trf199

Are you sure? I read that with both ports pressurised the ram would be static, although subject to slight creep, seeing as the surface area on one side of the ram is bigger than the other (because of the rod)

Yeah a 5 / 3 solenoid valve would be great (5 being 1 supply, 2 outputs and 2 vents). Trouble is they seem to be very expensive and not particularly small. Atleast the ones I've seen - I think they're considerably more complex.

Cool! Presumably you can control 4 channels at once?

That is true for the Fischertechniks solenoids because their deactived state is vented, but some valves are normally open, so in the stopped state pressure would be supplied to both ends of the cylinder and it would remain static. Maybe someone knows of alternative 3 port 2 way solenoid valves that are normally open, and not too expensive?

I'd brace an axle from both ends with a wheel on and put it on some digital scales, then apply downward pressure to it until just before it starts to bend, then you'll have a measurement of how much each will take. Then if you know your weight you'll be able to work out how many wheels and axles you'll need so they don't break.

Yeah, I've come across that before, but was a bit turned off by "due to quality concerns it is withdrawn from general availability". Interesting what it says about the power booster circuit to provide an initial power spike, and then reduced power (which would presumably sort out the floating cylinder issue on no power). I wonder how compatible it would be with the powerfunctions receiver

I've had an idea for a Powerfunctions pneumatic switch which makes use of a couple of Fischertechnik solenoid valves plus a couple of diodes. The FT solenoid valves cost about $14 each, and are nicely small. I've attached a drawing of the circuit, and how I think it would work. The limitation I see is that when the stop message is sent to the receiver the cylinder would be floating as opposed to frozen. You might be able to get around this with an alternative pneumatic circuit design? The FT solenoid valves connect the output to the input supply when power is applied and vent the output when power is not applied. Another idea I had is to supply the constant +9V and 0V to each solenoid valve, thereby making them permanently activated and the cylinder frozen. The C1 and C2 outputs could then be connected via pull down resisters to alternatively ground each solenoid and move the cylinder. I guess they would have to be power resistors? I think the solenoids draw 140 mA when active. Not sure if this particular idea is workable. If using the PF speed controller you get PWM outputs, and it may be possible to pulse width modulate the airflow, and thereby get variable speed cylinder movement. This depends on the reaction time of the solenoid and the frequency of the PWM I would imagine. If this all worked it would give 8 switched, variable speed pneumatic circuits to play with, enough to remotely control every function of the 8455 Backhoe loader with its 10 cylinders (I think 7 pneumatic switches + 1 for the compressor?). Please have a look at the diagram, and if you're electrically or pneumatically knowledgeable give your advice. I reckon that together we could design some cool pneumatic addons without waiting and hoping for TLG to do it. Thanks, Tom

For smoother pneumatics, could you not skip the air tank and only run the compressor when a movement is needed. This would give nice slow pneumatic movements, instead of unleashing 30PSI in the blink of an eye. You could also control the speed of the compressor motor to vary movement speeds. This of course depends on outputs used, NXT / PF etc.

Fantastic idea - go for it :). How about seeing how much weight each of your wheels / axles will take comfortably then dividing your weight by the weight each will take. If you have a separate trailer then get the front part to rest on the rear axle of the tractor to get more traction, or sit on the tractor directly.

Has anyone tried making their own cable operated linear actuators? When I say cable I mean string, but string operated doesn't sound as good :). It's possible to create a virtual pneumatic cylinder by pulling the ram from opposite ends of the barrel of your virtual cylinder. I think it is easier to route the cable around your design and get good torque than it is with LEGO linear actuators, and unlike with pneumatics you get accurate motor control. My other idea is to use Fischertechnik solenoid valves together with something like the Mindsensors NXT RCX motor driver, this would allow control from my NXT. Any thoughts? Tom