oldbots first laser 2.x build

[oldbot]

This log starts near the end.
The photo shows a piece of thermal paper darkened by the laser.
I made quite a few mistakes on the way to get here, and I will be adding to this post in the hopes that you can learn from my errors.
I mostly followed the design faithfully, but I will also document some changes I made that may be improvements.

I have over 30 year experience making high voltage power distribution splices, so I will start with the insulation on the laser tube.
Silicon tape is only rated at 8 KV per layer, so one layer does not seem like enough.
Plus we are insulating a protrusion that may work its way through the insulation as time passes.

I insulated the positive on the laser the same way I do a power distribution splice:
1) First I eliminated all air from the splice. Duct seal is OK for this.
the important part is to displace the air in the vicinity of the splice.
where there is air, moisture or even water can take its place.
I kneaded a bit of duct seal around the wires to reach into small crevices.
2) Next I wrapped the splice on the tube with Scotch 130C.
I went around the wire and around the tube as well.
This is good for 69KV per wrap, but I used four or five wraps.
The 130C stretches and eliminates air pockets.
3) The 130C will fall off if you don't seal it in place.
I use Scotch 33+ for this, another three or four layers.

The voltage will break down the insulating material over time.
So extra layers of insulation are a good idea.


Measuring the current.

I put a 100 Ohm resistor between the laser ground wire and the power supply.
I ran a couple of wires (one on each side of the resistor) from this to a voltmeter.
I get one tenth of a volt per milliamp.

My biggest mistake so far.

I bought the cheapest laser power supply I could find on Ebay.
The first time I applied power to it, I could smell ozone.
After a minute, a glowing crack appeared in the high voltage transformer.
If I had left it running, the results might have been quite spectacular.
Its in the trashcan now.
I should have know better.
The laser power supply is the one item I should not have tried to save money on.
The replacement from LO does not emit ozone or glow.

More later.

[bdring]

It seems power supplies are the number one failure item of the Chinese parts.

I design HV power supplies for big microwave tubes at my day job. I concur with all your statements. The number one thing to remember is air in the insulation is bad. You will have corona and the insulation will fail.

[oldbot]

Here is a photo of the failed power supply.
Note the brown mark on the inside on the white section below the "110V".
This was bright yellow-orange with the power supply on.

My plan to replace the 5K pot for laser power level.

I want to solve two problems: possibility of setting the pot too high and overloading the laser and the problem of it being difficult to adjust the pot.
I plan to replace the pot with a 6 pole double throw selector switch, ten 500 ohm resistors, one 500 ohm pot and one 2k resistor.
The selector switch will be able to select these six ranges:

0 to 500
500 to 1000
1000 to 1500
1500 to 2000
2000 to 2500
2500 to 3000

Resistances above 2800 ohms on my unit overload the laser, so setting a maximum possible resistance of 3k is good.
This will increase the resolution of the pot by ten times.

I am also considering replacing the 500 ohm pot with a ten position selector switch and nine 50 ohm resistors.
This will give me repeatability.
There would be about 50 useful settings.

I looked at the possibility of a ten turn pot with a ten turn dial, but the selector solution cuts off the dangerous top resistance range.
I also looked at the option of using a digital pot--which would give me more settings,
I may revisit this option if this configuration does not work well or I want the digital pot for computer control of the power levels.
The digital pot configuration could be set to exclude the useless low settings and the dangerous high settings.

Please ask for clarification if this is not clear.
I tend to write incomprehensible explanations at times.

[TLHarrell]

Digital control of power to the laser is usually by PWM.

If you have a 30mA panel meter, the 10 turn pot is extremely simple to use. Also, laser power is not entirely linear, and laser power effects on materials vary depending on the material, manufacturer batch, shop temperature and tube condition. Being able to set a 6 position switch doesn't sound ideal to me.

What are you using for a controller?

[oldbot]

I am using the Buildlog.net Interface/Laser Driver PCB Rev 3.3 with Pololu stepper drivers.
My software is EMC2 on Ubuntu.

I use Inkscape and Intellicad for the design work and PyCam to generate G-code.

I may end up with a DSP board, but not before I have tried to write some software and design come hardware.
Possibilities include:
programs to modify G-code,
adding some custom circuits that can interface with G-code
direct control of the laser driver board,
a custom parallel cable that will address both laser driver board and custom circuits.

I should look into pwm control.
This should be easy to implement with a microcontroller.

I was trying to explain a plan for six ranges, not six settings.
I expect that I will be letting other people use this laser, so eliminating the possibility of overcurrent on the controls is important.

I do expect to ruin a lot of material finding out what works.

[oldbot]

More on my experiences and mistakes building a 2.x

I soldered a buildlog.net laser interface board.
I ordered an interface board, watched a video, downloaded the BOM and ordered all the parts I did not have.
When the board arrived, I discovered that I had ordered parts for the wrong board.
However I had ordered the ideal board for my system.
So I ordered parts again.
I have accumulated quite a few parts, electronic and mechanical that I have been unable to use.
I have also had to stop many times because I found I needed a part I did not have.
If I could start over and buy a kit, I would not do it.
I learned a great deal from looking for the parts myself.
I found many valuable sources and websites.

Some suggestions for the interface board:
Lead free solder!
Lead fumes are very bad for you.
A fume extractor.
Not as important as lead free solder, but it does guarantee no smoke in your eyes.
A soldering station instead of a soldering gun.
This just makes soldering a little easier.
I upgraded all the resistors to metal film resistors and increased the voltage on the electrolytic capacitors.
This may have just been a waste of time, but it did make me feel better.
I have seen many carbon resistors and electrolytic capacitors let out their magic smoke.
(All electronic components work because they contain magic smoke. As soon as you let out the smoke, they stop working.)

The interface board packs a great deal of functionality in a small space.
It would be hard to come up with design improvements, so I will not even try.

More later

[oldbot]

I learned a great deal about CNC machine possibilities from this build.
The combination of makerslide, v-groove pulleys, custom parts, timing belt and idlers is particularly useful.
I plan to try to make a larger 2.x laser next.
After that I will try my hand at some new CNC designs using makerslide.

Something I did right

I got ball end hex keys, both T-handle and L shaped.
The T-handle hex keys are my favorite.
The bright red handles makes it harder for me to lose them on my workbench.
What I did wrong: I did not have these on the first day of the build.

Something I did differently from the 2.x plans

I put a shelf on the upper half of the frame for the laser power supply.
I drilled through the side of two Misumi angle brackets and bolted them to the makerslide.
This enabled me to bolt the shelf from the bottom.
The high voltage power supply lead is very close to where it connects on the laser tube.
This should cut down on induction from the laser pulses, although the long coil of cable is certainly not ideal for this.
The interface board is below the supply and at the opposite side of the frame.
You can see the 100 ohm resistor I used to get current measurements.
I am considering adding a voltage controlled oscillator (VCO) and a speaker so that I can hear the variations in laser power.

Aligning the mirrors.

I did my initial alignment with a laser pointer.
It took two hours, even after I tried to get the mirrors in the right place first with a 45 degree protractor and a dial caliper.
I have seen some very nice 50mm laser pointer holders in another build log.
I used some 20 mil pipe wrap tape, some fender washers and then some more pipe wrap tape.
It only took ten minutes ( and only one small hole in the worktable) to get the mirrors aligned with the laser tube installed.
While trying to find the lowest setting at which my laser would fire I discovered two things:
It is very easy to set thermal paper on fire.
It is even easier to burn a hole through thermal paper so that the beam continues on to the next mirror.
At least I did not need the fire extinguisher.

[TLHarrell]

With PWM control, you will set a max output for the laser (100%) in software. Set the software to a point where the laser tube will get around 14.5mA.

I just got a new digital soldering station this weekend and kicked my old $15 one to the curb. I've still got tons of old leaded solder around though. Lead fumes... yummy.

[oldbot]

Experimenting with the 2.x

Acrylic seems to cut very well with this laser.

Here is a 21 point star about 4" diameter cut from 1/4 acrylic.
16 ma, two passes, speed about 1 inch per second.

The edges are a bit sharp and brittle.
If I cut another one, I will cut a circle a bit inside the star to get rounded edges.
It will also be much easier to get the star out of the sheet.
It took longer to saw it out of the sheet than it did to cut it with the laser.

I am using a program I wrote in python to modify the g-code.
I call the program shortcut.
It revises the g-code so that every g-code cut is very short: .1mm for this star.
This adds quite a few lines to the g-code, slows down the x and y movements,
but evens out the power along the cutting path because there is little acceleration and deceleration.

I will post the code if anyone is interested.
It is not completely tested and may have a hidden bug or two.
The resulting g-code is too slow to cut paper, but seems to work well on everything from thin cardboard on up.