Will this burn up galvos?

I have a question on 2 signals:: Will the attached signals burn up any motors?

These signals are 'pre-buffered'...meaning it is not centered on 0V, and the gain is not correct. The current output is 0V to +4V.

I am asking for help on the timing and the 'noise' on the signals.

I am creating a new laser controller, and I want to make sure that A) The Speed of the switching signals and B)Noise on the lines are not going to blow the motors.

I have the first signal set up to output at 20k.

If anybody could please let me know if this signal is going to work with Galvo amps, I would really appreciate it. I do not want to burn up anybody's motors...

As far as the second signal...It looks like it is outputting at about 60k, but they are fairly small steps...

The Second attachment is a test for Rastering an image...Is this too much for galvos?

Thank you for any help.
Shaun

I think rastering an image might be out, I don't think any galvos can cope with that, except the vertical scan.

The ILDA test pattern uses fast jumps in the inner square but the amplitude isn't a full scale deflection. Most galvos can cope with one or two arbitrarily fast full scale deflections I think, so long as they aren't repeated enough to hammer them. WideMoves are forgiving, CT scanners probably not so. I don't know if you can be sure a designed scan is safe for all galvos, it might be better to put a slew rate limiter on the input of a specific drive board to do that in case of extremes in the scan signal.

Are you talking a 20khz square wave (first pic) or 20k points per second?

The scanner will not even think about doing a 20khz square wave forget about 60. We are lucky if we can get a diode laser to modulate correctly at 10k. A really good well tuned pair of scanners might be able to do a two or three k sine at a couple of degrees. Remember we are moving little mirrors that have inertia.

The noise does look a little bit high. Depending on what freq you are trying output you might want to think about a low pass.

I understand that the test output is unipolar but to play nice with everyone else you need to plan on a balanced bipolar output.

Chad

Quote:

---

I think rastering an image might be out, I don't think any galvos can cope with that, except the vertical scan.

---

Galvos can raster, granted, its not tv quality, but it does work and can be a pretty neat effect. Is this what you were refering to or am I missing something?

That's raster. It depends how fast you want to go, frame rate, line count. Whenever the subject comes up on alt.lasers the usual answer is that a spinning polygonal mirror is best for the horizontal scan, and a galvo or slow polygon mirror is best for vertical.

If you want TV quality, the vertical galvo can cope with the sawtooth scan but even if using triangle instead of sawtooth for the horizontal one you can't take the strain off to bring it in range of a current galvo, you'd still need nearly 8 KHz of triangle wave for 25 fps and 625 lines.

Doc;

Actually, if you *really* want to get good raster images, there's another solution for getting the high-speed horizontal scan rate... Use an AOM!

The problem is that you only get 1 to 2 degrees of optical scan :( , but it's damn fast!

NTSC video has a horizontal scan frequency of 14.7 Khz. No galvo on the planet comes close to that, but a multi-faceted spinning mirror can get close. But AOM's can reach that speed with no problem. You just need a really long throw!

And as you said, the vertical scan is easy. Even at 60 frames per second you're only talking about a 60Hz sawtooth wave. Your average 30Kpps galvos can manage that.

DZ:

I did my best to count the lines in that raster image... I came up with 48 lines, so at 15 frames per second that equals 720 hz on the horizontal, which is within the small signal bandwidth for 30Kpps galvos. I'm guessing that you can probably display that at around 3-4 degrees optical, right? Not exactly NTSC quality, but cool nonetheless.

Adam

No reason to use sawtooth just draw every other frame from top to bottom.

You need to study how video signals are formed... Even if you draw every other frame you still want to use a sawtooth.

You use a sawtooth wave because when you're drawing the picture you need to be slow enough to keep up with the color signal, but on the return scan you can move as fast as you want. All video, whether it's NTSC, VGA, or High-def, uses a sawtooth for both the horizontal *AND* the vertical sync.

In a standard monitor (or a TV, for that matter), you have two oscillators running. One controls the electron beam moving left to right. It's a sawtooth waveform, so it moves smoothly until it gets to the right, then while the beam is shut off it is zipped back to the starting position on the left. The other oscillator controls the vertical steering of the beam. It moves from top to bottom, again very slowly to give the horizontal beam time to draw all the way across the screen. This means that the lines that make up the screen are actually drawn on a slight slope from left to right. When the beam reaches the bottom, the vertical oscillator ramps back up to the top. The funny thing is that during this vertical retrace the horizontal oscillator is still running, so if you invert the intensity signal to the electron gun you can see the retrace line zig-zag it's way up the screen!

This is why a polygonal mirror works so well for raster scanning. The trace from left to right is at a set speed, but the beam returns nearly instantly as the corner passes the beam and the next facet comes into alignment. It's just about the fastest sawtooth you can get.

AOM's will work even better, since they are several orders of magnitude faster than standard video signals. (Even VGA only runs at 31.4 Khz on the horizontal.)

Adam

Nope. You don't want to use a sawtooth in this case. A saw tooth is used with TV because there are no physical limits and the video stream comes in the correct order. But if you are going to send the images to a scanner you don't have to do it that way because you can capture each frame, digitize it, and ship it to the galvos any way you please. It would be silly to start at the top each time because of the second frame you can just draw the image from bottom up.

The vertical scan rate isn't the limiting factor, the horizontal scan rate is. Even if you wanted to do 60 frames per second, you could still scan from top to bottom just like a TV does. So scanning top-to-bottom-to-top doesn't buy you anything.

Futhermore, there is a good reason for scanning top to bottom. Persistance of vision. Since the image will be displayed at a greatly reduced frame rate (because of the limits on the horizontal scan frequency) you want to always scan the same way. If you do frame # 1 top to bottom and frame # 2 from bottom to top, then you can develop flicker at the top edge of the picture because it will be dark for a longer period of time than the bottom will. (This will be reversed when you start displaying frame # 3.) The end result will be flicker at the top and bottom of the image.

This is a fundamental part of video; after all, there's no electronic limit that would prevent a TV from scanning top to bottom to top... It's all about consistency... About maintaining a constant brightness to the eye to limit flicker.

Adam

Hmmmm yea, i see your point.

AOM, nice idea, you can enlarge the whole thing with a lens later. So long as you don't have a lot of overlap between lines or excessive aspect ratio error, correction is doable, even easy. Although people who have tried might tell me different.

I suggested the sawtooth on vertical because it's easy, but mostly because that way the division of 1 by the frame rate is observed for all parts of the frame so any 'point' in the frame gets updated at equal rate and step at all times. With triangle in horizontal that's not entirely true but the time difference is very small, while a bigger problem gets solved, no instant transitions full scale on the X axis at high speed. Even an AOM might benefit from that given the extreme speed of that transition if sawtooth is used, especially if by doing it, you can push an AOM into wider deflections.

Main problem with triangle X axis scan is the scanned lines aren't parallel, which might be annoying.

I have a resonant galvo in my desk drawer, it resonates at 1/2 15570 khz, go figgure, so you scan in BOTH directions, and voila, you have laser video. You need to do some linerizing cause its a sinusoid, but its much easier then syncing a spinning polygon with 32 faucets or trying to get wide angles using lenses for a AOBD.

and yes, you can raster with regular galvos, it flickers somewhat.

heck for 7 kilobucks you can buy the whole laser video AO system from intraaction, seen it at a conference with a 100 mW yag running daffy duck cartoons off a vcr.

Steve Roberts

Would it be possible to use two spinning wheels that were geared together so that one spins a lot faster than the other? One wheel would give X and the other Y.

The only issue with gearing is any time you add mechanics into a system, you are going to get slop (especially with Gearing)...The whole system will bog down greatly.

Not to mention the fact that the scan frequency for Y is very low. Even at broadcast quality video you're only talking about 60 frames per second, so that's 60 hz on the vertical. No problem doing that with a galvo, so there's no reason to go with a clunky mechanical system.

On the other hand, the horizontal is running at 14.7 Khz for NTSC video. (31.4 Khz for VGA). That's a couple orders of magnitude faster than what most galvos can handle, which is why the spinning polygonal mirror becomes viable. It's a kludge for sure, but it *will* get the job done, whereas a galvo won't.

However, from a purely theoretical point, you *could* do it with two sets of polygonal mirrors that were geared together. It would be sloppy, but it would work.

BTW, one thing about the spinning mirror trick. If you have a mirror failure, it's pretty destructive. I had a grocery-store scanning mirror wheel fail on me a few years ago, and it was amazing (and scary) to watch. The mirror cracked, flew off it's mount (thus destabalizing the balance of the rotor), and in a matter of seconds the rotor tore itself completely apart - sending mirror fragments and rotor pieces flying in all directions. Lucky for me I didn't get hit by anything sharp, but I had to extract glass fragments from the sheet rock in the room after it was all over...:o

Adam