I believe I have found a solution to most issues arising from the illegality of exhaust restricting solenoids that has been a hindrance to all plans involving stopping a cylinder mid stroke.
Unfortunately, due to the apparent desire of the GDC to prevent the trapping of air within the pneumatics system, http://forums.usfirst.org/showthread.php?t=16642
the solution is slightly more complex than it otherwise would be, but that can be worked around.
The solution, in compliance with this years rules is as follows:
Connect one secondary regulator per cylinder to be controlled to the 60psi side of the pneumatics system.
.
Attach a servo to the knob that controls the pressure setting of the secondary regulator.
.
Connect one port of the pneumatic cylinder to the secondary regulator output (technically, a solenoid is unnecessary, but this seems to make them mandatory, use your best judgement in deciding placement, powering it and programming it are unnecessary.)
.
Attach an adjustable flow control valve to the second port. Leave it fully open at first and reduce the flow if the spring induces significant oscillations.
.
Place a variable force device (i.e. a spring) such that it applies force to the piston opposite the force resulting from the pressurized air in the cylinder.
.
Remember, the restoring force of a spring is kX. where k is the spring constant and X is displacement; and the force of the piston is (Pi*[bore/2]^2)pressure if the cylinder is extending and (Pi{bore/2}^2-{rod diameter/2}^2])*pressure if it is retracting.
.
Set the spring force equation equal to the appropriate pneumatic actuator force equation substituting in the correct values for bore, k, and, if necessary, rod diameter and the desired value for X. For unbalanced mechanisms, account for the force applied by the mechanism in calculations by adding this force to the appropriate side of the equation.
.
Solve for pressure.
.
Program the servo to adjust the regulator to the pressure value found using this method.
.
Note:constants are red, independent variables (inputs) are blue, and dependent variables (ouputs) are green.
I am not an official source of information, I cannot verify the legality of this method of cylinder control.
Did you actually try this? I can think of a few immediate issues, like valves need a minimum pressure to operate, and a servo has 170 degrees of throw, where a regulator has multiple turns.
I haven’t tried it, but I also think servo will not have enough power to turn a KOP regulator.
there are no valves that require actuation so that issue is avoided entirely. The only valve is a single solenoid that is open by default and never needs to be switched.
Not all servos have such limits. Sail winch servos can travel 3 full rotations, I believe, and some servos can rotate indefinitely, but have no position feedback (see point 3)
3)It doesn’t have to be a servo, a motor (or continuous rotation servo) with a potentiometer or encoder would also work with some extra programming. Gearing could also increase available torque for weaker motors/ servos.
No, I have not tried it physically, but the physics works out, at any given pressure, a certain pneumatic cylinder has a set force. Additionally, at any given displacement, a certain spring has a calculable restoring force, if these forces, along with forces applied by the weight of the mechanism (assuming an unbalanced mechanism) balance out to zero, the mechanism will be stationary (technically, it might exhibit oscillations, but by reducing the flow rate out the open port on the cylinder, they should be dampened to the point at which the amplitude is nominal).
However, I do have 2 corrections to make and they shall be made momentarily.
Well, it’s not much of a first post, but I just had to put in my $0.02 on this subject.
Now, why wouldn’t something like that work? This is just for the control of a single cylinder; the rest of the circuit continues past the black arrow. Each check valve prevents backflow from the accumulator into the cylinder, while simultaneously allowing venting of all pressurized areas from a single valve. (Note: I was intending to use a solenoid flow-control valve, but I seem to have lost the link.)
Based on what I have read about pneumatic schematics, this looks like a fancy way of operating a cylinder in 2 positions, out or in, since when one side is pressurized, the other would be vented to the atmosphere. I considered a similar approach myself, but abandoned it when I realized there was no way to make it work properly without complex programing or trapping pressurized air in one side of the cylinder. Also, I’m not sure if solenoid flow control valves are legal, but in this configuration they certainly are since it is a second solenoid and according to <R74> only one solenoid can be used to control every commanded action of a pneumatic actuator.
I started reading this, and it seemed so complicated I just gave up. I don’t typically give up at understanding things, but this just seems like so complex of a system, that perhaps the design should be reconsidered if it requires all this. Motors anyone?
After reading this a second time, I totally get it now. In theory, it’s actually a rather nifty idea. Not nearly as complicated as I thought. In practice, I think it would leave a lot to be desired.
The valve in the schematic above is a 5-port, 4-way, 3-position valve with the center position blocking both pressure and exhaust ports. This allows the pneumatic cylinder to be controlled similarly to a hydraulic ram, since one can shut off and trap the air inside the cylinder, preventing significant movement.
Of course, <R73> specifically states that this is not allowed, so one must be able to vent the trapped pressure at the same time as the rest of the system, from a single valve. The two check valves allow this. Each check valve allows air to flow (mostly) unobstructed out of the supply to each cylinder, but does not allow the high-pressure air to reach the cylinder supply lines. Since the pressure in the cylinder is lower than that of the accumulator, little air will be lost, due to the natural tendency for the higher pressure air to seal off the check valve.
During normal operation, this will cause the cylinder to act as if it were connected via a normal piece of tubing. However, in the event that we need to vent all of our stored pressure rapidly, the check valve will allow the stored pressure in the cylinder to flow straight out the dump valve.
Oh, I get it - the check valves allow the whole system to vent to atmosphere when high pressure is released, thus making the 3-way valve (center position blocking) legal.
I’d be prepared to explain carefully to the inspector why the high pressure side seems to be connected to a cylinder.
Just to clarify, the system explained in the first post and the schematic in the second are actually 2 very different methods of controlling cylinders. I imagine it would be very difficult to try to match the schematic to the description in the first post, since they don’t actually go together.
The bolded words are why it won’t work (at least not legally). <R65> and <R66> combine to heavily limit the types of pneumatic devices allowed on the robot. Check valves are not one of those allowed devices.
EDIT: A unidirectional inline flow control valve, when it’s closed entirely, will act like a check valve. If each check valve in my above schematic was replaced with one of these, it would perform the same function, and remain legal.
May I pose a practical question? Why do we think the GDC has this rule in place? Why not allow the use of additional pressure relief valves in line with the cylinder rod/head supply lines, mounted on your bot right next to the “main” relief valve? You could use quick-release toggle valves (e.g. http://www.mcmaster.com/#pneumatic-toggle-valves/=b7a2dk) to make it super-easy to relieve pressure in an emergency… One could even devise some kind of mechanical toggle that connects to the valves and flips all the connected reliefs at once.
So, is it a safety / emergency thing for use at the end of the match or in an emergency situation, or does the ability to trap air at mid-stroke during normal use pose some other safety concern?? I can’t imagine the latter, or there wouldn’t be much market for closed-center solenoids…!
<R73> The pressure vent plug valve must be connected to the pneumatic circuit such that, when manually operated, it will vent to the atmosphere to relieve all stored pressure. The valve must be placed on the ROBOT so that it is visible and easily accessible. If the compressor is not used on the ROBOT, then an additional vent valve must be obtained and connected to the high-pressure portion of the pneumatic circuit off board the ROBOT with the compressor (see Rule <R69>).
That’s why you couldn’t have multiple release valves. Unless you could find something that can simultaneously vent multiple outputs while keeping each output isolated, you cannot use your idea.
Roboman: I understand that’s the rule, I’m asking “Why do we think the GDC has this rule in place?”, and should it be reasonably challenged for next year’s rules? The single vent rule was further emphasized first in Lunacy with “all” underlined. 2008 was the first year they broke out mention of the valve separately; before that it was mentioned as part of the Nason main relief rule and wasn’t as explicit: “The Parker pressure vent valve must be connected to
a Clippard tank such that, when manually operated, it will vent to the atmosphere to relieve any stored pressure.”
The fact that they’ve refined and emphasized the rule leads me to believe there were safety issues at some events (perhaps in 2007 and again in 2009). I’m just wondering if they were “emergency” occurrences that required quick relief by one and only one valve, or if we could build safe (and potentially more functional) robots that had multiple relief valves that together are capable of relieving all stored pressure.
Based on what I have read, they want all pressure to be released with a single action, the opening of the plug valve. I don’t think the relief valve is involved in that much. The desire to release all the air with this single action stems from concern about people working around potential energy, in this case in the form of air pressure differential.
A secondary concern might be teams using isolating air from the rest of the system and compressing it to dangerous levels (in excess of 120 (or even 240) psi, if one designs the system “properly”) and releasing the potential energy in a highly energetic manner that could be destructive and dangerous, even if the system survives long enough for the release to occur on command, in the desired manner. Additionally, it could be possible to design similar, though not quite as extreme, systems accidentally, even if the team tries to stay within the rules. In most applications, the system would be designed with this in mind thus making the use of center blocking solenoids practical.
I understand, and always have understood, that you were describing a different system. I just didn’t want people to perceive either system as being more complex (or confusing) than it really was by trying to match the system I proposed to the system you provided a diagram of, since that is impossible due to the fact the systems are far from similar.
Also, ‘simpler’ is a matter of opinion, I consider mine to be simpler due to the fact that it doesn’t require very much attention in the code, all you need to do is set the regulator and let it fine equilibrium with the spring on its own. Also, it isn’t very complex physically either. I will draw a diagram and post it for clarity.
Stored pressure in a pneumatic system is inherently dangerous. If a part happens to fail, one must have a reliable method of quickly venting the system pressure, preferably from a single point. Requiring a single release valve is a much easier way of making sure all teams have this, rather than judging a pneumatic system’s safety on a case-by-case basis. I actually believe this rule is a good one. However, I also believe that the GDC should allow a broader range of pneumatic devices, such as check valves, for next year.
I understand your system, however, it seems a little jury-rigged, to be frank. Keep in mind that the force exerted by springs changes based on how much they are extended or compressed, unless you use a constant-force spring (obviously). Also, you’re relying on a relieving regulator to vent excess pressure in the cylinder. While these are more common than non-relieving regulators, the latter are certainly not rare, and could easily be confused with a pressure-relieving regulator.
In addition, to implement your system, you would need to figure out some way of controlling the regulator. You mentioned hobby servos, which could work, but they are confined to <180 degrees. Sail winch servos are not allowed, and still would not rotate enough to open and close the regulator entirely. They are also much more coarse in their movements, which could limit your true control over the cylinder.
Your system would have a variable force in addition to the variable stroke, since you’re varying the pressure, rather than the amount of air in either end. As you should know, reducing the pressure also reduces the force exerted, and the spring on the end that counteracts the cylinder’s rod will cause the net force to be near zero, since it’s stopping the travel mid-stroke. This is obviously not good for actuating an arm, or anything that will be exerting any sort of force.
Guys,
You are reading a little too much into this. The vent valve is to relieve system pressure so that the robot cannot move while in transport. There is no rule (except in my mind and those teams that design safely) that says the robot cannot move when energized. So many teams design just such a robot. When the pump comes on, things begin to move. Because the pits, the queue and transport areas are so crowded, we don’t want a accident to occur simply because a system has pressure and a team member inadvertently enables the robot. I have had my arm caught in a moving mechanism and seen a robotic arm swing out and knock over the pit table into the pit behind it at Champs In addition to the safety, many rules are in place to reflect good engineering practice (similar to wire color codes) and to prevent teams with no pneumatic mentorship from hurting themselves.
