Printable View
Can anyone tell me of a way to find out the PPS output of my PM25 control desk?
When I bought the desk I was told that it was set to 18k PPS and is not adjustable. I would like to confirm this is correct as I want to make sure that my ScanPro20's are being tuned to the right speed.
Carl
You'd probably need an oscilloscope to be sure.
Adam
PPS refers to how many PPS you scanners can draw the ILDA Test pattern..
The ILDA pattern is designed to be very hard to scan. It uses long swings and a complex array of designs to basically torture test your scanners into submission.
a 20kpps scanner could probably draw a circle at 40kpps (circle is a simple easy shape for a scanner)
so your question is slightly vague.. what do you actualy want to know about scanners?
Keeperx
My knowledge of how to tune scanners is very limited but do have some basic understanding that the ILDA TP is complex and designed to drive scanners quite hard at there rated speed.
WRT the PPS speed of the output from my desk if is as mentioned above of 18k PPs would scan a simple pattern with say only 10 points then the desk would scan that pattern 1800 times a second and a more complex pattern with say 100 points, then the desk would scan that pattern 180 times a second.
My question was to find a way to clarify the PPS speed of my desk and was thinking that only a scope as Adam has suggested would be the only way. Something I don't own. The next time I get to see Rob Stanwax, maybe he could help with this.
Carl
What is your goal? To tune your scanners to whatever speed the board puts out? I think I would just tune them to 20K.
Maybe I could - though I guess I wont be doing much scanner tuning this weekend :)Quote:
Originally Posted by Carl B
Rob
Sorry to split hairs here.. but when you say tune them to 20k, what are you talking about??Quote:
Originally Posted by carmangary
20kpps is 20-khz or 20,000 instructions a second
AKA 20,000 little sine waves flying thru the the galvos every second
so as i said.. a CIRCLE could be scanned much faster than the ILDA pattern because its a very gentle curve that repeats and the galvo's just hum along..
the ILDA pattern is harder so you have to to lower the Khz to compensate (less instructions per second).
so 20kpps doing what?
and FYI tuning scanpro 20s to 20kkpps(ilda) is fine if your gonna run them at 18kpps(ilda)
you do not tune TO a specific speed, you tune to a MAXspeed.. anything below the max speed is gravy
I'm talking about tuning to 20K with the ILDA test pattern. It is implied that when talking about tuning that the ILDA test pattern will be used.
I know YOU knew what you meant.. i was playing devils advocate so i we could help Carl better understand what he was actually doing:)Quote:
Originally Posted by carmangary
Oh damn... Now i'm curious. Are you on a promise or something ;) :DQuote:
Originally Posted by stanwax
Jem
hehehe I guess in reality I am!Quote:
Originally Posted by Jem
:)
Rob
Keeperx
When I talk about PPS, it is the position instructions per second that I am referring to and understand that 20k scanners scan a circle faster than the ILDA TP which explains why the circle scans without flicker unlike the test pattern.
Yours and everyone else's input on this issue is appreciated. My reason for my question was really for curiosity.
Jem
You know what they say about what curiosity did to the cat.:D
Rob
I meant the next time we meet for laser related reasons.:)
With out having the means to check the PPS rate of the PM25 my self I think I should take George's word on this as gospel.
Carl
rob said
aahhhhhhQuote:
hehehe I guess in reality I am!
:)
so you off to the fetish weekend again then rob....
all the best ... Karl
Karl, Do you know something about Rob that the rest of us should know about?:D
Carl
Karl you tawt - my spelling is getting worse
I have just wet myself laughing at that - reminds me I must put talc on the shopping list
Rob
:DCarl said
yep .... i know that he does not have many good photos of himself on the internet tha can be photoshopped into other photos ... :DQuote:
Karl, Do you know something about Rob that the rest of us should know about?:D
rob can you post a decent facial shot in your foto albums that i can use ....
did i just say Facial shot .... oooohhh errrrrrrr misus :D
Rob .... at least when you wet yourself ... no one knew ...... you just had a warm sensation in the rubber outfit :D
all the best .... Karl
This is not true at all. A sine wave at 20,000 hertz would be impossible for even 30K scanners to handle. The maximum *small step* bandwith of a 30K scanner is around 2500 Hz.Quote:
Originally Posted by keeperx
Those "points" in the point per second speed have limits on how far they can be from one another. In the case of the ILDA test pattern, the central circle is actually a series of 12 points forming a dodecagon where all 12 of the points are outside of the center square. (Actually it's 36 points, because the figure is traced 3 times in a row.) The scanners, of course, never get close to these points. That's why you see the image as a circle that lies completely within the square. Still, one trip around the circle represents 1 cycle (up and down, or left and right).
If you consider that when drawing the circle, the scanners are ballistic (that is, under constant, maximum acceleration the entire time, because they never reach their destination points), then you can see that it's fairly easy to calculate the maximum speed (in Hz) of the scanners. Assume that all you scan is the center circle. So that's 12 points for a single cycle. If you're running 30,000 points per second, divide by 12 points in the frame, and you'll see that you'll make 2500 trips around the circle in a second. Thus, the small signal bandwidth is 2500 Hz.
Of course, this ignores the fact that there is significant distortion between the 12 points in the ILDA file and the resulting scanned circle. So in reality the bandwidth is even lower than that. But it's a good starting point.
Bottom line: a 20Khz sine wave is an order of magnitude too fast for today's scanners.
No, it couldn't. When the scanners are scanning the circle in the ILDA test pattern, they are already moving as fast as they can. You aren't going to get any more speed out of a set of scanners once you tune them (correctly) to the ILDA test pattern. Also you don't "reduce" the scan speed when you tune to the ILDA pattern. You tune at the speed you will run at.Quote:
a CIRCLE could be scanned much faster than the ILDA pattern
There is a lot of good information about the ILDA test pattern on the Pangolin website, and some more over at the old LaserFX website. You might want to read up on the reasons behind the specific portions of the pattern.
As for testing the speed of Carl's laser bench, he's going to need an Oscilloscope to test it. And his reasoning is sound. If you tune at a high speed and then display at a lower speed you will get overshoot. In extreme cases the center circle of the ILDA test pattern will enlarge to the point where the scanners actually manage to reach the outer points of that dodecagon!
The opposite is just as bad - tuning to a slow speed and then displaying at a higher one causes undershoot. The scanners never reach *any* of the points in the image. Both of these cases are bad for the quality of your scanned image.
Adam
When we had a go at tuning Carls scanners (before he did some rewiring that affected the output considerably) I played the ilda tp through Mamba set to 18k. Carls concern now is that the PM25 is putting out the said 18k and I guess thats what he needs confirming. When he nexts visits me - dont worry Carl for casual visitors I pack the rubber away - I will be able to examine the board output with a scope and with a bit of luck compare it to a signal from Mamba, set to 18k, of the same frame.
Rob
Rob;
If you're good with your scope, then load the ILDA test pattern on Mamba and look for the center circle in the pattern on your scope. It's about 2/3 of the way through the waveform. (It will be visible on both the X and Y traces, so it doesn't matter which one you connect the scope to.)
Keep drilling down until you find it. (It will be 36 points clustered together that create a rough sine wave that is 3 cycles long.) Remember that the amplitude will be rather small compared to the rest of the signal.
Measure the time between peaks there just to confirm that Mamba is running at the correct speed. (And so you're familar with what it looks like so you can find it again.) Then load the ILDA test pattern on the bench and do the same. You'll know right away if the speed is off!
Easy, Peasy, Lemon Squeezy!
We did this at FLEM a long time ago. Bill Benner was trying to explain what sample jitter was, and some of us (Buffo sheepishly raises his hand) couldn't quite visualize what was going on, so Bill fired up my 'scope and zoomed in on the waveform coming out of Fred's Easylase USB DAC. And sure as hell, you could see the timing shift back and forth!
Adam
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
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
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
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.
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.Quote:
Originally Posted by Carl B
That shouldn't matter, so long as they are uncompressed on the fly and displayed properly.Quote:
The other thing, don't know if it of relivence but the imiges are stored in the desk in a compressed format.
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.Quote:
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.
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.Quote:
Originally Posted by keeperx
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).Quote:
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.
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)
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.Quote:
Originally Posted by keeperx
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
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...
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! :rolleyes:
James. :)
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. ;)Quote:
Originally Posted by keeperx
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.
No, not by a long shot. :o 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.Quote:
you realy know way too much about this subject...
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! :D
Adam
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
good thread, great job of laying it out, Buffo. Your time has not been wasted. Now if it could be stickied..
After paying Rob Stanwax a visit yesterday we were able to determine the PPS output of my PM25 desk and what a shock we got. It turns out the desk is only running at 13k PPS and not at 18k PPS as I was told. We managed to work this out by comparing a single frame of a Pango show that's in my desk and found the exact same frame in the show on Robs computer and then compared the two X axis outputs on a scope.
So it turned out that my concerns of the PPS output from my PM25 has bought out the true speed that it is running at.
And again must thank Rob for his help with this matter and the time he has spent tuning my scanners.
Carl
You are welcome....what else would I be doing on a Saturday :)
Rob
PS it was like UKLEM 3.25 at my house yesterday with Kevski and Carl there!
BASTARD !!
You had a UKLEM and you never invited me !!
And I was in St Helens yesterday too !!
:mad:
No way! My other half is from St Helens & I worked there for 10 years.
Richard if you are ever in the vacinity please drop me a line, the door is always open and there is always warm water for a brew - though it may be from the waste of the 168 :)
Rob
Hmmm..
Somehow your waste water doesn't sound that appealing. :rolleyes:
When i got my first PM25 it came with GD120 Galvo`s
That was a few years ago and it had no blanking.
I was informed that it ran at around 12k
Cheers
Well while at Rob Stanwax's Place on Sat I got the chance to connect my PM25 to Rob's SP168 white light that also has G120's and the scan quality was just gob smackingly fantastic. I wish I had took some pic's of text and griphics.Quote:
Originally Posted by starlogic
Carl