Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: ART on August 21, 2006, 07:30:04 PM
-
Hi Guys:
I promised soem photos of the digitized prop..
65000 points, 6 hours. .5mm grid..
Art
-
Is this digitizing done in a new function of Mach ?
Tell us about the process and the probe you used, etc.
-
Hi Benny:
I was a bit brief wasnt I? :)
This was my test of th enew 3dProbing plugin. Its included in the latest versions. To use it, just turn it on in config/plugins, then restart. A new menu item "3d probing" will appear.
When you select it, you must have zeroed at the front left of your surface, doesnt matter where Z is, but it shoudl be 0 or less. Set a width and length of the surface, and a depth
for maximum probe stroke. The clearance shoudl be 1/5 of full probe deflection horizontally or less. .1mm is used above, this is th epullout distance on a collision. The point cloud above was made in
6 hours and comprises 65000 points or so. The completed point cloud was 95000 in 8 hours.
This is reported now to be up to 10 times faster than the old methods of probing in MAch3, thats a huge increase. Very easy to run, very nice poitn cloud output. Im kinda pleased..
Art
-
Oh, heres the point cloud if anyone woudl like to play with it..
Art
-
Will this work with a g-rex also or is this a lpt thing only?
Chad
-
Hi Chad:
I havent tried it on a g100 yet, Brian reports some trouble.. Ill be working on that soon..
but it really needs the next firmware for the speed of update..
Art
-
Ok cool, Have you given any more thought on a offsets edge finding probe plug in? I am trying to improve the accuracy of the 0,0 process. It sure would be nice to be able to just enter a probe diameter, and then jog into the edges and set the g54 or others.
As usual thanks for the awesome software!
Chad
-
Probe design this was based on an idea I got from http://www.indoor.flyer.co.uk/probe.htm I have used this lots of times to good effect, although I have since purchased at auction a Renishaw probe.
The attached jpg’s depict my version this will scan to a resolution of 0.2mm, the model is in Solidworks if any one requires this let me know, if enough are interest I am willing to convert the components to IGES or what ever format is preferable and possibly produce some Dimensioned drawings, the body is made form Aluminium, the hub from Engineering Nylon, the pins are 3mm silver steel, the Balls are 5mm, the probe tip I made from a Carbide Burr, others parts from the scrap bin, hope this helps.
Shaun….
-
Shaun your pics make it hard to see the detail around the balls / switching area.
I know how they work. Looked at the probe site and have seen others.
Just curious how you have done the ball retention and pysical part of the circuit.
Great drawings by the way.
-
Here are the files in Solidworks format, IGES, I have also included a Solidworks Viewer file for people who dont have this software, and a brief notes file IGES files can be pulled into most Cad packages.
Hope this is of help.
Shaun
With electrical correction :-*
See Cad Drawings Below :)
-
Here are the files in Solidworks format, IGES, I have also included a Solidworks Viewer file for people who dont have this software, and a brief notes file IGES files can be pulled into most Cad packages.
Hope this is of help.
Shaun
Hello Shaun,
nice design and works :)
Many thanks for Solidworks files :)
Francis
-
Thanks for that Shaun
-
Here are the files in Solidworks format, IGES,
Shaun
Hello Shaun
I found a small electrical mismatch, the first picture's not electricaly correct the second picture's correct.
Hope that help
Francis
-
Yes I agree however in my defence I did re-draw the other day, about a year after making it and didn't check,
Thank you for the correction I will also correct the model
shaun......
-
Art here is the little pistol grip I was testing the digitizing plugin on. Please notice the two black lines with arrows. The line that runs the lenght of the grip has three arrows on it. The arrow at the bottom of that line is where we are having trouble. Everything has probed fine till we reach this point then the z axis does not lift after making contact, but backs up the clearence amount and stops. On all the rest of the moves where it would contact the surface from the side it would back up, raise up a little and try again. Once contact was made it proceeded along with no problems. The led shows the switch is broken right now but not once did it try to lift to clear.
Mike
-
Mike:
You need to straighten out that edge. Your accel is probably a bit low, so by the time the probe makes on that edge, 5 backups wont get off. Striaghten the edge so it doesnt hit so accutely.. IT will likely work then..
Art
-
Art sorry about being so long in getting back. I am about to become grandpa for the 6th time either tonight or in the morning. We have had lots to do this afternoon.
Now you know me well enough to know that what I about to say may not sound like my normal kind and gentle self.....but I'm still me.
It is going to be impossible for everyone who probes a part to lay that part so the program will work. If I make that angle more in line with the y axis, then the other side of the grip is going to replicate it so the problem will only be moved to the other side. I have lots of products that are much more complicated that this one and I can see lots of problems if I have to plan out the position of the part based on whether the plugin will work on it or not. Do you follow what I'm trying to say here?
My accel is about a high as I can get it with my current setup. This is a brand new setup for me using Mach 3 as I have used Mach2 till this plugin was available. Could debounce be part of the problem? I have not messed with debounce so it should be at default. I'm grasping at anything right now.
Mike
-
Mike:
:) , I didnt suggest straightening as a long term option. Probeing is something that will have to have a few "strategies". At the moment, we have only the Grid stategy. This will work for many things, but we'll need more to deal with other things. Your acute edge is a perfect example. A spiral strategy I think will fix that one, The point of the test so far is to find the weaknesses in it, and its doing pretty well, but I dont think there is any "one fits all" type of probing that will work. IT will take a few output types in terms of movements. The grid is fine, but I think Ill revise it to make it more easily rectangualr, then Ill add a spiral. In order to maintain safety, rather than try to override the pullup and pullout behaviours which seem to have proved themseves, I think simply different shape strategies will be the final solution. I shoudl have another one soon. :)
Thanks
Art
-
Thanks Art,
Now I understand much better. See I was under the impression that the square sequence probe with the current pullback and up system was it. Now I see that this is just the beginning. I'll help you in any way I can. Just tell me when.
Mike
-
Shaun,
Could you provide drawings of the probe build showing dimensions etc. It would be very helpfull for non Solid Works users and I am sure there is a few of us out there.
Thanks
Jim
-
Here are the Cad drawings for the Digitising probe in pdf format so everyone should be able to view print etc.
Shaun........ :)
-
Here is the dxf file for Prop.txt I translated with CamBam.
http://www.brusselsprout.org/CAMBAM/ (Free)
Paul, Central OR
-
Hello everyone -- I'm new to this forum. I just finished building my first probe and thought I would share some of my design ideas. I basically started with Shaun's plans using them as a rough guide for the overall size of things. I kept the circuit board and ball spacing the same size, but made a major revision to how the board and the balls are configured.
Here's what I did:
Made the housing out of two pieces. The top is shaped like the original housing except it is shorter. The bottom housing is shaped like the top, i.e., it has a hollowed out center. This new bottom housing combines the function of the "platern" and the "cap" into one piece.
The top has a hole for the arbor like the original. I made the arbor with a spring retaining extrusion similar to the one on the part known as the "hub" (refer to Shaun's drawings). The top and bottom housings were milled from HDPE.
In my design, I soldered the balls to the circuit board, holding them in place tighly against the PC board with a piece of metel tubing with an I.D. a bit less than the balls. This makes the assembly a bit easier than playing the dimple and ball puzzle game. This also eliminates any problems of contact between the balls and the circuit board. Soldering the balls would not be a bad idea in the original design too.
The PC board is then mounted upside down in the bottom housing. This eliminates the need to mill out the three cut outs for the pins on the circuit board because now the pins are on the top side of the balls and the PC board is on the bottom side of the balls. It also eliminates the need to mill the dimples for the balls as on the old "platern" (now part of the bottom housing).
The clearance between the protruding spring retainers (one on the hub and one on the arbor) was made small enough to keep the pins (and the hub) from rising far enough to allow the hub to turn and thereby displacing the pins from between the balls.
I used three countersunk through holes in the sidewalls (about 0.165" thk) of both the bottom housing and top housing to accomodate #2-56 screws and nuts to hold the housing halves together. The amount of the counter sink I used was enough to allow me to use some #2-56 x 1.25" screws I had on hand. They can be hard to find at that length, so anyone copying this idea may want to make the countersinks deeper. (I got the 1.25" screws at www.boltdepot.com, by the way.)
In summary, I made the assembly much simpler by soldering the balls to the PC board and flipping it upside down (with the balls on the top side and eliminating the pin cutouts), combining the platern and the cap into one piece, adding a spring retainer to the arbor and spacing the arbor and hub to keep the hub from over-traveling.
It is working quite nicely, but I need to cut some of the spring off because it is tighter than it needs to be. Just needs a little fudging.
Now I am looking for the easiest and cheapest way to get the point output of the digitizing wizard to a .dxf file.
-
That sounds good. Show us some pictures please. We like the sexy pics. ;D
Brett
-
So you want the sexy pictures, eh? :o Okay, but it may take me a day or two to get around to it. I should probably start a web page for my mill and lathe and put my pix there.
I am definitely going to build a second probe just to streamline the design. My first prototype is taller than it needs to be and that cuts down on the range. Another modification worthy of note is that I used brass rod instead of "silver steel" rod for the pins. I figure the forces involved are so small that brass will last a long time. The pins are simply pressed into holes in my HDPE hub and the fit is tight enough that they stay in place without glue which won't stick to HDPE anyway. Since my last post I clipped a bit off the spring and improved the sensitivity. I will fine tune the spring length for the second probe.
Since I will be building a second probe, I will document it in pictures throughout the fabrication. I will probably do some ACAD drawings of the new parts too -- maybe. In the meantime, I will take some pix of the first one and post them in a day or two as soon as I get the time.
Maybe I should mention my equipment:
Taig Mill (came with CNC ready motor mounts) with Xylotex 4-axis controller, Sherline rotary table, running the latest Mach3 on 1GHz Dell.
Grizzly Model G8688 7 x 12 lathe -- all stock equipment except for a quick change tool post from LittleMachineShop.com.
-
I started my second probe and redesigned my redesign and didn't get far before I decided to draw it up in ACAD before going any further. So this .dxf file shows everything except the stylus (will add one to the plans later). There are no dimensions on the drawing since this shows a complete unit, but it is a simple matter to check the dimensions on the drawing. I will extract separate drawings of the individual parts later and add dimensions to make a set like Shaun's. I put everything on a separate layer in this drawing making it easy to visualize in AutoCad by turning off the layer that holds the housing.
Some notes:
The spring is made from .025" dia. music wire formed by wrapping tightly around a 1/4"x20 threaded bolt. I chucked the wire and bolt in the lathe and turned the chuck by hand while holding the wire tight agaist the threads of the bolt. The number of coils, diameter and length in the drawing is accurate for a spring made this way on a 1/4"x20 bolt.
The balls are 3/16" diameter.
The pins have shoulders to make sure they extend the same amount out of the hub. The pin diameters are 1/8" and 3/32".
The screw that holds the arbor to the housing is #8-32x1/2" socket cap head and the cap is just the right diameter to hold the spring during final assembly. The washer is a #8 flat washer. the arbor shaft is 3/8" diameter.
There is a hole in the top of the housing for the wires to go through. A well-placed knot in the wires will act as a strain relief.
The three long screws are #2-56x1.25" and they serve to hold the PC board to the bottom cap as well as holding the cap to the main housing by virtue of two sets of nuts.
I still haven't taken any pictures yet.
Note (Added June 5, 2007): For my latest design, ignore the download here and go to my page at http://tetralite.com/mach/downloads.html
-
Here is an image made using my first probe prototype. The sampling grid was a 1 inch square with x and y steps of 0.005" (200 steps per inch and 40,000 points total) and the probe stylus is stainless steel with a tip diameter of about 0.027". The quarter was held in place with double-sticky-side fiberglass carpet tape. You can actually make out some of the weave pattern of the tape in the image. If I used a needle point stylus I think I could make a pretty good looking picture of a quarter! ;D Anyway, this demonatrates a darn good degree of resolution -- not bad for my first attempt at building a probe if I do say so myself!
(http://tetralite.com/mach/images/quarter02.jpg)
After the scan I took the point cloud file generated by the Mach3 digitizing wizard into Global Mapper (mapping is another hobby of mine) and applied one of the shading options (I used the "Gradient Shader") and then exported the JPEG image. Then I took it into PhotoShop and desaturated the image to remove the color generated in Global Mapper.
Here is a picture of my first protoype probe set up to scan the quarter:
(http://tetralite.com/mach/images/quarterscan.jpg)
-
Pretty nice job!!
Art
-
Looks good, you have done well in my book. You should be proud of yourself, I am. Are you going to make any to sell?
Brett
-
Thanks for the praise Chaoticone and Art!
Since I started my second probe my first one seems really crude to me now, but it works okay as you can see. If I were to use it in the manner it was intended (for scanning real parts probably with less detail than a quarter), I think it would suffice nicely.
I am considering making them to sell. I have no idea yet how long it would take me to make one (or several at a time) and how much I would charge for one. Any ideas/comments as to what a reasonable ballpark price might be would be most welcome. I wonder how big a market there is also. My lathe has no CNC on it but I'm considering a conversion. If I started making more things on it than I already do (not that much really) and started to sell them I think I would definitely want to automate the lathe a bit.
Btw, I have the meager beginnings of a web site up for my Mach3 related stuff at http://tetralite.com/mach -- not much there yet -- just started today.
-
Mark, As far as price I have no idea. The cheaper you can make them, the more you will sell but you need a profit margin. You can go fishing for free if you know what I mean. You could ask in the form of a poll here in the forum.
Brett
-
Brett,
Yes, there are many things to consider when setting a price. I will check what is available out there for starters. I am currently marketing a product that undersells all my competition (see my website for the TetraMouse.com). I like the idea of trying to keep the price as low as possible while still making a profit, but I have been told by several people that my product should cost more because people don't think it is good quality if it is too inexpensive. Lots to consider!
In the meantime I suppose I could consider taking offers and see what happens. Hmmmm...
-
I didn't read it but that mouse looks like a great product for a great cause. Good job. I have sent you a personal message.
Brett
-
Yes, I am very happy with it and it is extremely rewarding for me and my customers.
Replied to your message.
I have put another picture or two on my new still-rather-sparse Mach web page at http://tetralite.com/mach
Oh Gawd! another bunch of web pages to maintain!
So many toys and so little time, money, energy, etc.
-
Well, I'm on about the fourth revision of my probe design. The object is to make it simple enough to build easily while making it sensitive enough to create precise scans and measurements. On my prototype I learned that it is practically impossible to repeatedly build the probe and have the tip of the stylus be perfectly in line with the axis of the arbor and thus, the spindle. This misalignment means you can't use the probe for x,y centering the spindle or using it for edge finding with the intention of replacing the probe with a tool and having it centered on the spot you were shooting for. So I have come up with a way to make it possible to adjust the stylus tip to center. This is done by spring loading the circuit board so it is held against the top of the bottom cap and putting setscrews at three points 120 degrees apart and threaded through the bottom cap. Three springs are added; one on each of the 2-56 X 1.25" screws and tension on the spring is achieved by adding a nut on each screw to push the spring down against the PC board.
The set screws are located directly under each of the three pairs of balls that are soldered to the circuit board. By adjusting the setscrews, the plane of the circuit board can be adjusted so that the stylus tip becomes centered along the spindle axis. The setscrews are of course, accessible from the bottom of the bottom cap, so the stylus tip can be checked for runout and adjusted while mounted on the machine.
This picture shows the setscrew locations ("B" - 3 places) and the 2-56 x 1.25" screws ("A" - 3 places) that hold everything together:
(http://tetralite.com/mach/images/probebottom.jpg)
This picture shows one of three 2-56 nuts ("A") holding one of the three the springs in compression against the PC board. The "B" arrow points to the gap between the PC board and the bottom cap. This gap is adjusted by the 3 setscrews. In this picture it almost looks like the PC board is warpped, but that is just an illusion, probably due to looking through the clear polycarbonate tube (1.5" o.d., 0.125" thk wall) that I made the housing from.
(http://tetralite.com/mach/images/probepcbgap.jpg)
And here is a shot of the probe mounted in the spindle of the Taig mill showing the 1/8" phone jack with the plug plugged in. The jack fits nicely inside the housing luckily. Unplugging the wires makes it much easier to adjust the stylus tip to the center of the spindle axis because you can spin the spindle around. The adjustment of the plane of the PC board (and consequentially, the position of the sylus tip relative to the axis) is quite easy to do. It's a bit easier than chucking something up in a four-jaw chuck.
(http://tetralite.com/mach/images/probejack.jpg)
The stylus is female threaded and screws onto the screw that passes through the hub (the part that holds the pins that contact the balls). I did it that way rather than putting female threads in the hub because the HDPE doesn't thread well and the screw cap holds the spring in place at the top of the hub. This means that I can make various types of stylii with the female thread mount and easily swap them out.
I will post new plans when I get them drawn up.
-
Amazingly a digitizing probe patent has just been issued by the US patent office: 7,202,608.
It exactly describes the Renishaw probe. They claim their new and improved invention is to solder
a wire from ball to ball so there is no need to relay on the PCB to interconnect them. Talk about obvious knowledge.
The patent is poorly written too and easy to design around.
The easy link to patents: http://patft.uspto.gov/netahtml/PTO/srchnum.htm
Here is a good description by Renishaw:
http://www.renishaw.com/UserFiles/acrobat/UKEnglish/GEN-NEW-0126.pdf
Bertho
-
Hi Bertho,
Yes, I find that really amazing! If that's all it takes to get a patent, my version would be easy to patent with all the changes I made. I wonder why they didn't think of soldering the balls to the board. That is way easier than soldering a wire to the balls. Since I am publishing my plans right here and very soon on a web page, I will declare my design as public domain. The way I understand it is, if you publish complete plans for something and declare it as public domain, nobody can patent it. I don't hold much stock in patents anyway. All a patent does is give you the right to sue anyone that you think is violating your patent. From what I hear, you can expect to lay out $25,000 up front just to begin a patent lawsuit. So, in my opinion, it isn't worth patenting unless you think you are going to have a market big enough to expect to make millions off it. I don't think the market for digitizing probes is that big. So, just for the sake of public domain, I hereby declare my design (and as much of it as I have already published here and on my site) to be in the public domain.
I think my idea of mounting the balls on the top side of the PC board (not to mention soldering them to the board) makes more sense. And also my idea to make the plane of the PC board adjustable to compensate for the misalignment of various probes is a good idea -- it makes it possible to align your spindle with what you are probing rather than to wherever the tip of the probe is in relationship to the axis of the spindle. Without the adjustment, the probe is useless as an x,y positioning instrument and only good for scanning. It would take incredible amounts of precicion to manufacture a probe that is completely concentric to the spindle of your machine, especially if you have different stylii, so make it adjustable and save the headache!
I am considering offering probes of my design for sale at a really low price compared to what is out there now. The cheapest I've seen on the web so far are around $295. I think I can produce and sell mine for maybe $150 or less.
I just looked at the Renishaw paper you linked to. They obviously have a lot of time into making a well-built probe. It's no wonder they charge so much for them. I wonder if their probe tips are always exactly on the axis of the spindle they are mounted in?
Here's a 0.25 inch square section of the back of a dime that I scanned with my latest prototype. The scan was done using 0.001" steps in the x and y axes for a total of 62,500 elevation samples. The point cloud generated by Mach3 (using the digitzing wizard that is included with those other wizards, not the add-on one that you download from Artsoft -- I can't get that one to work at all!), was taken into Global Mapper and displayed in a 3-D projection with "Atlas Shading" which applies colors according to elevation like maps in an atlas. I also added some vertical exaggeration to emphasize the elevation differences (exaggeration here is 5.5 times). You can see the red color indicates higher elevation as you move toward the right front corner. This is because the thickness of the dime is greater toward the edges. You may also notice that the scan is cleaner looking in about the first half of the scan (the half closet to you in the image). Up about 3/4 of the way you can see where there is a stripe where the elevations were sampled wrong and there is a valley cutting through the oak leaves. I think I may know what caused this -- not sure yet though -- not going to try to guess the real reason at this point. I think this image demonstrates a good degree of accuracy -- close enough for any practical use of the probe in my opinion. We're talking errors less than the thickness of oak leaves on the back of a dime!
(http://tetralite.com/mach/images/dime.25sq.jpg)
-
Way to go Mike. I'll take one at $150.00. Thanks for the info you have declared as public domain as well. Good job. It is looking great and very funcitional. Keep up the great work.
Brett
-
Brett,
Well, I haven't committed to a price of $150 yet. That is just a guess. I *do* need to research my costs (time it takes me to make one mainly -- the materials don't cost much).
Forgot to mention that the tip I used for scanning the 1/4" square of the dime was made from an ordinary straight pin. In practical applications you would probably want a larger tip!
Which brings to mind a question that occurred to me last night -- How do they mount those ruby balls on the tips of probes?
Jeez! I sure like my adjustable PC board idea. Takes all the worry about precision out of the manufacturing as far as tip alignment is concerned.
(Correction: the name is Mark ;D)
-
LOL, Sorry Mark, I'm so bad with names that I can't even type them right. ;D No,no, you figure your price to make money, you can go fishing for free. I was simply wanting to be the first n line if you can sell them for $150.00. If it goes up, I will most likley buy one any ways. As long as I can get my accountant ( the wife ) to sign off. Anything less than my sock money, I can spend without approval. ;D
Brett
-
Ha ha! I've been called worse names than Mike! (with apologies to everyone named Mike)
Please explain how one goes fishing for free anymore! Again: Ha ha!
I keep forgetting to make money. But I am hoping I can offer these things at a low price. I think it is possible using the materials I'm using. Lots of plastic. Plastic is really fast and easy to machine. This will be a poor man's probe in that regard, but I think it is plenty good enough, at least for the hobbyist. I seem to get really good repeatability on measurements. I will have to do a lot more tests so I can come up with some good figures on how accurate it is.
-
LOL, Listen, break a limb out of a tree, tie a string and hook to it, dig up some worms or catch some crickets, or grasshoppers, put them on a hook, lower the hook in the closest body of water. If you can't find a tree limb, you can tie the string to your toe. ;D
I know what you are saying. I think it the screws for adjustment are great too. I hope you can offer them at a low price as well, I'm just saying, make it worth the effort.
Brett
-
, you can tie the string to your toe. ;D
Brett
You mean you actually use 13 lines at one time or do you just use one of your toes ;)
Tetra
nice idea on the probe, i had a different way of doing this but so far have not had time to finish it, but yours looks easier. It would however need a redesign of my probe from the ground up to do it your way :(
Hood
-
Brett, Thanks for the fishing lesson. Yes, seriously, I will try to make it worth my while.
Hood, Yeah, I kinda like it. I need to finish my plan drawings, which means I have to finish making my latest revisions on the prototype, which means I will probably end up changing the drawings again. But I think I'm really close to finalizing it. Trying to make it as simple as I can so it can be made easily.
-
Hi Mark,
I have been following this thread for a while. I appreciate the work you are doing and the time you are taking to get this thing working correctly and accurtately. One thing that has intrigued me about all this is the software you are using to manipulate your point cloud. I did a search and found it but would like to ask a couple of questions. I feel this is pertinent to this thread since it does us no good to get a point cloud and have to way to use it. I have imported into Rhino and have one product we modeled from those points but the time involved was too long.
Can, or would you be willing to do a quick overview of the map software that you use as well as the steps taken and the type of file you end up with? For instance, the portion of the coin that you captured. Is that a machinable surface that you created?
Mike
-
, you can tie the string to your toe. ;D
Brett
You mean you actually use 13 lines at one time or do you just use one of your toes ;)
Tetra
nice idea on the probe, i had a different way of doing this but so far have not had time to finish it, but yours looks easier. It would however need a redesign of my probe from the ground up to do it your way :(
Hood
Hood is your's the one that has all the electronics on the outside of the body of the probe? The one that looks like some sort of space gun or something? If that's your's, then I think the melding of the two designs would be a huge step forward. The one with the electronics stops the false triggers with the probe stuck in an open situation and crashes into something. I just read last night that another person on the zone had crashed their probe.......though I think it was due to another reason.
Mike (no offence taken on the name March) ;D
-
Mark,
Your design is very clever and it is nice to be able to calibrate it to zero. It might be ideal for mill.
It is amazing the resolution you got in the scan. I am also interested to learn more about you “mapper”
software but right now I got to learn how to use Mach-3 first. Life is full of compromises and it is especially true in engineering.
I suspect by adding the adjustments there is a tradeoff in stability and long-term accuracy.
There are simply more items that affect the absolute position of the probe tip.
A careful reading of the patent that I referenced, indicates that they know what they are doing.
I am sure the “insulating layer” refers to anodized aluminum to mount the balls repeatable and
they are soldered the as we discussed. Note, the balls are selectively gold plated but not the whole ball since the gold is too soft.
In other words, do we want adjustability and convenience or super accuracy and stability. A lot of these
probes are used on CMM machines were the absolute accuracy is critical. On a regular CNC mill the home
made probes might be more accurate than the mill can resolve. The high-end probes are looking
for 1 micron errors or less so the design rigidity is critical.
One possible compromise is to lock the balls in place and let the whole housing slightly move in
X-Y plane on the mounting spindle. It would preserve the rigidity and give a way to adjust it.
We only need to be able to move it a few mils or so.
Many years ago I did a single axis probe using 3 balls with a forth one resting on top as a pyramid,
shorting the three together. I electronically detected when the top ball was pushed off by a rod
from underneath the pyramid. It worked very well!
I enjoy challenges to push for high accuracy or other similar functions.
Bertho
-
The old Brown and Sharpe CMM I have has a 19.1mm (?) dia. refernece sphere that you use to calibrate the probe before measuring. Perhaps something similar could be done with Mach?
-
Hood is your's the one that has all the electronics on the outside of the body of the probe? The one that looks like some sort of space gun or something? If that's your's, then I think the melding of the two designs would be a huge step forward. The one with the electronics stops the false triggers with the probe stuck in an open situation and crashes into something. I just read last night that another person on the zone had crashed their probe.......though I think it was due to another reason.
Mike (no offence taken on the name March) ;D
No, dont think so, mine just has an internal circuit with a battery and LED for edge finding use and a jack socket for connecting to Mach if it gets used for probing. Heres a few pics.
-
Looks good Hood.
Brett
-
Looks good Hood.
Brett
It works great but afraid the first attempt at making up an adjuster was a mess, i eventually got it centered to within 1/2 thou but it was a slow process. I am going to do it a different way sometime but when I dont know.
Hood
-
Many responses here since my last one!
My probe definitely is not going to try to compete with the high-end ones, but I think it has enough accuracy for the everyday home mill. One test I am doing is running the "Center of Hole" routine that is built in to the latest Mach3 release. I just run the routine, reset x and y using the "Ref All Home" button and then run the routine again to check repeatability. After about the third time, I get errors no larger than .0002, and more often it is .0001 or even .0000 on the Mach DRO's. The first one or two repeats of the test usually shows a greater error than subsequent tests, so that indicates that the first test or two tend to have a greater error, becasuse the error is reduced every time by running the test three times. After three times I never see a difference of more than .0002. Yes, that's 10-thousandths. But not microns!
For those asking about Global Mapper, you can download a trial version that has crippled saves and exports, but it is good for evaluating the program. It is all mapping related, but it can manipulate elevation models in many ways and in reality we are talking about a kind of mapping anyway. When I open a point cloud file it asks which type of geographic projection you want to use. These are standard projection types that are used on maps and they usually take into account things like the curvature of the Earth (which is not enough to worry about in small areas) and how to display a curved surface on a flat one. You can even specify projections for some of the other planets! Depending on the projection system you pick, the point cloud file is opened using (usually) meters as the basic unit of measurement. If your original point cloud is in inches or millimeters it really doesn't matter because we are dealing with units and not inches or meters or whatever.
In the registered version, you can export to a huge variety of file formats for many programs including dxf points or converted to faces.
The trial version can be downloaded at http://www.globalmapper.com/ (click on "Download Trial" from that page). It can also be downloaded from the USGS at http://mcmcweb.er.usgs.gov/drc/dlgv32pro/. This version (they call it "DLVG32Pro" may be less crippled than the "trial" version -- I am not sure about that because I have the full version already and haven't bothered with the crippled versions. The large njumber of file formats that can be read by Global Mapper are listed at http://www.globalmapper.com/product/formats.htm. The point cloud files generated by Mach's wizard is a simple text file that is recognized by the Global Mapper program as an ASCII elevation grid file.
I haven't tried to use Global Mapper for generating a file to convert to G-code yet and haven't given it too much thought. I have so far used it for viewing point clouds with various shading techniques. On road trips I have a laptop computer with a GPS unit in the car with all kinds of mapping software including Global Mapper. I have digital elevation model files (DEMS) for the entire state of Washington (where I usually confine my road trip to). I also have geo-referenced topographic map files and other map format files and aerial photography for the entire state. These can be overlayed in Global Mapper to produce 3D images of the terrain (what it looked like a few years ago when the photos were taken. I can generate a picture of what it is going to look like around the next bend before I go there. Kinda fun. Anyway, Global Mapper is quite a program and could be useful for converting some formats into something useful for machining purposes, especially when dealing with images that you might want to engrave. I really haven't explored this possibility at all.
-
Some new pix of my probe (shaded in ACAD with the housing not shown so you can see inside):
(http://tetralite.com/mach/images/DigitProbe01.jpg)
In the above image you can see the springs that push the PC board down against the bottom cap. The setscrews used for adjusting the plane of the PC board thread through the bottom cap and push up against the bottom side of the board. Also note that I have added the jack for plugging in a 1/8" phone plug.
(http://tetralite.com/mach/images/DigitProbe02.jpg)
In this picture you can see the holes through the bottom cap for the setscrews and the long (2-56x1.25") screws that hold the probe together.
The setscrews are 6-32x1/4". The probe is shown here without a stylus which screws onto the 4mm x 0.7mm x 30mm bolt that runs through the hub (the part that holds the contact pins). I used the metric screw here because it was a better length than the cloest standard length #8-32 screw, so my stylus must have 4mm x 0.7mm female threads.
(http://tetralite.com/mach/images/FDRsEar.jpg)
And here is a scan of FDR's ear on the front of a dime. Notice the big peak on his ear lobe. This must have been due to a false activation of the probe, perhaps caused by a vibration on my mill's bench. The image was generated in Global Mapper and it extrapolates the elevations between sample points. That is why the sides of the peak on the ear lobe are tapered to meet the surrounding sample points. I haven't seen it do this before and I think it is because my latest assembly of the probe didn't put as much compression on the spring that pushes the pins against the balls making it too easily opened by vibration. The vertical exaggeration in this image is 10 times nomrmal. The scan was 0.12" by 0.12" with x and y steps of 0.002" making a total of 3600 sample points.
Below for download is my latest ACAD drawing in .dwg format. It is a three-dimensional drawing -- the same that I used to create the images above. If I save it as a .dxf file it is over 23MB. As a dwg it is 868KB. If you can open this 3D file in your CAD program you can measure everything or look at object properties to get dimensions which I have not put on the drawing. There are separate layers for everything.
-
The resolution is amazing! And the display is very nice after it has been run through the mapper.
How is the probe doing for measuring in X-Y planes and repeatability?
I saw a wizard for finding hole centers but is there an option to use it as a regular CMM function
where the ball diameter is subtracted from all measurements, X-Y-Z.
Even further, how do I go from either a point cloud or controlled data points to a 3-D CAD model?
TIA
Bertho
-
I can answer all three of your questions with confidence: (1) I don't know, (2) I don't know and (3) I don't know!
I haven't really done much in the way of testing yet. So far I have just tried to see how well it can scan very small items using a needle point for a stylus. Well, I have done the hole center thing I described above and those tests show repeatability easily within a thousandth of an inch. I get repeatability errors of about 3 ten thousandths maximum (I haven't done a lot of tests). All this is, of course, dependent to some degree on the resolution and accuracy of the milling machine itself, determined by leadscrew slop and stepper motor resolution and whatever else there is to take into account. But it looks to me that if you're not shooting for accuracy better than a couple thousandths or so, this probe should be close enough.
Modified entry: I have noticed that in my scans of quarters and dimes that every other line is shifted a tiny amount along the x axis. This must be, I presume, due to end play of the x-axis leadscrew (what else could it be?). I am glad my machine is that tight! Perhaps it should be noted that my stepper motor settings for "Steps per" (which is steps per inch) are at 32000 which means in order to move a stepper motor one step on my machine I would enter G1 X0.00003125 and the motor moves one step, assuming X is starting from zero. Whether on not the table moves depends on end play and which way it moved the last time it was moved, I suppose. So I am set up for pretty fine resolution.
I haven't seen any options for CMM function either.
I have been wondering about the best way to get from point cloud to 3D CAD models myself. I will be experimenting with that soon I think. Seems like I saw something about that in this forum somewhere.
-
Amazing. ;D ;D ;D ;D Looks real good to me.
Brett
-
Modified entry: I have noticed that in my scans of quarters and dimes that every other line is shifted a tiny amount along the x axis. This must be, I presume, due to end play of the x-axis leadscrew (what else could it be?). I am glad my machine is that tight!
Actually this is a function of the probe plugin for Mach3. I have probed several parts and I have the same thing.
Mike
-
If that is true that would mean that the probe plugin is generating errors in the output file that would amount to elevations being reported at the wrong (shifted by 1 x step) x,y point. Is that what you are saying?
-
I have begun assembly of several probes and when I have a few built I will begin offering them for sale on a web page at http://tetralite.com/probe. I have completed the time study and I have decided the price will be $150 plus a flat rate for shipping anywhere in the US plus a PayPal fee (orders will be only through PayPal). I will not be selling to foreign countries. I just don't want the extra hassle of filling out customs forms and calculating the shipping charges. I have enough of that to do with my other product (TetraMouse).
During this short initial run I will be refining my tooling (jigs and etc.). I am considering offering the probe without a stylus, which would mean it would be a do-it-yourself stylus situation. On the other hand, if I come up with a simple all-metal one piece stylus I will offer that. The problem is I can find no info regarding how to attach ruby balls to a stylus for the tip. Anybody know about that?
As soon as I have probes for sale I will put a PayPal link on the above mentioned web page. If I get more requests for probes I will begin another run right away. Lead times will be short -- a couple or three days probably. I will try to keep an up to date total of probes on hand on the web page. If I have to put orders on hold and anyone doesn't want to wait, I will refund their PayPal payment at their request immediately.
The first run is 5 units and two of them are already spoken for. I expect to put them up for sale within two days (June 7th probably -- maybe tomorrow the 6th).
-
Why not have them accept a standard styli? I've bought some from here before http://www.itpstyli.com for my old CMM.
-
That sounds like a good idea. I am taking a look at itpstyli.com. The first thing I notice is their styli have male threads. One of the things that makes my assembly economical is having male threads on the shaft coming out of the bottom of my probe. It is actually a 4mm x 0.7mm cap screw that runs through the hub (the part that holds the contacting pins). I am making the hub out of Delrin now (HDPE is just too gooshy) and it is easiest to make if I only have to drill out the center to fit the screw rather than tapping it. Then a nut is put on the end to hold the screw in the hub. If I tapped it, then I wouldn't have the cap head of the screw to hold the main spring from moving lateraly and I don't want to rely on tapped plastic because it is too easy to wreck the threads and they loosen faster with use. So I don't want to make the hub with female threads. So all I have to do is supply a threaded female coupler that will thread onto my probe's male thread and accept a 4mm male threaded stylus, which can be found at itpstyli.com.
Jeff Birt, thanks for your input. I have decided to include the threaded female coupler so people can buy a standard stylus. Also, if they wanted to use a stylus with a different size thread, they can make an adapter that has 4mm threads on one end (to fit the probe) and whatever size thread on the other end for their stylus. Or they can still just make their own stylus with femaile threads.
I still want to know how they attach ruby balls to styli -- can't find that yet!
-
Great lead Jeff!
It would be a good idea to standardize the mounting to fit a common probe tip to give people a choice for the future.
DIY ought to work if the rod tip is cut out to roughly fit the surface of the ball and then a microscopic
drop of high strength epoxy is used to permanently attach it.
Bertho
-
Is that how they attach those balls commercially?
Waitg a sec... what am I thinking!... I can just use a female threaded coupler in place of the nut that holds the screw in the hub. Then a 4mm stylus can just mount directly. (I could save a part if the hub didn't have to be an insulator.)
-
What format does mach 3 store the file as?
Artcam pro wants an ASCII G-Code toolpath format.
Is it compatible?
Thanks
-
Hi Mark,
If you want to make it as convenient as possible for people to buy styluses, you might want to
consider using 5mm thread instead of 4mm. I checked their web site:
www.itpstyli.com and there are 87 choices of 4mm tips but an incredible 379 choices of 5mm tips.
Hopefully it will help to make your probe sell more.
Bertho
-
WYO_123,
Mach3 stores the output of a digitizing file in ASCII text format. Each line of text represents a sample point. Here is an example of a tiny part of a file from a scan using the digitizing wizard that is clumped in with the other wizards:
0.00000,0.00000,-0.01859
0.00100,0.00000,-0.01859
0.00200,0.00000,-0.01844
0.00300,0.00000,-0.01847
0.00400,0.00000,-0.01844
These five lines define five points with x, y and z values delimited with commas, so the first line says x=0, y=0 and z=-0.01859.
As the program samples points, it increments the x location by the x step value you specified, taking a z sample every step, until it reaches the maximum x dimension that you specify in the set up. Then it increments y by one step unit and then, starting at the maximum value of x, it decrements the x value back down to x=0 taking a z sample at every step. Then it increments y and so on.
You need to turn this "point cloud" output into G-code and I haven't even tried yet and I don't know what program would be the best for doing that. If you have a program that accepts comma delimited x,y,z point cloud files and can convert them to dxf files, I think that might be a good way to go. I'm sure there are people on this forum that have some experience with turning point clouds into G-code. I haven't got that far yet myself. I do know that the Global Mapper program will accept point cloud files and turn them into dxf or JPEG image files and these can be converted to G-code with the right program, perhaps even with LCAM.
This is new territory for me also! Sorry I don't know more about it.
-
Bertho,
With my latest plan, it will be just as easy to make the probe accept 5mm styli as 4mm ones -- just a matter of hole and tap size on the threaded coupler. I will make that an option. ;D
-
You need to turn this "point cloud" output into G-code and I haven't even tried yet and I don't know what program would be the best for doing that.
I wrote an AutoCAD macro to create 3D faces from the point cloud. http://www.artsoftcontrols.com/forum/index.php?topic=2421.0
You can use a CAM program like MeshCAM to create the g-code from the .dxf of 3D faces.
-
I will definitely check that out, Gerry. Thanks! I found a point cloud macro for ACAD online somewhere, but they wanted money for their version that wasn't crippled. I have MeshCAM so I can try out your method.
WYO_123, are you getting this?
Today I am testing repeatability of my digitizing probe design down to resolutions in the 1 hundred-thousandths range (0.00001") and getting some interesting and amazing results. I am scanning an area of the table of my Taig mill as a test surface. When I get it all sorted out and interpretted correctly I will post the results on http://tetralite.com/probe
Update: I have posted some test results on that page. The probe seems to have measurement repeatability better than 0.0002" and actually somewhere around 0.0001". I also discovered the the table of my mill slopes down from front to back at the rate of roughly 0.0001" per inch (a real rough estimate).
Probes will be available for sale soon at www.tetralite.com/probe.
-
Gerry, you wrote your macro as an Autocad plug in, can your converter work outside the box such as a small exe file whereby you can run the digitized point cloud file through it and have it converted to 3d faces (3ds) type format? this would be a quick loading into most modeling software for clean up and toolpathing.
-
Not as it's written. It's written using VBA, and I'm not to comfortable with VB.net to do it. If someone wants to do it, I can share the code I used to create the faces. You just need to read the points and write a .dxf
-
Thanks Gerry
-
Tetralite
Yeah, I am getting this, I got the file of the prop to imported fairly well, but it looks like that file had some problems.
Can you send me one of your files so I can play with it, the dime one would be cool.
Have you decieded on a price?
Thanks
-
I'm thinking the price will be $150. I am doing extensive testing for the past couple of days and I am still not satisfied with the assembly enough to sell it. Have to work out some details. I am attaching the point cloud file for FDR's ear on the front of the dime. That is the one that I posted an image of earlier in this thread -- there is a giant pimple on FDR's ear lobe where an error occurred (a false surface detection). The area of this scan is (if I remember right) 0.12" by 0.12" and I had to scale it all up in order to work with it. I don't know if it will work with whatever you try to do with it.
-
I am attaching the point cloud file for FDR's ear on the front of the dime. That is the one that I posted an image of earlier in this thread -- there is a giant pimple on FDR's ear lobe where an error occurred (a false surface detection). The area of this scan is (if I remember right) 0.12" by 0.12" and I had to scale it all up in order to work with it. I don't know if it will work with whatever you try to do with it.
Is this done with the digitise wizard? My AutoCAD macro doesn't like it. All the sample files I had when I wrote the macro had a line that yours doesn't, at each direction change.
Here's a few lines of your point cloud.
0.10000,0.00000,-0.01087
0.10200,0.00000,-0.01087
0.10400,0.00000,-0.01087
0.10600,0.00000,-0.01087
0.10800,0.00000,-0.01087
0.11000,0.00000,-0.01087
0.11200,0.00000,-0.01087
0.11400,0.00000,-0.01087
0.11600,0.00000,-0.01084
0.11800,0.00000,-0.01081
My macro is looking for this line
0.12000,0.00000,x.*********xx
you go to this line
0.12000,0.00200,-0.01081
0.11800,0.00200,-0.01081
0.11600,0.00200,-0.01081
It appears that you're skipping a point each time you change direction.
-
I used the digitizing wizard that is clumped in with the other wizards under the "Wizards" menu in Mach3, version R2.0.065 and not the digitizing plug-in.
Yes, I too have noticed that it skips a point on each row, but the software that I am bringing it into (Global Mapper) does not expect any particular values in any particular order, but rather just interprets the points as absolute points with their respective elevations. It doesn't matter what order they are in because a location is a location and not dependent on which location you last looked at. I can mix up the lines in the point cloud file in any way and Global Mapper will still map each point independently from the others and end up with the same grid of x,y points at their respective elevations and use it to generate the output, which is a 3D image consisting of points and their elevations and interpolated elevations between the points. Obviously, this is not quite the same as interpretting points as adjacent locations like you would want to do in order to ultimately generate efficient G-code. To get to G-code would require another step for me that would go from point to adjacent point in the output from Global Mapper with no missing points in the grid. Global Mapper fills in the missing points as well as filling in the elevations between points, so I could generate a new point cloud by exporting one from Global Mapper that would tend to have smoothed out the original input and also filled in missing points. Of course, this doesn'tg really help the problem of the wizard skipping points to begin with, which it does in my copy of the program. You can see where the points are missing in the Mach3 toolpath display on my copy. There is a diagonal line from the end of one line to the beginning of the next line, instead of a perpendicular line like you would see if it didn't skip a point. So the wizard is in error and is decrementing or incrementing the x value (depending on which direction change you are at) at the same time it increments the y value to start the next line.
This file is probably not a good one to work with because the elevations are very small. The lowest point is (-0.012") and the highest point is (-0.006"). The image I created had a vertical exaggeration added to it in order to visualize and emphasize the elevation differences, while in reality, the max difference in elevation from the lowest to highest points is only 0.006". (I may have those numbers wrong, but it was something like that.)
-
Mark,
What input format are you using for global mapper. Seems they have a lot of them.
Roger
-
Hi Guys, you can also have mach add the axis letters to the point file. Then if you add a g1 to the top of the file you have a simple Gcode file. Then use a gcode to dxf converter to convert it to a dxf.
Or if the points were gathered in a specific order such as the wizard does then if you add the g1 and make it a gcode file then Mach will cut it exactly like you probed it. If the surface was deep you would want to multistep it down one depth at a time to prevent gouging. But with a surface as shallow as the dime it should cut ok.
But you will have to config to add the axis letters to the file.
Just a thought, (;-) TP
(;-) TP
-
Mark,
What input format are you using for global mapper. Seems they have a lot of them.
Roger
Hi Roger,
Under the File menu the click on "Open Generic ASCII Text File(s)..." and it prompts you for the location and filename of your point cloud text file. When you select your file it then comes up with a window where it asks for some settings. The first setting is "Input Type" and you want to choose "Elevation Grid from 3D point data". Set "Coordinate Column Order" to the "X..." option and set "Coordinate Line Prefix" option to "None..." (I think those are the defaults). You can click on the "Select Coordinate Offset/Scale" button to scale up the image. I usually scale things up with those settings. Leave the "Coordinate Delimiter" setting on "Auto" or select "Comma" -- it works either way. Leave the slider on the "Loose" setting since you are dealing with absolute points. Or experiment. Setting it tighter takes more time.
When it asks for a projection (in the next window) select "Equirectangular" from the list. This is because the sampling points are a square grid. The other projection types take into account various distortions in the various map type projections and have no use in this type of purely regular 3D plotting where we don't have to account for things like the curvature of the Earth.
There are many ways to project and scale and alter the display of the resulting image in Global Mapper and I can't even begin to go into what I know about it here. It just takes a lot of trial and learning and exploring.
-
After a year's delay from our last discussion here, I finally decided it was time to build a touch probe. For a long description see here: http://www.vinland.com/Touch-Probe.html
That also brings up a couple of related questions:
What is typical spring pressure and contact resistance? Mine is 280grams and I am surprised that the the total contact resistance varies between a few ohm up to several hundred ohms. The higher the spring pressure, the lower the resistance. I have carefully cleaned everything.
What is the typical contact point on the balls? The further down, the higher the contact pressure but it might also make the probe sticky. Any research papers or information on that?
TIA
Bertho
-
Very nice, Bertho! I like the simplicity. Your questions are ones I have been asking myself for some time. The contact resistance varies dramatically with different materials and the amount of time in use (number of actuations) according to my tests. I used a circuit board in my original design and soldered the balls to the board, using small drilled holes to align the balls. The biggest problem I had with ball and rod contacts was that after about 100,000 samples (that's 100,000 switch operations) the contact resistance would increase so much that the op amp circuit would start to fail to switch its output state. At first, when I started testing my probes they were great, but after hundreds of thousands of samples they started to fail to close enough to work reliably. I couldn't accept those statistics, otherwise I would be selling them for cheap! Logic tells me that the problem is that the contacts don't wipe (self-clean) and a buildup of carbon occurs. This even happens at the extremely low current I was using (less than 10 microamps). Cleaning the contacts with some fine steel wool fixes the problem for up to another 100,000 point samples, but I still find this unacceptable.
The spring pressure is also something to be reckoned with. If it is too weak, as you know, the resistance increases. However, making it too strong means you are applying more force to the probe tip and the object being scanned. Objects made of softer material may not stand up well to these higher forces. Also, for really fine scanning with a needle-sharp tip, you will definitely leave a mark on all but the hardest of metals each time a sample is taken. Of course hardly anyone will be using such fine point probes for everyday use. I use them when I am testing my probes in order to get the finest resolution I can in my test results. If you want to scan an object that is a model and made from wood or clay, then the spring force may become a factor due to deformation of the object.
I am currently working on new designs. One uses a piezo-electric sensor and the other uses miniature off-the-shelf normally closed switches. The piezo element tends to deform over time and I have scrapped that idea for now, although it did perform pretty well for a while. The switch system is working even better than I suspected it would. Instead of using three rods I have gone to five. It still uses the rods to re-center the probe after each point sample. I use five switches that are operated by a ring that lays on top of the five radial rods. My thinking in using off-the-shelf switches is that, being made with excellent switch contact materials, they will survive longer before their resistance goes too high. I would like to get 3 million operations between servicing the probe. For my off-the-shelf switch model, servicing would mean replacing the switches or perhaps the whole circuit board with the switches pre-mounted just to make it easier for the user. Or perhaps just call the probes disposable and toss them after 3 million point samples and buy a new one.
My goal is to build and sell them for under $100, with accuracy and repeatability of under 0.0002" and have them last for 3 million operations.
-
Thanks Mark,
What are the commercial probes doing that is different? "Normal" usage might not include vertical surface probing to generate a cloud. If they are only used for measuring features or references the number of actuation will be very low.
Bertho
-
Yes (regarding measuring and referencing), that's a good point! I don't really know how most people would use them, but as a hobbyist I think I would want to do 3D point clouds to copy shapes. Also, my tests have huge numbers of point samples and maybe the the ordinary, everyday scan that one might usually do would be a much larger grid with much fewer sample points. Maybe I've been making good probes all along and didn't know it!
From the information I've seen on commercial probes of the type that use balls and rods I don't think they are doing it any different basically. Maybe they are using some special materials? I just don't know.
Here is a link to an interesting PowerPoint presentation from Renishaw, a maker of probes. It does have a fair amount of information and its worth a look (if you haven't found it already):
http://www.renishaw.com/media/ppt/en/744f3d51d1b944caa6c8bba90515b9a8.ppt
-
I remember seeing the Renishaw presentation a year ago or so and I had forgotten about it. It is very well made!
Three items caught my attention:
They monitor the increase of resistance and at a defined point they turn off the current before the contacts actually open.
They are talking about 10g force. Very low!
They guarantee 10 million life cycles.
I think the normal industrial use for the probes is to verify that machined parts meet the specification. In other words, selectively measure a hole size and its location in reference to other holes and features. No scanning at all.
Bertho
-
You mentioned: "They monitor the increase of resistance and at a defined point they turn off the current before the contacts actually open."
AHA!!! That could be really easily done with a very small microcontroller chip with an A to D input (I have some and I have the capability of programming them). By stopping the current before the contacts fully open it would stop them from arcing, which causes carbon to build up. And indeed, the amount of force needed to move the contacts only enough to detect increased resistance would be very small, but of course, the machine does not stop instantly, so more force would continue to be applied until the Z axis drive stopped moving. But cutting off the current before the contacts fully open is very interesting. This is a very neat idea and I just might pursue it. The PIC12F615 or PIC12F675 microcontrollers are 8-pin DIP packages with a 10-bit A/D converter and could be built right into the probe. This would require three wires to hook it up: one for +5 volts, one for ground and one for the output signal which would not need amplification and could drive a parallel port directly. A tiny potentiometer could be added for fine tuning the trigger level but may not be needed. I might give it a try soon.
-
Hey Mark, was hopeing you would see that reply from Bertho. Hopefully this will be what you have been looking for. I have my fingers crossed for ya. :)
Brett
-
Bertho, Nice looking probe you have there too.
Brett
-
Now that I have point clouds from scanning I need to first make them easy visible to see any obvious defects and then be able to import them into a CAD program to scale, tweak and adjust the scanned data, and finally machine the updated data. With Mach-3, LazyCAM, MeshCAM, and Leading Edge I think I am OK on the output side but what suggestions do you have for the cloud conversion? I have seen:
Global Mapper: http://www.globalmapper.com/index_data.html which looks very good for the visualization and it is also tempting since I have been doing a lot of design work with GPS and:
VRMesh: http://www.vrmesh.com/default.asp which does the 3-D import and repair.
Both look like excellent programs but expensive to just play with. Any suggested low-cost alternatives?
TIA
Bertho
-
Here is an update on my scanning adventure:
After hunting high and low for software to check out my probe with, I finally found several free and extremely low cost solutions and I have updated the info with samples on the website. See:
http://www.vinland.com/Touch-Probe.html
Bertho
-
Hi Guys:
I promised soem photos of the digitized prop..
65000 points, 6 hours. .5mm grid..
Art
I know this is an old thread, can you tell me how you avioded the dreaded two point per probe bug? my points clouds are useless, my probe hence is not usable, need to probe some parts badely.
-
I am sorry but I cannot remember having any "2 points per probe" problem when I tested the probe and scanned the quarter.
Bertho
-
Sorry for the dumb question but, I have been looking over these pages and just do not understand the actual function of the ball bearings and the rod,
can someone please explain, thanks in advance.
-
A rod touching two balls forms an electrical contact from one ball to the other. The same is true for the other two rods.
If the probe tip is pushed, one or more rods will lift and the electrical contact will be broken.
-
Thanks !