Thanks for mentioning the dynaloy valve -- I'd not heard of it. Dynalloy's site is woefully short on data. If you happen to know how fast the valve is -- by any measure, to any accuracy -- I'd greatly appreciate hearing about it. (Also, what is the valve configuration? How many ports?) I'm not particularly worried about valve power, but am looking for both speed (over one cycle) and how-quickly-can-you-cycle-if-you-do-it-continually?
I forget the specifics, but since it's SMA based, you should assume the speed is proportional to the power being pumped through the valve. The configuration is essentially analogous to a pinched tube. Without power, a spring cuts off flow through the tube. With power, the SMA counteracts the spring, opening the flow. Therefore, in order to power a full stoke pneumatic pistion or air muscle, you'd have to have two valves. One to let pressure in, and one to vent pressure out. Of course, this only ensures actuation in a single direction, so for a true single DOF, you'd need two pistons or air muscles, requiring four valves. At this point, you shouldn't discount the power requirements. Typically safe power requirements are about 750 mA per valve, but at the speeds you're looking for, you would probably require over and amp.
For more specific answers to your questions, I recommend you contact Dynalloy directly. My original contact in the company was Daniel Dorn ( snipped-for-privacy@dynalloy.com).
You might get what you're looking for in SMA valves, but solenoids might provide faster action. If Dorn doesn't satisfy your interest, I suggest you check out Clippard
formatting link
They make a lot of pneumatic components, including solenoid valves, and their prices are almost reasonable.
I'm not quite as pessimistic about pneumatic robotics, though I admit there have been few commercial successes. I recently learned that one source of pneumatic servo valves (HR Textron -- who had a approx $200 valve) got out of the business, leaving (so far as I know) only derated hydraulic valves (even more expensive!) as alternatives for 4-port servo-type valves. > > Still, there may be opportunity in designing a valve that would fill the niche. Figers crossed, though I'm not going to hold my breath.
There are a few companies developing better solutions, but I haven't found anything anyone's actually selling. Viking Technologies
formatting link
is a good example. Their website's nothing but hype, with no details or information on even obtaining samples. Plus, they're only licensing the technology, and even then to select partners, so there's practically no way to find out about their product's actually capabilities. Robotics can be a harse mistress.
Curious: What are you studying, and how do you like UDel?
I'm in my final semester of studying computer engineering. I have no major complaints.
TIA > > Lawrence Pfeffer, Ph.D. (Larry), > Engineer, NIH/NIMH, Research Services Branch, > snipped-for-privacy@mail.nih.gov > 301.435.3530 FAX: 301.435.3537 >