| ??? 04/30/09 11:39 Modified: 04/30/09 11:43 Read: times |
#164961 - I don't quite see it Responding to: ???'s previous message |
Per Westermark said:
Richard Erlacher said:
[...] but since the voice coil must either be linear and have as long a throw as the image size demands [...] Don't you mean "must either be angular or have as long a throw..."? What I said was, the voice coil must either be linear and have as long a throw as the image size demands, perhaps several meters (not realistic!) or must be rotary, I doubt a linear option can be considered. Richard said:
Perhaps you can explain how this can be accomplished. Did you look at the link in my previous thread? There are a number of similar threads on the net, where the mirror is placed between the rim and a point on the cone. Rim? What point? I guess I'm having trouble visualizing this. The tolerances are normally relative, not absolute, since the system is multiplicative. You transform a small mirror movement into a large deflection of the ray, That's why the mirror movement would have to be controlled quite precisely. The error is multiplied by the ratio of incident to reflected displacement. In addition, the changing skew between the projector and the projection surface provides an additional complication. but the biggest problem is if you require a very large angle change for the mirror, to allow for a big image using a short throwing distance. This would also represent a problem with the size of the second mirror, since the required size of the second mirror will grow with the product of the distance between the mirrors and the modulation factor used. In the end, a wide-throwing solution is problematic with a speaker but also if you use a disk head assembly or other angular system - you need a tiny galvanometer or something capable of moving a very big mirror.
In my case, I had about +/-50cm writeable area at a 3m throwing distance. I got a repeatability of about +/- 3mm which was more than acceptable. The biggest problem was that the laser diode used was too weak. The visibility was more of a limiting factor than the physical bandwidth of the system. When I did play with this long time ago, I used four analog outputs (simple R-2R ladders) and a couple of comparators. Two analog outputs specified my target position in X and Y. Two analog outputs specified the speed to transfer from x0 to x1 and from y0 to y1. When both comparators toggled, I had reached my destination and could emit the next pair of locations and speeds. Having a way of controlling the sweep size allowed me to control the visible intensity of the traces, instead of having the loudspeaker cone move at maximum speed. And I didn't have a fast enough processor (or resolution enough of the analog outputs) to try to "walk" the ray between two points. The couldn't do the IDDA in real time. A big challenge when using an open-loop solution is to know exactly where the mirror is, and not run the system too fast. As I said, I've got a bit of a problem visualizing what you're proposing, though I did look at the link you provided. I don't doubt that you have a way that would work, but ... RE |
