Commercially available laser controllers

Electronics related to CNC

Re: Commercially available laser controllers

Postby fullspeceng » Thu Jul 01, 2010 4:36 pm

We've had our RetinaEngrave working for a while now but only recent have gone into production.

You can now see the final board here at the bottom: http://www.fullspectrumengineering.com/ ... 2-40w.html

The video and results are there also.

Cheers.
fullspeceng
 
Posts: 38
Joined: Thu Dec 17, 2009 10:13 pm

Re: Commercially available laser controllers

Postby lasersafe1 » Thu Jul 01, 2010 7:36 pm

I spoke with one of the major manufacturers today. He was understandably reluctant to give out too much information to a DIY hobbyist. His laser machines sell for $100,000, so it would not be a wise business move to teach people how to build their own. When I mentioned the grayscale, he asked me how fast my laser was. When I said 15us, he indicated that this is impossible for a DC tube. They use RF tubes because they have a faster response than the DC. I immediately realized my mistake. I have been thinking that my laser was fast because I saw the 15us pulses coming from the DSP and thought I saw an accurate burning on the item being engraved. This is not proof that the laser itself was turning on and off within that time period. For all I know, the laser didn't respond at all for that particular pixel.

I think there are a couple experiments that could be done.

1. Put a low value resistor as a balast on the ground return line from the tube and measure the pulse on an oscilloscope as compared to a command pulse in channel 2 of the scope. This would tell us if the power supply responds within XX microseconds of a command.

2. If the pulse is present on the ground return, then that means the plasma was indeed generated and the laser fired for that time period.

The alternative is to find a fast photodiode that responds to 10.6um light. I am not aware of any. I guess one could also use a conventional photodiode looking at the side of the tube to measure the visible light that appears from the excited nitrogen in the mix. Clearly if there is no visible glow present through the side, there is no infrared being delivered out the end.

Anybody else got any ideas? I'm sure it already exists in literature somewhere, but I haven't found it yet. Would have thought I could find it at Sam's Laser FAQ.
lasersafe1
 
Posts: 599
Joined: Fri Nov 27, 2009 8:23 pm

Re: Commercially available laser controllers

Postby lasersafe1 » Fri Jul 02, 2010 1:30 am

I just received a definitive answer on the power control issue from an expert in the industry. The high end machines really do set up to 256 power levels quickly as the head sweeps across. These machines are not cheap. The controllers are not available for purchase on their own.

It is very easy to determine if a machine has true power control. A true power control engraver will produce an output that looks like this:

pic961-carve-3d-laser-engraving.jpg
pic961-carve-3d-laser-engraving.jpg (26.9 KiB) Viewed 19979 times

There are no "pinholes" in the surface because the power turned down to only burn to a certain depth.

A machine that does not have true power control will produce an output that looks like this:
3D%2520Eagle.jpg

The pinholes indicate that the file was created from a dithered bitmap. The final cut depth is the same across the entire surface, but the spacing between the holes determines whether or not the surrounding material falls away. That eagle head is now essentially a big wood sponge and would crush if pressed hard enough.

It is truly a sad day :cry: . I was really looking forward to making some of those beautiful 3D engravings. This is beyond the price range of any common hobbyist. (Unless Bart or Henry or Marco hook us up with an awesome controller). :mrgreen:

The answer may lie in a controller that does many passes of X sweep without indexing the Y forward. The first pass would only fire on the areas that are full depth (255 or black). The second pass would fire on the areas that are 255 and the areas equal to power level 254. The third pass would fire on the areas of 255, 254, and the 253 depth. Etc etc until it has gone for 255 sweeps in that Y location. Then the Y is indexed one increment and the process is repeated. Yes, this is a mighty slow way of doing it, but it would work on systems that have slow power level controls. One could imagine that if certain levels of gray did not exist in the master file, then that line could be skipped as long as it still balanced the burn of the items of greater intensity. Using this method, a very low power laser could be used to get burn depths that presently only exist on higher power lasers or lasers that are rastered very slowly.

Yes, it would be slow, but it wouldn't cost $15,000 or more and it would get the job done.
lasersafe1
 
Posts: 599
Joined: Fri Nov 27, 2009 8:23 pm

Re: Commercially available laser controllers

Postby fullspeceng » Sat Jul 03, 2010 2:36 am

There are two ways to control laser power:
1) Adjust the pulse width -> limitation to on/off time
2) Adjust the current power supply -> reaction time limited

We can control both on our controller.

I'm 100% certain I can get true 8bit or even 16bit time trimming. The pulse width resolution on my controller is 10ns. The code is easily changed.

However, I have more pressing issues like getting our device driver installer finished for deployment and general CNC controller released which will be higher demand compared to graylevel engraving.

Currently I cannot engrave at the motor limit speed in order to get a deep engrave so timing is not a problem. I can just run the engrave slower and adjust the pulse width as a variable from 0-255.

RF lasers do not necessarily turn on/off that much faster. If you want fast switching you should use a Q-switched laser.

I have purchased a few Ytterbium Fiber Lasers and will eventually put out a product based on these lasers which will give superior engraving results but not so good at cutting.

Our controller is far superior to any available units for purchase. We have side by side comparisons with the Leetro MCP6515 as well as Newlydraw 1 and Newlydraw 2.0 and cannot be judged unless seen in person.
fullspeceng
 
Posts: 38
Joined: Thu Dec 17, 2009 10:13 pm

Re: Commercially available laser controllers

Postby lasersafe1 » Sat Jul 03, 2010 3:01 pm

fullspeceng wrote:There are two ways to control laser power:
1) Adjust the pulse width -> limitation to on/off time
2) Adjust the current power supply -> reaction time limited


Statement number 2 seems correct (Power is controlled by the 0-5V or PWM pin). Let's assume we have placed 2 volts on the analog control pin. When we take the TTL control pin to the laser ON state, a typical 40W tube will produce about 6-10 Watts. If we raise the voltage from 2 volts to 3 volts the output will rise to about 20W. The problem with the present DC tube and power supply is the fact that the turn-on point is not very reliable. It doesn't start at 0 and cleanly rise to 40W. It starts at about 8 W with a very noisy output and rises to about 35W maximum. I don't see how you could make an accurate 8 bit power control using this analog input pin.

Statement number 1 doesn't make sense to me. Assume we have the analog pin at 3 volts to get the 20W. If we take the TTL control pin ON to fire the laser for 1 millisecond, we will get 20W for 1 ms. If we turn it on for 500us, we will get 20W for 500us. The power density was the same in both cases. If the laser mirror were sweeping this beam at 1 meter per second, then the 1 ms pulse would make a line that is 1mm long and the 500us pulse would make a line .5mm long. The burn depth of these two lines would be identical because they both received the same watts/sec.

Using the above example (20W, 1M/S mirror speed, 1ms pulse, producing a 1mm line)
If I wanted that 1mm line to be EXACTLY 255 times deeper I have the following choices:
1. I could hit it 255 times with the same power (or)
2. I could drop the mirror sweep speed by 1/255 and fire the laser for 255 ms. (or)
3. Raise the 20W laser to 5100W. :shock:

That is why I proposed the multiple pass software. I could do it myself with CorelDraw, but imagine this process. Tell Corel to only show you the parts of the image with gray depth = 255. Process to laser.
Tell Corel to only show you the parts of the image with gray depth = 255 and 254. Process to laser.
etc. etc.
Since there could be errors after multiple indexing of the Y axis, I would prefer to do all the 255 passes on the present Y position before indexing forward to process the next line.
lasersafe1
 
Posts: 599
Joined: Fri Nov 27, 2009 8:23 pm

Re: Commercially available laser controllers

Postby Tweakie » Sun Jul 04, 2010 8:40 am

Statement number 1 doesn't make sense to me. Assume we have the analog pin at 3 volts to get the 20W. If we take the TTL control pin ON to fire the laser for 1 millisecond, we will get 20W for 1 ms. If we turn it on for 500us, we will get 20W for 500us. The power density was the same in both cases. If the laser mirror were sweeping this beam at 1 meter per second, then the 1 ms pulse would make a line that is 1mm long and the 500us pulse would make a line .5mm long. The burn depth of these two lines would be identical because they both received the same watts/sec.


Correct me if I am wrong but isn't this exactly the way in which microwave ovens control their power for defrosting and other reduced power settings. Mine turns on and off (at full power) with variable settings of duty cycle and this reduced duty cycle applies less energy over the same time period.

Tweakie.
Tweakie
 
Posts: 120
Joined: Tue Apr 27, 2010 5:17 pm
Location: UK.

Re: Commercially available laser controllers

Postby lasersafe1 » Sun Jul 04, 2010 3:24 pm

Yes that's fine for a microwave where you want an "average power", but not for a target under a sweeping mirror that can see every "instantaneous pulse" of a pulse width modulated signal. I think the results I posted a few posts back prove this. If the laser is set for a high enough power to burn 1/8" into the wood, then it will burn 1/8" down if pulsed for 100us or 100ms.

I think there are still people that are confused about the Chinese DC tube power supplies and their controllers. The PWM signal is applied to the analog input to create the 0-5 Volt potential for setting the tube power. A change to this PWM signal takes many milliseconds to respond, (citation needed) so it is generally only used to set the maximum burn depth before an engraving is processed. The actual engraving only steps the motors and fires the laser using the TTL control pin at any given spot for XX microseconds and then steps the motor again to the next spot. There is no PWM of the TTL control pin.

The RF tube lasers are a whole other beast. Apparently (from my limited understanding of their operation) they have the ability to change power levels on a pulse by pulse basis depending on the duty cycle of the incoming carrier frequency. One can easily perform a complex power control (AND) ON-OFF control using a single control pin. As an example, a common carrier frequency is 5kHz for the Synrad. A signal "tickle" pulse of 1us is applied at this 5kHz rate to keep the laser on edge and ready to fire. 100% duty factor at 5kHz is 200us, so this is the minimum hold time where the stepper controller needs to pause between steps if you want to apply 100% power on this spot. Now as the X motor is stepped to a new position, the controller simply needs to change the tickle pulse width to 100 us for that one spot to get 50% laser power hitting the target. It can then move to the next spot and apply a 50 us pulse to get 25% power hitting the spot. 8 bit power control can be acheived if the timer has the ability to produce pulse widths with less than 1 us resolution at the 5kHz carrier frequency. 780ns to be specific. The software could be amazingly simple using this laser. :roll: spoken as a non-software person :roll: You would simply look at the color of grey requested for this pixel in the file and apply 0 to 255 units of pulse at this XY location. Of course if there is time between step pulses, one could imagine that you can give two or three pulses of that given setpoint at any given spot location.

Since this site is mostly dedicated to home-built machines, we generally don't have the funds for more expensive RF lasers and the commercial controllers that do the 8bit grey scale power control are not sold separately from the expensive engravers. I am trying to suggest a way to use the cheaper CO2 laser and power supply to get a similar result to the more expensive systems. That being said, if someone makes me a controller for an RF laser that does the 8 bit power control on a pixel by pixel basis, I would probably buy it and fork out the money for an RF tube.
lasersafe1
 
Posts: 599
Joined: Fri Nov 27, 2009 8:23 pm

Re: Commercially available laser controllers

Postby Tweakie » Sun Jul 04, 2010 3:47 pm

Now I am totally confused LS1.
Did you not show in your measurements here viewtopic.php?f=6&t=36&start=20#p501 that the TTL pin was in fact being provided with a PWM signal ?.
Sorry to appear so dense but I would really like to understand all this and you are doing an excellent job as the teacher. ;)

Tweakie.
Tweakie
 
Posts: 120
Joined: Tue Apr 27, 2010 5:17 pm
Location: UK.

Re: Commercially available laser controllers

Postby lasersafe1 » Sun Jul 04, 2010 5:27 pm

Thanks for the compliment. I never envisioned myself as a teacher. I sometimes say things that exceed my knowledge and hope nobody catches me on it.

No, perhaps I didn't make it clear enough in that post. The PWM was going to the PWM input of the power supply to set the tube output power. The "bit banging" was happening on the TTL pin during the sweep of the motors to fire the laser at different X positions. The PWM pin remained at a constant pulsewidth and frequency during the entire engraving process, so whenever the laser would fire using the TTL pin, it would put out the power level set by the PWM pin. This is how the DSP controller that I own works. Perhaps other controllers have a different way of doing it. I haven't really sat down and thought about the dead time between steps during a raster. Perhaps someone has discovered that they can indeed take one step of the X motor and then fire the laser for varying amount of time while the motor is paused between steps. Perhaps we can really get the analog burn depth that we are looking for in those 3D carvings. I just haven't seen a single example from any of the Chinese lasers that proves it.

The first person that makes this controller will make a fortune.
lasersafe1
 
Posts: 599
Joined: Fri Nov 27, 2009 8:23 pm

Re: Commercially available laser controllers

Postby pixpop » Mon Jul 05, 2010 5:08 pm

lasersafe1 wrote:1. Put a low value resistor as a balast on the ground return line from the tube and measure the pulse on an oscilloscope as compared to a command pulse in channel 2 of the scope. This would tell us if the power supply responds within XX microseconds of a command.

LS1, on the Chinese power supply I had, there was already a low value resistor in the low side. I think I measured something like 22 or 27 ohms to ground. Presumably, they have some degree of closed loop current control, or overcurrent detector.
pixpop
 
Posts: 24
Joined: Fri Dec 18, 2009 7:47 pm

PreviousNext

Return to Electronics

Who is online

Users browsing this forum: No registered users and 7 guests

cron