow solenoid actuation from the robot cRIO
  • I have charged my son (in Electrical Engineering program) to come up with pressure and position transducers, force load cells, and a PC-based data recording system so we can quantify performance and compare designs.

If anyone has good ideas or would like to team up, please let me know!





Question: Where is your regulator? I see the pressure relief valve, attached to the compressor, with a pressure gauge. I presume that one reason for making this system is to mimic a pneumatic system on a FIRST robot.

Suggestion: Make the “stored” pressure (~115 PSIG) one color of tubing, and the working pressure (<60 PSIG) another color. While certainly not required, it can be a good feature when teaching students.

Also, I typically rig up a “pneumatic tree” around the regulator. It has a pressure gauge for the stored pressure, the regulator, and a pressure gauge for the stored pressure attached to each other with brass fittings. (Use double male brass fittings and “T” brass fittings.) This also makes the system easier to explain to the students.

I think you are likely going to run into problems powering the compressor with that power supply. The compressor can draw ~20amps on start up and ~10amps when running.

I’m also not seeing any sort of relay to control the compressor. The FIRST legal pressure switch can’t handle that many amps. You can use a common automotive relay to control it. wire the pressure sw inline with the relay’s coil and you’ll be good to go. Look for a universal relay often sold for driving/fog lights, they are typically rated for at least 20a and many are rated for 30a or 40a.

+1 on this!

Color coding can save you a huge amount of time when diagnosing a problem.

Personally, I would not only color code between stored pressure and working pressure, but between actuators too. (the tubing from one solenoid would be green, next one blue, etc.) That is, if you have that many colors of tubing :smiley:

Good call on the regulator. I added it in so the solenoids are at the working pressure of 60 PSI as required. I will experiment with the other suggestions. I wanted a nice “layout” where everything is visible. However, there will undoubtably be line losses that will drive us to us a more compact arrangement. This will be part of the learning experience.

AG

Thanks for the suggestion. I started simple, with a separate manual switch to shut off the pump when at pressure. The next step is definitely to add a relay connected to the switch.

AG

Gotta love that improvisation!

Nice work all around.

Have you considered putting something over the air tanks in case of shattering?

Avrum,

A few more tips:

  1. You should have minimal line losses. If the pressure drops too fast after you shut off the compressor, check for leaks. Put a little liquid dish soap in a cup of water, and spread the soapy water on the joints. Bubbles will appear at the leaks.
    a) For brass-brass joints, snug them up a bit more. CAREFUL: The same quality (malleability) that makes brass good at sealing joints, makes them susceptible to ruining them through catastrophic deformation. Happens when you twist too hard with steel wrenches.
    b) For plastic line joints, be sure the lines are cut clean and square, and inserted snugly into the quick-disconnects.
    c) For the M7 fittingat the solenoids, be sure to use an allen wrench in the center of the quick disconnect. (AM website: “Fastening method: Internal hex (use allen wrench)”)
    d) At any threaded joint, you may need to use small strips of teflon tape. Be very parsimonious on using teflon tape, as pieces can break off and get into the valves.
  2. Recommend that you build the pneumatic systemas close to the FIRST system as you can, including the compressor control system with pressure switch and relay hooked to a cRIO. This can give you the experience of building up a whole system to spec now, and you won’t have to learn it during the pressure of build season. If you can, work with a local team to borrow a few parts.

The blue “control panel” was cut from a plastic dustpan! It had the right size, and I liked the design: color, tilted surface. I shelled out for the $4 version rather than the dollar store one for added stiffness!

Kidding aside, I really want all aspects to have a “step by step” approach for ease of trouble-shooting, use, and educational value.

I see a three step approach:

  1. Simple: Push the button and watch the piston actuate.
  2. Moderate: Program the robot controller to send the signal to activate the piston.
  3. Advanced: Transducers to measure air pressure, position and force vs time, with data logging so we can analyse the mechanism for force/timing margins, and correlate with the design/analysis objectives.

AG

I wasn’t aware that this could be a problem! I am definitely interested in safety. Would a plexi-glass sheet be sufficient?

AG

Jim,

Thanks for the tips. We have “thin” teflon tape, which is a god-send. I will definitely work my way up to connecting the cRIO for control.

AG

Read this:::safety::
http://www.usfirst.org/roboticsprograms/frc/blog-plastic-air-tanks-important-safety-notice

Also, there was a picture posted (somewhere) of the plastic tank remnants. I have been unable to locate said picture. Also, there were a few threads posted on CD, commenting on the FIRST safety post.

The safety bulletin suggests that the tank should not be stressed, and in our installation it is securely mounted to the board. However, As weight is not a problem on this test bench, I will build either a plywood or sheet metal cover plate.

Thanks once again for the safety tip!

Update 27 Oct 2013

  • I have added a separate power supply for the pump: the “on board” one powers the relays and instrumentation
  • Per above suggestions, an pressure switch drives an automotive relay to cycle the pump

To do:
-The “air production module” is now complete. I will put a Lexan screen over the plastic tanks per the above safety discussion.
-I am now working on the “test bench” module. This is where you can quickly lash up different mechanisms and try them out.

  • I need some sort of “tinker toy” parts quick to assemble arms, gears, cams, grippers, levers, etc.
  • I need to have alternate inputs to allow solenoid actuation from the robot cRIO
  • My son (in Electrical Engineering program) is developing a data analysis systemm with pressure and position transducers, force load cells, and a PC-based data recording system so we can quantify performance and compare designs.

If anyone has good ideas or would like to team up, please let me know!