Any interest in a guitar specific CNC design?

[Mike.Mara]

So... I've been talking to Prostheta and we've both noticed a gap in the CNC world where guys who build guitars are kind of left out.

Most CNC's are either too big, too short or too wide. (In my opinion).

Would there be any interest from the guys here in designing a CNC for instrument related purposes?

I'm under the assumption that people will want it as cheap but as accurate as possible, and I figured a lot of minds are better than one.

I'm thinking as many off-the-shelf parts as possible and everything else being able to be made easily by hand... Or at least cheap enough to get a routing template laser cut.

If there is any interest, anyone care to throw out some ideas regarding linear rails? Work area, drive system, construction materials?

I'm hoping that if there is interest then we can make it to upgrade it self from whatever material to aluminium so we can end up with machines that are capable beyond their price range.

Mike.

[MiKro]

First thing is to determine the actual cutting area size needed and Z height? After that is established, then a plan of materials can begin to materialize. Also, what price range to start at? It is getting much harder to find an older PC that uses a parallel port. Most of the controller software still use this as do many of the Break out Boards that are inexpensive. Steppers, drivers and BoBs are not that expensive anymore. Accuracy is something that you will not get in a low end machine and this will relate to problems if wanting to slot fret boards with the machine.   The other major cost becomes CAD/CAM software. Yes, there are some low end packages that work  including Fusion360 to a point, F360 being Cloud based so that becomes a problem if the PC is used for both controller as well as the CAD/CAM PC. Watchdogs can be set to counter errant PC apps but are not always reliable during controller operations. These are just a few of the problems I see in designing an economical solution, as I have also thought the same as you that there must be a way to  help entry level with a GOOD machine at a minimal cost.  I still believe that no matter what it is going to be in the $2500+ range at a minimum without software. There are Chinese machines for less, but then the size and accuracy are a problem.

MK

[MiKro]

Actual cutting area sizes. I would think 900mm ,450mm, 150mm  This would be approx 36"x 18" x6" as a minimum.

Min on screws would be 1000mm, 600mm, 300mm. Question then, is 2 screws for long axis and slave them? Or a single screw with lower crossbar? Again cost is a factor. The screws are not that expensive, it is the bearings that add up.  Again if 2 screws for long axis then add extra electronics for that screw. An Ebay kit would cut this cost for screws and rails but is for a true 3 axis no slaved long axis. something like this.

https://www.ebay.com/itm/US-free-Square-Linear-rail-guide-3-Ballscrew-1605-400-700-1000-BKBF-Nut-CNC-Kit/272911706945?hash=item3f8ace2b41:g:PDYAAOSwrk5ZeFd3

mk

[Mike.Mara]

Well, personally I think 500mmX1100mm work area is adequate. That would allow Flying V/Explorer types as they're the widest I can think of, and through neck basses.

The PC is less of an issue now since you can get a UC100 USB (Which I currently use on a Windows 7 laptop with Mach3). I've not had any issues as long as I keep it only running Mach3 when the CNC is running.

I wouldn't say accuracy is un-achievable with a low end machine, mine certainly is low end and with rough calibration with a steel rule I got to within 1.2mm over a 711.2mm scale length and the intonation is perfect. This is also GT2 belt driven, cheap chinese electronics and a router with a decent amount of runout. So I'd say given some careful and thorough design work we could do better than that.

I also think that $2500 is a high number, unless you're talking HIWIN rails and ballscrews ect? I was thinking more supported linear rail and trapezoidal screws. Maybe prices are different in the US but over here you could get the rails and the screws for less than £250.

My initial thought was that the CNC would be capable of doing what we want it to, and something you can easily upgrade later if your needs change.

[Mike.Mara]

5 minutes ago, MiKro said:

Actual cutting area sizes. I would think 900mm ,450mm, 150mm  This would be approx 36"x 18" x6" as a minimum.

Min on screws would be 1000mm, 600mm, 300mm. Question then, is 2 screws for long axis and slave them? Or a single screw with lower crossbar? Again cost is a factor. The screws are not that expensive, it is the bearings that add up.  Again if 2 screws for long axis then add extra electronics for that screw.

mk

On a machine as low budget as possible, I'd think 2 screws on the Y, I don't think it's particularly needed but then you need a stronger frame for the bed which would probably cost more than an extra screw, driver and motor.

[MiKro]

2 minutes ago, Mike.Mara said:

Well, personally I think 500mmX1100mm work area is adequate. That would allow Flying V/Explorer types as they're the widest I can think of, and through neck basses.

The PC is less of an issue now since you can get a UC100 USB (Which I currently use on a Windows 7 laptop with Mach3). I've not had any issues as long as I keep it only running Mach3 when the CNC is running.

I wouldn't say accuracy is un-achievable with a low end machine, mine certainly is low end and with rough calibration with a steel rule I got to within 1.2mm over a 711.2mm scale length and the intonation is perfect. This is also GT2 belt driven, cheap chinese electronics and a router with a decent amount of runout. So I'd say given some careful and thorough design work we could do better than that.

I also think that $2500 is a high number, unless you're talking HIWIN rails and ballscrews ect? I was thinking more supported linear rail and trapezoidal screws. Maybe prices are different in the US but over here you could get the rails and the screws for less than £250.

My initial thought was that the CNC would be capable of doing what we want it to, and something you can easily upgrade later if your needs change.

I understand, I was simply stating something for conversation and a start to the question.

MK

[Mike.Mara]

Just now, MiKro said:

I understand, I was simply stating something for conversation and a start to the question.

MK

I wasn't having a go, sorry if I came across like that! I was just bouncing my thoughts back 

[MiKro]

Just now, Mike.Mara said:

On a machine as low budget as possible, I'd think 2 screws on the Y, I don't think it's particularly needed but then you need a stronger frame for the bed which would probably cost more than an extra screw, driver and motor.

Again that is the question on that? which is the lesser of two evils LOL!!. In reality an MDF machine supported properly will get the job done. 

[MiKro]

Just now, Mike.Mara said:

I wasn't having a go, sorry if I came across like that! I was just bouncing my thoughts back 

No problem that is what I am doing as well.

mk

[Mike.Mara]

Just now, MiKro said:

Again that is the question on that? which is the lesser of two evils LOL!!. In reality an MDF machine supported properly will get the job done. 

Exactly the idea! 

And well... Lesser of two evils indeed... I would think a rigid enough bed to take repeated Z axis poundings would cost more than the extras for 2 screws on the Y. What are your thoughts on that?

I've seen a few MDF machines cut aluminium to an acceptable degree so I think it may be a good starting point, everyone can work it, you don't need a CNC already ect.

[Mike.Mara]

Also as a side note, can you get metric sizes over there easily? I know it's difficult to get imperial over here... Wouldn't be very good if we can't get all the parts both sides of the pond huh? 

[MiKro]

58 minutes ago, Mike.Mara said:

Also as a side note, can you get metric sizes over there easily? I know it's difficult to get imperial over here... Wouldn't be very good if we can't get all the parts both sides of the pond huh? 

Yes and know, as far as Screws, rails and 80/20 Aluminum extrusion yes. Most of the plywood or MDF is imperial but really metric since most comes from China now. LOL. That would be the easy part to convert though.

[Mike.Mara]

1 minute ago, MiKro said:

Yes and know, as far as Screws, rails and 80/20 Aluminum extrusion yes. Most of the plywood or MDF is imperial but really metric since most comes from China now. LOL. That would be the easy part to convert though.

Awesome, well at least we have common numbers in most of the important parts.

An extra 0.7/1.5mm or so here and there isn't too much hassle.

[curtisa]

Hell of a challenge you're taking on there, @Mike.Mara. I'm currently at the tail end of rebuilding my CNC too (not my first foray into CNC, but certainly my first build) and I echo most of what @MiKro has to say about setting the primary goals and limitations.

The machine I'm putting together is about 600mm x 450mm x 100mm cutting size. Frame is aluminium (or aluminum if you prefer) 30-series extrusion plus some 10mm Al plate for other components, gantry is the single Y-axis drive variety with the underslung crossbeam, Hiwin-type linear rails, 1605 ballscrews, NEMA23 3A steppers, 1.5kW spindle with VFD, generic M542 drivers. Mike's estimate of $2.5K AUD is not far from what I'll likely end up spending once it's all done, including some kind of control box to house all the electrickery stuff, 90% of which was purchased from China and assembled from scratch.

Rigidity, slop and backlash is good, with the exception of the Z axis carriage, which was my own fault for purchasing a prefab unit in the hopes it would save me some time and money by not having to make one. Two unsupported 16mm rails with one bearing block per rail just isn't enough to take the weight of the spindle, and the whole thing visibly wobbles in the Y direction. Lesson learned - I'll be rebuilding a new Z axis from scratch using profiled rails with two bearing carriages per rail like I should have done to begin with.

I already use LinuxCNC on a PC with the baby CNC3020 I bought a couple of years ago, rather than Mach3. As Mike suggested, I initially thought I was going to struggle finding a PC with a parallel port, but the $15 PCI parallel port card I bought from eBay has worked perfectly. I'm not sure if they're as effective with Mach3/4.

If you work slowly and incredibly carefully, and are willing to prematurely age some router bits, aluminium will cut with a hand router. I've made up several adapter plates for the ballscrew bearing supports by creating templates out of acrylic plastic on the CNC3020, cutting the 10mm Al plate as close as possible to the dimensions of the plastic template and finishing off the Al shaping using a router fitted with a template bit. I could have used the CNC3020 to directly mill the Al plates, but the process is incredibly slow and the use of cutting lubricant makes it messy too. That's something you may want to consider if you're looking at making a machine out of MDF. It might be strong enough to mill aluminium, but as soon as you start spraying cutting fluid onto the workpiece, the MDF is going to soak it all up in the process.

[MiKro]

I am also in a major upgrade on my original machine.  Mine also uses a single long axis screw with a cross beam. Added extra cross to prevent racking and 4 linear hiwin rails just for that axis. there is no flex :).

Currently, I have about 4k or more in the upgrade. The difference is I am overbuilding mine. I have no flex so far in any axis that I am working on at present X and Y. I also made my Z axis as well with full tramming ability and expect it to be very tight in tolerance I used round 20mm linear here 4 bearings. If not it will be rebuilt again. I also went with a VFD 2.2KW Spindle. all new electronics, including PC ( I3, 3.6 ghz win7 8 gb ram), this includes an ethernet Smoothstepper controller attached to a 6 axis BOB, larger stepper motors, higher voltage digital drivers, Cable chains, will also be adding LED lighting as well, the works so to speak. I also upgraded my Aspire to 9.0.  I have repacked the ball nuts with oversize balls and one extra. So far I have my backlash on the ball nuts to max +/- 0.0008 ( cold).  My new spindle has about 0.0003 runout as it is a 4 bearing model. Also adding a 15lb air spring to the X axis to equalize the load. as well as redid the ball  screw ends to accomodate 2 angular bearings and  an adjustment nut with thrust bearings for zero screw movement.

So my primary goal is to be able to mill Aluminum and build a better more accurate larger CNC. I do have extra ball nuts and designed in the ability to use 2 per ball screw so I can use one as a preloaded anti-backlash nut if needed.

@ Curtisa the parallel port is still supported with mach 3 mach 4 is totally new and has a plugin for it. I went with a controller so as to avoid the P port as well as gain the higher Kernal freq. since I was limited to 45k on the old PC. I should be able to run in the 100khz or better with this , hopefully moving my rapids up into the 400+ in/min. My steppers are 600oz/in 5 amp with technically 2005 sized screws.. Running 48volts.

[Mike.Mara]

Don't get me wrong, I agree with both your points... For a machine with a great deal of accuracy, speed and rigidity that original number of $2500 minus software seems a reasonable number, hell, for just HIWIN 20mm rails and 20mm ballscrews here for a machine of the rough dimensions to make a guitar, you're looking at almost £900. That's no bearings or couplings either.

Just to make the goal clear, I'm not expecting to be able to split a hair or take more than a couple mm's off the stock each time.

I indeed agree with the concerns of the MDF and cutting fluid, or any moisture in general, so we'll need to come up with a solution. What do you guys think of sealing with shellac and finishing off with a lacquer to protect it? I know you can route aluminium by hand, but I'm sure I'm not alone when I say that scares the crap out of me. 

I had the misfortune of planning my first CNC as a scratch build and deciding half way through that instead of supported rails and trapezoidal screws that I was going to buy a "kit". So I sold all the rails and the ballscrew I had for the Z axis and bought one of those Openbuilds OX monstrosities. (Not that they're terrible in their own right, but I had the misfortune of the only person in the UK at the time selling the kits deciding he was an engineer and re-designing it, terribly, he clearly has no idea of even the basic forces involved and the "rails" were cut terribly, the whole thing was as floppy and about as rigid as damp cardboard.)

I think with some careful engineering, aluminium profiles, strategically placed angle iron and the supported rails the rolling frame should be rigid enough to deal with light aluminium cutting and hardwoods at a reasonable feed rate.

It should go without saying (definitely will to you guys) that it's not going to be a professional machine, more of an entry level machine with capability beyond it's price range. I have no doubt that for less than my machine cost, one could be built that would kick its ass in both rigidity and cut quality. My original plan for my CNC, the one I was building, was an MDF frame (which I still have most parts of) and when assembled, it was a damn sight more rigid than the one I have now.

@curtisa Hell of a challenge indeed! I know it can be done, there will be compromise and limitations but as long as we're realistic like you said... I don't see why it shouldn't be tried. I've seen some badly designed machines made of MDF do decent work and even cutting aluminium. My machine like I said, isn't the best, or rigid enough, or many other things... But it can cut aluminium (albeit extremely slowly) and hardwoods with a reasonable degree of accuracy. Enough that I used it to upgrade itself from completely useless to usable.

Price becomes a major factor though, if it can't be done reasonably for £1000 or less (minus electronics, software) then it's not worth it.

[Mike.Mara]

Something like this for the X axis for example, should be more than rigid enough to do aluminium, and costs under £200.

The drawing doesn't show the trapezoidal screw, bearings, lead nut or housing for the nuts, but the major costed parts are included, including a back plate of 10mm aluminium.

The max travel for that would be 400mm, but the price increase to get to 500mm is minimal, and if using the smaller bearings could be made to travel 450mm. Those linear bearings are oversized for that gantry, they're just what I had on hand to put in there.

I would imagine after the coupler for the motor and bolts ect. that it would come to around £250.

Thoughts on that style? Any pitfalls you can see?

[curtisa]

My gut feel is that if you want to design a CNC intended for DIY-ing by the guitar building community, irrespective of what it is made from, you'll probably need to make it as plug and play-able as possible. The person making it is unlikely to have access to another CNC or milling setup capable of making the parts with sufficiently tight tolerances that make up the CNC (otherwise they wouldn't be making it!). Factor in how accurate is accurate enough when measuring, making and assembling the various components, and what the minimum tool set is required in order to achieve it.

Most DIY guitar builders will also freely admit that their electronics skills are limited to following wiring diagrams, with little understanding as to why things are done the way they are done. Getting the steppers, power supplies, limit switches, drivers, interfaces etc working together will require a certain amount of knowledge as to how to make it all go, and what to look for when it doesn't. The unitised nature of larger/more elaborate control setups also place an additional demand on wiring abilities - knowing what gauge of wire to use for steppers to minimise voltage drop, what insulation level for the incoming mains, the importance of earthing, the dangers of exposed mains connectors are all things people who only DIY guitars are unlikely to know about or understand. If you can specify an all-in-one self contained control system it's likely to work safely right from the get-go for nearly everyone.

There's also only two CNC motion control applications (that I'm aware of) out there - LinuxCNC and Mach. LinuxCNC is free but will require a willingness to play with Linux. It's also not necessarily a solution that works straight out of the box either (I've installed LCNC onto two PCs - one worked flawlessly from the start, the other took lots of experimentation, researching forum topics, hunting for BIOS settings, writing scripts, modifying boot parameters etc to get it to play nicely). Mach may be slightly more user friendly and runs on Windows without too many foibles, but will set the user back a couple of hundred smackers

[curtisa]

17 minutes ago, Mike.Mara said:

Any pitfalls you can see?

Only as I've said above - factor in how the DIY builder will cut and drill the necessary components that may require a certain degree of minimum tolerance. The two endplates with the bearing mount drill holes ,and ensuring baseplate extrusion is cut squarely at the ends where the endplates butt up against are good examples.

[curtisa]

5 hours ago, MiKro said:

@ Curtisa the parallel port is still supported with mach 3 mach 4 is totally new and has a plugin for it. I went with a controller so as to avoid the P port as well as gain the higher Kernal freq. since I was limited to 45k on the old PC. I should be able to run in the 100khz or better with this , hopefully moving my rapids up into the 400+ in/min. My steppers are 600oz/in 5 amp with technically 2005 sized screws.. Running 48volts

I knew the parport was supported in Mach3, just wasn't sure if a plugin PCI parport would cause issues. I'm not pushing a spindle around over large distances for production runs, so blisteringly fast speed isn't big on my list of priorities and I can live with direct parport control for now. 

[Mike.Mara]

Very good points!

I've been spoiled by Misumi and their cutting service, they've always had square ends. Misumi Europe anyways.

The end plate tolerance I think can be done reasonably if you order a laser cut template. (For those guys without a CNC).

I totally agree about the electronics, that's something I've not considered a problem since at this point it's almost second nature to me. A ready to go system I'm guessing is going to be a budget blower for a lot of people.

With a digital caliper and a marking gauge I think most people with a pillar drill should be able to get pretty accurate holes for mounting everything, or am I way off base here?

Cutting may be a different story... I know here in the UK most suppliers that I use deliver parts ready to use. That may be a different story everywhere else though.

I guess this is why I started this thread lol... So people like yourself can poke holes in the idea so we can try to make it as straight forward as possible.

I mean there is always the option for those who can't do it themselves to outsource to a water jet cutter or someone with a mill, although that will add some to the cost. I guess as long as we try to keep the parts as small as possible that would reduce the cost somewhat.

Mach3 I've found is a bit of a learning curve but once you've got the hang of it its super simple... Yes it does cost, but I think as long as we lay out all the costs up-front then there won't be any nasty surprises for anyone. I know there were for me what with CAD and CAM, not to mention Mach3.

[curtisa]

14 minutes ago, Mike.Mara said:

With a digital caliper and a marking gauge I think most people with a pillar drill should be able to get pretty accurate holes for mounting everything, or am I way off base here?

Sounds good. I'd also add a square, router with pattern-following bits, brad point and forstener drill bits (if we're still looking at an MDF-framed machine), steel ruler big enough to cover the longest edge, a compass or trammel. Most of these items a DIY guitar builder will likely have access to anyway.

 

23 minutes ago, Mike.Mara said:

I mean there is always the option for those who can't do it themselves to outsource to a water jet cutter or someone with a mill, although that will add some to the cost. I guess as long as we try to keep the parts as small as possible that would reduce the cost somewhat.

If you can keep it within the realm of 'doable by somebody with a modestly equipped workshop', it is more likely of success. The added cost of third-party cutting and shaping is likely make it a non-starter compared to buying a CNC6040 off eBay.

Perhaps the JGro CNC might be a good basepoint to get some ideas from?

[Mike.Mara]

All good tools to have for this!

I think a mix of MDF for the more complicated parts and aluminium for the parts that are less so should work a treat. (Like the back plate on the X if it were aluminium it would only need a series of loose tolerance holes for the extrusions to bolt to. A mm here or there makes no difference.)

I was hoping to use a fair bit of aluminium profile for the base, so it's just bolt together and the rails bolt right on too.

I agree it should be kept as doable in a home workshop as possible, the more off-the-shelf parts and bits people can make themselves the better.

I'll take a look at the JGro

[Mike.Mara]

Well, I had a little play and came up with something like this: Not complete obviously.

The brown parts being MDF.

The cost for framing, aluminium, rails, leadscrews + nuts and couplers is right around £580. I'm going to take a wild guess and say the rest is about £200 for the T-nuts, bolts, bearings, shaft collars ect. (Not ballscrew bearings).

So should be just shy of £800 for all the mechanical parts. Electronics + spindle + USB driver (incase people want to use laptops) ~£430.

So in total about £1,250.

Work area is ~X 500 x Y 1100 x Z 150. (Z could be raised or lowered by adjusting the length of the Y plates.) Machine footprint 1400x720x430 minus electronics.

Like I said, just rough... Just testing an idea to see if you guys see anything I don't or if you have a better idea.

(20mm rails on X and Y, 16mm on Z. Extrusion is 40X80mm, MDF can be either 18mm or 3/4" without any modification.)

[MiKro]

My Z axis.