diegocn
07-11-2009, 04:04 PM
Hi, I just read many topics from people saying all things from pneumatic platforms, that are not precise, and not encouraging people do to in this way.
Then, I want to publish my position about it, and some knowledge.
When I read these comments or projects, like all projects
are with 5/3 simple valves. To implement a PID control (or simple servo
control ) people try to use this valve in a "PWM" ( talking about the fluid itself
and not the electric control of the valve ). Pneumatic is a fast system,
and good simple valves go to a maximum as 15ms to open and close, so
the maximum switching frequency you get is ~65Hz. This is extremely low,
and even, impractical, as you need 100x or 1000x more than this.
The solution, as the industry know for years, and I dont know why sim
builders don't go this way (about later), is proportional 5/3 valves.
Proportional directional valves control the flow of the fluid, precisely from
0% to 100%, the majority can position the spool in any flow option in
~5ms.
The fact of precision with this ( that of course is dependent on other factors,
like the actuator feedback and PID control ) are this:
Positioning precision of ±0.032 in or ±0.81mm ( source: http://www.hydraulicspneumatics.com/200/Issue/Article/False/7223/Issue )
This is an exaggeration to a motion platform, and we can be very happy
with lower values from lower priced valves ( Of course this precision is reflected in velocity control to, resulting in good washout ).
Velocity and force dont need to be explained, but, just for complementing:
A cylinder with 2 inch internal diameter, result in an ~141 Kgf actuating force.
A cylinder with 5 inch internal diameter, result in an ~883 kgf actuation force.
(Calculated considering 100psi pressure and ignoring friction, a think all professional and some hobby compressors today go to values equal to 116psi or greater.)
Only for curiosity, a 5 inch internal diameter cylinder with 116psi pressure result in an ~1024 Kgf actuation force. Sufficient for a real Boeing 747 cockpit ;)
Other solution, a bit new to industry, is hydro-pneumatic control.
Knowing some more advanced products, I draw this idea.
http://img6.imageshack.us/img6/7808/hidropneumatic.th.gif (http://img6.imageshack.us/i/hidropneumatic.gif/)
This is an tandem cylinder, its like two cylinders with one rod.
The orange part is actuated by pneumatic, equal to a pneumatic
cylinder. The blue part is filled with hydraulic fluid, and have a
flow control valve, that controls the passage of hydraulic fluid from
the cylinder.
So, this is in fact, a pneumatic cylinder with an hydraulic brake.
You can control the pneumatic cylinder with a simple 5/3 valve,
but the brake cylinder is that control the rod, just opening and closing
the flow control valve, that can be a ball valve actuated by a small motor.
You don't need so fast response because the hydraulic fluid is like incompressible, giving more predicated response. Air is the inverse. So,
when the hydro valve is closed, even external force (like anyone
jumping in the cockpit ;) do not move the cilinder. In an only pneumatic
cilinder, the proportinal valve and the PID control need to be fast to
compensate the external force, what is possible with no problem, too ).
Only to be clear, I don't invented the hydro-pneumatic cylinder, just
think in the idea of using the principle, on tandem pneumatic cylinder.
Some links:
Enfield Technologies, very good about pneumatic servo control:
http://www.enfieldtech.com/
http://www.youtube.com/watch?v=Bk6eLsqvBxc
(A good video from Enfield Technologies showing anything like a 3 dof
plataform ;))
Good page with pneumatic myths:
http://www.hydraulicspneumatics.com/200/Issue/Article/False/7223/Issue
Thanks,
Diego
Then, I want to publish my position about it, and some knowledge.
When I read these comments or projects, like all projects
are with 5/3 simple valves. To implement a PID control (or simple servo
control ) people try to use this valve in a "PWM" ( talking about the fluid itself
and not the electric control of the valve ). Pneumatic is a fast system,
and good simple valves go to a maximum as 15ms to open and close, so
the maximum switching frequency you get is ~65Hz. This is extremely low,
and even, impractical, as you need 100x or 1000x more than this.
The solution, as the industry know for years, and I dont know why sim
builders don't go this way (about later), is proportional 5/3 valves.
Proportional directional valves control the flow of the fluid, precisely from
0% to 100%, the majority can position the spool in any flow option in
~5ms.
The fact of precision with this ( that of course is dependent on other factors,
like the actuator feedback and PID control ) are this:
Positioning precision of ±0.032 in or ±0.81mm ( source: http://www.hydraulicspneumatics.com/200/Issue/Article/False/7223/Issue )
This is an exaggeration to a motion platform, and we can be very happy
with lower values from lower priced valves ( Of course this precision is reflected in velocity control to, resulting in good washout ).
Velocity and force dont need to be explained, but, just for complementing:
A cylinder with 2 inch internal diameter, result in an ~141 Kgf actuating force.
A cylinder with 5 inch internal diameter, result in an ~883 kgf actuation force.
(Calculated considering 100psi pressure and ignoring friction, a think all professional and some hobby compressors today go to values equal to 116psi or greater.)
Only for curiosity, a 5 inch internal diameter cylinder with 116psi pressure result in an ~1024 Kgf actuation force. Sufficient for a real Boeing 747 cockpit ;)
Other solution, a bit new to industry, is hydro-pneumatic control.
Knowing some more advanced products, I draw this idea.
http://img6.imageshack.us/img6/7808/hidropneumatic.th.gif (http://img6.imageshack.us/i/hidropneumatic.gif/)
This is an tandem cylinder, its like two cylinders with one rod.
The orange part is actuated by pneumatic, equal to a pneumatic
cylinder. The blue part is filled with hydraulic fluid, and have a
flow control valve, that controls the passage of hydraulic fluid from
the cylinder.
So, this is in fact, a pneumatic cylinder with an hydraulic brake.
You can control the pneumatic cylinder with a simple 5/3 valve,
but the brake cylinder is that control the rod, just opening and closing
the flow control valve, that can be a ball valve actuated by a small motor.
You don't need so fast response because the hydraulic fluid is like incompressible, giving more predicated response. Air is the inverse. So,
when the hydro valve is closed, even external force (like anyone
jumping in the cockpit ;) do not move the cilinder. In an only pneumatic
cilinder, the proportinal valve and the PID control need to be fast to
compensate the external force, what is possible with no problem, too ).
Only to be clear, I don't invented the hydro-pneumatic cylinder, just
think in the idea of using the principle, on tandem pneumatic cylinder.
Some links:
Enfield Technologies, very good about pneumatic servo control:
http://www.enfieldtech.com/
http://www.youtube.com/watch?v=Bk6eLsqvBxc
(A good video from Enfield Technologies showing anything like a 3 dof
plataform ;))
Good page with pneumatic myths:
http://www.hydraulicspneumatics.com/200/Issue/Article/False/7223/Issue
Thanks,
Diego