Hello,
Name is Josh Lotts, I run the fastest growing info hub for DIY lasers on facebook - The Laser Guild
That being said, I have been building lasers for 8 years now. Well before it became a popular DIY project and have not stopped learning since.
Naturally gantries require automation and said automation requires motors.
I’ve built machines that use everything from 2-3 phase stepper systems to current clearpath hybrid monsters that are achieving well over 2kmm/s with 5G+ loads.
Also, ironically I am a long time RC buff with airplanes, helicopters, drones and cars since 1991. That being said, I was around for the brushless/lipo game changing modernization of the industry.
Ultimately, this has brought me here.
For those with such RC models, it’s undeniable how powerful those brushless motors are and it had me thinking about implementation for CNC years back.
Unfortunately I am not a programmer or chipset builder. (I’m learning)
Low and behold! ODrive filled in what I wasn’t capable. This is great news for me.
I’ve been watching the start till now with this production and am happy to say am finally in a place where it would make sense to move forward on this.
What I would like to do is build a unit featuring these drives and am going to need a lot of input along the way. If I am able to get the performance specs I am after, this may very well change a vast majorities opinions on what the standards should be with affordable options such as ODrive. Anyhow, I’ll post more later, just thought I would toss out an intro, my intentions and purpose being here. Look forward to hearing from you guys!
Hi! And thanks for responding.
And to your ideas of lasers, yes maybe for most, and no that doesn’t really apply to me.
I guess I should have made it more clear on “what type” of lasers we’re talking about here. I personally don’t really mess with the little solid state diode units, especially from bozos that decide “Laser Pecker” is a good idea for a name… FACE PALM 2000!
Generally I don’t tend to go under 70 watts with any tube I personally use unless its an RF of fiber unit.
My typical builds for myself and those whom I build machines for tend to utilize RECI, SPT or GSI 80-120 watt Co2 Glass tubes that have a frequency range of 20khz. At that rate of modulation you can take advantage of up to 850-900mm/s of transmission speeds before you start to outrun your pulse rate. However, I also have a coherent 70 watt RF unit that has a modulation frequency of 50khz and is good up to 2500mm/s+.
So probably not what you had in mind for poking fun of the speeds with!
We’re talking about a system that can tattoo your grandma before she spits out the dentures!
Ultimately what I’m wanting to achieve may be currently unrealistic with these drive/motor setups within reasonable cost, and maybe it’s not. That is what I am here to find out.
Anyhow, what I am really after more than anything past 800mm/s is high G accelerations such as those found from my Clearpath builds that can achieve over 2k mm/s with 5G loads while swinging a one lb. carriage, nozzle assembly.
Now lets get a little technical on specifics of what I’m working with currently and where I am headed.
Being that I’m heavily involved in constant discussion and coordination with a great number of the industry leaders in parts supply, such as "Trocen, Ruida, Teknic, LightObject, Lightburn, American Photonics, and few others, I have special access to things that are brand new or am participation in BETA development scenarios with, I tend to get cool $#!^ and have the honored position of nudging the industry in certain ways to further the standards of current products and offerings.
All that being said, I am building a unit as we speak that consist of an ultralight nozzle that Steve Walters of American Photonics is developing in coordination with my suggestive feedback and ideas. SWEET!
To put into perspective the gravity of what I am doing you need to understand that most nozzle systems/carriages on the X axis that are currently being produced via companies such as cloudray can weigh as much as 620 grams on their carriage.
AP’s ultralight nozzle riding on a PBC linear RR30 “Redi-Rail” weighs 287grams! Not only that but it features precision centered optics that are glued in place with a completely sealed head ZnSe pass through window before refracting off an Si plated mirror down through a ZnSe meniscus lens. You will also see that is has a cleverly placed heat index tape for measuring thermal exchange to tell you if the lens needs to be cleaned without having to tear it down to realize something is off. All optics are military grade and are rated for 8000 watts of optical transmission. BAD @rse is a phrase that comes to mind!
So with that being said, now that I’ve effectively eliminated 2/3 of the typical dynamic loads on the axis, I am in search of an optimal method of motor transmission that will give me my cake so I can eat it to. Thus, here I am!
I would like to feature a machine with these drives and brushless motors sending this thing to the moon with performance. What say you guys? Anyone have any experience driving a laser with acceptable results?
What’s this going to cost, you have me thinking about laser weedcutters now. Imagine a farm robot with a single-axis as you describe to sweep the laser focus back and forth across a crop row… what is the power consumption of the laser when not cutting but ready to pulse, and is it going to be better than my other approach of using 5KW of LED light in a row unit.
Haha, I have thought of this before actually! But I can’t imagine it realistically being a good choice for the public at least. I have a feeling it would be fairly frowned on from the fire code folks!
A Co2 tube can range from 35 watt glass unit from $150 to $100,000 depending on power and modulation type.
The unit I am building now has a cutting area of a little over 5’x5’ and will cost about $5K for the entire frame, ready to run mechanically. But that is before you add the cost the tube. I am considering making this a dual tube unit with both an RF and glass tube Co2 combination. 130watt glass and 70 watt RF. Figure in a factor of about $5800+ for those. But you can do a lot with a 70-100 watt glass tube which you can obtain for $450-$700 depending.
Here is a photo of the first unit I ever built. It was originally a 1200x600mm Lasersaur that I eventually tore down and replaced all the guts with retrofitted parts sourced from various locations. This unit can be replicated for about $4k+ give a little ready to run with a 100 watt Co2 tube. It can cut up to 3/4 wood if needed.
so as I understand it, your question has nothing to do with lasers per se. You are asking how much acceleration and speed can we get from the ODrive with a 1 lb load on the end of a belt?
The answer comes from the torque that the motor can apply, and what the ratio is between motor torque and force at the load, due to your pulley sizing.
Without doing any calculations, I can confidently say that you will be able to accelerate from 0 to 2000 mm/s in a fraction of a second, if you pick the right motor and encoder. It will easily outperform any stepper motor for this kind of task.
You might want to think about the lateral stiffness of the belt though - as you push up the acceleration, obviously you increase the strain of the belt. If it is too springy then your positioning accuracy will be low, and you may get control instability.
As a very rough calculation: Assuming the ODrive and motor can put 200W /mechanical/ power into the system (which with most motors, it comfortably can), that equates to accelerating a 1kg load at 100m/s/s (~ 10G) at 2 m/s (2000 mm/s)
(google dimensional analysis is my friend)
Sure, I understand the motor size/specs has everything to do with that, but I only seen one for sale. So that is what I am referring to before I jump into finding my own sources.
Understand I am not playing with simple open loop steppers here.
I am looking for a potential more affordable route spanning from Teknic clear path units.
These are no slouch. They can pull over 3000mm/s with 5 g loads for there Nema 23 HD series. And that is will a one pound load. These use very technical acceleration curve algorithms (RAS). I realize the curves would be something that I would likely need to develop myself, but that is fine I suppose.
What I want to do is put this system in a package that is more or less developed into a drop in unit for DIY builds. I have thousands of people following what I do on a regular basis and this holds a lot more potential for revenue draw for O-Drive than anyone may realize if I can get someone to hear me out with a seriousness.
Thank you for your response and the link.
Here is a look at what Odrive would be competing with for my desired applications.
Haha, nothing. Because I wouldn’t! That would be far to dangerous to have any hand in on my behalf for someone else.
Not because telling you not to do it, but we’re talking about a class 4 laser. One fraction of 1 second will blind you for life if it refracts and hits your face.
These things are in the 10.6um wave length range. In other words, they are so far into the IR spectrum you can see nor detect them with anything until it’s on fire!
Also you need to understand that in order to cut something like you’re talking about, you would need a long focal lens, such as a 4"FL unit for getting through something like a wet tree that large. It would require multi sweeping with possibly a galvanometer head 6-10 inches away from the surface you want to slice. You would need at least a 130 watt Co2 for this. They are powered with a 28kv transformer to about 28mAh. You might want to consider looking into a highspeed liquid cooled spindle with a long endmill bit on a gantry arm. That would by far be the most effective way to go about it outside of a clamping chainsaw setup.
You should totally build a high performance laser with ODrive!
In terms of input, I don’t have much to say other than try it. You seem to have a good handle on torque and speed requirements, so I’m sure you can pick the right motor. Or just pick up our D6374 motors and NEMA face plates, they will let you get started very easily.
Encoder is very important. In the future we want to offer an RS-485 absolute encoder solution, but for now probably the easiest to get started is a good incremental encoder with index. Or if your cable from encoder to ODrive is very short you could use an SPI absolute encoder.
In the future we might offer a totally integrated motor+ODrive solution, and then it would be very similar in integration level to a ClearPath.
