Notifications
Clear all
Accessories
6
Posts
2
Users
0
Reactions
5,929
Views
Topic starter
Hi,
Just finished making the hardware for my tool length sensor. It uses the classic 6 balls, 3 rods technique and from preliminary tests appears to have a sensitivity and resolution in the sub 10 micron region. When mounted on my mini mill with a DRO reading to .01mm, I can repeatedly go between on to off states without the DRO changing by one digit.
The 6 off 1/8" balls are polished phosphor bronze which solders easily to the PCB. These are located in holes drilled with a 2mm carbide drill. Placement accuracy was assured by using a precision rotary table with 1/20 degree scale markings, and the PCB was centered to better than 20 microns using a dial indicator and a 4 jaw chuck.
The rods are polished silver steel for the moment, pending me finding some gold plated rod of suitable diameter. The originals are just a press fit in the hub and so easily changed.
Overall sensor diameter is some 28mm and the height is 33mm. Over travel is about 5mm to give some safety margin when fast probing.
Time has come to learn how to programme UCCNC macros - I'll use one of the published M31 ones as a starting point, no point in re-inventing the wheel.
All the parts:-
Partially assembled:-
Finished Sensor:-
Hope these are useful - it was an interesting break away from my Stepcraft.
Peter
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
Posted : 25/09/2016 6:36 pm
Awesome - interesting to see how accurate it is.
Posted : 27/09/2016 10:06 am
Topic starter
Hi,
Totally amazed!! It all worked first time - even the default M31 macro in V1.2xxx did the job without tweaking apart from adjusting the default sensor height. Setting this could be quite tricky - perhaps an electrical contact between the test tool and pcb laminate might be as good as any to get within a few microns.
I think the default macro needs some "safety" features though - what should happen if the sensor is already tripped or not plugged in when invoked? Once tripped under the "fast" approach, would it be better to retract a mm or so and then do a "slow" final approach before updating the Z axis zero? It all depends how quickly the UC100 responds to the trip signal - I guess it's pretty damn quick (microseconds??)!
There are examples of other more sophisticated M31's on the Forum - these are certainly worth investigating as I can probably do a cut and paste job to alter functionality without having to get my head around C# syntax.
One happy bunny!!!
Peter
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
Posted : 27/09/2016 1:27 pm
Topic starter
Hi,
Now done a few accuracy and repeatability tests on the sensor and M31 macro, and so far it has proven amazingly consistent.
The target surface is PCB FR4 laminate fixed down with double sided tape. Probing sets the Z offset to exactly the same value each time (as it should), then the tool is lowered until electrical contact is made with the PCB copper.
Over a dozen or so repeat operations the contact reading between tool and PCB was always better than +- 0.01mm - more often or not between 0 and -0.01mm. A complete cycle included homing the system and then returning to "zero".
I think I have confidence not to open another can of worms by double probing - the default Z speed seems to be well within the reaction time of the system, so if it ain't broke don't fix it!!
Peter
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
Posted : 27/09/2016 6:26 pm
0.01!
Nice....
yes always test manually by pressing the button before offering it up to the tool 🙂
Posted : 28/09/2016 10:24 am
Topic starter
Just to complete the picture, I've done some more tests to find the repeatability and hysteresis of my homemade probe.
I used a 10:1 lever arrangement to step down the Z axis movement and have found both the repeatability and hysteresis to be of the order of 2-3 microns.
I'm well pleased with this and will happily use the sensor for all my future tool set ups. Using it to set up for PCB milling, it consistently sets the tool about 10 microns high - one .01mm step down always brings the tool into contact with the copper. Cant argue with that!!
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
Posted : 29/09/2016 6:41 pm
