Official Luthiers Forum!

Good to hear Laguna's service has improved.

I don't have experience CNCing shell, so not sure about the requirements -- Andy may chime in on that. I think the issue is whether the bits will break -- a 0.002" repeat means the bit may see up to 2x that (from +0.002 to -0.002) much material at times when the code didn't intend for the bit to cut that. My guess is a 0.002" repeat would be ok if you go super slow feed and very high (50k+) rpm, but I'd bring my finest design and hardest material to the vendor so they can do a test cut with a 1/64" bit.

I believe K2's 0.001" repeat refers to the XYZ axes less router.

_________________
David Malicky

Take a close look at the z height of the Laguna tool.

I have a local friend with a Laguna IQ Pro. The z height is minimal and only allows a neck carve to be done in two operations...from each side. This works out for him because of the negative draft in his heel design (you have to approach the neck carve in this way with a negative heel draft) but it limits him from ever doing necks in one operation if he wanted to redesign the heel.

Repeatability of .002" is fairly common with table top cnc routers. Larger, production type cnc's are around .0003"...some even finer than that. The .002" routers are commonly used for inlay. The finer the better, of course but the typical high rpm of a router allows for less stress on a tiny cutter when engaging that .002 "slop". Once you determine your safe feed rates you'll be able to predictably use a typical table top router for inlay.

Tool changers aren't exclusively beneficial for high production. Having options for which tool to use will inevitably lead to better choices and better quality. Without a changer, you'll often be tempted to minimize the number of change-outs by using less tools. As an example....on a fretboard I do several preparatory cuts that establish sharp corners and edges prior to cutting the top and side surfaces. These cuts are often made with reverse direction cutters. The tool paths are designed with the wood grain in mind...controlling the direction at which the flute attacks the grain. This guarantees no tear-out. It's a minor programming hassle but without a tool changer...I'd never do it.

_________________
I read Emerson on the can. A foolish consistency is the hobgoblin of little minds...true...but a consistent reading of Emerson has its uses nevertheless.

StuMusic

Thanks, Dave and Stuart.

Stuart, the z height on the Laguna (both the HHC and the Pro) is 6" clearance, with 5" cutting height. I'm surprised your friend can't clear a heel with that, as I would have thought 5" cutting height was enough.

I like the tool changer idea, but there is a pretty large price jump to something with a changer. I may go without one for now and (gulp) buy another tool down the road if it comes to that.

Thanks again!

Stuart, one more question: I would love to hear any feedback you can pass along from your friend with the IQ Pro -- what he uses it for, what he likes, what he doesn't like, etc.

5" of cutting height lets you fully work on a 2.5" tall part at a maximum, or 2.5" minus the radius for a ballnose cutter. The spindle needs to be the tool length above the part for the tool to clear.

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

He's using it for fretboards, headstock facings, bridges, and much of his inlay work. He also does some of his headstock profiles and tuner holes with it. I made him a routing and drilling fixture that he uses for another model.

He's learning still. I expect he's dreaming of more applications but his modeling and tool path generating experience is limited so it's a work in progress for him. I know he wants to make his necks with it but he's not quite there yet with the computer work.

He doesn't have any criticisms of the machine and I haven't enough experience playing with it to really render an informed opinion. We did have to tear out the table top and replace it with a spoil board surfaced by the machine itself. I imagine that would be necessary to do with about any machine though.

_________________
I read Emerson on the can. A foolish consistency is the hobgoblin of little minds...true...but a consistent reading of Emerson has its uses nevertheless.

StuMusic

Great info. Thanks, Stuart.

I think the repeatability thing is more of a hedge by Laguna so that people are happy with what they get. I looked over the specs and with the hardware they have, I'll bet that precision is closer to .001 than .002. I'll bet that it's really their accuracy that's closer to .002.

The machine has ballscrews on all axis which should be good for better than .001. The round rails might not be as good as the profile rails on the long axis but they are supported so should be ok. The thing is that aligning all that is difficult and I'll be that's where the .002" thing comes in. I spent a lot of time aligning the rails and table on my machine and I'm probably only good for about .002" myself. Precision is under .001 though so in the end it all pretty much works out.

It sounds that for the price, it's a pretty decent turn key system. It looks like it moves a bit slow and the control software might have some limitations but I imagine that it would work fine for inlay. Would be curious if the control software has S-curve acceleration. Mach 3, which I use, does not and having it would definitely lead to fewer broken bits.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

Is there a standard definition for repeatability ? From my looking into machines vendors seem to use different methods of measurement, some coupling repeatability with homing position accuracy.

Just to muddy the water, Camaster's Stinger I (tabletop) comes standard with 5" gantry height but they have a riser option that can add 2". They don't spec the Z travel but I'm guessing it is probably 5" too. They list repeatability at .003 but when I emailed their sales guy (J) asking how they define it he went silent on me so that's a pretty big red flag not to answer a simple question but perhaps they very overwhelmed there.

On the whole metric controller thing has termite (mike) confirmed for you that Rhino and madcam will work properly when switching back and forth between units in a model ? I image a workflow where you design your parts in imperial units then change the units definition in Rhino, let it scale everything, do your cam programming in metric and post, then change the units back to imperial and save etc ? It would be great if he could help you understand the workflow. If Stuart's friend is dealing with that, maybe it's not such a big deal but best not to be surprised.

Not being a machinist by trade, I don't have a real answer for that but I expect that for industrial machines, repeatability actually means repeatability under normal operating conditions. e.g., if you start from scratch two different mornings you can expect your parts to be within .0003" of each other.

For the not-so-industrial machines, I think you're right that it's a combination of "slop" in the system, which in most cases is pretty low, combined with homing, accuracy in the rail alignment etc.

Practically speaking, I think these machines perform better than the specs. C5 ballscrews are toward the lower end of the accuracy scale and even they are good for about .0011 over 48".

What does this mean? I think it means that with a ballscrew machine, you can hold tolerances between parts under .001" if you're careful. Some of the things to consider:

1) Eliminate any backlash in your system - this will break bits and cause problems
2) Set your zero (or home if that's what you're doing) and do all of your operations with that zero.
3) if you're having problems with fit between an inlay and a pocket, use THE SAME zero for the pocket and the inlay. This means centering the workpiece under the bit (which is at 0,0) rather than jogging the machine to find the zero marked on the workpiece.

By doing this, any positioning errors of your machine will be minimized.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

I think the way to look at "repeatability" is...how close does the machine actually get to where you told it to go?

_________________
I read Emerson on the can. A foolish consistency is the hobgoblin of little minds...true...but a consistent reading of Emerson has its uses nevertheless.

StuMusic

Stuart, I'm not so sure that what you're saying is what they're quoting. It probably is for an industrial machine but not for these little ones.

e.g. With the equipment I have, I've got my machine trammed/squared/aligned to under .002" (Better where I could). I don't know the grade of my ballscrews but let's just say they're good for .001". So grand total, my error could be up to .003" from where I commanded the machine to go.

However, my homing system is accurate to well under .001" and because of the nature of my drive system (i.e. zero backlash ballscrews, profile rails), my repeatability is very close to my homing accuracy with any additional error coming from the my stepper driver (have a micro stepping driver and holding anything under a half step is somewhat arguable though when just playing with a dial indicator I can subdivide .001" into 4 discrete movements).

What this boils down to is that my accuracy has a range of about .003" but my repeatability has a range under .001". Accuracy vs. Precision.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

I should add that while I don't have a rack & pinion machine, it's my understanding that unlike ballscrews, there is inherent "slop" in the straight geared systems that most lower end R&P systems use that does affect repeatability and accuracy.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

Most of the router industry specs performance with at least: repeatability, rapids and cut speeds, and for steppers resolution. They generally also disclose something about the specs of the parts the machine is built with, and to Andy's point you can estimate what a well designed and built machine should be capable of. Laguna has no published performance specs, and regarding components other than steppers, ball screws and some kind of linear guides on the long axis there is little in writing to go on. If the machine they send you doesn't work as well as you want it to, there isn't a discrepancy to the published performance to help you. It's a matter of expectations. For a blowout priced Chinese router what should you expect and is that a good match for what you want to spend that money on. I'm not negative about Laguna, I have an LT-18 bandsaw and it's OK and they have been OK to deal with, but nothing more than OK.

Other router companies have told me that their repeatability spec was not based on what the machine will do but rather a performance spec that they feel comfortable it can achieve and that they are willing to stand behind (lawyer number). If that is the most meaningful parameter they'll spec I wish there was an accepted definition and testing method.

The main losses of accuracy in routers aren't due to the drive components (assuming ballscrews and rails), but the frame. Cutter deflection can be huge, things flex when reversing direction, things sag, etc. On industrial mills, they use a test called a ballbar that you'll get the results of with your machine. That's really the standard for testing accuracy and precision, but I'd bet my hat that 90% of companies selling CNC routers don't even have a staffer who knows what a ballbar is.

A ballbar test shows the aggregate of all errors, so things like frame flex, etc, will show up. A poor man's ballbar test, for me, would be if they gave you these specs:

-Ballscrew, rail, etc part spec. accuracy
-Tramming/squaring/parallellism accuracy of axes
-Straightness of rail mounting
-Deflection of the table with 100lbs in the center
-Deflection of each axis during a full-speed turnaround
-Deflection of the spindle during some 'standard' cut (1/4" endmill going through MDF at 100 IPM and 10000 RPM)
-Following error running a 'test' G-code file

When it comes down to it, I think stiffness is easily the biggest source of errors on light machines, so get the beefiest frame you can. At the same time, you can never really win with chasing errors. My old machine (a 1994 Fadal 4020 mill) had mechanical accuracy unattainable on a router: it was 10 tons of precision ground cast iron with cooled 2.5" ballscrews laser calibrated at the factory. It also couldn't interpolate a circle accurately at feedrates over 25% of max with the original controller, because the servo loop wasn't fast enough to account for the inertia of the table.

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

+1

Flex in the frame, especially the gantry causes nothing but frustration if you're trying to do accurate work. Not a big deal if you're a hobbiest wanting to do a few 3D carved signs etc. I bigger deal if you're trying to make parts fit and holes line up.

Andy is pretty much correct in his assessment. You can model & carve 3d archtops with Aspire using the 2 rail sweep or other methods, but it's probably a lot easier (and more accurate) to model it in something else like Rhino and then import the model. It works great that way and it's a lot easier (for me at least) than working directly from Rhinocam or another CAM program within Rhino. There are some limitations in machine strategies but for me it works really well.