Loading ...
Sorry, an error occurred while loading the content.

Re: [SeattleRobotics] testing robot

Expand Messages
  • David Buckley
    Peter Air fittings (and water fittings come to that) are rated to 10 bar ie 150psi, above that you are on your own. Fancy expensive compressors will still have
    Message 1 of 20 , May 16, 2013
    • 0 Attachment
      Peter
      Air fittings (and water fittings come to that) are rated to 10 bar ie 150psi, above that you are on your own.
      Fancy expensive compressors will still have trouble delivering 10 bar - as it's the limit for fittings.
      If the ram never leaves the thing you are testing you can avoid a shock load.
      A DIY air compressor will do and you will maybe get 8 bar out of it, it will be noisy. A Hydrovane will not be as noisy but will be very expensive.
      Industry standard is 24v for valves, you will not need massive valves, ones with 1/8 BSP fittings will do.
      You will need three port valves (inlet, outlet and exhaust)
      You can probably use 4mm air hose, it will slow things down a little. 6mm hose will almost certainly be fine.
      Get a PLC programmable in a high level language - forget ladder logic.
      Or choose you favourite micro and hook up a ULN2003 or similar to drive the valves, don't forget the back emf diodes!.
      Important - use Adjustable Air-restrictors to reduce the air-flow to minimise impact loads.
      100Kg (220lb) you are going to need at least a 2" dia (5cm) cylinder, you can get short stroke ones.
      1000Kg use a lever arm with a >100mm stroke cylinder - you could use the same cylinder.
      DAvid
       
       
       
      ----- Original Message -----
      Sent: Thursday, May 16, 2013 3:50 PM
      Subject: [SeattleRobotics] testing robot

       

      I want to fatigue-test some load cells. For instance, apply 100kg for 1
      second 50,000 times and apply 1000kg 1000 times; total travel 10mm.

      In our local Ikea store, there's a demonstration device that uses pneumatics
      to open and close a cupboard door all day long. That's the sort of thing we
      need.
      http://www.youtube.com/watch?v=fPTyGGGaWq8
      (OK, I know that's a demo not a real test.)

      I can find lots of pneumatic stuff in catalogs, e.g.
      http://uk.rs-online.com/web/c/pneumatics-hydraulics-power-transmission/pneumatic-cylinders-actuators/pneumatic-compact-cylinders/

      But it's very hard to get started. What should we buy? What pressures do we
      use? How do we do the sequencing: via electric solenoid valves or pneumatic
      logic?

      What I need is a beginner's guide. Is there a website or a book? A DIY
      explanation?

      Peter

    • Peter Balch
      ... Perhaps it does. That s one suggestion I ve had from my wife s cousin who has done exactly that for small engineering firms. Once again, I d probably use
      Message 2 of 20 , May 16, 2013
      • 0 Attachment
        > Wouldn't this be a good application for a cam mounted on a gear-motor with
        > a
        > spring to moderate the force? Then for the higher force, just substitute
        > a
        > heavier spring. This provides the advantage of not having high impact
        > loads
        > that could result from a pneumatically-actuated cylinder, plus you don't
        > need a control circuit, solenoids, valves, etc. to cycle the mechanism.

        Perhaps it does. That's one suggestion I've had from my wife's cousin who
        has done exactly that for small engineering firms.

        Once again, I'd probably use the crank or cam to lift the fence post rather
        than apply the force directly.

        When I start looking at gear-motors and cranks that can manage 100kg over
        10cm, I get into foreign territory. How many h.p. is that? How
        over-engineered must it be. You generally buy the motors and gearboxes
        separately but if they're from the same "series" then they bolt together.
        What gear ratio? Well the gear ratio depends on the motor speed which
        depends on the torque and the torque depends on the gear ratio. Hmmm. It's
        not obvious what to buy.

        Then I'd need to get the crank or cam made. What length do we need this
        week, or next month?And the engineering for the test rig becomes more
        complex.

        I can see a lot of advantages to pneumatics. It fits togther like Lego or
        Meccano.

        All through this project we've realised that we don't know what we're doing.
        And even at the testing stage, I think we can be pretty sure that we don't
        know what we want to test. OK, it's clear this month - but next month, and
        the month after ...

        What I'd like is to make a test rig bolted together out of Dexion with a
        piston (or two or three) we can move from place to place. That way, we can
        rearrange the parts next month to do a different job.

        People have said "why not get a local engineering firm to make a such and
        such". But that's what I'd like to avoid. We'd have to give them drawings
        then wait days for them to make it. All of which implies that we actually
        know what we're doing. Bolt the Dexion together ... oh, no, that part needs
        to be much longer ...Easy.

        Similarly "why don't you get a local university engineering department to do
        the fatigue testing?". Well, I'm pretty sure that half way into it, we'd
        realise we'd specified the wong test or we were collecting the wrong data.
        If we can keep it all in-house, we can recover more quickly from our
        incompetence.

        >> Wouldn't this be a good application for a cam mounted on a gear-motor
        > That's brilliant! Extra points for KISS.

        Except it isn't that much simpler. Not once you start looking into the
        strength and speed of the motor, getting the crank made, mounting it all,
        etc. With a piston, it's already a linear actuator and the "power" part of
        the motor is somewhere else.

        > I'll bet good money 50K cycles with high force peaks
        > will break something in your rig

        Yes, probably. But that applies equally well to a gear-motor system.

        I've tried to find the plastic-deformation force on a 50cm beam of Dexion
        and it doesn't seem to be given anywhere. But applying what I remember from
        high-school physics stress/strain implies that a Dexion box would be
        over-engineered by about 20-to-1.

        > Wouldn't a cam + motor + spring be dramatically cheaper, too?

        Somewhat. But heavy duty motors and gearboxes aren't cheap and then we'd
        have to pay an engineering firm to make whatever converts rotary to linear
        motion.

        As far as I can see, pneumatics is more like Lego than engineering. Lego is
        good when you don't know what you want.

        Peter
      • phoneguin
        Sounds like a windshield wiper motor would be better suited. Instead of pushing windshield wipers, have it push your fence post up and down. Much easier than
        Message 3 of 20 , May 16, 2013
        • 0 Attachment
          Sounds like a windshield wiper motor would be better suited. Instead of
          pushing windshield wipers, have it push your fence post up and down. Much
          easier than pneumatics.

          Ed

          > Hi Chuck
          >
          >> I'd recommend you stick with ordinary "shop air" pressure ~100 psi where
          >> everyday valves &fittings are happy.
          >
          > OK.
          >
          > That's about 7 bar right? 7 atmopheres. 0.7MPa.
          >
          > I wasn't sure whether I should be sticking to 3 bar.
          >
          >> F=P*A tells you you need a large cylinder dia
          >
          > Maybe.
          >
          > At the moment, our "press" has a large fence-post. It's hinged at one end
          > with the hinge clamped onto a heavy-duty workbench. The load-cell is
          > placed
          > near the hinge and various weights are hung on the far end. The moment
          > ratio
          > is 10-to-1. So we can hang a 100kg weight on it to put 1 tonne on the
          > load-cell.
          >
          > My idea was a piston that raised/lowered the far end of the fence post by
          > 10cm. 100kg over 10cm is a fairly modest piston. The force on the
          > load-cell
          > depends on the weights we hang and so is independent of the piston, air
          > pressure, whatever.
          >
          > 100kg is 1kN. If the pressure is 0.7MPa then that's 14 sq cm (unless I've
          > slipped a decimal point). which is 90mm diameter.
          >
          > Here's a typical cylinder (from a rather expensive supplier)
          > http://uk.rs-online.com/web/p/pneumatic-compact-cylinders/7035577/
          >
          > 14 sq cm times 100mm stroke is 0.14 litres at, say, 0.5Hz is 4 litres per
          > minute which sounds like a pretty small compressor.
          >
          >> cylinder dia for 1000 kgf, I suggest you look at an "air spring"
          >> actuator.
          >> http://www.eskimo.com/~wfd/Catalogs/Firestone/EMDG.pdf
          >
          > I wondered about bladders. What are the advantages and disadvantages?
          >
          >> If a "pneumatics" search isn't finding the tutorials you want, you could
          >> also try "fluid power".
          >
          > A "pneumatics tutorials" search does find tutorials but they're not at the
          > level I need. For instance, these two are probably excellent
          >
          > http://www.digital-circuitry.com/DOC/ELECTRONICS/Manuals/Pneumatics/Course%20Books/Text-Pneumatics%20Basic.pdf
          > http://www.fluiddraw.de/fluidsim/download/v3/hb-eng-p.pdf
          >
          > But I don't have time to learn all about pneumatics. I just need the one
          > circuit: lift - wait - relax - wait - repeat. It seems like it must be one
          > of the most widely used circuits.
          >
          > But I look at the circuit diagrams and they're full of symbols that I
          > don't
          > understand.
          >
          > If I were trying to do something similar with electronics I could find a
          > hundred examples of NE555 circuits on the web which would show me exactly
          > what to do. I wouldn't need to start from stratch with a basic electronics
          > book and learn how a transistor works then design the circuit myself.
          >
          > I've tried finding a consultant we could hire: on Monday, "these are the
          > components you must buy"; on Friday "I'll come in a plug them together for
          > you". Such people probably do exist but I can't find one.
          >
          > It's all quite frustrating and a bit puzzling. I don't understand why it's
          > so difficult.
          >
          > Thanks
          >
          > Peter
          >
          >
          >
          > ------------------------------------
          >
          > Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
          >
          >
          >
        • twcarroll@...
          Peter, From my past experience with aerospace testing, I would go with a pneumatic cylinder set-up. a short 6 diameter cylinder has a 28.26 sq. in. surface
          Message 4 of 20 , May 16, 2013
          • 0 Attachment
            Peter,
                 From my past experience with aerospace testing, I would go with a pneumatic cylinder set-up.  a short 6" diameter cylinder has a 28.26 sq. in. surface and a 100 PSI supply would give you over 1000 kg. (2200 lb.)  Pressure regulated from 2826 pounds down to 2200 pounds of force is easy. A simple pressure gauge could let you read the pressure strokes. One solenoid valve would compress the cylinder and a second would release the pressure.  It only has to travel less than a half inch.  Cycling back and forth would take a simple microcontroller and I would use a mechanical counter instead of the microcontroller, just for safety in remembering the cycles.  A couple of hundred dollars could do it, plus building the framework.  A much smaller cylinder would handle the 220 pound test.
                 A windshield wiper motor would not have the strength and I don't think a cam would last very well.
                 Good Luck,
                 Tom C.
             
            In a message dated 5/16/2013 2:12:38 P.M. Pacific Daylight Time, ed@... writes:
            Sounds like a windshield wiper motor would be better suited.  Instead of
            pushing windshield wipers, have it push your fence post up and down.  Much
            easier than pneumatics.

            Ed

            > Hi Chuck
            >
            >> I'd recommend you stick with ordinary "shop air" pressure ~100 psi where
            >> everyday valves &fittings are happy.
            >
            > OK.
            >
            > That's about 7 bar right? 7 atmopheres. 0.7MPa.
            >
            > I wasn't sure whether I should be sticking to 3 bar.
            >
            >> F=P*A tells you you need a large cylinder dia
            >
            > Maybe.
            >
            > At the moment, our "press" has a large fence-post. It's hinged at one end
            > with the hinge clamped onto a heavy-duty workbench. The load-cell is
            > placed
            > near the hinge and various weights are hung on the far end. The moment
            > ratio
            > is 10-to-1. So we can hang a 100kg weight on it to put 1 tonne on the
            > load-cell.
            >
            > My idea was a piston that raised/lowered the far end of the fence post by
            > 10cm. 100kg over 10cm is a fairly modest piston. The force on the
            > load-cell
            > depends on the weights we hang and so is independent of the piston, air
            > pressure, whatever.
            >
            > 100kg is 1kN. If the pressure is 0.7MPa then that's 14 sq cm (unless I've
            > slipped a decimal point). which is 90mm diameter.
            >
            > Here's a typical cylinder (from a rather expensive supplier)
            > http://uk.rs-online.com/web/p/pneumatic-compact-cylinders/7035577/
            >
            > 14 sq cm times 100mm stroke is 0.14 litres at, say, 0.5Hz is 4 litres per
            > minute which sounds like a pretty small compressor.
            >
            >> cylinder dia for 1000 kgf, I suggest you look at an "air spring"
            >> actuator.
            >> http://www.eskimo.com/~wfd/Catalogs/Firestone/EMDG.pdf
            >
            > I wondered about bladders. What are the advantages and disadvantages?
            >
            >> If a "pneumatics" search isn't finding the tutorials you want, you could
            >> also try "fluid power".
            >
            > A "pneumatics tutorials" search does find tutorials but they're not at the
            > level I need. For instance, these two are probably excellent
            >
            > http://www.digital-circuitry.com/DOC/ELECTRONICS/Manuals/Pneumatics/Course%20Books/Text-Pneumatics%20Basic.pdf
            > http://www.fluiddraw.de/fluidsim/download/v3/hb-eng-p.pdf
            >
            > But I don't have time to learn all about pneumatics. I just need the one
            > circuit: lift - wait - relax - wait - repeat. It seems like it must be one
            > of the most widely used circuits.
            >
            > But I look at the circuit diagrams and they're full of symbols that I
            > don't
            > understand.
            >
            > If I were trying to do something similar with electronics I could find a
            > hundred examples of NE555 circuits on the web which would show me exactly
            > what to do. I wouldn't need to start from stratch with a basic electronics
            > book and learn how a transistor works then design the circuit myself.
            >
            > I've tried finding a consultant we could hire: on Monday, "these are the
            > components you must buy"; on Friday "I'll come in a plug them together for
            > you". Such people probably do exist but I can't find one.
            >
            > It's all quite frustrating and a bit puzzling. I don't understand why it's
            > so difficult.
            >
            > Thanks
            >
            > Peter
            >
            >
            >
            > ------------------------------------
            >
            > Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
            >
            >
            >




            ------------------------------------

            Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links

            <*> To visit your group on the web, go to:
                http://groups.yahoo.com/group/SeattleRobotics/

            <*> Your email settings:
                Individual Email | Traditional

            <*> To change settings online go to:
                http://groups.yahoo.com/group/SeattleRobotics/join
                (Yahoo! ID required)

            <*> To change settings via email:
                SeattleRobotics-digest@yahoogroups.com
                SeattleRobotics-fullfeatured@yahoogroups.com

            <*> To unsubscribe from this group, send an email to:
                SeattleRobotics-unsubscribe@yahoogroups.com

            <*> Your use of Yahoo! Groups is subject to:
                http://docs.yahoo.com/info/terms/

          • Kevin Ross
            I also agree that pneumatics is going to be the easiest way. You will need a fairly large diameter piston to generate 1000kg of force using off the shelf air
            Message 5 of 20 , May 16, 2013
            • 0 Attachment
              I also agree that pneumatics is going to be the easiest way. You will need a
              fairly large diameter piston to generate 1000kg of force using off the shelf
              air fittings. You are also likely to need a larger compressor than you are
              thinking. At 50,000 cycles for 1 second, you will need to cycle twice
              (compress/retract each take a cylinder load of air). Compressors have a
              series of different ratings you need to consider, most important is the
              expected duty cycle. They will get very hot if you exceed their duty cycle.
              I would expect this to take a larger (60 or 80 gallon) unit to keep the
              thing from running every couple of minutes.

              You could use the same large cylinder and have two regulators and two
              solenoids. One regulator might run at 10psi to generate 220lbs of force, the
              other would run at 100psi to generate 2200 lbs of force. That would make
              this mechanically much more simple. At 100 psi you are going to need a 5.75"
              or larger piston, so 6" it is OR you can use a multiple chamber unit like
              this
              http://www.bimba.com/Products-and-Cad/Actuators/Inch/Compact/Low-Profile/Flat-One-Multiple-Power/
              assuming you don't need full power to retract.



              -----Original Message-----
              From: Peter Balch
              Sent: Thursday, May 16, 2013 7:50 AM
              To: SeattleRobotics@yahoogroups.com
              Subject: [SeattleRobotics] testing robot

              I want to fatigue-test some load cells. For instance, apply 100kg for 1
              second 50,000 times and apply 1000kg 1000 times; total travel 10mm.

              In our local Ikea store, there's a demonstration device that uses pneumatics
              to open and close a cupboard door all day long. That's the sort of thing we
              need.
              http://www.youtube.com/watch?v=fPTyGGGaWq8
              (OK, I know that's a demo not a real test.)

              I can find lots of pneumatic stuff in catalogs, e.g.
              http://uk.rs-online.com/web/c/pneumatics-hydraulics-power-transmission/pneumatic-cylinders-actuators/pneumatic-compact-cylinders/

              But it's very hard to get started. What should we buy? What pressures do we
              use? How do we do the sequencing: via electric solenoid valves or pneumatic
              logic?

              What I need is a beginner's guide. Is there a website or a book? A DIY
              explanation?

              Peter





              ------------------------------------

              Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
            • Charlie H
              Peter, Have you thought about using a motorized automotive scissor jack? They have plenty of power and and the linear range is greater than you need. You would
              Message 6 of 20 , May 17, 2013
              • 0 Attachment
                Peter,

                Have you thought about using a motorized automotive scissor jack?  They have plenty of power and and the linear range is greater than you need.  You would need to drive the motor (typically 12 vdc) with some electronics, of course.  I've seen these online for < $100.00.

                Charlie


                On 5/16/2013 2:10 PM, Peter Balch wrote:
                 

                > Wouldn't this be a good application for a cam mounted on a gear-motor with
                > a
                > spring to moderate the force? Then for the higher force, just substitute
                > a
                > heavier spring. This provides the advantage of not having high impact
                > loads
                > that could result from a pneumatically-actuated cylinder, plus you don't
                > need a control circuit, solenoids, valves, etc. to cycle the mechanism.

                Perhaps it does. That's one suggestion I've had from my wife's cousin who
                has done exactly that for small engineering firms.

                Once again, I'd probably use the crank or cam to lift the fence post rather
                than apply the force directly.

                When I start looking at gear-motors and cranks that can manage 100kg over
                10cm, I get into foreign territory. How many h.p. is that? How
                over-engineered must it be. You generally buy the motors and gearboxes
                separately but if they're from the same "series" then they bolt together.
                What gear ratio? Well the gear ratio depends on the motor speed which
                depends on the torque and the torque depends on the gear ratio. Hmmm. It's
                not obvious what to buy.

                Then I'd need to get the crank or cam made. What length do we need this
                week, or next month?And the engineering for the test rig becomes more
                complex.

                I can see a lot of advantages to pneumatics. It fits togther like Lego or
                Meccano.

                All through this project we've realised that we don't know what we're doing.
                And even at the testing stage, I think we can be pretty sure that we don't
                know what we want to test. OK, it's clear this month - but next month, and
                the month after ...

                What I'd like is to make a test rig bolted together out of Dexion with a
                piston (or two or three) we can move from place to place. That way, we can
                rearrange the parts next month to do a different job.

                People have said "why not get a local engineering firm to make a such and
                such". But that's what I'd like to avoid. We'd have to give them drawings
                then wait days for them to make it. All of which implies that we actually
                know what we're doing. Bolt the Dexion together ... oh, no, that part needs
                to be much longer ...Easy.

                Similarly "why don't you get a local university engineering department to do
                the fatigue testing?". Well, I'm pretty sure that half way into it, we'd
                realise we'd specified the wong test or we were collecting the wrong data.
                If we can keep it all in-house, we can recover more quickly from our
                incompetence.

                >> Wouldn't this be a good application for a cam mounted on a gear-motor
                > That's brilliant! Extra points for KISS.

                Except it isn't that much simpler. Not once you start looking into the
                strength and speed of the motor, getting the crank made, mounting it all,
                etc. With a piston, it's already a linear actuator and the "power" part of
                the motor is somewhere else.

                > I'll bet good money 50K cycles with high force peaks
                > will break something in your rig

                Yes, probably. But that applies equally well to a gear-motor system.

                I've tried to find the plastic-deformation force on a 50cm beam of Dexion
                and it doesn't seem to be given anywhere. But applying what I remember from
                high-school physics stress/strain implies that a Dexion box would be
                over-engineered by about 20-to-1.

                > Wouldn't a cam + motor + spring be dramatically cheaper, too?

                Somewhat. But heavy duty motors and gearboxes aren't cheap and then we'd
                have to pay an engineering firm to make whatever converts rotary to linear
                motion.

                As far as I can see, pneumatics is more like Lego than engineering. Lego is
                good when you don't know what you want.

                Peter


              • Peter Balch
                Charlie ... I think that it would be too slow for this particular project. They seem to take tens of seconds to extend. But I could imagine it being a useful
                Message 7 of 20 , May 17, 2013
                • 0 Attachment
                  Charlie

                  > Have you thought about using a motorized automotive scissor jack? They
                  > have plenty of power and and the linear range is greater than you need.

                  I think that it would be too slow for this particular project. They seem to
                  take tens of seconds to extend. But I could imagine it being a useful device
                  elsewhere.

                  I wonder what their design-life is. Probably not 50,000 lifts. And probably
                  not continuous operation.

                  Ed

                  > Sounds like a windshield wiper motor would be better suited.

                  I wonder what the torque is - wiper blades are hard to move by hand. And
                  they can opereate continuously for many hundreds of thousands of cycles;
                  they have to, particularly in Seattle and here in Scotland.

                  Peter
                • Peter Balch
                  Kevin ... I was expecting to use a 10-to-1 lever so the piston has to lift 100kg over a distance of 10cm. ... Maybe. But the piston would be under load (100kg)
                  Message 8 of 20 , May 17, 2013
                  • 0 Attachment
                    Kevin

                    >You will need a
                    > fairly large diameter piston to generate 1000kg of force using off the
                    > shelf
                    > air fittings.

                    I was expecting to use a 10-to-1 lever so the piston has to lift 100kg over
                    a distance of 10cm.

                    > You are also likely to need a larger compressor than you are
                    > thinking. At 50,000 cycles for 1 second, you will need to cycle twice
                    > (compress/retract each take a cylinder load of air).

                    Maybe. But the piston would be under load (100kg) so it could return
                    passively.

                    > Compressors have a
                    > series of different ratings you need to consider, most important is the
                    > expected duty cycle. They will get very hot if you exceed their duty
                    > cycle.

                    I hadn't considered that.

                    At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                    is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                    cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                    is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                    Would that be enough?

                    > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                    > thing from running every couple of minutes.

                    That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                    > You could use the same large cylinder and have two regulators and two
                    > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                    > the
                    > other would run at 100psi to generate 2200 lbs of force.

                    Interesting idea. I was assuming that the cylinder lifted a lever which
                    passively dropped to provide the force. The force would depend on what
                    weight we hung on the lever.

                    You're suggesting that the cylinder provides the force directly (or via a
                    lever) and the force is determined by the air pressure.

                    How accurately can you control the force that way? Doesn't friction and
                    stiction affect the force? And how accurate is the air pressure?

                    It would be nice if we could use the same rig for calibration as well as
                    fatigue testing so a force accurate to a couple of percent would be good.

                    Peter
                  • ListServ
                    Clippard has a pretty complete product line, and some good documentation on pneumatics. Look at their knowledge base and catalog for what is possible.
                    Message 9 of 20 , May 17, 2013
                    • 0 Attachment
                      Clippard has a pretty complete product line, and some good documentation on pneumatics. Look at their knowledge base and catalog for what is possible.
                      http://www.clippard.com/

                      The symbols you see are likely pneumatic ansi symbols. A google search will describe them. If you can read a schematic, the pneumatic drawings should be understandable.



                      On Thu, May 16, 2013 at 8:50 AM, Peter Balch peterbalch-at-btinternet.com |Listserv| <lf4fhwmmot@...> wrote:
                       

                      I want to fatigue-test some load cells. For instance, apply 100kg for 1
                      second 50,000 times and apply 1000kg 1000 times; total travel 10mm.

                      In our local Ikea store, there's a demonstration device that uses pneumatics
                      to open and close a cupboard door all day long. That's the sort of thing we
                      need.
                      http://www.youtube.com/watch?v=fPTyGGGaWq8
                      (OK, I know that's a demo not a real test.)

                      I can find lots of pneumatic stuff in catalogs, e.g.
                      http://uk.rs-online.com/web/c/pneumatics-hydraulics-power-transmission/pneumatic-cylinders-actuators/pneumatic-compact-cylinders/

                      But it's very hard to get started. What should we buy? What pressures do we
                      use? How do we do the sequencing: via electric solenoid valves or pneumatic
                      logic?

                      What I need is a beginner's guide. Is there a website or a book? A DIY
                      explanation?

                      Peter


                    • Kevin Ross
                      Compressor duty cycles are often times specified in 10 minute intervals or in 1 hour intervals. You need to check with the specific model you are looking to
                      Message 10 of 20 , May 18, 2013
                      • 0 Attachment
                        Compressor duty cycles are often times specified in 10 minute intervals or
                        in 1 hour intervals. You need to check with the specific model you are
                        looking to use. A common value is 30% / 10 minutes, which means you can run
                        it for about 3 minutes then it needs to cool down for the next 7 minutes.
                        There are a lot of variables in this. You are also going to find compressors
                        which claim to have a 100% duty cycle. Be very wary of that claim, it isn't
                        actually true in most cases and especially at the pressure you are going to
                        need for your device. (We have 100% duty cycle compressors for use on FIRST
                        robots, but if you attempt to run them for too long they heat up and fail.
                        You have to let them cool down, meaning they are not really 100% duty cycle!
                        It is marketing BS.)

                        I see you have some of the math below. I don't really understand the details
                        for your design so I am not able to verify your numbers at the moment, but
                        you seem to get the idea. If you are generating 2200lbs of force with a 10cm
                        stroke, my guess is that you are going to use much more air than what you
                        have stated. I base this on a gut feeling from a guy who does a lot of
                        pneumatics on FIRST robots and watches the compressors heat up very quickly.
                        However, if you do the math, you get to tell me I am wrong and I won't feel
                        bad at all.

                        Pneumatics are a full of little variances. If you need more precision, you
                        should consider hydraulic. Air acts like a spring, and is full of all sorts
                        of surprises. It is not instant power, nor is it linear. It is, however,
                        easy to work with and pretty reliable once you get it setup. The regulators
                        do a good job at not allowing the pressure to exceed a certain amount, so
                        they are usually pretty accurate at limiting the pressure.

                        Friction will be an issue in any device you make. The fewer moving parts,
                        the fewer sources of friction.

                        -----Original Message-----
                        From: Peter Balch
                        Sent: Friday, May 17, 2013 11:43 AM
                        To: SeattleRobotics@yahoogroups.com
                        Subject: Re: [SeattleRobotics] testing robot


                        > Compressors have a
                        > series of different ratings you need to consider, most important is the
                        > expected duty cycle. They will get very hot if you exceed their duty
                        > cycle.

                        I hadn't considered that.

                        At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                        is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                        cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                        is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                        Would that be enough?

                        > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                        > thing from running every couple of minutes.

                        That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                        > You could use the same large cylinder and have two regulators and two
                        > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                        > the
                        > other would run at 100psi to generate 2200 lbs of force.

                        Interesting idea. I was assuming that the cylinder lifted a lever which
                        passively dropped to provide the force. The force would depend on what
                        weight we hung on the lever.

                        You're suggesting that the cylinder provides the force directly (or via a
                        lever) and the force is determined by the air pressure.

                        How accurately can you control the force that way? Doesn't friction and
                        stiction affect the force? And how accurate is the air pressure?

                        It would be nice if we could use the same rig for calibration as well as
                        fatigue testing so a force accurate to a couple of percent would be good.

                        Peter



                        ------------------------------------

                        Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
                      • Kevin Ross
                        Compressor duty cycles are often times specified in 10 minute intervals or in 1 hour intervals. You need to check with the specific model you are looking to
                        Message 11 of 20 , May 18, 2013
                        • 0 Attachment
                          Compressor duty cycles are often times specified in 10 minute intervals or
                          in 1 hour intervals. You need to check with the specific model you are
                          looking to use. A common value is 30% / 10 minutes, which means you can run
                          it for about 3 minutes then it needs to cool down for the next 7 minutes.
                          There are a lot of variables in this. You are also going to find compressors
                          which claim to have a 100% duty cycle. Be very wary of that claim, it isn't
                          actually true in most cases and especially at the pressure you are going to
                          need for your device. (We have 100% duty cycle compressors for use on FIRST
                          robots, but if you attempt to run them for too long they heat up and fail.
                          You have to let them cool down, meaning they are not really 100% duty cycle!
                          It is marketing BS.)

                          I see you have some of the math below. I don't really understand the details
                          for your design so I am not able to verify your numbers at the moment, but
                          you seem to get the idea. If you are generating 2200lbs of force with a 10cm
                          stroke, my guess is that you are going to use much more air than what you
                          have stated. I base this on a gut feeling from a guy who does a lot of
                          pneumatics on FIRST robots and watches the compressors heat up very quickly.
                          However, if you do the math, you get to tell me I am wrong and I won't feel
                          bad at all.

                          Pneumatics are a full of little variances. If you need more precision, you
                          should consider hydraulic. Air acts like a spring, and is full of all sorts
                          of surprises. It is not instant power, nor is it linear. It is, however,
                          easy to work with and pretty reliable once you get it setup. The regulators
                          do a good job at not allowing the pressure to exceed a certain amount, so
                          they are usually pretty accurate at limiting the pressure.

                          Friction will be an issue in any device you make. The fewer moving parts,
                          the fewer sources of friction.

                          -----Original Message-----
                          From: Peter Balch
                          Sent: Friday, May 17, 2013 11:43 AM
                          To: SeattleRobotics@yahoogroups.com
                          Subject: Re: [SeattleRobotics] testing robot


                          > Compressors have a
                          > series of different ratings you need to consider, most important is the
                          > expected duty cycle. They will get very hot if you exceed their duty
                          > cycle.

                          I hadn't considered that.

                          At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                          is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                          cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                          is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                          Would that be enough?

                          > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                          > thing from running every couple of minutes.

                          That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                          > You could use the same large cylinder and have two regulators and two
                          > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                          > the
                          > other would run at 100psi to generate 2200 lbs of force.

                          Interesting idea. I was assuming that the cylinder lifted a lever which
                          passively dropped to provide the force. The force would depend on what
                          weight we hung on the lever.

                          You're suggesting that the cylinder provides the force directly (or via a
                          lever) and the force is determined by the air pressure.

                          How accurately can you control the force that way? Doesn't friction and
                          stiction affect the force? And how accurate is the air pressure?

                          It would be nice if we could use the same rig for calibration as well as
                          fatigue testing so a force accurate to a couple of percent would be good.

                          Peter



                          ------------------------------------

                          Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
                        • Kevin Ross
                          Compressor duty cycles are often times specified in 10 minute intervals or in 1 hour intervals. You need to check with the specific model you are looking to
                          Message 12 of 20 , May 18, 2013
                          • 0 Attachment
                            Compressor duty cycles are often times specified in 10 minute intervals or
                            in 1 hour intervals. You need to check with the specific model you are
                            looking to use. A common value is 30% / 10 minutes, which means you can run
                            it for about 3 minutes then it needs to cool down for the next 7 minutes.
                            There are a lot of variables in this. You are also going to find compressors
                            which claim to have a 100% duty cycle. Be very wary of that claim, it isn't
                            actually true in most cases and especially at the pressure you are going to
                            need for your device. (We have 100% duty cycle compressors for use on FIRST
                            robots, but if you attempt to run them for too long they heat up and fail.
                            You have to let them cool down, meaning they are not really 100% duty cycle!
                            It is marketing BS.)

                            I see you have some of the math below. I don't really understand the details
                            for your design so I am not able to verify your numbers at the moment, but
                            you seem to get the idea. If you are generating 2200lbs of force with a 10cm
                            stroke, my guess is that you are going to use much more air than what you
                            have stated. I base this on a gut feeling from a guy who does a lot of
                            pneumatics on FIRST robots and watches the compressors heat up very quickly.
                            However, if you do the math, you get to tell me I am wrong and I won't feel
                            bad at all.

                            Pneumatics are a full of little variances. If you need more precision, you
                            should consider hydraulic. Air acts like a spring, and is full of all sorts
                            of surprises. It is not instant power, nor is it linear. It is, however,
                            easy to work with and pretty reliable once you get it setup. The regulators
                            do a good job at not allowing the pressure to exceed a certain amount, so
                            they are usually pretty accurate at limiting the pressure.

                            Friction will be an issue in any device you make. The fewer moving parts,
                            the fewer sources of friction.

                            -----Original Message-----
                            From: Peter Balch
                            Sent: Friday, May 17, 2013 11:43 AM
                            To: SeattleRobotics@yahoogroups.com
                            Subject: Re: [SeattleRobotics] testing robot


                            > Compressors have a
                            > series of different ratings you need to consider, most important is the
                            > expected duty cycle. They will get very hot if you exceed their duty
                            > cycle.

                            I hadn't considered that.

                            At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                            is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                            cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                            is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                            Would that be enough?

                            > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                            > thing from running every couple of minutes.

                            That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                            > You could use the same large cylinder and have two regulators and two
                            > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                            > the
                            > other would run at 100psi to generate 2200 lbs of force.

                            Interesting idea. I was assuming that the cylinder lifted a lever which
                            passively dropped to provide the force. The force would depend on what
                            weight we hung on the lever.

                            You're suggesting that the cylinder provides the force directly (or via a
                            lever) and the force is determined by the air pressure.

                            How accurately can you control the force that way? Doesn't friction and
                            stiction affect the force? And how accurate is the air pressure?

                            It would be nice if we could use the same rig for calibration as well as
                            fatigue testing so a force accurate to a couple of percent would be good.

                            Peter



                            ------------------------------------

                            Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links
                          • Peter Balch
                            Thanks Kevin. I think I m beginning to understand what we need to buy but the valves are still a mystery. If I want active retraction of the cylinder (i.e. a
                            Message 13 of 20 , May 19, 2013
                            • 0 Attachment
                              Thanks Kevin.

                              I think I'm beginning to understand what we need to buy but the valves are
                              still a mystery. If I want active retraction of the cylinder (i.e. a
                              double-acting cylinder) then maybe I'll need a 5-port 2-Way, 3-way or 4-way
                              solenoid valve. Or maybe a 5-port air-spring valve operated by two 3-port
                              2-way solenoid valves. Or maybe just use two 3-port 2-way solenoid valves.

                              When I look at the catalogues, there's an enormous range of valves to choose
                              from with no clear way of making a decision.

                              Then I look at the photos of the valves and realise the connectors for the
                              hoses need to be the right type. Are there different ranges? Which is bect
                              for what we want to do? Presumably, it depends on pressure and airflow. And
                              so on and so on.

                              I can see what I need in theory and I can see where to buy Stuff. But
                              there's still a big gap between those two.

                              Peter
                            • mark@kenworthymachine.com
                              Rotary screw type compressors are designed to run continuously. I don t know of any type designed for continuous service. Connected by DROID on Verizon
                              Message 14 of 20 , May 19, 2013
                              • 0 Attachment
                                Rotary screw type compressors are designed to run continuously.  I don't know of any type designed for continuous service.

                                Connected by DROID on Verizon Wireless


                                -----Original message-----
                                From: Kevin Ross <kevinro@...>
                                To:
                                SeattleRobotics@yahoogroups.com
                                Sent:
                                Sun, May 19, 2013 05:25:46 GMT+00:00
                                Subject:
                                Re: [SeattleRobotics] testing robot

                                 

                                Compressor duty cycles are often times specified in 10 minute intervals or
                                in 1 hour intervals. You need to check with the specific model you are
                                looking to use. A common value is 30% / 10 minutes, which means you can run
                                it for about 3 minutes then it needs to cool down for the next 7 minutes.
                                There are a lot of variables in this. You are also going to find compressors
                                which claim to have a 100% duty cycle. Be very wary of that claim, it isn't
                                actually true in most cases and especially at the pressure you are going to
                                need for your device. (We have 100% duty cycle compressors for use on FIRST
                                robots, but if you attempt to run them for too long they heat up and fail.
                                You have to let them cool down, meaning they are not really 100% duty cycle!
                                It is marketing BS.)

                                I see you have some of the math below. I don't really understand the details
                                for your design so I am not able to verify your numbers at the moment, but
                                you seem to get the idea. If you are generating 2200lbs of force with a 10cm
                                stroke, my guess is that you are going to use much more air than what you
                                have stated. I base this on a gut feeling from a guy who does a lot of
                                pneumatics on FIRST robots and watches the compressors heat up very quickly.
                                However, if you do the math, you get to tell me I am wrong and I won't feel
                                bad at all.

                                Pneumatics are a full of little variances. If you need more precision, you
                                should consider hydraulic. Air acts like a spring, and is full of all sorts
                                of surprises. It is not instant power, nor is it linear. It is, however,
                                easy to work with and pretty reliable once you get it setup. The regulators
                                do a good job at not allowing the pressure to exceed a certain amount, so
                                they are usually pretty accurate at limiting the pressure.

                                Friction will be an issue in any device you make. The fewer moving parts,
                                the fewer sources of friction.

                                -----Original Message-----
                                From: Peter Balch
                                Sent: Friday, May 17, 2013 11:43 AM
                                To: SeattleRobotics@yahoogroups.com
                                Subject: Re: [SeattleRobotics] testing robot

                                > Compressors have a
                                > series of different ratings you need to consider, most important is the
                                > expected duty cycle. They will get very hot if you exceed their duty
                                > cycle.

                                I hadn't considered that.

                                At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                                is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                                cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                                is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                                Would that be enough?

                                > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                                > thing from running every couple of minutes.

                                That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                                > You could use the same large cylinder and have two regulators and two
                                > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                                > the
                                > other would run at 100psi to generate 2200 lbs of force.

                                Interesting idea. I was assuming that the cylinder lifted a lever which
                                passively dropped to provide the force. The force would depend on what
                                weight we hung on the lever.

                                You're suggesting that the cylinder provides the force directly (or via a
                                lever) and the force is determined by the air pressure.

                                How accurately can you control the force that way? Doesn't friction and
                                stiction affect the force? And how accurate is the air pressure?

                                It would be nice if we could use the same rig for calibration as well as
                                fatigue testing so a force accurate to a couple of percent would be good.

                                Peter

                                ------------------------------------

                                Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links

                              • Mark Kenworthy
                                That should be “any other type that is rated for continuous service”. Our rotary screw compressor has two operational modes. One where it just runs
                                Message 15 of 20 , May 19, 2013
                                • 0 Attachment

                                  That should be “any other type that is rated for continuous service”.  Our rotary screw compressor has two operational modes.  One where it just runs continuous and bleeds off excess air, and another mode where it cycles off when reaching the preset pressure level.

                                   

                                  They are fairly expensive, as compressors go, so you might want to try to find a business that has one and work out a deal to do your testing there.

                                   

                                  Mark

                                   

                                  From: SeattleRobotics@yahoogroups.com [mailto:SeattleRobotics@yahoogroups.com] On Behalf Of mark@...
                                  Sent: Sunday, May 19, 2013 8:13 AM
                                  To: SeattleRobotics@yahoogroups.com
                                  Subject: Re: [SeattleRobotics] testing robot

                                   

                                   

                                  Rotary screw type compressors are designed to run continuously.  I don't know of any type designed for continuous service.

                                  Connected by DROID on Verizon Wireless



                                  -----Original message-----

                                  From: Kevin Ross <kevinro@...>
                                  To:
                                  SeattleRobotics@yahoogroups.com
                                  Sent:
                                  Sun, May 19, 2013 05:25:46 GMT+00:00
                                  Subject:
                                  Re: [SeattleRobotics] testing robot

                                   

                                  Compressor duty cycles are often times specified in 10 minute intervals or
                                  in 1 hour intervals. You need to check with the specific model you are
                                  looking to use. A common value is 30% / 10 minutes, which means you can run
                                  it for about 3 minutes then it needs to cool down for the next 7 minutes.
                                  There are a lot of variables in this. You are also going to find compressors
                                  which claim to have a 100% duty cycle. Be very wary of that claim, it isn't
                                  actually true in most cases and especially at the pressure you are going to
                                  need for your device. (We have 100% duty cycle compressors for use on FIRST
                                  robots, but if you attempt to run them for too long they heat up and fail.
                                  You have to let them cool down, meaning they are not really 100% duty cycle!
                                  It is marketing BS.)

                                  I see you have some of the math below. I don't really understand the details
                                  for your design so I am not able to verify your numbers at the moment, but
                                  you seem to get the idea. If you are generating 2200lbs of force with a 10cm
                                  stroke, my guess is that you are going to use much more air than what you
                                  have stated. I base this on a gut feeling from a guy who does a lot of
                                  pneumatics on FIRST robots and watches the compressors heat up very quickly.
                                  However, if you do the math, you get to tell me I am wrong and I won't feel
                                  bad at all.

                                  Pneumatics are a full of little variances. If you need more precision, you
                                  should consider hydraulic. Air acts like a spring, and is full of all sorts
                                  of surprises. It is not instant power, nor is it linear. It is, however,
                                  easy to work with and pretty reliable once you get it setup. The regulators
                                  do a good job at not allowing the pressure to exceed a certain amount, so
                                  they are usually pretty accurate at limiting the pressure.

                                  Friction will be an issue in any device you make. The fewer moving parts,
                                  the fewer sources of friction.

                                  -----Original Message-----
                                  From: Peter Balch
                                  Sent: Friday, May 17, 2013 11:43 AM
                                  To: SeattleRobotics@yahoogroups.com
                                  Subject: Re: [SeattleRobotics] testing robot

                                  > Compressors have a
                                  > series of different ratings you need to consider, most important is the
                                  > expected duty cycle. They will get very hot if you exceed their duty
                                  > cycle.

                                  I hadn't considered that.

                                  At 700kPa that's a piston area of 14sq cm which at one operation every 2sec
                                  is 4 litres/min. (I think my math is right.) Even if I assume a 10% duty
                                  cycle that's only 40 L/min. A small compressor you'd use to paint your fence
                                  is 150 to 200 L/min. Even a desktop compressor can give 70 L/min.

                                  Would that be enough?

                                  > I would expect this to take a larger (60 or 80 gallon) unit to keep the
                                  > thing from running every couple of minutes.

                                  That's 200 to 200 L. Seems quite large if I'm using 4 L/min.

                                  > You could use the same large cylinder and have two regulators and two
                                  > solenoids. One regulator might run at 10psi to generate 220lbs of force,
                                  > the
                                  > other would run at 100psi to generate 2200 lbs of force.

                                  Interesting idea. I was assuming that the cylinder lifted a lever which
                                  passively dropped to provide the force. The force would depend on what
                                  weight we hung on the lever.

                                  You're suggesting that the cylinder provides the force directly (or via a
                                  lever) and the force is determined by the air pressure.

                                  How accurately can you control the force that way? Doesn't friction and
                                  stiction affect the force? And how accurate is the air pressure?

                                  It would be nice if we could use the same rig for calibration as well as
                                  fatigue testing so a force accurate to a couple of percent would be good.

                                  Peter

                                  ------------------------------------

                                  Visit the SRS Website at http://www.seattlerobotics.orgYahoo! Groups Links

                                Your message has been successfully submitted and would be delivered to recipients shortly.