Actually Building the Machine!

[Nena]

We shall see, today i manualy made a 45 deg cut in the bottom plate from the head, this way the two bolds arent in the way when turning the head. I didnt find any bolds to use so need to order them. This is how it usely goes... You need something you dont have, need to order it and you need to wait a week or so... This week im going to test the ramp speed of the motor and work it out to get it running at any desire degree, hopefully i can intergrate the electronics into my main testing program and get all 4 axis to work. Still need to do much before i get that working but i can try things out on the cutting head.

Also i had a nice idear for automatic dopping. Im going to use a Sherline ATC (automatic tool changer) for the cutting head. And to change the stone im going to use another one and rotate the complete thing if the second dop is mounted on the stick and remove the dop with an induction heater that i can controll by software. It also need some ajustments on it but i give it a try if it works.

Also need to design how to mount the atc onto mij cutting head mount also to set the acuracy up to a few thousands of a mm. Probably with 12 set screws in two layers of 6. Also need to figure out how to controll the air preasure for fastning and releasing the chuck. If that works i could use a simple loose lap and try something out before im going to start on the lap part.

Manny things need to be done but im getting there

Greetings,

Christian

 

[vonG]

The automated dopping feature is really sweet, but would it not be better to get the complete setup running first? I mean, why spend money on special things if you dont know if the machine will cut the stones as you like in the first place. When the simple solution work, then you can spend money on the special upgrades.

I would go for the simple solution first and make that work. It isnt that hard to dop the stones manually or change the dopstick manually either. You still have to clean the machine and laps after each stone or lap change anyway.

Just my 2 cents.

 

[ken2m]

I had a video sugested on youtube which had a dual lap setup with 5-6 stones being ground at a time (identical cuts on stones) having a multiple lap setup eliminates the problems with changing laps. Both laps rotated at the same time, so only one drive motor. You would have a reduction in machine cost, and all laps would remain running as close to perfectly as possible. Ken.

 

[Nena]

Hi Ken2m,

I saw the youtube video, but the thing is i also want to cut on multiple laps. Lets say first cut at 300 second cut on 600/800 then to 3000 and finally to 50.000 or 200.000 grid. Thats more than two laps. So i want to build a automatic lap changer, the laps are all mounted on a pre measured chuck to the hight will not be an issue for the first tests, i could alway's measure out the laps when it is in place. But also the machine in India is made for production and speed. Thats not the way i want to go but there is alway's an possibility for such as.

The first thing i need to do is write the first setup page for setting up the cutting head. Where i can set the zero position and set all the indexes. I finally know my main index now and that is 256 ^ 2 so thats an index of 65.536. As i have 80.000 steps per revolution i need to share some steps to the sensor. All these kind of functions need to be made for the machine. It is also a must to do otherwise the machine will never be cutting a stone

Here is a test video that i just uploaded.

[video=640,480]https://youtu.be/njnfqJNiQhk[/video]

So i have a start for the head but now i need to proper setup/program it all. This takes time. I also want to include some test functions for turning the head etc. I just hope it all works as i think it will But there are more than one solution to it all.

Yesterday i re-installed my machine computer and i installed Linux Debian 9.1.0 for running the complete system. Also Linux can be controlled by SSH (network controll) And as i dont want to program on my room upstairs i can program on the couch behind the tv and relax

Hope to show you more when i got the setup ready and the main design of the machine control. Also i will make a new video then and do some tests. The next things are all other axis.

VonG:

I know, the first things im going to do is program some software to setup the machine and do some tests with it. But if i see the video's online i know the machine will work. Even i if i need to calculate everything before cutting and dont use sensors. But in this case i use sensors, but dont know if everything is going to work as planed! I do my best to get it all to work.

I want to do some tests if the main part of the software works. But there are so manny difficulties for 'just' do a test. There are so manny variables to the machine and i need to handle those each. This will take some time to cover them all.

Greets,

Christian

 

[Nena]

Yeah!

Did made a start with the main software for controling the machine! Till now its only the login system but it is a start. I already have done some homework for the drawing software (3d) you can see that in some earlier posts. But hope to do the setup page for the first motor soon! Why the login?! Becourse it is on a open network. If some one is going to war drive and is hacking the network they cannot power the machine if they dont have a login of the machine. The complete machine can be controlled by an ipad or laptop so i dont want someone to peek in!

Christian

 

[Sodabowski]

...and I thought I was crazy to build a computer-controlled faceting machine! Feeling less alone now!
I'm still sourcing parts for mine but might start machining the custom parts this winter. Your machine is blatant overkill for me (although you might end up with a very sturdy industrial-level machine that could make tons of money) but the software you've been writing really caught my attention! Looking forward to more both for the rest of the construction AND the software!
Cheers from France

Thomas

 

[Nena]

Thanks,

Are you also doing everything yourself?! Also the software?! It is a pain in the *** but ill go on with building and do the software, just made some charts for the motor turning. The first image is a indexing the second is the motor curve but the motorcurve is still not oke. Im getting there but it will take some time! Im interrested what your budget for your machine is and how your going to get the precission about your machine! And how you are going to manage the motors!

Greets from the Netherlands.

Christian

 

[Sodabowski]

I've gone almost the same way as you : industrial parts (linear bearings, rails and supports, linear threaded rods with nuts, and so on) as much as I can do and homemade software, but for the stepper motors I'm going the cheap way for starters. All the specific parts will be homemade from aluminium (putting the finishing touches on the plans at the moment).

I'm kinda copying the UltraTec design, only with stepper motors and without a full-steel structure for now. I'm not really concerned about utmost precision (0.17 for starters) since I'll mainly use it to cut simple designs (SRBs, emerald/baguette cuts), and then see if it can do better as is or if an upgrade is needed : there will be a classic index wheel with a positive lock actuated by an electromagnet (stepper to rotate the index wheel). No automatic transfer (but I thought about it, no simple solution yet), single lap (same thing, I also thought of the jukebox solution), Ethernet control via an Arduino or a Raspberry depending on my mood and the ease of programming and interfacing it with PHP and LiveCode

The first version will probably be a semi-automatic controller, where I enter the angle and facet index so the machine autorotates the stone into position, then do the cut by hand (which would already be a tad faster than doing it all by hand).

To control the actual cut, I'll be using strain gauges, although the idea of a MEMS microphone or accelerometer sounds great (I have both in stock).

Greetings

Thomas

 

[Nena]

Sounds nice but raspberry isnt going to work... its unstable for the motors, and how do you going to hold the axis? I use harmonic drives to hold everything in place. So i have 80.000 steps per revolution thats 0,0045 degrees per step. The sensors are all high resolution etc... it just didnt going to work with a small machine.

Greets.

Christian

PS What is your budget?! What are you planing to spend on your machine!?

 

[Sodabowski]

BTW here's an old post of my design (modelled in Cinema4D, I don't have SW yet):

https://www.prospectingaustralia.com/forum/viewtopic.php?pid=123485#p123485

 

[Nena]

What do you think your overal accuracy will be?! And why arent you using block guides?! Hope it is all working out for you. I have seen many machines but the accuracy is very poor at the end. Mayby you should redesign your setup. As it looks to me you will get some troubles with holding the head. Your design is for small stones but i wonder if its going to work for fully fecet the stone. How about 90degrees?! And is it possible to cut stones under de 1mm? The goal with my machine is cutting stones under the one milimeter up to 120mm in diameter. (max 2.2kg stones) and the complete setup will be smaller then 1/100 of a mm. Till now im about 0.008mm for the complete machine. Linear rods going to set stuck in a few times with cutting stones. I dont think thats a good solution to use.

 

[Sodabowski]

The plan is stones between 0.5 and ~30mm, but the aim really is to do melee stones fast & acurate enough
If it works as expected and the machine pays back, I'll make something better - but not for anything bigger.
The angles will vary from 90 to -30 degrees (120 total), again if this needs more precision than the original I will change the motor for a more precise one. If the hardware flexes too much I'll also get bigger linear rods (trapezoidal rods) and ball screws (I made the plan for both). But really, I don't plan on using it for anything else than classic, "easy" designs, like the SRB, emerald cut, and so on, on small stones under 2 carats. For fancy stuff and bigger and real quality gems, I plan on investing in an UltraTec when money permits. On my side money really is the limit!

 

[Nena]

Yeah the money... Here is the same. The machine is costing me about 40k and a lot of hours. The precission 0.22uM. But it isnt 'just change a motor' friend. Hope you manage what you want! Realy do! The most i've seen made a build but they where stuck there. Im taking in total 3 years to complete my machine. And still need to do a lot. How you going to level everything?! Do you have the equipment to set everything straight?! My machine can all the designs you find, even own made and even rounded facets. Lenses, drills and end mills, thin sections etc. The movement is 360 deg. Also fansy cutting should be on my machine. Im going for it.

 

[Sodabowski]

I'll set it on a flat granite stone then shim the hell out of it to level it, checking with a DTI gauge Then for the vertical axis I will optically align it (by interferometry, where it helps to be a physicist!). I prefer to start medium and eventually build something bigger if needed!

 

[Nena]

Hi Folks!

Again a update! Sorry its taking so long but it is still a lot of work! I just finished the cross table for the machine and hope to go build it soon... Still need to do a lot but im getting there, its taking a while but i wont quit! The left table is the table for the lapp and the right table is for the high speed spindle (24000 rpm) so that i also can use it for other things like fansy cutting or for cutting the dop sticks etc. Here is a drawing of what i have now.

Speek to you later!

Christian

 

[Sodabowski]

You're positively nuts

Been working on the software on my side, I have all the foundations for the CAD software done, and most of the gem scanner software done too.

TTYL

 

[Nena]

Nice! Did you manage to get a 3d scanner?! But why am i positively nuts?! And what about CAD?! Why not program your own stones in your programming language?! Just some questions!

Christian

 

[Sodabowski]

Haha no, there's a far easier and cheaper method of scanning the stones than spending a fortune on a top notch machine that really is meant for high-end gemstones like diamonds and top clarity emeralds - I'm only after the overall shape of the rough, to optimize yield and adapt the faceting model to fit it and get the most of the rough.

My solution takes a rotating platform (as the big boys' version), but only an electroluminescent panel to backlight the stone uniformly, and a mere webcam (640x480 is waaaay enough) to take pictures of the stone at defined angles. The software will pilot the rotating platform (with the stone lightly dopped to a needle, stuck to a stepper motor, even low resolution since I really only need pictures every 10 or so). For every picture of the stone, rotated along the vertical axis at a given angle, the stone's contour is extracted (hence the EL panel to have a uniformly flat white background) and vectorized into a set of points, which are then merged to build a 3D mesh model of the stone. The vectorizing part is done and works great (and rather fast for an interpreted language), I'm currently working on the reconstruction of the whole model via 3 and 4 point polygons (3 points for the scanned points along the axis, 4 for all the others). Each picture gives two profiles at opposed angles (along the axis of rotation), a half-rev of the stone giving all the needed views to fully vectorize the stone's external contour. Any voids or crevices will have to be roughed out beforehand because this fast and cheap scanning method doesn't allow for concave shapes to be scanned (I don't really care, since most rough is convex).

I built my 3D gem modeler as a polygonal boolean modeler (hence the polygonal scanner and software), which is enough here. I haven't finalized the output view yet (and will post some screenshots to show you peeps) but it definitely works and simulates a gem cutting machine with similar controls (index, angle, mast height), which will allow for easy export of the designs to the de-facto standard printable sheet.

I'm programming everything with Livecode (formerly known as Runtime Revolution), which is multiplatform, meaning MacOS, Linux,Win, and even several mobile operating systems. It's a high-level interpreted software, which runs fast enough for the very low workload needed here - and I'm doing it on a very old G4 processor Mac Mini.

I think you're nuts because you're really building the ultimate gem cutting machine! ]
Building a faceting machine takes a nutcase (and I'm one), but THIS !!! You're really the epitome of nutty engineering and I have the utmost admiration and respect for what you're doing! :Y:

 

[Sodabowski]

You know what, I'm creating a new thread under the lapidary topic to avoid mixing up things with yours

 

[Nena]

Hello Folks!

Its going to look more and more a machine the cross table parts are being made right now and im trying to put all cables to the motors/sensors etc. But aint there yet! All good things come slow?! Or how the say it?! lol Hopefully soon more!!

Greets, Groetjes,

Christian