Working out PPS

[Carl B]

Hi Adam

The idea of comparing the ILDA TP on the scope from a PC and my desk is a good idea in theory. I Unfortunetly don't have the software to update the flash memory of the desk and the ILDA TP is not one of the 400+ images that are currently stored on it. The other thing, don't know if it of relivence but the images are stored in the desk in a compressed format.

How about the use of a common ilda frame that I do have in the desk and also Rob has stored in Mamba? edit : As Rob said above.

Carl

[keeperx]

Biffo i think we are talking apples and oranges here.

a perfect circle can be drawn without any points at all by sending the sine and cosine to to the x and y respectively.

in that case, its very easy for the galvos to move thru the pattern very quickly as they are not jumping around and they do not have to deal with as much inertia.

CARL: make sure you are using the actual ILDA pattern and not one that has been optimized. The ILDA pattern is "un optimized" for a reason

[mixedgas]

No, Buffo is correct, a closed loop galvo scanner will not do 20 khz sin/cos at more then ~.25 degree. 20 Kpps ilda works out to a 2500 hertz small signal response.

Steve Roberts

[keeperx]

oh i agree with him there... i was wrong about the hertz but i was just pointing out that a scanner has an easier time drawing a circle than a complex frame because of the physics involved.

[buffo]

I Unfortunetly don't have the software to update the flash memory of the desk and the ILDA TP is not one of the 400+ imiges that are currently stored on it.

Ah... Ok - that might make it more difficult then. Mainly because some controllers don't send the point data directly to the scanners as it's represented in the ILDA file. In the case of a very simple file with only a few "lit" points (say a frame with 4 beams in a fan), some controllers will move the scanners around a bit more while the laser is blanked to ease the strain on the scanner bearings. I know Pangolin does this, and since your control desk is a matched set (hardware and internal software), it may do it as well. However, I'm just about positive that Mamba Black / Easylase does not do this. So as long as your control desk also doesn't do it, you're still OK.

The other thing, don't know if it of relivence but the imiges are stored in the desk in a compressed format.

That shouldn't matter, so long as they are uncompressed on the fly and displayed properly.

How about the use of a common ilda frame that I do have in the desk and also Rob has stored in Mamba? edit : As Rob said above.

Yeah, that's pretty much your only option. It should work though, assuming you can rule out any funny business from the control deck. I'd take several measurements with different files to confirm the timing is constant.

Biffo i think we are talking apples and oranges here. a perfect circle can be drawn without any points at all by sending the sine and cosine to to the x and y respectively.

If you do that the scanners will *not* run any faster. At best, you could get them to match the speed they're running at when they scan the circle (really a dodecagon) in the ILDA test pattern.

in that case, its very easy for the galvos to move thru the pattern very quickly as they are not jumping around and they do not have to deal with as much inertia.

The points you send to the scanners have nothing to do with the inertia of the scanners, rather, they control the acceleration of the scanners. Inertia is a function of the mass of the rotor (a constant).

Momentum, on the other hand, is a function of both mass *and* velocity. However, when the scanner is ballistic, it's velocity is at it's maximum. (And it's acceleration is also at maximum.)

When the scanner is not on it's point, it will draw more and more current in an effort to get to the point. The magnitude of the "error" between the input signal and the position signal controls the current draw, and thus both the acceleration and velocity of the scanner. If the error reaches a certain margin (which is relatively small), the scanner will be drawing maximum current.

You can test this by gently pressing on one of the mirrors with your fingertip with no input applied. There will only be a small amount of resistance until the mirror *just* begins to move. Then you'll hear the scanner start to make noise as it draws more and more current to re-center the mirror as you continue to press. It doesn't take very much movement of the mirror (maybe a 10th of a degree) to make the scanners squeal as they draw max current trying to re-center the mirror. That's how sensitive the position feedback loop is.

In order to get maximum current draw by using a sine wave, you would need to be continuously ahead of the scanner by a significant margin. (IE: a larger amplitude sine wave, denoting a larger diameter circle than you could possibly scan)

i was just pointing out that a scanner has an easier time drawing a circle than a complex frame because of the physics involved.

It is true that circles are *quicker* to draw than squares as a general rule, because the scanner can maintain a high angular speed without having to change direction abruptly. But in the case of the ILDA test pattern, the 12 points that make up the center circle are not an issue, because the scanners are never fast enough to get to them before they head off in the next direction towards the next point.

As I said above, to get the same result with a sine wave would require one of equal magnitude (IE: denoting a circle much larger than the scanners could possibly hope to draw) as the original dodecagon. And even in that case, the best you could hope to achive is *equal* speed. Once the scanners are ballistic it doesn't matter what the underlying point information looks like. They're already running flat out, and you're not going to get any more speed out of them.

Another way of looking at it is that you're using the mechanical limits of the scanners (inertial limits, if you will) as a smoothing filter to soften those 12 hard corners into smooth curves. And in fact, this is exactly the theory behind the practice of "pulling points" to make artwork look better with less flicker. Basically, to make a *quick* circle, you use fewer points and place them much wider than they should be, and rely on the fact that the scanner *can't* reach those points, so they just make a really fast loop (or elipse, or whatever shape results from the position of your pulled points), allowing the curved shape to be drawn as fast as possible. When done correctly, you can drastically reduce the point count of a frame (so you can scan it more times per second and reduce flicker) while maintaining all the features of an otherwise complex image.

Adam

[keeperx]

well i stand corrected.. i was not taking into account all the available information...

thats what happens when you give a smart guy just enough information to make himseld look stupid...

but now that i know all that, i can totaly see what your talking about..

thanks for the info.. you realy know way to much about this subject...

[James Lehman]

I think what you really want to know is the maximum change in voltage per time period.

The PM25 is some kind of digital device, so it must have a DAC. That DAC might have a variable output clock or maybe not.

None the less, there is still a time quantum between clock ticks. The maximum voltage change between these ticks will determine the maximum angular velocity that is expected of the moving mirrors.

Velocity has a lot to do with inertia. It might be easier to go faster in the same direction than it is to stop and go the other way.

These factors must play a role in the way the vector data is arranged and optimized for display through a clocked DAC.

I know this doesn't really answer your question at all!

James.

[buffo]

thats what happens when you give a smart guy just enough information to make himseld look stupid...

Nah... You weren't stupid at all. I had the *same* misconception at one point. Scanner operation is kinda weird when you first look at it.

Believe me, I struggled with the question of how to equate scan-speed to hz for *months* before Bill Benner unintentially provided the answer while replying to someone else right here on PL. (It was one of those "aha!" moments for me...) Here's the thread if you're interested... It's a pretty good read. The meat of the discussion surrounding scanner speed starts at around post # 33 or so, but there's a lot of good info about scanner design in general at the beginning of the thread.

you realy know way too much about this subject...

No, not by a long shot. There are lots of guys here on the forums that have me beat by light-years... I just keep reading and experimenting in the vain hope that I might catch up to some of them one day. But I also try to spread the knowledge I do pick up to everyone else.

Bill has been a major source of my knowledge over the last couple years. Every time we have a Laser Enthusiast's Meeting, he's there, and he always has good information to share. That's how I learned how to tune scanners... Watching Bill do it and listening to him.

Before Bill, it was Sam's Laser FAQ. I still keep a complete copy on my computer at home and at work, and when I'm bored I'll go back in and read a section just for the hell of it. There's *tons* of great information in that thing! (Much of it written by our own Steve Roberts, aka MixedGas) Now if only I could get Sam to come to SELEM next year!

Adam

[Carl B]

Hi guys

Thanks for all the advice and info on this issue. Between me and Rob I think we will be able to work out the PPS output of my PM25 the next time we meet up to try give my scanners a retune this time with the amps wired up with diferential signals.

If anyone still has more advice and info they want to share on this subject, please feel free to post in this thread. I will be keeping my eye out for more info.

Carl

[drlava]

good thread, great job of laying it out, Buffo. Your time has not been wasted. Now if it could be stickied..