Getting More Power and Cutting Accuracy Out of Your Home Bui

[dirktheeng]

I was going to ask this on Dirk's build thread, but I'll ask here!

What's the significance of tracking direction?

Strictly speaking, you could easily do PPI without having such high bit counters or trackign direction, but I built this with engraving in mind which will be much eisier and faster by having counters to count the entire length of each axis both up and down (which means I need to track direction).

[dirktheeng]

All,

when i wrote the article I was on vacation visiting family. I didn't have acess to my CNC computer to get the latest code, but it is up there now. The biggest difference is in the distance traveled function. Without the serial communication on, the loop executes so fast that the algorythm thinks that the laser head has stopped between pulses when it is supposed to be on and all it does is flash the laser at incredibly high rates when it's on. The code allows the loop to execute a bunch of times b/4 saying that the laser has stopped. Still fast enough to avoid a "drill hole".

[dirktheeng]

All,

I did a bunch of testing with acrylic cutting today with the new PPI system. I am 100% convinced that it is much safer. I could not sustain a flame on the underside of the cut as long as my duty cycle was low.

I did several tests. The first tests I did was to take the table out and secure the acrylic to the table frame. I tried to get a spontaneous flame by setting the PPI pulse width very high and the power to maximum and changing feed rates. I could not do it. I got a bunch of smoke, but nothing ignited on its own. (this does not mean that it could not happen, just that it is not likely).

That really confused me because I would almost always get a flame with the continuous on control and my old table. Setting the pulse width to something like 1s is the same as full on with a PPI control, so that should have done it. So I thought, well maybe the danger comes in when the laser is cutting and hits the table or something stuck to it. I put my new table back in (which is one of the aluminum light vents) and tried again. On a clean area of the table, I could not get a flame. Then I set up the cut to go over a piece of plywood that was stuck to the table from all the cutting I did for Christmas presents this year (I still hadn't cleaned up everything yet). Sure enough, I got ignition and the flame was sustained as long as the laser was on.

Next I set the PPI settings back to what I would normally use to make a cut, 3ms pulse width and some reasonable PPI and feed rate (like 400 PPI, and 200 mm/min). I passed over the same area and when it hit the wood, the flame ignited, but did not sustain after the laser moved off of it. Further the flame did not seem to be very hot as it did not produce any bubbling of the acrylic on the bottom like it did before (it seemed to be pulsating and not like a candle flame like before). I tried with both 1/8" and 1/4" acrylic, same results.

Moral of the story: when cutting acrylic use a non-flamable table material (like aluminum) and always clean any gunk off the table from cutting things like wood. Always clean up cut pieces that fall through the table as the laser can burn those too. Set up a PPI system either through EMC2 like Ben or by hardware like I did. Run the material at a slow feed rate with PPI to reduce duty cycle to keep flammable gas concentrations low and reduce heat input to the material.

PS: I ran 1/4" acrylic at 50mm/min, 1200 PPI, and 3ms pulse width and got some stunning cut quality results. Sides were nice and smooth and there were no bubbles from overheating the material, not even in sharp corners. The only distortion on either surface was very small marks from where the laser crossed a rib in the table. The piece came out with no effort and was so tight and accurately cut that it squeaked as I took it out. It was done in a single pass and there was no flame problems. I am continually amazed at the difference PPI makes in cutting. I spent an entire day trying to get the settings right with just "on/off" control and I couldn't even come close to these results, especially with one pass. Granted the feed rate is slow so it takes time. That said, I couldn't even think about running the laser that slowly without PPI as it would just melt/bubble the plastic. To get good results with PPI, raise the PPI, lower the pulse width, and lower the feed rate.

[dirktheeng]

This is a post by educa that will fit better here. We were talking about safety and related issues to cutting acrylic in Mark's thread. I will respond here to keep Marks thread on topic.

Hi Dirk

This is my first post here on the forum, but I intend to post quite a lot.

I am looking for info to build my own laser cutter just like you guys all do.

The aim is to build a laser which accepts 121x61cm wood plates and acrylic, so its a rather large one. about 4'x2'

Currently I don't know if I'll go for 40 , 60 or 80 watt. The idea is currently to go 60 Watt from cnccoletech.


I found info about your PPI system and will try to understand your exact implementation. I'm also a atmel avr programmer (sometimes also arduino, but if needed I do also plain AVR which is much more powerfull for timing stuff)

If I understand right , then you have found out that the ignition of a laser is a lot more powerfull then the power it puts out in the rest of its burn cycle, so its better to ignite the laser for very short (3ms) pulses and instead of cutting you actually are perforating a row of holes... making the laser work less, provide more power and cut more clearly.



If I calculate right, then when you set a feedrate of 50mm/min at 1200PPI and a pulsewidth of 3ms, then the laser is only burning in about 118 milliseconds per second it is working and the other 882ms it is not burning at all. That means the laser is only on in 13.4% of the time and the rest of the time idle, making it run cooler and in meantime more powerfull ? Am I right with my calculations?


Now can I ask you how many steps has your machine for moving 1 inch in X/Y ? I suppose its the same steprate to ease things?

Also, I can clearly understand that, for example if you have 2000 steps per inch, you can count the steps on X or Y and know exactly how far the head has traveled, but what do you do then with the following example?

1) If X increases by 2000 steps, then the head moved 1 inch

2) If Y increases by 2000 steps, then the head moved 1 inch

3) If X AND Y increase by 2000 steps, then the head moved by 1.4142 inch !!!!!!!!!!!!!!!!!!!!!


Does your system count for that too ?

Kind regards,

Bart

Bart,

You are basically right with your calcs, but your expression of duty cycle is not quite right. At 50mm/min and 1200 PPI, the period between pulses is 25.4 ms. With a 3ms on time, the duty cycle is 3/25.4 or 11.81% So yes, the laser will run cooler even with higher average power per pulse obtained through pulsing.

With regards to your 1), 2), 3) scinarios. Yes, it will handle any mix of X/Y signals. I do this by using pythagorean theorum when I count total distance traveled.

I am very glad that you are thinking of doing a laser project. It is very fun and you will learn a lot! I would definately reccomend going to a PPI system though, it is safer, more efficient, and more accurate. That said, you may want to wait to impliment this until I can move this over to an FPGA. I'm ordering one this week and will begin to develop the cores needed to to both PPI and laser engraving in one system.

[dirktheeng]

All,

I began to mess about with getting good PPI results and try to come up with a recipe to get the best possible results with about any material out there.

With PPI, there are a couple things that will result in bad results:

1) supplying too much power to a material
2) not having high enough PPI settings

Let's look at each of them individually:

1) too much power: When cutting a material, the energy from the laser gets split between vaporizing the kerf and heating the sorrounding material. Generally speaking, supplying too much power causes your cut edges to be sub-optimal. This is espeicially difficult with plastics such as acrylic as they have a tendancy to liquify/sublimate/boil near the cut if the power is too high. There is a finite rate at which energy can be disipated throughout the material as it is being cut and a finite temperature above which bad things start to happen. This is different for each material. Theoretically, the best cut will result when the power supplied to the material is near zero as any material not vaporized will have a very long time to disipate the heat. Practically, this is undesirable as the cut speed would have to be nearly zero. That said, one wants to dance on the edge of disaster, so to speek by supplying enough power to cut at practical speeds but not enough to damage the material while cutting by overheating. This is why PPI is great because you can very tightly control the power going into the material and keep it consistant and low over a wide range of feed rates.

That said, you can also control the linear energy density while keeping the power put into the material constant!!!! The linear energy density is strongly correlated to the cut depth. You do this by inversly controlling the relationship between PPI and Feed Rate. That is, if you find a good Power level for cutting a material, you can maintain that power level and increase your cut depth by increasing the PPI setting and proportionately decreasing the feed rate. What does this mean? Consistant results between various thicknesses of the same material. That is truely awesome!

I found that acrylic cuts well with a delivered power of less than 4.7W or so and even in sharp corers it doesn't have issues with overheating. At a 3ms pulse with 1200PPI and a feed rate of 50mm/min, the linear cut energy density is about 5.67 Jouls/mm cut. That was enough to cut 1/4" acrylic well with very nice results.

2) not having high enough PPI: The PPI settings can effect the smoothness of your cut edge. I found that 1000+ PPI for acrylic gives VERY nice results. I'm not sure why yet, but these settings seem to produce even smoother edges than just on/off control.

So, how do you get your PPI settings right without knowing what to set them at beforehand? Well... a bit of guided trial and error. Here's what I have tried so far

1) Set the power to 100%, the pulse width to 3ms, the PPI to about 400, and the feed rate to about 400
2) Test cut a piece. Look at the cut. If you haven't cut through the piece, double the PPI and halve the Feed Rate. This will give you higher linear energy density and maintain the same power setting which results in deeper cuts
3) Test cut another. If you got through, look at the edges, double PPI and half the Feed Rate and try again. If the edges look burnt or too dark for your liking (wood) or show signs of melting/bubbling especially around sharp corners (plastic), reduce the feed rate. This keeps the linear energy density the same and reduces the power put into the material.
4) Once you get the material delivery power set, look at the edges. If you want smoother edges, proportionately increase the PPI and decrese the Feed Rate. Do the opposite if you want rougher edges. Keep in mind that this will affect your cut depth. If you speed up the Feed Rate, it will not cut as deep and may not cut all the way through if you are workign with thick materials. If you want smoother edges but don't want to decrease your feed rate, you can try to mess with the power setting and decrease that as you increase PPI. That also helps decrease power delivered to the material, but the relationship is not linear because you are now fiddling with the transient response of the laser. Generally, I don't mess with that and leave it set to 100% power unless I am working with thin materials.

I've attached a spreadsheet that does the calcs for you so you can mess about with it as well. Nothing really complex, just som simple calcs.

[educa]

Ow,

I calculated if the head moves 50mm/min, then that 50/60 = 0.83333333333mm/s

So that is 50/60/25.4 = 0.0328083989501312 inch per second

That results in 0.0328083989501312 * 1200 = 39.37007874015744 pulses per second


If each pulse lasts 3ms, then these 39,37007874015744 pulses per second make the laser go on for 118.1102362 ms and the other 881.8897638 ms its off.

So I had to divide 118.11 by 1000 and not by 881.89, that was indeed an error by me.


You system really looks nice. However, I must say 50mm/min or even 200mm/min is quite slow. Does the system also work if you go ut to for example 600mm/min ? Thats more like 10mm/sec and thats a speed which my cnc router also likes.



Can I also ask you what this does about the width of the cut? I am currently reading your whole buildlog (so I'm still ok for some hours of reading fun), but I dunno yet what laser wattage you have and also not what focus distance.
I understand that for a deepre depth of cut, i need a longer focal distance, but that will also make my cut wider. For engraving its then better to have a shorter focal distance? So what would be the best compromise then ?

By the way, I converted my CNC machine to laser engraver using a 0.7Watt diode laser some weeks ago and posted a movie on youtube.
The GOOD result combined with the ultra low speed of a laser diode of that power rating makes me want to build the ultimate co2 machine.

Movie:

[educa]

Dirk, you say above

3) Test cut another. If you got through, look at the edges, double PPI and half the Feed Rate and try again. If the edges look burnt or too dark for your liking (wood) or show signs of melting/bubbling especially around sharp corners (plastic), reduce the feed rate. This keeps the linear energy density the same and reduces the power put into the material.

Don't you mean that if edges look burnt, you have to increase the feed rate instead of reduce? Won't reducing feedrate provide more power to the material?

[dirktheeng]

Ow,

You system really looks nice. However, I must say 50mm/min or even 200mm/min is quite slow. Does the system also work if you go ut to for example 600mm/min ? Thats more like 10mm/sec and thats a speed which my cnc router also likes.

Can I also ask you what this does about the width of the cut? I am currently reading your whole buildlog (so I'm still ok for some hours of reading fun), but I dunno yet what laser wattage you have and also not what focus distance.
I understand that for a deepre depth of cut, i need a longer focal distance, but that will also make my cut wider. For engraving its then better to have a shorter focal distance? So what would be the best compromise then ?

Thanks, I have put a lot of work into it and now it is paying off.

The speed question is a loaded one. There is no simple answer to that. You posted this before I could finish the post about getting good results with PPI. I can certainly make cuts faster, but nothing comes free. Getting good results with acrylic can be tricky because it melts at a low temperature, which means you can't put much power into it and expect the edges to be nice and crisp without bubbling/melting. Unfortunately, this means going slow. This is also why acrylic doesn't machine well... if you aren't careful you'll melt the chips that come off and end up with terrible cut results. If you can increase the heat transfer and disipation of the system, then you can cut much faster.

I can also cut acrylic faster, but it is not reasonable to expect such high quality edge results.

Also, you really have to watch out for sharp corners. The power delivered per unit volume of material is high at those points even if the cut energy density is linear. Accelerating into and out of corners will help with that.

With other materials, like plywood, I can cut very fast. I routinely cut 1/8", 3 ply baltic birch at 400-800mm/min.

Going to a higher power laser may help. You really want to get a faster responding laser though to get speed with PPI. I can't go much below 3ms pulses with a DC discharge laser because it doesn't turn on fast enough. Getting shorter pulses with a higher power laser would be beneficial to cut speed and power control with PPI. You would need a Q-switched laser and/or an external AOM to get much faster than what you can do with the DC discharge, but you are talking serious money for one of those.

With regards to focal length, it isn't that important. I go with short focal lengths because I can rely on self focusing in the material (that is the laser interacts with the off gasses to self focus) and that keeps the beam very narrow throughout the cut. A few guys at the lab I work at didn't even use a focus lens to cut 2x4 lumber during tests and the kerf width away from the impact point was consistant and narrow. It's a weird thing, but it does happen. I have found that the beam actually gets significantly less than 25 microns as it pierces through the material even though the beam spot is 150-200 microns.

[dirktheeng]

Dirk, you say above

3) Test cut another. If you got through, look at the edges, double PPI and half the Feed Rate and try again. If the edges look burnt or too dark for your liking (wood) or show signs of melting/bubbling especially around sharp corners (plastic), reduce the feed rate. This keeps the linear energy density the same and reduces the power put into the material.

Don't you mean that if edges look burnt, you have to increase the feed rate instead of reduce? Won't reducing feedrate provide more power to the material?

No, you want to decrease it. PPI is somewhat counter intuitive. If you decrease the feed rate with PPI while holding everythign else constant, you lower the duty cycle. Since the pulses are the same width in time, this reduces the net power going into the material. If you want to increase the feed rate and keep the power the same, you have to proportionately decrease the pulse width, but you can't get it much below 3ms without q-switching the laser.

You could also just reduce your power setting on the laser, but that is a bit more tricky. Plus, if you double the feed rate and half the power setting, your cut energy density goes down by a factor of two. You really want to try to lower the pulse width.

[educa]

how many watts laser do you use? Where bought?

Acrylic sounds nice and even at lower speeds (I understand you a very happy with the ppi cuts of acrylic on your machine?)

Also, if you can cut 1/8" plywood at 400+ mm/min then this is also with PPI ?