RDWorks Learning Lab 117 Hunting those Phantom Pixels
Vložit
- čas přidán 23. 07. 2024
- If you own a Chinese laser cutter this little series of videos about me learning how to use the free software provided, may solve the problem of trying to learn from a virtually unreadable manual.
I am nothing to do with RD Works, I am not an instructor and I am no expert. This series will document the essential bits of many hours of trial and error - Věda a technologie
Great videos and a great teacher, audio is clear and crisp and the videos are rock solid. Thank you for all your time and effort it took in creating these video tutorials.
It's been a while since I have thanked you for your superb series on this laser cutter. I have learned much, and have tried many of your ideas. Thank you!
Hey Russ. I spent some time chasing a similar niggle i had with my my machines, the cause of the striations that the laser leaves on the edges of the material being cut.
it was suggested that this was also caused by steppers or belts or resonance, kerf etc, however in my case I ruled out the steppers/belts/pulleys by simply disconnecting the belt and moving the head manually with a weight and a pulley to get a constant speed held the pulse button and released the weight,
I found that the striations were still there exactly as they were when cut using the stepper, I also ruled out it being kerf marks caused by material removal in my mind as they were identical on both wood and acrylic, and the striations were the exactly the same if the material was cut clean or only 50% through
Granted it only gives you a single line to examine and not a block engraving like in the video but it does help to eliminate or not the steppers/belts and pulleys in one go.
From what i read at the time the steppers do have resonant frequencies towards the slower speeds depending on the size etc and that's where the reduction drives come in like on the lightblade allowing the steppers to spin faster avoiding the juddering or resonance of slow movements and speeds.
i'm interested to see where this rabbit hole leads you as I believe it will answer my questions as well
good luck
You really should get a prize for the level of detail which you're going into here! I was interested to see the "curtain" effect you described on the background of a rastered area - I've never experienced this (but then I haven't done much rastering), but I *have* observed a seemingly gradual increase or decrease in depth as the rastering progresses. For example, suppose you rastered a solid square at 100% power - the rastered area might be deeper at the back than at the front, as if the power is slowly drifting. I've meant to investigate this further, because it is a bit of a pain, but never got around to it. It's the main reason why I decided to keep my older laser cutter, one of the really basic 40W 20x20cm machines, because it always did perfectly uniform rastering.
Another strange effect I noticed tonight is that, at least with my tube and the lifetime it has, 50% power seems to give more power than 100% power! I rastered two areas with the different power levels, and that done with 50% was considerably deeper. Extremely odd. My only guesses are the tube is coming to the end of its life and is behaving weirdly, or that the higher power is exciting a different mode and is no longer producing a perfect Gaussian beam.
It's a great pity there's no detector which both is sensitive to 10.6um and also has a fast response, so you could actually look at the laser power in real time....
Hi Lindsay
I have dropped you an email on this power issue but could you also tell me if there is a makers label on your tube and what it is?
There is a way to monitor power continuously but the answer is not in watts.
see czcams.com/video/2SaOo5nFGrU/video.html at 25 minutes The power across the beam is not uniform but a Gaussiin type shape. At the edge of the beam the temperature is fairly low. If you clamp your themocouple into a fixed position then you can use the temperature at that point as a means of checking power variation live without significantly affecting your output power.
Best wishes
Russ
Amazing Work Russ! I subscribed to your channel because of the CNC machine, and although this still interests me, your systematic analysis, logical detailed characterization of behavior and brilliant problem solving skills are fascinating and THE MODEL. I agree that the stepper motor could be the source of the behavior, but also think it is possible that the resonant characteristics of the system beyond the driver could cause the behavior - even with perfect logical stepper motor behavior. Back to the stepper motor theory, one question that I had was whether all steps are equal when microstepping - specifically, i wonder if the in-between steps generate the same torque - and if not, perhaps this could be a contributor to the resonance anomalies. Fascinating stuff and thanks for sharing this with us!!
Hi Steve.
My goal is to understand the operating boundaries for the machine rather than to pimp it up. The fact that I used the new belt as a tool by which to prove a point and that it has also given me and improved performance, is an accidental win. It means that I can now avoid that narrow band of jittery performance. So from that point of view the work has been a great success. There are still some unanswered question which I will explore in the next session. Will it fix my curtain effect? will it help my 3D engraving lines?.
It's is really an irrelevant question about micro stepping performance because although I would like to know and understand the theory behind when and how it changes from micro steps to whole steps. Or does it gradually change from x10 to x8 to x6 etc until it drops into whole steps? I say irrelevant because it does not change the parameters that I have now mapped.
Still lots to learn and fun to be had.
Best wishes
Russ
Nice video Russ.
I have been trying to eliminate the wavy bottoms on deep pocket engravings for a couple years now. The best I have found is to put a pair of lines about 20-40mm outside of my photo so the head is a more constant speed. It reduces the amplitude of the waves on the ends, but it's still there. I have my fingers crossed for finding the true culprit, at this point I'm pretty convinced that would require a motor/controller overhaul to a servo system. I hope that I'm wrong and there is a simpler solution, but time shall tell. In the mean time here's to burning away more material.
Today I did some testing with regards to vibrations in the head. I found that I got much less vibrations when running the speed at a multiple of the step length. I think it's because the controller isn't trying to take fractional steps. For example my step length in the x-axis is 11.25 um, so I ran some test cuts with the speed at numbers that were multiples of that, and speeds that did not divide evenly, and each speed that was a multiple of the step length resulted in a much less vibration than speeds that were not multiples. I think the fractional steps could explain the cyclical vibrations. Might be worth checking out, and the math is simple: eg. speed 45mm/s devided by the step length 0.01125mm = 4000 steps. But 40mm/s divided by .01125mm = 3555.5555556. So when it attempts to output the fractional steps it would cause a more twitchy response.
Then again, there are a lot of variables involved, and it could just be coincidence for the speeds I tried... and my step length is a number that can go to whole numbers pretty easily so it may not work for everyone.
Very nice work as always, Will it help if you draw the lines using one direction only in order to eliminate the vibrations that occur when the direction is changed?
I think if you add some delay before starting every line it can help in damping the vibrations at the beginning and eliminating unnecessary vibration that could affect the accuracy of your tests.
Hi
I have already answered this question once and it seems that You Tube has posted it to some parallel universe!!!
I purposely chose bidirectional scanning to copy my previous work. The direction of scan is obvious because the burn spikes are not symmetrical and have motion characteristics. Yes, there is a small offset between the scans but this is not a problem. There is no vibration at direction change just deceleration and acceleration. The overshoot after the beam switches of is already quite long and is proportional to the run speed. For example at 20mm/s its about 5mm and at 400mm/s its about 50mm. That's plenty of time to get rid speed variations and you can see that in the uniform nature of the speed tracks burnt into the acrylic. I will try unidirectional scans in the next session because I like the pretty gear tooth pattern that it produces and will make it easier to determine the frequency of the stop/start speed spikes..
All best wishes
Russ
Hi Russ, here is a great article that describes the operation of the stepper motors, this one refers to using 10 microsteps and you will see the reason why it is not advantageous for accuracy to go above this for basic steppers. www.geckodrive.com/support/step-motor-basics.html to keep costs down (Simpler circuits), most manufactures choose to use binary multiples like 2, 4, 8, 16, 32 etc. more microsteps will give smoother operation but at the cost or torque but will not get more accuracy
it also describes issues at low speeds so may help you in your search for answers.
also, one thought to add is that at low speeds, the stepper will move to its next position far faster than the actual mm/sec you have dialed into your test. but then the controller will wait an appropriate amount of time before issuing the next step. at higher speeds the next pulse will come before the motor has stopped at its new position, producing a smoother rotation, the rotor will follow the moving magnetic poles and accelerate to keep up with it. again, at the lower speeds the motor rotation will easily move to the new position before the next pulse arrives.
So until your pulses (Step time) exceeds this fast move, the motor will always stop or at least slow down between steps and in reality will not move faster until it is exceeded
Regards Peter from czcams.com/users/thebreadboardca channel.
Hi Peter
Thanks for that link
I think my drives are set up with X10 multipliers , so 2000 steps per rev. The goal in all my explorations is to establish operational boundaries rather than to upgrade the machine. I am trying to wring all the possibilities out of this machine by understanding the minutia that others are not bothered with. Now that I have bounced off it, I am interested to learn more about micro stepping. I have made logical assumptions about what I have seen and heard but it does not really matter if I understand because the important thing is that the new belt has firstly proved the point by changing the low speed characteristic and secondly created a well defined jitter zone that I can now steer clear of. I still have work to do in the next session because I have the Lightblade machine that operates with vastly different mechanics. My China Blue machine has the X belt driven directly from a 16 tooth pulley on the stepper. By default, slow mm/s = VERY slow stepper speeds. This means lots of micro stepping. The Lightblade has compound gearing between the stepper and the X belt so my guess is that the stepper may already be above that critical jitter zone for all but the VERY slowest speeds.
The point you make about the rotor waiting for the next signal pulse at slow speeds is clear to be seem in my acrylic speed traces
Thanks for all the great information
All best wishes
Russ
I'm curious if an ink pen mounted to the side of your laser head would draw the same cavitations? Fluid dynamics seems to be a possible explanation to the ripple.
Sadly an ink pen has no 3D abilities. Drawing a straight line that stutters will still only draw a straight line.
I know the patters I produced might seem like a wave patterns, but there is no fluidity involved in their production. It is a shape carved away by simple sublimation. The vapour is at the surface and can immediately escape unimpeded. The idea of fluid dynamics is NOT ridiculous though, because if you captivate expanding hot acrylic vapour in a long thin tube you do get some interesting fluid dynamics going on. However, even better is the fact that acrylic records those effects for you to see. Long Live Acrylic....the best test material ever. see czcams.com/video/bnwXuJz-tXg/video.html
Maybe a nutty idea: take the motor out of the equation - pulse the laser while manually moving the laser head with your hand. You might even be able to get the head to start sliding and release it and let it slow down on its own. Or, pull the head with a weight and a pulley or something.
Fascinating video as always, thanks!
Of course, the belt would probably need to be disconnected.
Hi Hirudin
Others have suggested similar and initially I thought it would be a good idea. Another guy had disconnected his belt and mounted a pulley over the edge of his machine so that he could hang weights onto a cord pulling the head.
As you would expect , no cut variation and definitely no "dwell spikes." Interestingly he reports that it did not cure the striations on the edge of acrylic. I am already 100% convinced the marks are caused by steppers non-linear micro motions at certain speeds. By removing the sponginess in the system we can now see the distinct performance band where this occurs. There is still the possibility of other resonances beyond the belt but I'm ending my chase there. My goal is to set performance boundaries for this machine and not to alter/upgrade. Unfortunately I broke that aim here by adding a new belt, although I excuse that as an exploratory tool rather than an upgrade.
Thanks for the contribution as ALL thoughts and ideas are valuable because even the wacky ones can spark inspiration. That is one of the reasons I share my methods and procedures as well as the successes and failures. If my logic is awry then there will be help out there to teach me.. Remember I'm climbing my way up the knowledge ladder but I'm no expert, I still have a lot to learn and UNDERSTAND....they are not the same thing.
Russ, why not disconnect the X axis drive so the laser head remains stationary and move the acrylic by another means, perhaps manually, to determine if the pixels go away. My non-China blue machine does not produce these same artifacts even with stepper motors and glass reinforced rubber timing belts. Just a thought but it may help answer some questions and bring other questions to the surface.
This China Blue machine is a direct drive to the X drive belt via a VERY small 16tooth drive pulley. By default that means that slow speed means VERY slow rotation of the stepper. It is not until I get beyond 60mm/s that my jitter disappears. If you have a Red Sail clone design where there are intermediate drive belts between the stepper and the X belt, then your stepper is automatically running 6 to 10 times faster to start with and you may never be in this jitter zone. My Lightblade machine is just such a design and I will get round to doing the same tests on that soon. AS you rightly say, that may raise it's own questions.
All best wishes
Russ
Hey how are you. So I bought the new laser tube which you had recommended. The tube was too big for my machine and had to return. Finally got the new tube and installed. It worked for one cut. Now my screen wont turn on. Any idea what it could be? The light turns on in the cutting area, but the laser beam wont go on and the screen doesnt turn on
Hi Pamela
There is a socket on the white control unit into which the screen plugs. I would first unplug and reeplug that because it sounds the most likely problem.
When you power your machine on, does it run to one corner and set the machine 0,0? That one action tells me just how much of the machine is working. If you press the test button on the power supply does the tube light up?
Speak soon
Russ
SarbarMultimedia morning. When I power the machine on only the light switch will turn on. It doesn't run from side to side. The test button does not turn on either. It did work when you had previously had me test that. I had the laser tube installed yesterday and it worked for one cut. We installed an extractor (which is not plugged into the machine) and now it won't work. The only thing that will turn on is the light switxh
Hi Pamela
I need to get you to send me pictures and I need to send you data. We can only do this by email. I tried to personal message you but it seems to be blocked. You can either message me or send your email to me via this comment box. The good news is that any email address immediately goes into my you tube spam box where it will never be seen by the public
Best wishes
Russ
Hello,
i am engineer and i was made two laser machines with different designs.
On the second desing i had the same problem. I was found the problem in the lens position, the lens is too far from the linear holder and goes like string oscillator pushed from the stepper motor vibrations.
For the second machine i was bougth laser head with long nozzle which is not good, the same like on the video.
But the first laser head with short nozzle goes better cut and engraving with minimum spikes and wave cutting.
The other problem is the position of the belt, below the linear holder it is not good at all, have to be on linear holder. The linear holder itself is another chienese long topic.
Best Regards.
trinamic make great stepper driver controllers that run super smooth- id look into that if i were into it.
friction makes micro steps practically useless for positioning by the way- since micro-steps have such weak forces on the effector. its practically just a form of sound dampening between full steps- but the motor will tend to lock to the nearest steps- A proper stepper controller can do this much smoother. it comes down to matching the wave form to the type of motor etc etc- and if paired incorrectly can cause all kinds of weird noises and behavior like skipped micro-steps etc.
Hi
Thanks a lot for suggesting a better drive controller. However, it is not my goal to make a super performing machine but to determine the performance boundaries of the machine that I bought. In this case I had to spend a small sum to change the belt but that was more as an exploration tool than a machine upgrade. It has reduced my "jitter" zone to a very narrow speed band that I can now avoid with confidence, but I still want to prove that the frequency of those speed spikes is somehow related to the native 200 steps. It's all part of the enjoyment I get from this machine, to use shed technology to to do investigatory work that would normally require sophisticated measuring tools and instrumentation. The other issue with this particular machine is the fact that the belt drives directly from the motor via a 16tooth pulley, there is no gearing system that would let the motor run 4 to 6 times faster than this direct drive system. I will eventually perform the same tests on my Lightblade machine which does have a compound belt drive system between motor and head. Still lots to learn!!
All best wishes
Russ
understood- :) keep it up
These lens adjustble laser heads had another very big problem. If you put lens with Focal Length with exampe 76.2mm and if you adjust the lens on these length from the upper mirror, you can kill the upper mirror with the small percent of the laser beam defraction from the lens.
Best Regards.
Hi Chudomira
I greatly appreciate your experience of the very small mechanical design issues that may be contributing to some of the pattern problems I have identified on my machine. Obviously they all emanate from the stepper motor because that is the only mechanism that is creating the micro vibrations. I accept all the mechanical leverage issues that you have identified and I think they are all worthy of being investigated on this machine to see if I can do anything simple to mitigate such issues. It is not my intention to solve any of the problems I find because I am not trying to make a super machine. I am instead trying to understand A) if there is a problem. B) what is the cause of the problem. C) what are the boundaries of the problem
With understanding I have choices. For example, as you have seen in the video, I have decreased the zone and the pattern of the vibrations. It is now in a very precise zone between 40 and 70mm/s thus I hope to be able to prove in future sessions that by staying out of that zone I have improved my engraving. It would appear that if the issues you identified are present in my machine then they are all bundled up in that "nasty" zone. After all, the patterns I am recording in acrylic include ALL issues not just stepper weaknesses. This is exactly the goal of my experimentation to identify operating boundaries so that I can make informed decisions on how to best operate the machine.
I could spend huge amounts of money trying to improve the machine but as I work my way through the machine I conclude that you are getting a fantastic little CNC machine for very little money. Having owned two large 3kw metal cutting industrial lasers for 10 years, I realize that this is not industrial standard engineering and that there are many compromises that have been made BUT for what it is, we should not complain. Hence, it is my strategy to understand rather than fix it. I have been to trade shows and examined the design concepts that the professional machine builders such as Trotec and Epilog use, and they in a different league .....but that is why you will be paying 15 to 20 times as much for something that does the same job but a bit better.
Again, sincere thanks for you comments and I shall be bearing them all in mind as I experiment further. It may be that something as simple as your damping weight around the nozzle will produce great benefits and help reduce my "nasty" zone further. All suggestions are welcomed.
Best wishes
Russ
.
Hello,
thank you for your answer and your videos, i am very pleased with it. You can try to engrave by Y axis. These will not improve something significant, but will shows how the mass reduce the vibrations. Often i am egraving by Y. If the stepper motor do not holds the bridge, you can reduce max. acc. from the software.
Best Regards.
To reduce these vibration to minimum, you can buy transmition belt x3, you will increase the speed of the stepper motor and then the vibration effect. But on these blue chienese laser have to me mounted on the right on the moving bridge (i do not know what exacly is the name), these will be some problem. These will decrease the maximum speed of the machine.
Best Regards.
Not sure if this will help you but I recall that when I spoke to Trotec about my machine they told me that the chinese glass tubes are not capable of producing pulses as fast as a high quality sealed air cooled tube like Synrad. This also makes servo motors excessive in china machines since they will not be able keep up with the faster servo motor anyway.
This might be wrong, I am not sure but if I am not mistaken that is what they told me. :)
I'm pretty sure Synrad are not producing DC sealed glass tubes. Their specialization is RF excited tubes at about 4 to 6 times the price of sealed glass tubes, I consider them as an industrial solution rather than affordable for most enthusiasts. Their operating technology does allow for continuous pulsing whereas there are special operating modes for a glass tube where pulses can exist naturally or be forced by the controller. The Chinese design is elegantly simple and very basic with lots of small design compromises. But for 90% of the capability of an expensive industrial machine you are getting a lot for very little because there is no R&D team, no marketing costs and no service backup team to support. With a Chinese machine you will be your own 24/7 support tech.
Stepper motors are very reliable and pretty fast. They can be made to perform better by adding a rotary shaft encoder to guarantee position accuracy . They will never beat an analogue servo system, only on price.......but do they need to? That's all part of the 10% design compromise.
Best wishes
Russ
this looks like comparing cars by looking at their airbag-system ...
Also, with these vibrato effect you lost 5-10% from the power efficiency into the cut zone, because the beam steps on the material with fast and slow speed steps into the slow cutting speed. I was attached 50grams mass near to the lens and i was make to cut completely the same thickness with the same power.
Best Regards.
Hi Russ,
I think you are experiencing mid band resonance. The main springiness in the system is actually the stepper motor. It produces a torque proportional to the sin of the displacement angle from where it wants to be. For small displacements sin x ~= x so it looks like a classic spring.
The mass of the load plus the inertia of the rotor form a simple harmonic oscillator with the spring constant of the motor. The resonant frequency is constant, so as you sweep the step frequency it comes in and out of resonance. The oscillations are superimposed on the linear motion, so as you move faster they get recorded on the acrylic further apart. Just like an oscilloscope looking at a constant frequency waveform with a variable timebase.
Another possibility is your laser PSU has ripple on its output. Have you looked at the beam current on a scope? E.g. put a scope across the ammeter terminals if it has one.
Hi Thanks for the great explanation. There is no intention on my part to upgrade or change my motor of drivers. My hobby (apart from the video ) is the engineering challenges of finding the limits and shortcomings of what I bought. Ok, I changed the belt, but that was almost a tool to prove a point which I feel was successful. I have now defined a performance zone between 40 and 70mm/sec that, if possible, I should avoid. That is a much suppressed range than existed with the rubber belt, thus I now possess knowledge I didn't have before. I do not use the machine for business purposes, it's just a means of keeping my two remaining grey cells exercised. I think it's pretty amazing that I have a 3D performance storage tool in acrylic it can do so much more than a scope
I realise that steppers are digital and proper servos are analogue so it's good to understand that there is springiness in the jump between steps.
Thanks again
Best wishes
Russ
After initial rise time and overshoot the power is very steady. Nothing that would cause these spikes. They are purely down to motion because they have direction to the shape that reverses with the scans.
Thanks for the thought
Best regards
Russ
We need someone with a high-speed camera to shoot this at 1000fps so we can see what's going on :-)
I tried 240 fps but that's nowhere near fast enough. That is where acrylic does such a fantastic job, just like footprints in the sand. Tells me so much more about where the power is being applied and at zero cost. Getting hi tech answers with low tech methods is all part of the fun I get from this machine. I got refunded on my accelerometer and looked at several other technical solutions for tracking the detailed motion of the head including LVDT and laser distance measuring. All WAY too expensive for this sort of shed engineering. In the end a piece of stainless hypodermic needle tube and a piece of acrylic provided the answers in a more positive and at ZERO cost. I LOVE acrylic
All best wishes
Russ
I shall do some tests with my chinese dragon as I have a camera that will do 960fps and see if that's enough to prove anything ;-)
Fantastic. Thanks Tony
Great video Ross, In my experience yes, it is the motors but also rails may have a part in it. I made this video to help on the troubleshooting of the issue, the same type of engraving using servos czcams.com/video/JJs6K9E-VQ0/video.html
Hi Abraham
The link is your personal video editing link and will not work for others. Go to Video Manager and click on the edit button for the video you wish us to see. There you will find the link to send us.
Thanks
Russ
Hi Russ.
Rob here. Do you have an email address I can chat t you on?
Hi Rob
Check your You Tube messages
Russ