Fiber Laser Learning Lab 05 Let's Test Some Pulses

[@ThingsWhichArentWork]

I'm so glad you did fix all the problems on your old China Blue machine Russ - I binge watched all of your videos whilst I was waiting for my own one to arrive so when it did I was fully aware of exactly what to expect. Took me about six days to rewire it, align all of the optics, install a milliammeter and sort out the steps/mm. I would have taken me months if I hadn't had your help, so a big thank you for your last video series from me. Cheers!!

[@kingjamez80]

Another extremely satisfying technical video. Really appreciate your analytical approach!

[@jurgenblust2042]

Thanks, Russ, for another interesting episode! :-)

[@randystockton]

I love watching these videos... But it makes my head hurt! lol

[@chrismicklitz6361]

I don't know why the first dot is different either, but the RMI UM-Laser I use at work has a setting for it. It says, "first pulse suppression." The choices for it are either on or off. I am standing by to be educated on this phenomenon. Thank you.

[@SarbarMultimedia]

Hi Chris I have learnt a bit more since doing this video and I am now convinced that the dot is the LAST dot in a train of dots. As the line starts, there is indeed a dot but the increment to the next dot, when the frequency is high, is parts of a micron, and thus there will never be a visible complete dot until the last dot at switch off .Its exactly like the pattern you see with a welded line. I was puzzled at the time as to how the first pulse could be different to those that followed because the is no storage mechanism in the laser that would cause this, I clearly state that I am the student here and at the point I noted this oddity, I made an illogical conclusion. All part of learning.!! Best wishes Russ

[@johnrevill9592]

Another great video Russ. There are a lot of setting in the EZCAD software. Not sure if the version supplied by Lotus Laser supports them all. One of the options is drill mode. This moves to a black pixel, stops and fires for the specified ns, then moves to the next black pixel. This eliminated the oval, or football field shaped burn area because the beam is always stationary when it fires. There are also timing setting such as StartTC, OffTC, EndTC and PolyTC. This allows you to accurately control the timing of the laser to prevent overrun and or underruns. Personally I use a combination of timing and drill mode, but I have had a lot of problems with consistency using the EZCAD software when running at high DPI and speed and have recently moved to MarkingMate software which also required a controller board change. Oh, and 2ms:2ns is 1,000,000:1

[@SarbarMultimedia]

Hi Thanks for the details John, I am not digging into the software too deeply at present although I have fixed the bidirectional scan offset with the TC variable.. I was aware of drill function but I purposely forced the software into a difficult situation by choosing to draw pixels as lines by defining a fixed scanning offset of 0.1mm. I think this clearly showed an error of some sort in EZcad because the scanning increment did not match the vertical pixel increment , even though they were supposed to be the same. . You will note the missing dashes in some patterns . I think that is because the scan lines did not match the pixel vertical pitch and the scan lines dropped perfectly BETWEEN the bottom edge of one set of pixels and the top edge of the adjacent row and did not switch the beam on and off. The other reason was to see the power being put into such a short line by each pulse profile. Drill would only put a single pulse onto a pixel , at 850 kHz I was putting several hundred pulses onto each pixel. I still have a lot to learn/discover. I like to learn slowly and methodically, making many mistakes along the way.. Thanks for you invaluable comments , please don't stop Best wishes Russ

[@thebeststooge]

I am still fascinated how these do colors from a binary material. Good work, Russ.

[@SarbarMultimedia]

All will become clear in later episodes when I investigate black marking of clear anodised aluminium. If that is baffling you, just wait until I start creating colours on stainless steel.. I am not just trying to find all the colours I have seen others do, that would be easy to look at someone else's parameters. No, I shall gain no understanding from that, I must try to decode how it happens in a logical step by step way. I have done a lot of hunting and can find nothing by way research papers on this subject. Lots about the growth and composition of the oxide layer that forms on stainless but no details of how the colours occur. I feel this may turn into another "hunt for the smallest dot" saga. Best wishes Russ

[@alexismartial63]

Now this is real suspense, better than an Agatha Christie 😅👍 So eager to find what could explain these strange behaviours..

[@SarbarMultimedia]

The answer lies within EZcad. My chequerboard pattern is a manually created (dot by dot) 245 ppi bitmap with 0.1mm pixels vertically and horizontally. I have set EZcad to index vertically 0.1mm every scan. When EZcad has imported the 254ppi bitmap something has gone slightly awry and I suspect the image import is not scaled quite correctly. Lets say the pixels are now 0.099mm square instead of 0.1mm . I have instructed EZcad to scan at 0.1mm increments so as the scanning moves up/down the image the scan lines will move in and out of synchronism with the pixels. Most of the time the on off switching of the laser will be cleanly controlled by the pixel edges but when the software scan line encounters the chequer junction where the bottom of one line of pixels is almost in line with the top of the row below, it gets confused and sends out a continuous signal. I have seen this aliasing effect in other softwares. Lotus Laser advise that for this sort of pattern I should be using vector lines. I know that and I also know why I designed the pattern the way I did. A vector image would not show this small math error in the software. When I get time I will experiment with the scanning increment to see if I can find one that perfectly matches the pixel size. You can sleep at night now you know. Best wishes Russ

[@JERRYIRWIN41]

HI Russ, Brilliant as usual with clear detailed maths.. Oh Dear, have you left some graffiti (WOW) on the table... !!!

[@SarbarMultimedia]

Hi Thanks for the kind words but you must remember that the maths has to be simple or my two remaining grey cells protest!! Best wishes Russ

[@martinandersen7955]

Correction 2mS / 2nS = 1.000.000

[@SarbarMultimedia]

Thanks Martin. Lets see how many others spot my mistake. Best wishes Russ

[@arpanchoudhury_]

Thank you sir again for explaining everything /\

[@lazerusmfh]

I’m wondering if the reason you’re getting power spikes at the beginning of the start of a row is your laser on time delay is too long

[@SarbarMultimedia]

Thanks for the comment. You see shortly that I now have a proper 400x microscope to help me with my observations. It turns out that I was misunderstanding what I was observing. That blob is the beam diameter footprint as the beam switches off. It is not a power surge or pulse. Mistakes are all part of my learning. Best wishes Russ

[@lazerusmfh]

SarbarMultimedia ahh so your off delay is not set aggressive enough? I have a similar laser, but mine is a max photonics mopa 30w and have seen similar things

[@macswanton9622]

You have us on tenter hooks- did you find a gremlin?

[@SarbarMultimedia]

Not exactly. But I am convinced it's a problem with EzeCAD bitmap conversion . When I did the same pattern with vector dashes I did not see the same issue because there is no defined fixed pitch. It could well be a small maths issue with the Y pitch because it only needs minute scaling problem when importing the bitmap image and after a certain number of fixed Y pitches, the scanning line that defines the switching signal for the laser, falls BETWEEN the dashes (in Y) for many pitches before it drifts back onto the dash pattern and starts switching again. I may get time later on to investigate this further. I have seen similar problems with RDWorks and early versions of Lightburn. I'm not over worried about it because what I did was VERY unusual and probably never been tested like this. Hope this relieves your tension Best wishes Russ

[@MrT0mmygun]

You have great machine..why you use old EZCad 2? There is EZCad 3, which is another planet

[@SarbarMultimedia]

EZCAD 2 is not my choice.As I mentioned in the video, Lotus Laser clients are usually high volume small product range users. A bug free stable version of software that runs the machine reliably is all they need. They do have EZCAD 3 but found that everytime it was up issued there were bugs that did not previously exist. For their customers, a frozen version of EZCAD 2 was the safest option and has been rebadged as LotusMark. At the moment I don't need to do anything fancy, I just need to drive basic machine functions and understand how the various parts of the technology work. Thanks for the comment Best wishes Russ

[@TheDopalgangr]

He would also have to change the controller card since the one that works with v2 does not work with v3

[@hoocli]

Do you have this file and will you be able to share it

[@SarbarMultimedia]

Hi Clinton I'm sure I can find it somewhere in my archive, In the meantime youn will need to send me your disguised email address, This used to be a simple task but You tube have implemented algorithms that recognize simple masking and then prevents you sending a comment. The way that works is to bury your details in a sentance or two like this. Start with john and a dot followed by smith. Use the standard at symbol and then g==m)a i*l with a dot and a com. I will reply and also delete your comment Beat wishes Russ

[@hoocli]

@@SarbarMultimedia thanks! It's pretty easy. Just my first name and last name with a .c. in the middle with a g==m. And of course com

[@shadowrob7709]

In short, the hardware cannot deliver the specified parameters, which is the case with all Chinese machines. No wonder because you pay 20 times as much for good beam sources. My f-theta lens costs more than a China machine, Scanlab Galvo is the same. A good zoom beam expander costs over 2K.

[@SarbarMultimedia]

After "playing" with and understanding the intricacies of software and hardware of this machine I was able to reproduce this matrix correctly.. Your assumption of this being a cheap Chinese machine is incorrect. It is designed and manufactured here in the UK bu a specialist company that creates bespoke laser systems of all sorts for industries here and in Europe.. They use Coherent. JPT, Raycus, SPI sources ( dont know them all). The majority of systems are the simple Q switch type but about 10% are MPOA systems. They do not use Chinese F theta lenses because these are industrial machines that have to be accurate and reliable.. I eventually overcame the dashed pixels by using drill mode. Using a FIXED X and Y DPI for image scanning also fixed the odd lines I was getting. Please remember this is a learning exercise for me and at this point I am finding lots of interesting anatomies that my weird tests are revealing. I am sure if you ran the same tests on your MOPA kit you would get the same interesting results. I know that this hardware is only limited by the EZcad software and not by component specifications. The fact that I see problems is all part of the self-education and discovery process.. Best wishes Russ ps I was loaned the machine for 1 year and that period has passed and I no longer have it, Instead I have a cheap 20 watt Q switch Chinese laser and it's limitations are too many to describe.

[@shadowrob7709]

@@SarbarMultimedia A quality optical Q-switch costs more than all the China machines put together. JPT fiber module Coast in purchase under 2k Euro for 30W module. A really good F-theta lens costs over 3k Euro. I have several laser systems and also like to test these China fiber lasers. In continuous operation none of the machines lasts longer than 6 months. Burned out fibers, Q Switch inefficient, galvo distortion and power loss. I have as an example laser systems from Photon-Energy. 2 of the systems one is 18 and other 20years as, run per day min 4 hours through since then. All the China machines promise more than they can do and if you try to take advantage of it, the lifetime will be massively shortened.