Scan Issue: Skew

[JasonB]

Hi guys,

I've been struggling with this for weeks. I guess you could say I am an experienced amateur and am really out of ideas. My new PT-20 scanners for a satellite beamer PJ project constantly project images in the form of a parallelogram. I have successfully home built another RGB PJ and was able to align it, it is possible something is off and I can't recognize it. This PJ does not yet have MM1s for alignment so maybe it is a physical setup issue. The PJ is projecting directly onto window shade with no angles, camera shot is taken looking at the image from a position directly above the PJ. I've swapped the DAC (Riya Light to Ethernet Dream), same issue. Then swapped the scan set (not amplifiers) with another PT-20 set that works great, same issue. I've triple checked beam entry and it looks square to me.

So what is left other than software, alignment and amps/tuning? I dug up my documentation on the PT-20 tuning pots and re-read Adam Burns scanning tutorial and Steve's procedure he posted on the scannermax thread. Thinking there is nothing to loose at this point it is now time to start turning pots to learn how to tune. It was pretty clear what the pots changed on the projected image but none of them would help square up the image.

Finally I switched from LSX to Spaghetti to explore different options and discovered that Spaghetti's skew correction when fully applied straightened out the image.... What does that tell us, any ideas? Could it be misalignment of the beam entering the scanners that I can't recognize? As always, thanks for the help.

-Jason

[Datsurb]

I did experience something similar, and that turned out to be the wireing. Not that i had connected anything wrong, but the cables i used was of poor quality, and i believe it was interference or inducted interference that caused the strange skeewing. I ripped out most of the cables, separated the cablegate for power and signal. The result was good. No more strange skewing after that.

It might be something similar for you.

[drm916]

Wiring, tuning or possibly the scanners are not centered in the block and properly positioned.

[absolom7691]

Be sure to check the ground on your ilda cable (pin 25). My scanners freaked out once before and I found that the ground wire on pin 25 broke at the solder cup on my db25.

[Bradfo69]

My suspicion would be a wiring or ground issue too.

I think the skew correction in Spaghetti may be something like the geometric adjustments you can make in Pangolin's software or what you would do with a UGC. Sometimes where the projector is in relation to the surface it's projecting on requires some fine tuning to make it square and "look right". You get that with video projectors and have a "keystone" correction, when perhaps the projector is physically much lower or higher than what it's projecting on.

I don't think relying on the skew adjustment maxed out is the best solution. You may really want to go through and check each connection and particularly ground connections as was mentioned above.

[tribble]

Silly question... does the image distortion change if you slow down the scan speed? It might be helpful if you projected an image we all know what it should look like, like the ILDA test pattern.

[JasonB]

Thanks for the ideas everyone. It was the wiring! All the comments about wiring and grounds led me to the problem. I had the XY signals to the scanners wired single ended rather than differential. When I chopped into the PT-20 cable I discovered that the 'black' conductor on the signal input connector for + - GND (red, white, black) was actually a heat shrinked shield. So I wired the X+, X- to +and - rather than + and gnd(shield&gnd).

Interestingly enough, I copied the wiring from my first PT-20 scanset which was wired RED and BLACK, white (-V signal in) was left unconnected. Probably because the scanner was set up for simple 1 V peak-peak for use on an audio output (1/8" phono connector) running some old Laser Show Designer 1000 software that outputs signals over the PCs audio output, and the connector was supplied using RED and BLACK going to 1 Vpp signal... So they used ground/shield as 0VDC.

One a more interesting note, the old PT-20 scanset tuned at the vendor for 1 Vp-p performs better than my new one in terms of image quality. The amp boards are a bit different (old one had larger heatsinks) but the components and pots are the same.

Soooooo.... on the old scanset is it possible that when the guy turned down the input gain to accept 1Vp-p rather than 5Vp-p he was able to tune the amps better because they are running in the small signal performance region? In other words, does decreasing the gain allow for 'tighter' tuning because the scanner throws are less?

Thanks again for the help guys... And I have convinced the wife that I am attending SELEM next year... Can't wait for that!

-Jason

[innazone]

Incredibly helpful post Jason, I had the same problem, wired s- to the black lead on a PT 40 by mistake and had a skew which is now solved.

Nick

[JasonB]

innazone, thanks for the post. It is nice to know that following through with a solution post on an issue helped someone else.

-Jason