Using a laser to build a CNC router

[Lee DeRaud]

A thread awhile back about the "DIYLilCNC" finally prompted me to do the CNC router build I'd been contemplating for a couple of years. The main breakthrough was figuring out how to make the dimension-critical parts with the laser: I laminated them out of three layers of 6mm MDF. That allowed me to use the laser for two operations that it normally would not handle well, (1) making countersinks for the screw holes and (2) "drilling" the edge holes for the cross-dowel bolts. Other than that, it's a fairly conventional MDF/skate-bearing design, albeit a bit heftier than most. Cutting area is roughly 32"x17"x5", sized to fit some odd space constraints in my garage.
228067228066228068

[Michael Hunter]

That looks a very neat job. Nice to see it painted too, rather than the MDF coloured machines you normally see.

[Neal Schlee]

Nice Build, Lee!

Now, when can we get the files? :)

Neal

[matthew knott]

I like that a lot! wish i had the time to make one!! is that a porsche lurking in the background?

[Joe Hillmann]

Do you mind telling what that cost you as far as all the electronics go? I have thought about building one using the laser to make the parts as well but since I didn't really know what I would use it for other then a couple of whims I had and I don't know much about programming I decided that for now it was a bit to expensive of a toy.

[Lee DeRaud]

Now, when can we get the files? :)Real soon now: they need some cleanup, as I was designing and building it sort of on the fly.

[Lee DeRaud]

Do you mind telling what that cost you as far as all the electronics go? I have thought about building one using the laser to make the parts as well but since I didn't really know what I would use it for other then a couple of whims I had and I don't know much about programming I decided that for now it was a bit to expensive of a toy.Heh. Well, it's a lot cheaper than the laser was...

I've got close to $3K total into mine at this point (pretty much ready to use), but I could probably start from scratch and duplicate it for about $2400. (Note that probably 15% of those numbers are shipping and sale tax.) The power supply, motors (4), cables, and driver electronics totaled out to about $750 of that. But I didn't really attempt to economize on the electronics, on the theory that they can be recycled to the next machine if there is one, or easily sold for a good price if I pull the plug on the whole thing. I'll post a complete bill-of-materials a bit later.

Another thing to consider is that the workflow for a CNC router is insanely complicated compared to shoving bits at the laser from CorelDraw. I've killed most of the last week just installing demo versions of various CAM software packages and working through their tutorials to figure out which one to buy. I suspect that, if I wanted to cut the pieces for this machine on this machine, it would take me longer to convert the files to the right format that it took me to build the thing in the first place.

[Joe Hillmann]

Do you mind giving a little more of a rough break down of the cost? Not down to the nut and bolt level, it is just that when I was looking I had thought the once I had the electronics (the ones I was looking at were in the $700 range) I didn't think the rest of the parts would be that expensive, and I am just wondering what parts were the most expensive.

I also should add that it is a very nice looking machine.

[Doug Griffith]

Nice. I recommend VisualMill for a CAM package. It's intuitive and not too pricey for what you get. It's the same as RhinoCam but your not tied to Rhino which in my opinion isn't very intuitive. I'd personally avoid learning CAD/CAM packages dedicated to signs. Especially since your into building things like CNCs.

[Lee DeRaud]

I've attached the bill-of-materials for the hardware and electronics (including elementary software). It breaks down roughly as:
Electronics (including router and some probably unnecessary frills) - $1100
Hardware (nuts, bolts, steel stand, drive screws, bearings etc) - $750
Software - $325 and up (free demo/trial versions are available, sufficient to shakedown the machine)
There's maybe $150 of MDF, paint, and glue in it.
228077
(Note: these prices do not include shipping & taxes.)

[Lee DeRaud]

I recommend VisualMill for a CAM package...not too pricey for what you get.Neither are Ferraris, but $1K is a bit over-budget to run a hobby machine.:eek:

[Tony Joyce]

Be sure to look at Cut2D($149) and Cut3D($299) by Vectric. Good entry level with an upgrade path to more powerful packages.

[Lee DeRaud]

Be sure to look at Cut2D($149) and Cut3D($299) by Vectric. Good entry level with an upgrade path to more powerful packages.They're the leading contenders at this point. The others I've looked at in this price range have some odd UI quirks. I'll probably get Cut2D first, then maybe Vectric's bitmap-to-3D app (PhotoVCarve?), as I'm not doing traditional 3D modeling.

[Steve Clarkson]

Nice job Lee.....very impressive. How much would it cost to buy a comparable CNC?

[Lee DeRaud]

How much would it cost to buy a comparable CNC?I have absolutely no idea. The design constraints are roughly that of the Rockler CNC Shark Pro Plus at $4K (plus router, plus...): how close I got to that goal remains to be seen.

[William Adams]

Way cool!

I'm currently considering a Shapeoko as a starter unit as the best balance of cost and capabilities. Apparently it's fairly easy to expand on one axis, so will want to upgrade to at least 12" on one axis though.

[Mike Null]

Lee

Very nice work. Keith is high on Vectric software.

[Keith Outten]

Lee,

I agree with everyone, you did a nice job on the CNC build. Soon you will get the fever so bad you won't believe how much time you will be spending with your new machine.

I am in fact a Vectric Software fan. Vectric delivers the stuff that works and its easy to get started with their software. I was lost when I started up my first CNC Router and I struggled with the software until I purchased VCarve Pro. After upgrading to Aspire when it first started shipping I was able to do more with my machine than I had expected.

If you put the business aspects of CNC Routing aside these machines are the most "Fun to Run" I have ever seen.

Hope to see your work in our CNC Forum soon :)
.

[Ross Moshinsky]

Very nicely done. I will say I'm a bit surprised by your budget line though. $3000 you can do something with 8020 extrusion and a superior rail system. You won't get the dual drive though.

https://www.finelineautomation.com/store/show/FLA100-00 would be an example of what I'm referring to.

[Joe Hillmann]

So now we need to see pictures of things you make with it.

And thanks for the price breakdown. I guess the little pieces of hardware add up quick.

[Richard Rumancik]

. . . I will say I'm a bit surprised by your budget line though. $3000 you can do something with 8020 extrusion and a superior rail system. You won't get the dual drive though. . . .

Ross, the link you supply shows a decent 8020 frame, but I'm not sure what you are suggesting with regards to Lee's budget. He says he can duplicate his for $2400. (When you design something from scratch there are always a few parts left over from things that didn't quite work as originally planned.) The Fine Line product in the link is $1599 but there are no electronics, no software, no router, and no motors.

Can you buy something similar off-the-shelf rather than build it? Sure . . . the discussion of build vs. buy goes on all the time on the CNC forums. It always depends on what people's objectives are and their tradeoff of time vs. money. Many people enjoy the challenge of the design and fabrication exercise; other people just want to make stuff with a purchased machine. To each his own I suppose.

[Lee DeRaud]

Very nicely done. I will say I'm a bit surprised by your budget line though. $3000 you can do something with 8020 extrusion and a superior rail system. You won't get the dual drive though.

https://www.finelineautomation.com/store/show/FLA100-00 would be an example of what I'm referring to.Yup, a complete kit for $1600, for some arbitrarily low value of "complete kit". I see this all the time in CNC forums: somebody will say, "That machine only costs $1600.", then you look closer and realize it takes another $1000-$2000 to actually get it to a point where you can make something useful with it. Look at the bill of materials I posted and start subtracting out the things that kit is missing: my equivalent of that $1600 kit runs under $750.

Is my machine as good as that kit? Almost certainly not. But I ran the numbers and couldn't get a FLA machine configured to "turnkey" configuration for much under $4000. And most of the point of the exercise was the process of designing and building it: if I'd wanted a turnkey system, I would have just bought one.

[Lee DeRaud]

And thanks for the price breakdown. I guess the little pieces of hardware add up quick.The little pieces of hardware are bad enough, it's the big pieces that people tend to forget about that will hurt you: the router, the router bits/mills, the software, the computer to run the software, and the stand to put it on so it doesn't sit on the floor looking stupid.

An entertaining (and/or scary) exercise is to take that list to the hardware aisle at Home Depot and start pricing out the 25lbs of nuts and bolts.

[Doug Griffith]

I hear ya.

You are still going to need a 3-axis controller and software which adds even more cost.

Mach3 is a good starting point.

[Lee DeRaud]

A further note about the files: anyone looking at the files in their current state is likely to brainlock like a parent assembling a bicycle on Christmas Eve. I could post Corel files for the MDF cuts of the 'as-built' machine today, but I'm not sure how useful they'll be without a lot of top-level drawings and assembly notes, most of which doesn't exist at this point except in my head. (My mental 3D visualization skills are extremely good, but my 3D CAD skills are virtually non-existent.) Anybody still want them?

The path I took to get to this point ran through this book: http://www.buildyourcnc.com/Book.aspx. My machine has roughly the same layout, at least from the table up, but much beefier structure. Some of the odder aspects of the construction methods are explained there better than I could in posts of reasonable length.

[Ross Moshinsky]

Yup, a complete kit for $1600, for some arbitrarily low value of "complete kit". I see this all the time in CNC forums: somebody will say, "That machine only costs $1600.", then you look closer and realize it takes another $1000-$2000 to actually get it to a point where you can make something useful with it. Look at the bill of materials I posted and start subtracting out the things that kit is missing: my equivalent of that $1600 kit runs under $750.

Is my machine as good as that kit? Almost certainly not. But I ran the numbers and couldn't get a FLA machine configured to "turnkey" configuration for much under $4000. And most of the point of the exercise was the process of designing and building it: if I'd wanted a turnkey system, I would have just bought one.

$2,258.00 with electronics, Mach3, and the full kit. Figure $200 for shipping you're at $2,458. Router is about $150. Mount is another $70. Limit switches, wires, and e stop are about $100. That puts us at about $2,778. That leaves enough money for cable management, a few bits, and a spoil board. Obviously doesn't cover dust management or the CPU/monitor. These CNC's can run on older XP PCs no issues, so you're looking at about $200-300 for a CPU and monitor. Even if you go crazy and get a new laptop for $500 you're still not that much over $3000. If we really want to talk about money, you can find an 8020 supplier locally and have them cut the extrusion to size. You can also have a local metal supplier supply you with the CRS rails. A day on the drill press and you've knocked conservatively $200 off the price.

I'm glad you enjoyed building your CNC but I think people looking at it thinking "wow that is cool" should understand that for a very similar price, you can get an all metal machine that is bigger and badder. One paying job could pay the difference between the two machines as 60ipm vs 150ipm is night and day. I've yet to hear anyone say anything positive about their MDF CNC other than it was a great learning experience for their next build.

[Doug Griffith]

...but I think people looking at it thinking "wow that is cool" should understand that for a very similar price, you can get an all metal machine that is bigger and badder.

Bigger and badder is relative. I work with numerous "real" CNCs and can say that you get what you pay for. 20k+ in the used market will get real work done. In my opinion, there is no CNC machine capable of real production work for the $3,000 you mention. Mach3 is also just controller software. What about the controller? That can easily run a grand with the 3 required stepper boards. Plus, any CNC that runs a router is more or less useless. They are loud. They burn out. They have a lot of run out. A real CNC has a spindle. The cost of that plus the speed controller can be a large chunk of money. It looks to me like Lee wanted to build a CNC using his laser. That he did and it looks like he did it well.

[Ross Moshinsky]

Bigger and badder is relative. I work with numerous "real" CNCs and can say that you get what you pay for. 20k+ in the used market will get real work done. In my opinion, there is no CNC machine capable of real production work for the $3,000 you mention. Mach3 is also just controller software. What about the controller? That can easily run a grand with the 3 required stepper boards. Plus, any CNC that runs a router is more or less useless. They are loud. They burn out. They have a lot of run out. A real CNC has a spindle. The cost of that plus the speed controller can be a large chunk of money. It looks to me like Lee wanted to build a CNC using his laser. That he did and it looks like he did it well.

1. Did I say biggest and baddest or did I say bigger and badder? Why are we talking about "real" cnc's here? We're talking about $3000. It's not a production machine but you can do pay work on it. Considering MDF machines can't hold up to cuts of any significant depth at any real speed without bending, my statement is more than valid.

2. Go on the FLA website. It's very clear you get a very good 3 axis kit for $519. It includes a Geicko G540 and 3 NEMA 23 380oz motors. It's the standard package for a entry level CNC. The Geicko G540 is actually considered one of the best stepper driver packages out right now.

3. Why are you talking about spindles when Lee's MDF machine clearly uses a wood router?

As I said, I think $3000 is a lot of money for a MDF router. $3000 gets you an all metal machine which is bigger and badder. If you disagree with that, that's your choice. There are probably a thousand people on CNCZone who agree with me.

[Rodne Gold]

I think it's a great educational excercise , and a job well donebut an expensive one IMO
I saw some really nice machines , all configured with some good parts (italian spindle - not sure which one, good lead screws and bearings etc)) for about the same money in china... all seemed to work off mach 3 (prolly pirated) or Autocad/Phcad or Ezicut...
I saw em working and they were pretty good
the machine with a 3kw spindle was around $3700 complete , not sure of exact size but it was at lest 1200mm x 600mm or bigger
Yeh , I know , talking China is considered unpatriotic and sino stuff is considered junk , but I have been in CnC since 1986 , had one of the first cnc routers here (Isel from Germany) and run a serious tekcel machine , but if I were looking for another one , I would readily buy one of the chinese machines.

[Scott Shepherd]

Yeh , I know , talking China is considered unpatriotic and sino stuff is considered junk , but I have been in CnC since 1986 , had one of the first cnc routers here (Isel from Germany) and run a serious tekcel machine , but if I were looking for another one , I would readily buy one of the chinese machines.

I'm not sure too many people on this forum have ever said the reason to not buy Chinese is due to patriotism. The issue most people have with it is exactly what we've seen play out since your endorsement of them. We've seen dozens of people buy them, then we see daily threads, and I do mean just about daily where people can't figure something out, can't understand how to use it, want to know the electrical configuration, didn't get parts, had to fix things on their own, and a host of other issues like that.

It's not about patriotism, it's about buying a machine you can order, get, plug in, and use. From the many posts that have come up since you bought your machine, I can't say than too many of them fit into that category. Even you, after having the machine for a VERY short time were already talking about buying new RECI tubes. To you and other owners, you seem to be perfectly happy working on your machines and trying to make them better. I have no issues with that at all, however there are many other people that don't have the background and don't want to have to work on their machines to the level some of you do.

That doesn't mean suggesting a newbie with no mechanical or electrical experience should shy away from a Chinese machine means I'm "Patriotic". It means I'm giving good advice to someone that doesn't have a tinkering background, which might save them a ton of time and trouble.

It has zero to do with country of origin, it has all to do with the level of tinkering you'd like to do with your machine once you get it and the number of posts with new owners having all sorts of issues speaks for itself.

I don't believe you'll see a single thread in the years this forum has been up where anyone had to have Epilog, Universal, Trotec, or GCC video tape them putting parts in a box for proof they did ship them. THAT'S the kind of thing some people have no desire to deal with. Nothing to do with country of origin, in my opinion.

[Doug Griffith]

Sorry about that. I was focusing on your last paragraph as a whole and understood it as being pointed to CNC newbies who were looking for a little insight into how much CNC you get for the money. I didn't want them going away with an idea that it's cheap to get into. You get what you pay for.

[Lee DeRaud]

I'm glad you enjoyed building your CNC but I think people looking at it thinking "wow that is cool" should understand that for a very similar price, you can get an all metal machine that is bigger and badder. One paying job could pay the difference between the two machines as 60ipm vs 150ipm is night and day. I've yet to hear anyone say anything positive about their MDF CNC other than it was a great learning experience for their next build.Well, aren't you just a little ray of sunshine. The key phrase in that post was "one paying job": anyone who's been paying any attention at all to my posts for the last seven years knows I'm not doing this as a business. Sheesh.

[Ross Moshinsky]

Well, aren't you just a little ray of sunshine. The key phrase in that post was "one paying job": anyone who's been paying any attention at all to my posts for the last seven years knows I'm not doing this as a business. Sheesh.

Let's just re-cap here.

My first statement was "Very nicely done." After that I simply suggested that $3000 is a lot of money for an MDF machine when you can get an all metal machine for around the same amount. Then you went on to question the integrity of my statement by essentially saying I was pulling numbers out of the air. I proved that I was not. Please explain what exactly done that you deem offensive?

Also, there are a lot of people on here who don't know much about CNC routing but maybe always wanted one. They see your thread and say "maybe I should do that!". All I did was open their eyes to the fact that there are a lot of different options out there and for similar amounts of money, you can get a far beefier machine. I was fairly sure that was the spirit of an online forum. Opening people's eyes to different things.

[Kim Vellore]

Lee,
I am also looking for ideas to build a CNC, though a tiny one with tight tolerances, The longest axis you have you are using 2 motors to drive that, or one motor tied to the other with a belt?, do you see any problems with it, if one of the motors looses steps it will be bad.
Thanks
Kim

[Lee DeRaud]

My first statement was "Very nicely done." After that I simply suggested that $3000 is a lot of money for an MDF machine when you can get an all metal machine for around the same amount. Then you went on to question the integrity of my statement by essentially saying I was pulling numbers out of the air. I proved that I was not. Please explain what exactly done that you deem offensive?Ok, let's recap a bit less selectively:
1. I agree that $3K is a lot of money for an MDF machine, but I also said that $3K was what I had spent so far, not the cost of duplicating it. That included a bunch of R&D in the form of paid-for plans (including several extrusion and steel designs), leftover and tossed parts from false starts and dead ends, plus a ton of extra shipping for parts purchased onesy-twosy because I was designing as I built. I also priced out half a dozen "kit" machines, including the ones from FLA.
2. As I also pointed out (and you apparently ignored), that $1600 FLA kit corresponds to roughly $750 of my machine's bill-of-materials. Where I come from, that does not qualify as "similar amounts of money".
3. I never said or even implied that you were "pulling numbers out of the air". What I did say is that, on an apples-to-apples comparison, you cannot build a ready-to-run metal machine for anything close to the same price. Even your pricing ran over $3K without any CAM software.


Also, there are a lot of people on here who don't know much about CNC routing but maybe always wanted one. They see your thread and say "maybe I should do that!". All I did was open their eyes to the fact that there are a lot of different options out there and for similar amounts of money, you can get a far beefier machine. I was fairly sure that was the spirit of an online forum. Opening people's eyes to different things.I would hope people would see this thread and say "maybe I can do that!"...which is not the same thing as "maybe I should do that!". One would hope anyone silly enough to embark on such a time- and labor-intensive project will do the required research on their own. But telling them that they're better off buying a kit doesn't really push the hobby forward. If you want to "open people's eyes", maybe you should encourage them to do something they can actually learn from and enjoy, instead of trying to suck all the fun out of the process.

(Yes, hobby. "OMG, he used the 'H' word!" And I would like to take this opportunity to say that anyone who actually needs a CNC machine for their business should bloody well go out and buy a ready-to-run system, not some erector-set kit.)

[Lee DeRaud]

I am also looking for ideas to build a CNC, though a tiny one with tight tolerances, The longest axis you have you are using 2 motors to drive that, or one motor tied to the other with a belt?, do you see any problems with it, if one of the motors looses steps it will be bad.Two motors, slaved together in the software. If a motor is losing steps, that's a problem that needs solving regardless, but in any case both of them will tend to lose steps together anyway, usually as a result of hitting something. It's "bad" in the sense that you lose time/material and the gantry might need to be squared up, but it's not like it destroys itself or anything, and you'd have to re-home under those conditions anyway.

On a smaller machine, I'd use a straight 3-motor approach...no reason not to.

[Kim Vellore]

Thanks Lee,
Your thread has rekindled my desire to build my CNC. I have seen many kits but does not fit my needs, so for my unique application doing my own seems like a very satisfying endeavor.

Kim

[Michael Kowalczyk]

Lee,
Don't forget about Artcam 2012. They have a 149.00 version that may do what you need but you can upgrade the modules as you need them.

[Lee DeRaud]

Don't forget about Artcam 2012. They have a 149.00 version that may do what you need but you can upgrade the modules as you need them.I've gone with Vectric Cut2D/Cut3D, along with Accutrans ($20) to cover point-cloud conversions and things like that. The UI on the Vectric products matches up better with my mental processes (insert joke here): they don't have the radical clash with Corel-think that some of the other CAD-oriented ones do.