Setting up the DIY CNC to Laser as well (Recommendations for hardware)

Hello I have been playing with my Atomstack that has been modified for bigger laser and such. I just bought Atomstack R1 pro rotary and have high hopes for it.

But its winter here in Canada that machine is in the cold workshop.
But years ago I built a 36"x39" DIY CNC that lives inside for the most part. It was built to be a plasma cutter/spindle/vinyl plotter. By swapping out the head and the table.
It uses (x4) NEMA 23 steppers (Y & A axis, and on the X and Z axis) and the control box I built has the drivers and the power supplies in there. I use aviation plugs for “Box to machine” connections. There are limit switches on ever axis and the homing will home each Y and A stepper independently so its always squared up.
I use the same control box to run a Mini mill converted to CNC with NEMA 23 steppers as well.
I am controlling the whole mess with a PoKeys57 controller that is both usb or ethernet connected to a laptop. I run Mach3 for both machines and just use different “profiles” for each setup. I have seen mach3 & lasers… not looking to go that route. I love Lightburn for the most part.
So lets say I wanted to control the CNC with Lightburn like I do on the atomstack. Minding that I want to use the new R1 rotary on there as well.
What main board would be recommended to run the (x4) NEMA 23’s and still be able to run the rotary?
I don’t know if the PoKeys57 could be a good option? Controlling laser power and such…
I am guessing I would need to get new drivers and maybe make a “laser” specific control box. But best case scenario is using my current setup with a few mods or tweaks. To be honest I am planning on a Co2 build in the near future possibly with the same NEMA 23’s. So would love to hear your thoughts.

I’m not clear… are you wanting to diy a laser and want a suggestion about the controller or do you want to use your existing cnc machine to drive a laser by modifying the controller?

Generally most grbl control boards should work. The dsp models have more controls or control the laser using both control signals.

How deep are your pockets for this?


Currently I would like to control the DIY CNC to control the laser.
If I have to build a new control box with drivers and power supplies. Then I would like this to be the controller for the future DIY co2/rf laser.
So I spent the better part of yesterday searching for options.
I seen that the MLK DS32 has y1/y2 axis so that will take care of the slaved “A” axis. ← more on this below. I think I can use this to control drivers for the NEMA 23’s. Not overly worried about Wiring and soldering cables and plugs.
But if I want to do it right the Ruida RDC6442G seems to be the higher end ($$$) direction.
Since the controller will eventually run a co2 or a rf laser. It may be best to go big now.
But the one thing in the back of my mind is how the rotary hook up to a 4 axis controller if I am using 2 for the Y and 1 on X & Z axis?
For the simple diode laser it seems we unplug the y axis and the rotary becomes the Y. Great except that I need to home and power my Y & A steppers in order for the Gantry to hold perfect square to the wasteboard & grid.
How do the RM or Omtech Co2 laser unit use a rotary? is it because on the Y they use only 1 stepper? Freeing up the 4the axis for a rotary unit?

As for how deep the pockets are… LOL at moment very empty been off work for 6 months on sick leave. Got some work lined up but not as much pay there.
Maybe my thoughts are deeper than the pockets at the moment. But always good to plan many options on paper, and I tend to work towards one in some way or another.

My co2 is a 5030 type configuration, commonly called a China Blue.

Generally the machines using a Ruida, require separate motor driver modules for each axes. They are not expensive…

Mine uses these… along with NEMA23 type motors. This is the motor driver manual.


The Ruida will drive a co2 or rf machine. I don’t know if the gcode can handle the rf machine, but don’t know.

If you’re going to get an rf machine, I hope you are working, you’ll need the money :face_with_spiral_eyes:

Here is a manual for a 6442… they claim…

The dc excited co2 uses two control lines, allowing actual power control… One line for power and one to cause it to lase. The RF only uses pwm, so no power control as with a dc excited tube. If it matters to you.

Most of these use the Y axes to run the rotary. With a 6442 I think that’s the only option. If you cough up more money, resulting in a 6445, I think that model supports a rotary off the U axes… I have a 6442g.

There is no limit switch (can’t home it) on a rotary so you have to figure out how you are going to actually use it…

Good luck


All Awesome information thanks Jack.
RF is a 5-10 year upgrade dream LOL but want to pre plan that one day I may be able to afford the RF module…
The “newer” 6445 looks like it isn’t much more $$. But good to know it “could” operate a RF one day.
The goal of building a DIY laser is to slowly make the space for it firstly HAHA!.
Once the real estate is determined I will have a oversized linear rail driven “Diode laser”.
Then work up to a co2 tube and onward… as the story goes. But build it to have a 60-100 co2 ability.
Based on the future plans of the control box I think it would be in my best interest to get the Ruida controller first and foremost.
I actually need to upgrade the drivers in the CNC box since the MIni mill has the 4.2A steppers, and I am pretty sure the current drivers are under powered. So maybe these will get donated to the DIY laser pile. I also have a 1300mm linear rail so I am off to a good start.
Would a nema 17 (also in spare parts pile) be to small to lift a 4’x2.5’ table? It doesn’t need to do this fast or often, so it might be okay?

I built a homemade cnc router/plasma. Went with the router. So it’s heavy duty. I run a standard arduino controller nema 23 and generic GBRL. I mounted a laser on it. Only possible problem is maybe speed. Lasers like speed, routers like rigidity.

So can a diode laser be run on the 6442 or 6445?
Will there be any benefit in the speed side? Or is it still limited by the 20W laser module and the weight it needs to throw around with a diode module?

So this is what I was thinking. Get a 6445 and use it on the DIY CNC. Side build the larger frame and machine and get it moving around with the 6445. I would be at the end of the major stuff in the mean time. As in I would need to use the 20 W diode on it for a 6-12 months. Then I will either build a co2 or RF box for it.
But bigger on a slow diode seems counter productive. Unless the 6445 and linear rails have a decent speed & accuracy improvement already.
In my case its comitting to the space it will take to build it.

CNC covers about everything, even the spouses Subaru.

I don’t think the Ruida are a good choice for a milling type machine, if that’s what you mean. The Ruida does well with lasers but don’t know if you have the necessary control with Z to do a milling machine type operation.

Not clear what you mean… If it’s got a 20W laser it’s got 20W…

Sure they are handled by the Ruida by setting the tube type to RF.

This affects not only acceleration, but how big your stepper motors need to be. Within reason you can swap out something with more mass and lower the affected axes acceleration values.

This is from the preview, the red shows the overscan. Acceleration values of 40,000mm/s^2 compared to 6000mm/s^2 on the right. Notice the actual job times in the preview.

Both types have their pro’s and con’s. Make sure you know what you are getting and what to expect.

One advantage of the tube is that it can be current controlled, varying it’s output wattage. Require liquid coolant, have a much slower response time and are relatively low cost.

RF and solid state types are either on or off. A 100W RF laser will lase at 100W, no matter the percentage of power. If you want 20% power, then these lase at 100% power for 20% of the time. Have a very quick response time. Costs are high, but the ‘tube’ is metal, so it can be re-charged and put back in service.

Good luck – We want to see photos :crazy_face:


Thanks again jkwilborn.

What I meant about a 20W laser is that because of the lower power I imagine it wouldn’t be able to engrave at the speeds a Co2 or RF would do.

Couldn’t The Z axis of the CNC could be controlled the same as a motorized table on a Co2 machine? I’m assuming this is the “Z axis” still. But really I could set it manually then power the stepper so it holds that height while being thrown around.

In setting the tube type to RF for a diode will I still be able to control power settings on the Diode laser?
Will the Ruida alone be able to increase the speed of the diode engraving when compared to the MLK DS32? If so then I will look for the controller on a black friday deal and start there.
If not then it is 2020 and linear rails first.

I obviously bungled that explanation…

High power is only good for cutting… if you wish to cut through thick pieces of material, I’d suggest a higher power laser. Most of my thing is engraving. I have a 40W, it cuts well up to 6mm if I need too.

A 20W is going to cut like a 20W. Slowing it down isn’t going to make it cut like a 40W, so I don’t think I follow your reasoning here.

The Ruida control signals set for RF (non pre-ignition) mode is the same as an led.

I assume you know what a pwm control signal is…

With an led/rf machine, it turns on when the pwm state is high and off when low. It’s either on or off, there is no 1/2 state. This means the device output power is 100% or 0%.

If you have ever looked at a voltmeter at the pwm (ttl) signal, it will read 2.5V at 50% pwm. We know the it’s at 0V or 5V state only… the meter is giving you rms (root mean square) voltage.

The same thing occurs with an led/rf machine.

When a glass tube lases, it draws a much current as the supply will produce. The pwm is used to control that current limit but applying it to the lps IN terminal. The pwm, on a dsp it two signal lines. Pwm and L (laser enable).

Compare both machines operation when set for 50% pwm

The led/rf will lase at 100% power for 50% of the time.

The glass tube will lase at 50% power for 100% of the time.

Since the game with lasers is what happens when the beam strikes the material, I purchased a glass tube laser.

You would think…

A Ruida comes from a dictatorship/communist country. They don’t tell the western people anything about anything, let alone technology… So it’s all reverse engineered.

Dsp types, in a nutshell, many use the extra axes for specific purposes. Such as, the other axes may only need to move between layers. So it’s not directly available.

With the DCS32, it’s open source and schematics are easily available.

If you use one of these and wire it up with a manual pot, you have lost the advantage of it being a glass tube. It is run like an led, it’s on 100% at full power or it’s off.

I did find this but don’t know if it’s applicable, they claim it works…

You could also just run L through the front panel laser on switch, letting the pwm still control current and on/off.

The simple facts are that the Arduino based controller runs at 16mHz. most of it’s instructions are a byte. But if you make it only 8mHz that’s still 8 million instructions a second. The mks32 board runs at ~230mHz.

Simple fact is these are pretty much sitting there waiting for the mechanical operations to finish… the mechanics take for ever to occur compared to the controllers speed.

You will also be waiting for the mechanical to complete…

Nothing will increase the speed of the job with a laser… it’s all laser or tool dependent.

About the only way to speed up any kind of laser is to use a gavo head.

Make sense?


So what makes the Co2/RF so much faster? Is it just the power behind it that allows it to do that?
I was not aware that a RF worked this way, I have a friend that has a very expensive laser (RF) his works are always impressive.
I also will need to run this in a unheated workshop (unless I am working in it and still only around +7C in winter) So no liquid cooling was a main reason that I was drawn to his setup.
In the end it is what it is most likely Co2. But could the Ruida 6445 run dual laser sources but different types?
Spit balling here but if you had a Co2 and in distant future added a RF to shoot opposite direction of mirrors and either to the same head somehow or to a separate head. You could have best of both worlds?

Anyway I pulled the trigger on 2020 and linear rails. I think I will get the frame size figured out and then work on designing the tensioners and brackets for Nema’s. I am guessing that these really don’t need to be high Amp/torque? Looking at some 3A motors & drivers.
Get the motion control figured out and make a oversized diode laser.
Sure am considering the ruida 6445 they have the 6445S with power supply for less than $400CAN on ali express today…
As far as I can tell the Main difference between S and G is G has dual head capable but also seem like you never know what you will actually get. Still like the 5" screen

It depends on how much machinery they must sling around and how fast you want that to happen.

NEMA 17 motors seem adequate for low-end / small diode lasers, but typical CO₂ machines use NEMA 23 motors with a NEMA 34 (or larger) to raise the platform.

In all cases, folks tend to run the motors much hotter than seems prudent in a non-industrial setting. Although more current = more torque, the power dissipation goes up as the square of the current, so there’s not much point to cooking the poor things.

CNC machines (tend to) run slowly using leadscrews for better positioning accuracy under load from the cutters. Laser cutters run much faster and accelerating heavy gantries requires much more torque than you’d expect.

CO₂ lasers run much faster than diodes (because they produce more beam power) and require heavier structures, so you probably need larger motors than you expect.

At the same power levels nothing. Most dpssl, right now reach around ~40W.

Most of the time is mechanical, waiting for the head to get there practically speaking.

Gantry machines take longer to move the parts around, whereas a glavo only has to move the mirrors a slight amount and there is very little mass to them. This makes the movement the quickest.

Many times an rf machine will include a beam expander.

Probably a good idea without some kind of heat.

How could you mount this stuff in there and be able to have two flat paths to the heads? There are multi-tube machines out there you can find. Never seen one with only a single head. So you’d need a controller that supported both an rf and glass tube laser.

How fast do you want to accelerate the mass on your head? How much is the mass on your head?

I believe so, I saw a list at one time, but don’t remember the details.

Good luck


Thanks Ednisley and Jkwilborn.
I think at the moment the pockets can only fund nema 23 3A and drivers. One will be a dual shaft to pull the longer X axis back and forth. X & Y axis’s on linear rails.
I don’t know how heavy the head and mirrors are yet so you are right that I could be surprised.
I plan on using a single 2020 rail for gantry and it is going to be less than 36" long.
I do know that if I use the diode on this it will need to run more conservative speeds so as to keep things cooler slinging all that extra weight around.

One thing I never considered as a added hidden cost is a DSP license for Lightburn is more than the G code version I am currently running.
Are al the Ruida 6445S compatible with Lightburn? I see they say they are but then only talk about RD Works in the specs on the unit I am looking at…
Just don’t want to buy a Ruida that is “bricked” to anything but Lightburn. As it is the only program I know at the moment. So far it is working well for my needs.

The Ruida is a communist product, as is the RDWorks software. They don’t tell anyone anything.

I believe Lightburn has to reverse engineer many of the mechanisms within the Ruida to code for it. It’s a proprietary language, unlike gcode.

All same model numbers, supposedly are electrically the same. Firmware is different. Some handle two sources, not one.

The dsp type, including the Rudia are considered the best laser engraving controllers around… Of course that an opinion to many.

With a Ruida you have only RDWorks or you pay extra for a dsp enabled Lightburn license.

I don’t (won’t) run Windows, so if I want to use one of these I have no choice but Lightburn. Same with most of these fiber machines…

I’ve see a wide range of Ruida devices, that all seem to work.


With a Ruida you have only RDWorks or you pay extra for a dsp enabled Lightburn license.

But this is a lightburn purchase not a purchase that needs to be made when buying a 6445S or G controller.

I otherwords, I can consider the one I seen for less than 400 and I will need to deal with Lightburn license once I am ready to fire everything up.

Yes, you can use if without ever purchasing Lightburn if you want to use RDWorks.

The total cost will involve the extra cost of a dsp license from Lightburn if you wish to use that software.

You can buy it, install it and then get a 30 day free trial…? In the end you’re probably stuck with the dsp.

Lightburn has an interesting license. If you purchase it, there is no limit in time. You only have to update your yearly costs, if you wish to get a newer version.