Machsupport Forum
Mach Discussion => Video P*r*o*b*i*n*g => Topic started by: resonance on August 26, 2015, 05:39:48 AM
-
Hello!
I'm into the mouthpiece business and I have made many so far using mach 3 with a couple of lathes.
Now I bought a 4 axis mill that I'm using for engravings and light 2.5 sculpting on brass plates.
From time to time I have customers asking to copy their mouthpieces and most of the times I ask a friend of mine to make scans with a "arm"specifically made for this kind of jobs.
Unfortunately he lives far away and timing can be an issue for some customer.
I then thought to make it on the mill or on the lathe.
Can mach 3 do a scan of a curved shape?
What probe is best for this job?
Since a mouthpiece is a simple Mix of curves revolved on one axis I need the probe to scan only on two axis: let's say I want to measure the width at different depth values.
Can I then obtain a cloud of points that I can then transform into a complex curve?
The minimum diameter is 3.5mm.
Can the program compensate for the probing tip diameter?
Thanks!
Inviato dal mio iPhone utilizzando Tapatalk
-
From a Trumpet player ........... :o
In my lathe screen I created some buttons that will capture the manually probed X and Z points and then can create a version 12 DXF
file from the points or draw the lines. Using CAD you can create curves based on the points. Mouthpiece exterior profile is easy.
The rim and end edge contours require more attention to duplicate.
So it's manual probing. I use axis MPG's to assist in touch off along with a light indicating probe. This allows control over
accuracy and location of the points ( +-0.0002" repeatability). Plus I can make custom probe tips and no compensation is
required. If auto probing is used, ie; G31 then a very slow probe rate is required ...like 1 or less ipm. Thus manual probing as it's quick and easy.
Just a thought... maybe make custom visualizers such that they match the mouthpiece being used! ;)
I researched probing for the lathe and not much out there or you get high end and it's expensive.
Just some thought's,
RICH
-
RICH I have what you need to do auto 2d probing for turn. No point in doing it manually when you can push a button and watch with a cup of Joe in your hand.
It can be done in XZ linear fashion or a XZ Radial approach that is tunable so you could probe around a corner very easily. Do it all from one setup.
Linear probing of curves is not that accurate due to tangent error of the tip and approach angle. You can ether do all the math to corrct OR work smart and avoid the needed corrections.
Just a thought, (;-) TP
-
Hello, today me and Richard spoke a while on Skype about probing manually or automatically. He also pointed me to the lathe probing wizard you uploaded to the forum (thank you BR549!!!).
It took me a while to understand why I couldn't write the data to a file but finally I made it!
Since I don't have a probe yet I switched the spindle index interrupt to the probe one and make a few tests faking contacts by manually moving the spindle. It worked and now I can easily get data if I buy or make a probe.
The kind of shape I need tip scan brings me to the next problem: is there a way for Mach3 or any other software to CORRECT my cloud of points taking into account the tip reasons of the probe?
The other option would be to use a tip so small to almost eliminate the problem mechanically but it would be sharp and I can't scratch Mouthpieces that cost up to 400$...
-
IF you use a small spherical tip made of polished RUBY you should NOT have a scratching problem. Tip size around .050"/1mm. Even that small you wiil get tangent tip error BUT they will be very small. IF you are referring to tip radius comp Mach3 has that built in. But it only accounts for tip radius NOT tangent error.
(;-) TP
-
Hello!
Unfortunately my products need to be insanely precise and I can't exceed one or two hundredths of a mm.
Another option is to use a VERY thin internal grooving insert. Due to the shape I need to scan I just need one corner of the insert to touch the surface and, maybe, it just takes the insert to be gold plated and wired so that when it touches the silver/good surface I scan the contact will be closed.
Does it make sense?
-
IF you are talking .01-.02 mm (.0004-.0008) then you are NEVER going to get that from Mach3 probing a shape such as that. And without temp compensation you will never cut or measure it accurately.
Just a thought, (;-) TP
-
You mean the probing routine can't obtain such a precision or the progress aren't that precise?
-
I always say that accuracy will only be as good as the machine " system ".
By that I am meaning:
How repeatable is the machine ( it cannot be as good as the calculated resolution)?
How repeatable is the probe?
How repeatable is actual cutting?
SO
Any movement, sensitivity of a probe, and any actual machining, etc, etc will influence / determine how accurate the information will be.
Thus it's all that makes up the routine / process.
RICH
-
The lathe consistently cuts with one thousand of a mm precision, servo motors and zero backlash bell screw. It's insanely precise, beyond what I expected!
I can't say for the probe, I still have to buy it.
I spent the morning with that wizard for lathe and, in fact, is very easy to use and seems to compensate correctly for the tip radius!
Can't wait to have a real probe to test precision and repeatability.
-
Update: probe arrived, connection made, stylus installed (Renishaw, 15mm long, 1.5mm ball).
I had to install ProbeIt Turn (30 dollars VERY well spent!).
Here are the problems I had to solve: the software can't handle both big diameter ruby tips. 3mm with a step over of 0.5mm won't happen.
If you need a small step over like 0.2 or 0.3mm you need a 1.5 or 1.0 ruby tip.
The software is very easy to use but it needs many points and very close together to behave in a reliable way.
Export is both points and lines Ina DXF file, easily imported into any cad program.
Installation is very easy if you know how to solder two wires and how electricity works... I don't so I asked a friend and watched him doing it. It was easy...
You have to map one of the input channels to Digitize. The probe is always short circuited unless you open it by pressing the probe. This is for safety, makes sense...
The logic of the ProbeIt Turn is simple and clever: you first touch off the piece telling the wizard how big is the radius it's touching. Then you tell it how to move (clock or counter-clock wise) and then it will start probing. Every touch it makes it will move further on the side. If it senses the surface is curving it will try to guess the corner radius and thus it will make a "side step" that will follow the surface to be always square to it.
Problems: the stylus is always prone to some bending when it touches the surface, this depends on the stylus length, thickness, probing feed and probe sensitivity.
All of this can be solved by a little bit of math: you prove an external diameter, then an internal diameter. You calculate how smaller the distance is and this is the error. Divide it by two and make an offset of the scanned profile. Then simply translate the profile towards the center by that amount.
If you have backlash just enable the correction under Mach3, it will be cured properly.
The final test was to cut a mouthpiece with my lathe and then immediately scanning it.
I made a few cross checks and the measures are off by less than 1/100th of a mm!
There won't be many trumpet players here but this procedure can be useful for other purposes as well.
Have a good day!
Inviato dal mio iPhone utilizzando Tapatalk
-
Thanks for the update. Was wondering how things were going.
Did they fix the problem with the new lathe and is that what you used?
I downloaded ProbeIt but have not tried it. It will not work with my new screen set and have not contacted them to see if I would be able / allowed to change the scripting
if purchased.
You can do some calc's to find how much over travel post probe touch off due to Debounce and Deceleration.
I use a slow federate of 1 IPM for probing and based on that the calculated overtravel into the material is 0.0000390" and repeatability here is +- 0.0002" with
a "fixed probe". The probe I had for years just is not as repeatable and accurate as the fixed probe, and on the small lathe, the probe needs to be re-positioned.
I'll send you info, from my screen manual, via Skype for doing the calc's.
BTW,
Can you make a custom mouthpiece such that I don't need to practice, don't have to expend much energy to play those double high C's, and sound like Maurice
Andre? :D
RICH :)
-
Regarding precision: a trumpet mouthpiece is between 5 and 15 arcs depending on complexity so, in the end, you don't need so many points to infer what's the shape you are copying. What's more important is the offset derived by how much the stylus bends before you have a contact.
This, as you wrote, takes a bit of calculation but it's not that hard.
Let's meet on skype, thanks!
Mouthpiece: I had it, then I exchanged it for a new lathe from Germany that is still not working... I'm am pissed of, badly.
Inviato dal mio iPhone utilizzando Tapatalk
-
Here are the problems I had to solve: the software can't handle both big diameter ruby tips. 3mm with a step over of 0.5mm won't happen.
If you need a small step over like 0.2 or 0.3mm you need a 1.5 or 1.0 ruby tip.
Inviato dal mio iPhone utilizzando Tapatalk
Tony,
Glad you got all sorted and are making useful measurements! You are correct about the probe ball size relative to stepover... but the issue is really just trigonometry related. As the ball becomes large compared to the stepover, any 'error' in how squarely the probe hit occurred relative to the surface of the part translates into large relative error of where the touch is calculated to have happened. That error then compounds in the following moves. Thank you for reporting out your results and feedback.
You're also right that deflection errors can be calculated out. The 'mill' version of ProbeIt can take it out by probing a ring gauge, but not the 'lathe' version. Main reason is that you can really only hold the probe one way in a mill, but you can orient a probe in several different ways on the lathe... tip along x, tip along z, held at an angle, etc... so I didn't want to add an additional complication on the lathe version (which was really just an experiment)
Rich,
Any easy workaround to a screenset conflict is to just make a new profile by duplicating your existing profile, and revert that copied profile to a standard screen set and radius mode (if diameter mode is your standard work enviornment). You can then use the wizard from that profile without issue.
regards,
Eric Brust / CraftyCNC
-
Your lathe may be able to make a cut that accurately, but one minute later holding the part in you hand where it warms up all bets are off. I'd spend my time finding out the limits of accuracy required, in other words what the customer could detect, then work from there. In the real world accuracy is seldom the issue, and repeatability is much more important.
-
Thanks Eric! I think I might achieve even better results with the mill version of your wizard. Mill unfortunately is out of order at the moment so that's why I've resorted to the lathe.
Gary: in my. Business the amount of material you take away during the final polishing is more than the precision I get while making the measure so there is no big concern. But yes, the better the measure the better the result. :)
Inviato dal mio iPhone utilizzando Tapatalk
-
To measure and cut exactly teh same probe and cut with the same radius tip. IF you are measuring with a 3mm probe tip then cut with a 3mm radius cutter. NO trig involved (;-)
(;-) TP
-
Eheh, unfortunately i need to use a 0.2 or 0.4mm radius tip for cutting. :)
Inviato dal mio iPhone utilizzando Tapatalk