Laser Projector (Red, Green, Purple) "Laser 3D Party Light" from Spencer's

[tribble]

OK, more teardown pics. I know DogP already posted a nice set of these, but let's just consider this another sample for comparison.

First up, galvo insides and plugs:



Scanner amp bottom:


Scanner amp Y half (both halves look identical):


Scanner amp top:


MCU board:


MCU closeup: Maybe someone can put some of those numbers to use?


.

My other big project for the day was replacing the violet diode with an Osram PL450B from Jordan DTR. The PL450B is a 3.8mm diode; SO TINY! Thankfully, DTR sells these mounted and wired, and the mounts are a PERFECT FIT for the Spencer projector diode mount heatsinks. This is a DTR 3.8mm mount in the stock heatsink/mount:


I LOVE THE PL450B BLUE. Can I say that enough times? I LOVE THE PL450B BLUE! So much nicer than the 405. Leaps and bounds better. Makes a nice crisp white, and no resistor surgery required. The red seems to be the weakest of the lot; I guess that's just the way eyes work. I was able to get a nice balance by turning the red up as high as it would go, then twekaing the green and blue to make a nice crisp white. Aside from an alignment, I think replacing the 405 with a better blue is the biggest bang-for-the-buck improvement we've found to date for this little guy. DogP, would you agree? That is, until you find some magic that you can work in that galvo amp. :-)

Sadly, all of this activity means no time for additional scanner tuning or post-tuning test image pics. I will try to do those in the next few days, incorporating any galvo-whisperer tips Steve can pass on. I read the Adam Burns tutorial maybe a dozen times in places, but I sense there's more to be learned.

'night... I'm pooped.

[tribble]

Oh, one more thing. The stock aluminum LD mounts are screwed/glued to a PLASTIC adjustment mount. The play in the plastic is removed by the heap of epoxy-like hot-glue they use to lock everything in. I found that heating the plate from below greatly simplified removing the glue. I scored the glue between the plastic diode mounts and heated it up from the bottom and was able to pry it up with a carpet knife. The glue gets softer and less brittle when heated, and it lets go with a lot less effort and sudden jerking of tools.

Here is a photo of the bottom of the mount. You can see the two screws holding the aluminum heatsink; little to prevent forward/back rocking motion. Also notice the two plastic ears that keep the plastic mount from sitting flat on the aluminum baseplate; this is how they adjust the alignment, by tightening the front or rear screws as needed and then waiting for the glue to set! Left-right alignment is just play in the screw holes.



Here is the pool of glue under the mount. You can see it took the finish off one of the screws (the stripped one, BTW... quality through and through...)

[mixedgas]

Galvo wiring:

So let me do the educated guess, two wires for the drive coil, one for the led power , ground for the led is the cable shield, and the two photodiode currents run down a shielded cable?

Does that check out against the little PCB on the galvo body?

Would be very interesting to see if the unknown pots are the scale factor for the feedback.

Would you object to wiring a say 47 ohm resistor in series with the led and grounding the amp inputs so you could see if the mystery Pots change the led current?

My bet is a gentle bend of the galvo led leads with pliars (power off of course) shows up as a slight change in the distortion in the image. Blocking either of the light paths while running can have un-intended consequences.

Steve

[tribble]

I'll see if I can find a way to measure the IR LED current without getting things in to a state I can't clean up... :-)

Do I need the 47 ohm resistor for some specific reason, or can I just measure the current directly in line with my DMM?

I could not see the galvo PCB traces by just taking off the back cover. If I pull out those three screws visible in the picture, am I going to be able to put things back together? Or is there any super-sensitive stuff hanging out in there that will surprise me? I've only had this thing for a few weeks, and it's a joy as my first projector of any sort, so I don't want to kill it. :-)

[tribble]

Oh, I forgot to mention on the galvo amp board... all of the pot adjustments are "Increase == clockwise."

[mixedgas]

OK, looking at the board.

Any way you can check the LED current is fine. First Do No Harm is a pretty good rule.

They do have 8 op-amps per channel, so if they use them all, its not a bad design. However if 2 are burned doing difference inputs, that might be cutting it close on offsets in the feedback loop.

The TDA2030 is a 14 watt driver. That is one speed limit. I do see a snubber on the outputs, which is good. I do see a current sense resistor, which means you have both current derived feedback and the sensor feedback. The current derived feedback really helps with the coil inductance and slew rate.
The good news is the power traces are fat or even Phatt, more then generous.

The LM324 is not the op-amp of choice. However I see them all the time in low cost laser gear for two reasons. 1. cheapest op-amp available anywhere. 2. It will work off big supply rails without the voltage regulators other scanner amps need for their op-amps. It has wicked crossover distortion and wicked distortion near the rails, so that explains why your full field pattern has distortion at the top edge and some other places.

The lack of regulators makes me wonder about the led current even more. I suspect the two TO-22 packages are either VRs for a reference for the leds, or are transistors for part of the led current source. Whats the part numbers on the TO-22 three lead package?

What does "PS" mean on the abbreviations?

As I suspected, the Op=Amp supplies are not well decoupled. That explains part of the poor linearity, and some of the crosstalk. With the output stage so close, you really want to float the op-amps on their own clean supply.

See how much the last page of this lines up with what you have, which I guess is a highly trimmed 6850 clone.
Drawing number D006-1852 is what you want to look at.

http://www.skywise711.com/lasers/scanner/scanner.html
Thank you Skywise for hosting that for me all these years.
Ignore the LM22SDA stuff, it will just get you in trouble, the amp feedback loop design is way different.

So a lot of what I see in your image now comes down to the shaft core, magnetic system, and quality of the bearings.
Those are not sealed bearings, so the lube will evaporate over time. I'd make sure the bearings stay lubed with a tiny amout of oil say every six months or if you see the galvos really slow down. Has to be a really thin oil or it will cause severe bearing drag. If the oil gets up between the shaft housing and the rotor, the galvo is trashed, so the last thing you want to do is lube carellessly.

I commend you on a excellent first job of tuning a amp that has had 50% of its parts chopped out.

Amazing they could simplify it that much, which speaks highly of the Cambridge basic design. And speaks highly of the Chinese engineer, he chopped it about as far as it can be chopped.

Take a look at http://www.laserfx.com/Backstage.Las...Scanning3.html

PS, Next time you tune, disconnect the blanking and turn the lasers on with a pull-up or pull-down resistor. Or recolor the whole frame and save it as a differnt file. You can see a lot more with the ILDA pattern if you can see all of the hidden lines.


Steve

[tribble]

Whats the part numbers on the TO-22 three lead package?

They are our old friend, the 2N3904! :-)

What does "PS" mean on the abbreviations?

"picture scale." I used the same abbreviations from the chinese amp ebay pictures that DogP posted.

I commend you on a excellent job of tuning a amp that has had 50% of its parts chopped out.

I can't take credit for the tuning given in the pictures I posted at first; that was the factory tuning! When you see what I've done, you may not want to commend me any more. ;-) Where I am at now is better circles and linearity in my hypocycloids, but at the cost of some corner definition in the LaserMedia pattern. I am still going to have another go at it, and when I'm satisfied I can do no more, I will post those results.

Thanks for the other references; I will dig in to those when I have a chance. Holiday is almost over for me, and that means hobbies slow way down. :-(

[mixedgas]

Ah, the led currents will come from the 2N3904, as they sure as heck would not work for the core temp shutdown.
Steve

[DogP]

I uploaded all of the text pics, a text file describing which images are of what, and the .ild files for test images into this directory:

http://www.likesgadgets.com/489143/

I'm missing the .prg, which I already re-wrote, but I can re-create it and post that too. I bet your first guess about what it looked like is correct anyway. :-)

Cool... I'll give them a shot when I get a chance.

OK, more teardown pics. I know DogP already posted a nice set of these, but let's just consider this another sample for comparison.

Nice pics!

Aside from an alignment, I think replacing the 405 with a better blue is the biggest bang-for-the-buck improvement we've found to date for this little guy. DogP, would you agree?

I Agree 100%.

The LM324 is not the op-amp of choice. However I see them all the time in low cost laser gear for two reasons. 1. cheapest op-amp available anywhere. 2. It will work off big supply rails without the voltage regulators other scanner amps need for their op-amps. It has wicked crossover distortion and wicked distortion near the rails, so that explains why your full field pattern has distortion at the top edge and some other places.

I was wondering about the LM324s as well... in one of my previous projects, the LM324 was used (in an audio circuit), and sounded terrible. Simply having a better chip (uPC324) improved the quality. And I was able to even improve on that, swapping for some nice op amps from Digikey with a compatible pinout. But they were run off a single supply, so I went with nice single supply rail to rail op amps. This one is running off dual supplies, so I can't use the same ones (I think I had ordered a few bipolar ones to test as well, if I can find them).

But it may be worth simply swapping the op amps to something nicer, since they're cheap and easy to swap. I think being able to adjust the high frequency damping would be important as well though, since increasing the bandwidth of the op amp doesn't help if the signal bandwidth is limited by a LPF. The same thing was done on the LM324 audio circuit... since the op amps were so bad, they used a LPF that started rolling off around 4KHz. Increasing that, and installing nicer op amps made a HUGE difference.

DogP

[tribble]

By The Way... when I had the back off of the X galvo, I noticed that a strand of glass fiber on the edge of the blanking plate was loose and flapping around. It was just a thread... but I removed it. I have not seen the X-axis jitter since!