I am still working on this. Have some personal issues to handle.
Need to thicken the outer ring and figure out some feeds and speeds.
more to come.
No file updates this time, just a picture.
will work on Bill’s idea.
we can do this with the coaster idea and still make usable pieces using Bill’s idea.
Just turn the disk 90* and mill 2.
Yeah, my thoughts were Benchy is supposed to be tough to print, forcing evaluation of all aspects of your print process. The same thing should be true with a MPCNC/Low Rider test, it should be technically difficult but possible when things are all working right. If we have a known good gcode file then it tests the assembly, once the known good file cuts correctly the user generates their own version and verifies it still cuts right, showing their CAM skills are as good as their machine. 
Yup, I am still to timid to test my own skills like that.
Who is going to be brave enough to go first?
I guess if I clean off my table, double check the the endstops and give it a try…
What would you consider amazing reasonable tolerances in say HDPE (most accurate easy to cut material I have on hand in two colors). 0.15mm on each part for a total of 0.3mm (0.0118") or is that being too generous? Maybe a test with inlays that step down from 0.5mm gap to perfect fit?
I think there can be specific sequences that can highlight specific imperfections if present, and it’s helpful to incorporate as many of these as possible. Some of these sequences might not be easy to generate from CAM and might be simpler to just write by hand.
Lately I have been contemplating the problem of determining precise tool offsets with a simple and painless process, and I made an attempt at a test pattern that would tell me some things. The g-code is simple, hand-written:
G92 X0 Y0 Z0
G1 Z2 F300
G1 X0 Y20 F1000
G1 X0 Y20 Z0.5 F300
G1 X20 Y20 Z-1.5 F300
G1 X20 Y0 Z0.5 F300
G1 X38 Y0 Z0.5 F300
G1 X38 Y20 Z-1.5 F300
G1 X58 Y20 Z0.5 F300
G1 X58 Y38 Z0.5 F300
G1 X38 Y38 Z-1.5 F300
G1 X38 Y58 Z0.5 F300
G1 X20 Y58 Z0.5 F300
G1 X20 Y38 Z-1.5 F300
G1 X0 Y38 Z0.5 F300
G1 X6 Y32 Z0.5 F300
G1 X26 Y32 Z-1.5 F300
G1 X26 Y52 Z0.5 F300
G1 X32 Y52 Z0.5 F300
G1 X32 Y32 Z-1.5 F300
G1 X52 Y32 Z0.5 F300
G1 X52 Y26 Z0.5 F300
G1 X32 Y26 Z-1.5 F300
G1 X32 Y6 Z0.5 F300
G1 X26 Y6 Z0.5 F300
G1 X26 Y26 Z-1.5 F300
G1 X6 Y26 Z0.5 F300
Which generates eight L-shaped cuts in a pattern like this: [attachment file=97879] (I found if I rub the surface with pencil first, then it’s easier to see where the cuts are.)
There are a lot of measurements that can be made from a test cut like this. But my point is not to propose this for benchy, rather it’s the more general concept of creating patterns that highlight specific defects. In my case I am thinking if I use one tool to carve (or draw) the upper-left and lower-right L shapes, and another tool to carve the lower-left and upper-right L shapes, I can take measurements and get pretty accurate with tool offsets.
I agree that it would be desirable to have a finished product that’s attractive, useful, and branded. I just think it’s also helpful to try to think about fiducials or something that can be incorporated to make evaluation easier.
I guess I should sort of explain my distance from this one a bit, maybe you all can work me through it. I do love all the ideas, but… I know when something like this comes up there is going to be a whole new flood of tuning questions. Just like people asking “what tolerances does your MPCNC/LowRider do”. I am not interested in helping people tune out the last 0.1mm from a cut, it is next to impossible over the internet.
For most people none of this matters, slap down some old plywood from the backyard and carve your kid’s name in it, win! I do think there should be some sort of milling test like the Crown, just a basic cut, since most use the crown itself as a milling test, something a bit more advanced, coaster, would be cool. We all know that different sized machines, different spindles, endmills, material, etc, is going to complicate me spitting out some golden code that will work on everything.
I am also interested in something like Bill is recommending, something like Maker Muses’ Tolerance torture test, so it is very clear that it is a brutal test and give you a sense of where you are at, for this I do not think Code should be supplied, but a good model/DXF should be.
So a learning path something like, Crown, a more accurate plotting test that can be measured for accuracy, coaster/LCD caps that can be measured for accuracy, Bills torture test?
Sorry for the long post, I am working on the instruction updates and new control page ideas and this is fresh in my mind.
Oh, and I really really, like the idea of some sort of functionality even if it is not perfect. Milling takes a long time, might as well get some sort of benefit out of it.
I agree. For this there is no magic gcode. Material, Spindle, bit, and the mpcnc build all play a part in the feed and speeds.
A good video tutorial and a DXF of something simple to produce (coaster, logo plaque).
The coaster has a lot of angles and geometry in it already.
It just needs predetermined measurements to let you know you are in the ball park.
Anyone looking for $10,000 accuracy can go buy a $10,000 CNC and have at it.
However, wouldn’t it be cool to find out that if everything is put together correctly that you are getting that level of accuracy across builds? That would be a mile stone. Yes, probably unlikely. Who knows?
Import attached as mm to Estlcam and try it out?
5" x 10" material minimum
Scale of each coaster is 4"
My Grid is 1"
Should have some hangover on material
Use Foam board first to test.
Rename to BenchyV1Logo.dxf
BenchyV1Logo.dxf_-1.txt (56.6 KB)
If it were to be a carving test we could angle the cuts for the inlay so most likely they’d fit fine, just not completely seated or slightly too seated if the size were wrong. That’s going to require a tool change though, since you’d want a V bit. Maybe for a V2 version where the technical side gets more difficult. A Benchy style part should be simple to try, but hard to master. Remember that if you get close you can always manually do some sanding to get things to fit and you will still have the object to show off. So, I’m still suggesting the same thing… A void cut and an inlay cut sideways such that once you clean the tabs off you just rotate 90° and drop it into the void where it fits perfectly. I can’t think of any way to do it without needing to cut tabs off, and that opens the door to a little fudging on the final quality.
I don’t think success is measured in your tolerances as much as in the fitment. If the whole thing is done at 97% of perfect, it still fits together and you get your V1 Benchy coaster with bragging rights. When you measure it you’ll know more about your machine, because it did fit together, but was slightly the wrong size. If your inlay doesn’t fit then you have outside vs. inside cut issues and might look at that 3mm bit that’s actually 3.175… Anything that leaves you with shapes that are wrong means you have something like a slipping pulley. Over time it turns into a diagnostic tool. “Oh, I saw someone last year that had exactly the same problem, check to make sure you don’t have one 20 tooth wheel instead of 16 tooth like the rest.”
The biggest issue we’re going to see is not every machine is the same. Unless we state the benchy should be done in foam, we’re going to have a lot of failed cuts. Something my machine can cut isn’t necessarily something Bill’s machine can cut. Maybe my machine is a few inches smaller, or larger. That’s going to change the speeds and feeds. So now we can’t all use the same gcode. It works for the crown test, plotting has no cutting forces. Now if the benchy is just the dxf files, that would be great!
That does sound great Bill. I think we are getting close. Part 1 is actually a basic test, lets say Jamie’s coaster with dimensions. So you cut out the first part, get all your feeds and speeds right to get the best dimensions you can. At this point you might have 2-3 coaster, fully functional even if they are off a bit. For most that is a great place to stop. For the brave Part 2 would be the “inlay” cut 90 degrees off, just to prove your build is awesome. Now the part 2 DXF can be available in multiple clearances…right?.. At this point you can use these to test your skills. I think.
To make this easier, just using the logo symbol with no sharp corners would be cool.
Each dimension is an even metric number so all axis can be tested at once. You will know which direction the inlay is cut by the grain.
We can make them stack-able to test a flip cut (test 3) and then V-bit chamfer the cut and add the logo with small text and points (test 4 tool change)
I agree. I’m willing to test whatever you come up with.
Would this work with dual and non-dual endstops?
The inlay idea is good too. It’s a bit advanced, but can teach a lot of concepts.
My thoughts were always a multi-step process, of different cuts, one more advanced then the previous. By the end, one will have learned a lot.
Is it safe to assume single flute flat tip bit, 1mm DOC per pass, 10mm/s feed, total depth 12.7mm, 5% step over, would be safe for any machine and wood, excluding metals.
There has to be a safe zone in the numbers that could transcend any build.
All tests first done in foam of course.
I like the idea of many coasters each learning something new. I have another idea to add to the mix. When we start flipping them over to make the stackable we could also have a tutorial on how to create a logo or signature file that can be carved onto the bottom. Granted it might be better to just make a stamp. And that tutorial is already made.
That is funny you mention the signature. I was thinking recently the exact same thing.
I have had a lot of things going on and have not had time to start really cutting wood.
But, that is on my agenda. Stamp is ok, but I think carved is better, won’t rub off.
Thinking about the tool changes, with the entire process, we could be looking at 3 different tool changes.
Flat tip - carve out the center and depth, no tabs on the top side
V-Bit - for the V1 logo
Bottom Side
V-Bit for the personal logo/signature
Flat tip - carve the stacking groove and then the full cut with tabs
Still working through the processes.
Also, thinking that this would only be a DXF file and let the user create the tool paths
We could also advise on how to create this in Inkscape.
With a guidance on speeds and feeds and DOC’s, but not set in stone.
No gcode file provided.
While there is the tutorial on the side plates for the LCD.
I was reading through it, it’s bit more advanced.
I think this Benchy is more introductory to beginner+ process.
Yup, not quite complicated enough yet Ryan. The logo surface is down from the outside ridge as you show, but the area outside the logo should be pocketed down some from there and the center area inside the logo (which will be used for the inlay) should be pocketed even deeper. When the user moves beyond the inlay to the other side we’ll cut the outside to match the red part of the logo depth and do something wonderful and difficult as the signature. (maybe the tiny MPCNC drawing you’ve been using with a V bit?)
Sorry for resurrecting an old thread, but I am new to CNC and I was thinking, are there any torture tests like there are in 3D printing to help tune/debug/optimize your build? I came across this thread and thought, great, just what I was looking for. And then it just peters out. Did anything ever come of this?
The crown with a pen is decent test for no-load accuracy, since it’s easy to see if the lines don’t meet up.
For cutting under load, the cnc race is probably a good test and illustrates an average decent starting point compared to pushing a machine to its limits.
In both cases there are enough examples of what other people have been able to achieve, so you can get a feel for where you are as far as build quality.
