DT 40 Pro set or......

[steve-o]

If you can find the money

Ahhh, yes, the money. mmm. If only i was born rich instead of so talented hehe

[The_Doctor]

Yeah, was a big decision for me, getting them.. Glad I did though, despite their shortcomings with small movements. I got my money's worth. Could be worth putting on the eBay watch list...

[buffo]

BINGO!

Bill, if you were a dancer at a strip club, I would have thrown you a twenty for that last post of yours! "Small Signal Bandwidth..." That is *EXACTLY* the information I was looking for. And though I hadn't heard the term mentioned here before, I instantly understood what you were talking about when you said it.

In retrospect, the answer really was right in front of us. I knew that at maximum acceleration the galvos would be ballistic (just like they are when scanning the circle in the ILDA test pattern), but I didn't think about how the circle in the ILDA test pattern is actually made of just 12 points... If I had made that connection the conversion would have been obvious.

Either way though, thanks for clearing that up!

Adam

[steve-o]

Uhhh- I hope he takes that 1st sentence as a compliment.. otherwise he might not be coming back here 00
]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]................................... -

[buffo]

I think Bill knows I didn't mean it offensively. Seriously though, that question has been nagging me for the better part of a year now. It's come up here on the forums several times, and I've never gotten a decent answer before now.

Adam

[Pangolin]

Hi guys,

Just a few quick points

There's no reason why a slightly different number of points (anywhere between 8 and 24) might not have been used, so long as one circle's worth was drawn in the same time.

The main thing to understand is that when the servo system is tuned properly, it behaves as a second-order Bessel low pass filter. The points in a circle attempt to present the scanner with a sine-wave. 12-points ordinarily wouldn't make a very nice sine-wave, but due to the scanner's low pass filter response, the scanner will produce a circular image.

What Doc said above is somewhat true that it doesn't matter how many points a circle has, but there are a few elements to keep in mind. First, the fewer the number of points, the greater the harmonic distortion. If the number of points is too low, it comes back to haunt you -- the servo has to work extra hard to filter out the out-of-band signals. Second, this may be obvious by now, but you can't just re-do the circle part of the ILDA test pattern and expect it to still work at 30K. That pattern requires a certain ratio of number of points to K. If you re-do the test pattern with 24 points in the circle, you will need to run the pattern at 60K even though the small signal bandwidth being tested would still be only 2.5K.

but I didn't think about how the circle in the ILDA test pattern is actually made of just 12 points...

Actually the circle portion of the ILDA test patern has multiple iterations of a circle. There are probably around 36 or 40 points in that area. But there are only 12 points PER circle (the circle is drawn over and over again to get the scanner up to speed before turning on the beam, and keep it going the same speed after the beam is turned off).

if you were a dancer at a strip club, I would have thrown you a twenty for that last post of yours

Ooooo. Big tipper!! But how is it that this was the first thing to come to your mind?


Bill

[The_Doctor]

Second, this may be obvious by now, but you can't just re-do the circle part of the ILDA test pattern and expect it to still work at 30K. That pattern requires a certain ratio of number of points to K. If you re-do the test pattern with 24 points in the circle, you will need to run the pattern at 60K even though the small signal bandwidth being tested would still be only 2.5K.

True, I tend to overlook that because I was working with wave data, with the sample rate rather than a point rate. If I chose two points closer together in space, they'd also be closer in time (and if on a linear gradient I'd not consider them as separate points until the gradient changed). The thing about that is, it can interpolate within the same timeframe which is another way of reducing unwanted harmonics, as what stepped gradient changes there are can be smaller on curves, within the allowance set by the sample rate. I'll still use a stepped change though, as calculating nice curves takes so much time, and the scanner's inertia is part of a filter that has no CPU overhead.

Actually the circle portion of the ILDA test patern has multiple iterations of a circle. There are probably around 36 or 40 points in that area. But there are only 12 points PER circle (the circle is drawn over and over again to get the scanner up to speed before turning on the beam, and keep it going the same speed after the beam is turned off).

I should have said twelve points per turn.. I noticed from a file that James Lehman made for me to test with, that there were 3.25 turns, I think. I got the feeling that this total could have been made smaller with the WideMoves, but I guess it's chosen to allow time for all scanners that have been tested with it.

[buffo]

Ooooo. Big tipper!! But how is it that this was the first thing to come to your mind?

Ha! You got me there Bill... I honestly don't know either. I can count up the total number of strip clubs I've ever been in on the fingers of one hand. I can't imagine why that thought popped into my head. (And no, I'm not ready for a Freudian analysis of this!)

But I did appreicate the information. We've kicked that topic around several times here on PL without ever really answering it. And the number has real meaning, knowing how fast (or slow) the galvos actually move has huge implications for other parts of the projector system. (Laser blanking speed, for one!)

Adam

[werner1619]

Hi buffs,

I feel realy stupid asking the following qeustion after reading all the posts in this thread, but can some one explain to me how to propperly tune scanners to ILDA 30k?

I think I have PR scanners, got them from Lumina, after reading about ILDA 30k test patern the first time I tried to reset the scanners in hope of better performance and since then the scanners have never been the same again!!

Another stupid qeustion, Lets assume your scanners are set/ tunned to ILDA 30k, how do you ensure that your software controlling the scanners are doing the right thing/ sending the right signals to match the tuned state of the scanners??

Hoping to get this ILDA test patern and setting under the knee once and for all....

(thanks for not laughfing)

[buffo]

Hi Werner;

Actually, your question isn't the least bit stupid. It's actually a damn good one. It's a bit hard to answer, even.

Tuning galvo amps is as much an art as it is a science. There are several different ways to procede. This link is a good place to start reading about the procedure. Here is a more step-by-step approach. Both links have a lot of background information on what various areas of the standard ilda test pattern are designed to test.

Having said all that, when I watched Bill Benner adjust the galvos at our gathering in Florida (FLEM), I noticed that the first thing he did was to turn off the blanking, so you saw the entire pattern, not just the lines in the picture but all the re-trace lines too.

Then he adjusted the gain first, which is different than the procedure I linked to above. Basically he first made sure the x and y gain were about equal (the circle was round and not elliptical), and then he started tweaking the x and y gain up and up until the circle stopped getting larger. That's when you know the scanners are at max acceleration (ballistic), which is how they should be for the circle. Then he proceded to adjust the damping; often looking at different parts of the pattern to get it right. There is a lot of trial and error involved, and it's a time-consuming process. (Bill is a pro, and he still spent at least a good 30 minutes doing it.)

Hopefully Bill will chime in here with some of his other tricks. I remember he was using one of the re-trace lines near the bottom of the pattern to adjust something, (high frequency dampinig, maybe?) but I can't remember it now. (This is why the next time I see Bill adjusting a set of scanners, I'm going to videotape the whole damn event and study it!)

Normally, once you get them dialed in you don't have to mess with it again. But getting them dialed in can be a pain - especially if you've never done it before.

As far as knowing that your software is right - well, that's harder to tell. If you're running Pangolin, you can check the setting, and as long as it's set for 30K you should be fine. Same thing with Mamba Black, LDS, Full Auto, etc... However, there are some laser show packages out there (like the Alphalite) that either have slight variations in the timing, or that need to be calibrated for the specific computer you are useing before it will work right. So the answer to your second question is hard to say without knowing what hardware and software you are using.

Either way though, good questions!

Adam