DT 40 Pro set or......

[The_Doctor]

I understand, perhaps I wasn't clear. Just as a motor has a max RPM, I thought scanners on a set single input wave would have a maximum scan back-and-forth rate. I guess not. Apples and oranges I suppose.

Ok, you could try the simplest form, two sine waves in quadrature (90 degrees phase shift). That draws a circle, and if you speed that up gently, you'll see the circle break up with erratic rings accompanied by overall distortion of shape, and hear the scanners start to scream. Do at your own risk. That will show a direct upper frequency, but it's still highly dependent on scan angle, and is harder to derive an equivalent point rate from than the trapezoids I described earlier. Those would have a much slower top speed too.

[steve-o]

OK I'm going to be honest here, I'm not testing scanners- I dont even own a set I build most of my own stuff and I was just curious about the scan speed of pro-made galvos. I'm probably gonna do it again and diy. Sorry.
No, I'm not sorry. That's my hobby and my preference. It's my party and I can build if i want to -and yes I realize they wont even be 1/10 as good as the 'big boyz' but I'm not pro, only do one show a year and dont have Pangolin/ DT40s/ at the top of the to-do list. (blasphemy-I know) I just like lasers. always have. My first he-ne was 20 years ago, and I still love piddlin' around with 'em the best I can.

[buffo]

Hey Steve-o;

I don't know about your comment that "home made galvos won't be 1/10th as good as the big boys..." I remember the home-built scanners that this guy built, and from the pics on his website they look to be about on par with the 15kpps galvos that Dave and Aijii sell on E-bay. (In fact, the guy said he can run his home-built scanners at 18Kpps....)

Personally I'd rather not attempt to build my own scanners, but I don't doubt that it is possible to build something that works quite well. Sure, you might not get "top" performance, but if you get 75 percent of what the "big boys" have, and you've done all the work yourself, I'd say that's a *huge* accomplishment. So go for it!

Don't ever apologize for having the drive to do something yourself. Hell, if it wasn't for Patrick Murphy and Bill Benner building a new laser show hardware controller and writing the custom software to go with it, we'd never have Pangolin! (And I promise you they never apologized for doing it themselves...)

Finally, for what it's worth - I have always wanted a rule-of-thumb conversion for Kpps speed to Hz for the scanner when moving at a given scan angle. And, like the Doctor posted, the answer is always a fuzzy "it depends"... Now, I understand the issues that he raised in his posts, and they're valid, but you can control for nearly all of them.

Thus the question remains: Assuming you are running a set of galvos within their normal operating range (that is, they're not screaming, the pattern is not distorting, the coils are not overheating, and you are not in danger of damaging the scanners), if you set your scan speed at 30 Kpps, what is the *maximum* speed that the scanner will move (either in degrees per second, or in Hz), assuming 8 degrees optical scan angle? The answer will be specific to a single scanner, I know, but it would be enough to make some comparative calculations. It would be useful information for some of the other discussions we've had here on PhotonLexicon. (A second question would be, what is the galvo speed with the same conditions above if you are scanning at the maximum scan angle for the galvos - say 60 degrees, instead of only 8 degrees.)

Adam

[Pangolin]

Hi guys,

A few quick points. Some may be related to Steve's questions, and some may not.

First, you really can't equate squarewave performance to "K" rating. This is because a squarewave requires maximum acceleration, and then maximum deceleration from the scanner. If the servo and scanning system is designed properly, you can produce squarewaves at any scan angle up to the scan angle rating of the scanner. The frequency limitation of the squarewave will have essentially nothing to do with "RPM" or similar factors, but will probably have everything to do with the heat limitations of the system. When the scanner is producing maximum acceleration and maximum deceleration, heat will be generated for those brief intervals. If they happen often enough (i.e. high enough frequency), then you will reach (or exceed) the heat dissipation capability of the system. Also, it might be obvious, but you will be able to scan a smaller image faster than you can scan a larger image, before the heat dissipation capability is reached.

So unfortunately, from a squarewave perspective, we really can't give a difinitive frequency or size that will apply to all scanning systems. A General Scanning model G120 can handle scanning squarewaves at 1kHz or more, and at its full rated angle, but moving magnet scanners could never do this because of the way heat is generated and dissipated in these systems. Also, a Cambridge model 6800 can scan squarewaves wider and faster than Chinese scanners can, even though the Chinese scanners are theoretically copies of the 6800, because of the way the coil is manufactured and then bonded to the stator. So we can't even say that a particular style of scanner will do the same job...

Really, a squarewave is not a good way of rating a scanner, other than to say "will it produce a critically damped step yes or no". Regardless of the size or type of scanner, I usually use a 50Hz squarewave to test scanners, just to calibrate critical damping. (If you look at the picture I posted earlier, you will notice that the function generator is set for 50Hz.)

Also, while on the topic of "K", you find the small signal bandwidth of the system by dividing the "K" by 12. For example, 30K equates to a small signal bandwidth of 2.5kHz. This is because there are 12 points in the circle portion of the ILDA test pattern, AND because the square is 70% of the ORIGINAL size of the circle. So if the circle is inside the square, it has been reduced to 70% of it's original size, or 0.707, which constitutes the -3dB (power) point, or -6db (gain) point. Eyes might be starting to glaze over about now... so just suffice it to say that 30K is a small signal bandwidth of 2.5kHz, and 60K is a small signal bandwidth of 5kHz.

However, just because you have your scanners tuned for 30K, and thus, you have a small signal bandwidth of 2.5kHz, doesn't mean that the scanning system will produce a 2.5kHz circle at any angle. There is also an angular limitation which is related to heat and to the power supply capability of the servo driver. That's why when using the ILDA test pattern, you are supposed to only use it at a scan angle of about 10 degrees. Scanning the pattern wider will require more power from the servo, and greater heat dissipation from the entire system. Eventually the servo will "run out of power supply" and the circle will stop getting larger, and will deform.

There is a related question / factor, which is "well exactly how large is this small signal". That's a point I am always making to scanner companies. At Pangolin, we have some scanners that can scan a 15mm beam, and have a small signal bandwidth of 5kHz. Yes, that's right, a scanning system capable of moving a 15mm beam at "60 K". BUT, at 5kHz that "small signal" is pretty small... It can only produce a 5kHz sinewave at 1 degree or less...

One more thing while on the topic of "K" and bandwidth. Just because a scanning system is tuned for 30K, and has a small signal bandwidth of 2.5kHz, doesn't mean it won't scan any faster than 2.5kHz. The mirror will still produce some motion out to around 7.5kHz for systems tuned to 30K, and out to around 10kHz for systems tuned to 60K. So this relates to the "highest reproducible frequency".

Anyway, there are a number of ways we rate scanners, but for lightshow folks, there are only two meaningful measures. One is "small signal bandwidth". This is probably the most meaningful measure and the ILDA test pattern can help to determine this. The second is "heat dissipation capability". We have a test pattern which can help determine that as well, but it is a bit more abstract in the sense of what it buys you in terms of imagery.

Best regards,

William Benner

[steve-o]

Thanks William!
Very informative. Exactly the information that I was looking for!
Steve

[The_Doctor]

I don't think anyone advocated square waves. I set out by saying that they were probably NOT the best source, as my first sentence in this thread. I'm beginning to wonder if there is really a flaw in my input here, or if it's just coming from the wrong person.

[steve-o]

I really don't know how to answer that Doc. All I wanted to see is some numbers in Hertz instead of 'points' for a ballpark scanning frequency. I know that X-Y scanning is a lot more complicated than that. I didn't see any "flaws" anywhere.

[The_Doctor]

Ok. The main concern I had was that my input might be seen as somehow wrong, corrected or contradicted by Bill's. If you review it though, you'll see a lot in common. While Bill links the small signal bandwidth to the circle's 12-point form in the ILDA pattern, he says that the size of this small signal depends on scanners (available power, inertia, heat dissipation), so it can be difficult to pin down. He's given you some specific numbers, but even that /12 might be considered as a convention, because the circle is run fast enough to smooth those points into a round form, not the 12 sided polygon you see in a chart of the ILDA pattern. I think that's why he says that the small signal bandwidth is the most meaningful, because point rate is a derivation of the choice of 12 points in that circle. 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. 12 is a good compromise, as less might mean that at slow speeds you'll see more polygon than circle, and more would not help it look any more circular.

I like Buffo's point about being able to contol for nearly all the variables. The more things are the same, the easier you can test the changes. If you start with the maximum speed at which you can safely draw a small circle of a chosen scan angle, you can then take the perimeter distance, divide by twelve (ILDA convention), then you know roughly how many points you can make and how fast, and how far apart you can make them. Even this isn't easy to predict because some scans reduce the point separation to navigate awkward turns, but it should give you a working average.

[steve-o]

Thank you Doctor,
Helpful info also. I'm now getting a good grasp on the fundamentals and intricacies of scanning principles.

[The_Doctor]

If you can find the money to get some WideMoves, you can pick up loads. I should know. I thrashed mine, only partly by accident, and they work as well as ever. They deserve their rep for taking a beating well. I'm glad I learned from those and not from a pair of new-and-subsequently-destroyed CT scanners. I still don't know how WideMoves do take it so well, given that they're not the best at getting heat out. Might have something to do with the thin mirrors and mounts. While not ideal, their pliability makes a weak point that bends rather than breaks, so people tend to back off well before serious hammering occurs.