I intend to experiment with the topic but would like some input on the matter also.
When you do a ramp test the laser burns the entire length of the test line albeit different amounts of bad to excellent. When the laser moves over a bump in the substrate whether it is .5mm or 2mm the distance from nozzle to substrate obviously changes, normally the FL becomes shorter from nozzle to plywood in my instance. If I increase the power over my setting of Max & Min 35% to lets say 45% and instead of adjusting nozzle distance from the lowest point I adjust to the highest point, again supposing it wouldn’t be more than 1mm. In my ramp test for the 63.5mm lens and lens tube I had at least .75 inches of a fine line burn/cut. Wood I used for ramp was about 18" long, I only cut the test line 10" in length with the end elevated 5". Without calculating I think that .75" would translate to at least 2mm difference in distance from substrate to nozzle. I am using 10mm setting from nozzle to substrate. If I reset the distance from nozzle to 8mm shouldn’t I have a good cut over the 2mm bump as well the level portion of the plywood?
I may be all wet thinking this way, I am determined to find a way of cutting through the plywood even if it isn’t exactly level throughout.
If I am climbing a dead tree let me know.
I think most people do it somehow this way, either proven or unproven. If we take a 50.8 lens then we have a “focus depth” of approx. 20mm +/-. If we then say only 50% of it, 10mm, then there is still a lot of tolerance to cut 3-5mm material with.
A few mm difference of the perfect focus point is not the big problem, but of course not perfect either. But it takes some power, I could not have done the tests (so nicely) with my K40 and not at all with my 5.5 diode laser.
What is more crucial, in my opinion "is the combination of power and time.
The picture shows a small test with 6 different combinations that I can use to cut 3mm plywood with.
(the bottom line of text is only a line / font power / time test)
The back is without fire / soot marks, I used my magnetic punk spikes and “normal” air.
I have done (again) a little test 
to check what I wrote in my last post.
The focal point is on the surface of the 3mm plywood. It will probably be possible to continue in both directions, one or two mm more, but the usable 12 mm focus depth in this small experiment is impressive enough (for me).
But everything has its price, a beauty award I will not be able to win with the cuts and measure of the pieces also change along the way.
Within +/- 3mm there are no problems.
Bernd,
Thank you for your reply.
I cut two puzzles today. First measured 8" x 10". Used speed = 20mm/sec, Max and Min Power = 45%, 2 passes, nozzle distance set from slightly higher center of plywood @10mm. Center area did not cut completely leaving minute tabs on the photo side that isn’t worth messing with. Remainder of puzzle cut correctly. This piece of plywood had not been treated with water on the bare side.
Second puzzle measured 15" x 12". Speed 20%/sec, Max & Min 45%, 2 passes, meter reading 14mA, nozzle distance 8.5mm. I had sprayed water on the bare side of this plywood and allowed it to rest for about ten minutes. Almost perfectly straight and flat. Water evaporated about half. Placed in laser and ran it within next ten minutes after checking my settings etc.
All pieces fell free after the second pass. I had set the program to run three passes. I allowed it to run enough at three that I could see two were
sufficient. I then stopped the laser from burning on the third pass for most of the puzzle.
I am beginning to believe the plywood cuts better after I have sprayed with water. Maybe it relaxes, lol. I am using 5mm ply from Lowe’s with three plies of what they insist is Poplar, it definitely isn’t poplar. One side is luaun other, who knows. I haven’t cut across any voids.
Questions?
I think you got bad plywood and as you say yourself, it’s probably not poplar plywood. You should be able to feel it on the weight of the plates.
It does not sound quite right that more than one pass should be used, for 5mm plywood of a reasonable quality.
Your water treatment of plywood is interesting and I will also try it when I get very uneven material. Fortunately, this is rare because I have a good supplier of “lasser suitable plywood”. My plates are 600x400 or larger, if they are larger I cut them myself suitable for the machine or the tasks.
Tomorrow I will do some tests in 5 mm, I will probably also make a small puzzle to learn a little more about this topic
I cut that ply one pass without the photo
It is very light weight
Test of 5mm mixed plywood (poplar + BB)
The day started with a video from Russ Sadler where he tries his hand at puzzles. Although I admire and respect Russ a lot, his suggestion how to cut plywood puzzles is not exactly my cup of tee. It is possible that if many puzzles were to be produced, his basic idea of a suction and workpiece capture device could be developed. But there must be better and easier ways to do it.
However, today I did not manage to make puzzles, I have had a small production of “serious” items to supply.
I managed to do a material test in P / BB 5mm plywood.
The material is 5 layers, 3 birch and 2 poplar layers and weighs only 4.6g / cm2 (BB weighs 8.5g / cam2)
As with birch plywood, many variations of power / time are also possible with poplar plywood.
I chose a “standard” settings for me, for just this material I have in stock, which is higher than expected and also higher as needed. The reason for this is that even though pretty much all items fall out by themselves, there are some that hang in very thin birch wood fibers on the back. To make puzzles, it is not acceptable, there should be no finishing other than color processing if desired.
All subjects emit no soot on a white cloth although some look quite dark in the pictures, they are actually nicely light-brown.
I also tested different focus distances, 50.8, 53.8 and 47.8mm, measured to from the wood surface to the lens bottom edge. - It can be done, but the nicest result is achieved by the correct focus for a 2 "lens, (from the material surface to the lens).
The test is not entirely adequate but directional for me and the material I have today.
For puzzles, I will again set it up on my spike magnets and tape both sides. (if it is not a mini production 5+)
My “puzzle” - favorite
… there was just a little overlap
What are your settings and it is properly understood that you have a 60Watt tube?
The tube is obviously only 50 W
I have some things to do today. I will try to make some cuts this afternoon
… it’s most, but what did you buy it for?
60W Omtech
By measuring it is only 50W
Maybe 55W at most
I knew that when I bought it
It’s mine too. You can compare my power% with my mA and see if your tube is in the same area.
I often write down the two data, to keep an eye on my tube. It is already a little bit weaker than when I bought it, although I always take great care of my tybes.
I have been trying to get a clean cut with 20mm/sec ,Max & Min power at 45%, 14mA.
I will try your setting of 25mm/s- 60%- and see what I get for amperage.
I watch my tube values also along with the temperature the cooling water which I make it a point to not go over 70 - 75 degrees F.
I have a digital thermometer in the water tank. I haven’t come up with enough cash to buy a chiller yet. That is on my list.
I have choices slightly lower values, 18-20 degrees Celsius, but as long as it does not get above 30 degrees C it should be fine.
Ps. Beware of air bubbles !!
I have an obsession with the bubbles. I may restart everything several times until the bubbles are all gone. lol
…the same here
Bernd, you are probably in bed now but I will continue.
I received a small order of 3mm Baltic Birch and 3mm MDF.
Both cut well without my photos attached. Photo paper is commercial Epson Roll , it is heavy and I don’t remember the actual weight/thickness.
I am using 2 1/2” lens with suggested Nozzle distance of 10.5mm although I have experimented with 8.5mm which gave me excessive smoke and wide kerf on top side. Ramp test showed at least -1.5mm to + 1.5mm in a distance of 1”/25.4mm.
Center of that one inch measured 10.5mm from nozzle to substrate. I will run another Ramp test tomorrow. Have been experimenting with distance from photo side from stainless steel mesh I have resting on original aluminum bars.
I have one serious obstacle to overcome. When cutting a large puzzle some parts fall prematurely in a strip before they were completely cut causing some of the puzzle to not get cut completely. What I have found is the photo side of work piece must rest on the mesh throughout the cut process. Obviously that enables another possible problem, flash back from the mesh. I have two pieces of the mesh so I can have one in a cleaning process while using the other.
Omtech 60/50W
2.5” lens
Speed 20mm/sec
Power Max 55%, Min 55%
Single pass
Air Assist 25psi
3mm Baltic Birch and/or MDF (MDF smokes more than the BB.
I have a trial of aircraft plywood @ 3mm ordered, may not receive before some time in March.
Good morning 
It’s great that you put so much effort into getting a good result. Firstly, you learn a lot about your machine and the different materials and secondly, you should probably get the best possible result in the end. But not to forget, the purpose of this forum is this kind of discussions and exchanges of experience and to help each other, I enjoy it.
Just for clarification, 2.5 "lens with 63.5mm focal length, in a nozzle tube from Cloudray - I suppose, is that right?
I have no experience with it. Does photo paper on the top (or bottom) affect plywood cutting?
Could you add very small bridges to prevent the single parts from falling out ?, it is not optimal with any kind of finishing when there are so many parts, but if it is not for commercial use, it may be an acceptable compromise.
With 3 and 4mm BB plywood, I once did an interesting test to find the optimal way to cut this material. I found that I have the ability to cut very fast with lots of laser power or very slowly with very little power. The two main differences were, speed and how the surfaces looked. I will try again tonight to verify this little test. - I report back with the result.
Sorry, I didn’t mean to isolate anyone from the conversation.
Lens and lens tube from APC.
I have tried the photo paper on both top and bottom with best results when paper is on bottom.
I spent most of a sleepless night trying different ideas in my mind. The pointed apparatus won’t work since I need to support each puzzle piece. Going to try some different ideas for supporting pieces that won’t exceed .75 inches across both ways. Support will need to be metal.
I enjoy figuring how to accomplish something more than manufacturing but this is one time I am considering making for sale.
I have varied speed and power quite a bit. The need for “catching “ the pieces before they would fall to the mesh presents several unique problems of which I haven’t resolved.
It seems higher power makes for much darker edges while higher speed keeps the edges lighter. Definitely a fine line. I have run some tests with different speed and power. Need to do it again.
Thanks to you bernd, and all the other contributors to my endless questions, I appreciate very much.
I have done the little test that I talked about, and the interesting thing is that at very low speeds and equivalent power, there is the least soot on the subjects. I have not optimized the font, it actually soots more than the cut itself. The back is completely clean, with all settings. I’ve used my punk spikes magnets on the perforated plate. If I had to make a lot of puzzles I would probably make a fakirbed with very thin pins at the same height and with very little distance to each other.
(3mm normal plywood)