What's new
What's new

2D drawing to g-code.

emtor

Plastic
Joined
Dec 11, 2008
Location
Norway
I'm a total newbie at cad, cam and cnc.
Here's my situation: I make knives, and these of course need handles.
I make typical Sami-culture knives and these handles are not simply a round piece of wood. The fact is, they have a quite complicated shape, so complicated in fact that making them by hand is very labour intensive and very difficult to make them look good. It also takes forever to make them.
In order to let you get an idea of what I'm up against let me show you an image of a typical Sami-style knife:

niibi.jpg

Let's say I make a drawing of the profile of this handle, like this:

kniv.jpg

The drawing shows (from top down) the profile viewed from the front (pointing the blade towards you). Then the knife is seen from the side, and lastly seen from the top.

Would a cad/cam-whatever software make enough sense of this 2D drawing to produce g-code or do I need to make a 3D file in cad software?
I do hope I don't need to learn how to make 3D files in a cad software.

I'm grateful for any hints.
 
Hi emtor:
OK where to begin?
Yes, you need to have a 3D CAD model if you wish to make complete handles to net shape or near net shape on a CNC milling machine.
You will than take that 3D CAD model and design toolpaths onto it allowing you to take chunks of your raw material, fixture them onto your machine and apply cutters to them to chop away the excess.
Depending on the materials you choose and your cutting strategy, you can get parts that will require only finishing (sanding polishing coating) to become complete handles ready for installing onto your blades, or, if you design the process correctly, to shape raw blanks that are already attached to your blades.

All of this will cost a lot of money, and your expected outlay will be in the low tens of thousands of dollars if you want a low budget solution and in the high tens of thousands to low hundreds of thousands if you want a state-of-the-art "John Grimsmo" "spend like a maniac" kind of solution.

You have to make the decision whether your sales volume can support this kind of expense; just don't kid yourself that buying the goodies solves the problems.
You have to learn how to do it and judging from the nature of your inquiry, you will have to learn from the ground up, so expect heartbreak and frustration before the clouds lift and the sun shines on your promised land.

You will lose: A lot of cash and a lot of hair getting this to work from where you're starting out.
You will gain: an autonomous or nearly autonomous process that will take a lot of labour out of each knife and improve your ability to make a knife to a price point.

There is another way you may wish to consider:
The advent of CNC has made pantographic tracer mills totally obsolete, so very high quality models in great working condition can be picked up for truly dirt cheap.

A Deckel KF1 or KF2 used to be a premier machine costing a fortune; now you can score one for much less than the cost of decent CAD software alone.

It requires two things not needed by the CNC approach:
1) It needs a physical model for you to copy.
2) It needs an operator to run the machine.

It does NOT need:
1) A big investment in CAD software to buy AND to learn
2) A big investment in CAM software to buy and to learn
3) A big investment in a CNC machine to buy and to learn

It will make handles just as nice as those you can make with the CNC toys and it will make them probably just as fast, but it won't make them overnight while you snooze, and it won't make them if you are not standing in front of the machine working the handles.

So that's it in a nutshell; your choice depends mostly on your appetite for risk and your patience and willingness to take the time to learn what you will need to know to become successful...it's not rocket science but it's not trivial either.

Last; remove right now from your head, the notion that moving in this direction will magically take away the need to attend to the handle making part of your business completely...it will not.

If you want to make and sell quality knives you will still need to finish the handles ; sanding, polishing, detailing, coating and all that good stuff.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining
 
All of this will cost a lot of money, and your expected outlay will be in the low tens of thousands of dollars if you want a low budget solution and in the high tens of thousands to low hundreds of thousands if you want a state-of-the-art "John Grimsmo" "spend like a maniac" kind of solution.

I think you might be overstating the case a little bit.

Let's assume he's doing a wood glue-up and winding up with a piece of square or round stock.

Throw it in a $5k Tormach 440, in a cheap vise, facing vertically (knife hilt up). 1/2" bull nose end mill, an adaptive clearing to get most of the stock off, a spiral contour op to finish and he's there. 10-20 minutes of machine time. Program it in Fusion for free. Call the total investment about $6k.

IDK if the business case can support that, but we're hardly talking "low tens of thousands" levels here.
 
A pantographic tracer mill is much like the small copying mills used when people are copying fishing lures. Quite simple machines, but as you say it requires a model to copy. If you look at the scetch you'll note that the transition from the handle to the "knob" at the rear end is curved.
This curve is the problem,-it's very difficult to make and to make it look pretty. Making a durable model in for example metal will be time consuming and difficult. Same thing for wood or any other material,-very tricky. There are knifemakers here who manages to make this transition to perfection but they spend weeks and months and these knives cost up to 2 to 3 thousand dollars and end up in exibitions or museums.
I've looked at other possibilities that don't involve cnc-machines. Some people build small grinders that grind cams on camshafts for smaller engines. A model cam is mounted between two lathe chucks and a wheel rests upon the model cam. On the same axis the camshaft to be ground is mounted between two other chucks. The model cam makes the wheel rock back and forth and makes a grinding disk follow the movement of the wheel when the assembly is rotating. The grinding disc could be replaced by a router and the router assembly could be made to move along the knife handle guided by a steel guide having the same profile as shown in the drawing I posted (side view).
Perhaps this will be an easier and cheaper solution.

I know I won't be able to buy a cnc machine, and I don't think I would need a regular cnc machine either, I'm envisioning a small wood lathe with some minor changes to mount the knife with a raw handle already attached with the motor removed and replaced with a stepper motor to make the knife rotate nice and slow. The router would only need to move in two axis,-up/down or back and forth depending on how the machine is designed,- and sideways along the length of the handle. I have a good grasp of electronics so that won't be a problem.

A purely mechanical machine based on those camshaft grinders have a drawback: I would only be able to make one type of handle unless the profile cam and guide is replaced, but a cnc machine could make all kinds of handles.

It's a choice between being shot or hanged :ack2:
 
I think you might be overstating the case a little bit.

Let's assume he's doing a wood glue-up and winding up with a piece of square or round stock.

Throw it in a $5k Tormach 440, in a cheap vise, facing vertically (knife hilt up). 1/2" bull nose end mill, an adaptive clearing to get most of the stock off, a spiral contour op to finish and he's there. 10-20 minutes of machine time. Program it in Fusion for free. Call the total investment about $6k.

IDK if the business case can support that, but we're hardly talking "low tens of thousands" levels here.

If you read my last post you may already have figured out that I'm playing with the idea of making a simple and specialized cnc myself. Linear bearings, shafts and lead screws plus stepper motors can be bought off ebay. I also have a friend that teaches machine work at the local high school. Tons of lathes, milling and cnc machines there. He could surely help out.
I just might be able to pull it off if I'm lucky.
 
I would go with what emplmex said except I would look at a copy lathe, not a mill. These have produced millions of parts that are basically round which is what you have. Instead of a high speed spindle, the spindle rotates slowly and a motorized cutting head, think similar to a toolpost grinder, does the work.

Tom
 
I would go with what emplmex said except I would look at a copy lathe, not a mill. These have produced millions of parts that are basically round which is what you have. Instead of a high speed spindle, the spindle rotates slowly and a motorized cutting head, think similar to a toolpost grinder, does the work.

Tom

The problem is that the handle isn't round,-or to put it this way: -The back (top) has a certain radius, the belly has a very much smaller radius and the sides have a very much larger radius,-almost flat but the sides still have a radius. It's more like an elongated egg shape. Could that be done with a copy lathe?
I see no other option than to use a cnc machine or a mechanical machine based on a cam and a guide.
 
Hi again emtor:
Your principal problem is probably going to be developing the 3D model; you'd be surprised just how difficult it is.

As an example from my own recent experience; one of my customers is a sex shop and as you can imagine, I make a lot of weird shit for them; mostly organic shapes.
The last one was a titanium dildo and I can tell you it took a LOT of modeling to get a decent looking result.
I've got a half dozen 3D printed versions that were "not quite right" before I got the CAD model crafted to my satisfaction. The design ended up taking about 60 hours of my time, and there's unsatisfactory dildo models littering my shop, arousing snarky comments whenever other customers show up.:D

All of them looked great in CAD but they all failed the 3D print test, and the failures were sometimes quite subtle; not dissimilar to what you're describing as the subtle design details of your knife handles.

In fact, I find it normally MUCH easier to sculpt my shapes physically, then have them digitized, and I'm exploring how best to do that process with my customer right now simply because it's such a challenge to do it all digitally for the shapes he needs.

When it needs to be eye sweet or sweet to the hand, there's still no substitute for a physically crafted master model...if it needs to be perfectly symmetrical you can do that part much better digitally, but that's just a matter of cutting the scan in half digitally and mirroring the better half.

Returning to the physical shaping of the knife handles; have you considered mounting a Sherline lathe on the work table of a Deckel KF 1, mounting another on the drag table, slaving the spindles together with steppers or even a flex shaft and treating the whole works like a copy lathe?
It would be a robust way of getting the mechanics without having to make a CNC build and then having to program it.
Yes, you'd have to make a physical pattern, but once you've made it, you're all set to duplicate it as many times as you want.

If you insist on going the digital route, you could CAD design it, then 3D print it in one of the new hard carbon fiber materials, doll up the fine details by hand as needed and run it as your master.

You'd be sure of getting a handle that's just like the master; so all you'd need is a really good master, but you only have to make one, so once you do, you're golden.
The nice thing about using the KF 1 as your machine platform is that it's a quality piece of kit; not some cobbled together POS you have to fight with to get a decent output from.

You could do the same with a copy lathe if you can find a reasonably small one.

Once you've got a half million bucks to spend, you can buy yourself a nice Mill-Turn and run the digital models directly, but by then you'll be miles past needing our advice!!

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining
 
How many handles do you think you will need in an average year? What about subbing it out? Your first handle may cost you over a thousand dollars in paying for the modeling but after that someone who already has a router and is used to doing 3d work should be able to make them for a reasonable price if you are willing to buy 10-20+ at a time.

Unless you are making thousands per year I suspect this could be cheaper than buying a cnc software etc. Then spending the time to learn how to use it. If your plan is to make a product all the time spent making a tool to make the product is time not spent making your product.
 
Lots of good advice here,-thanks a lot.

I have been convinced that cnc is not for me. It will take too long to learn how to make 3D files in a cad software. I have no problems with any other software like photoshop, video editors etc. but I have downloaded cad software a while ago just to try it out and I didn't manage to do anything useful with it. I quickly realized that it would take forever to master such skills.

I will have to try a mechanical approach to solve my needs without having to make a model.
Explaining my idea is like trying to explain in words how to tie shoelaces, but I'll give it a try never the less.
Imagine a very simple lathe where you can mount a handle blank.
In addition to this the "lathe" has an extra spindle below the main spindle.
These two spindles move syncronous and very slow, and the bottom spindle has a sort of camlobe mounted on it that is equal to the cross sectional shape of the knife handle.

Then comes the assembly on which the router will be mounted.
First there is a base plate which moves along the length of the handle with the aid of linear bearings, steel rods and a lead screw.
Another plate with linear bearings etc. is mounted at a 90 degree angle on top of the first plate, which will make it move either towards or away from the workpiece. The movement of this second assembly is governed by a guide that would look something like this:

guide.jpg

When the first (bottom) assembly moves along the length of the handle the second assembly is forced to follow the shape of the guide as it moves along.
This assembly would need springs to ensure contact with the guide.

On top of the second assembly sits the plate that the router will be mounted on. This assembly will be equal to the second assembly and it's movement is governed by the "camlobe" mounted on the extra spindle on the "lathe".
The camlobe will look something like this:

lobe.jpg

This last assembly will also need springs to ensure good contact with the "camlobe".

As the whole assembly moves along, the lobe will push on the top assembly more and more toward the edge of the plate. I may need more than one lobe on the axle to ensure proper operation.

The router mount will also have a leadscrew so that I can adjust the distance from the routerbit to the the workpiece.

The two spindles on the "lathe" will be coupled together with sprockets and chain or wheels and a belt of some sort. This could be driven by a windshield wiper motor which could be supplied with a DC-squarewave voltage with variable
pulse width to adjust speed without loosing torque.

I hope this explanation makes sense. I should have made a drawing but I'm really lousy at drawing machinery.
 
How many handles do you think you will need in an average year

Not too many, as this is not really a bussiness but rather a hobby with economical benefits.
These knives are very popular particularly in the Sami areas of Norway, Sweden and Finland.
Hunters use them, reindeer herders use them and they are given away as gifts when Sami youngsters go through the religious ritual of what we call confirmation. Babies get baptized, then at the age of 15 their faith is confirmed in the church even if most of them have no faith at all. Families spend fortunes on their young during this ritual and a sami knife is optional and the nicer the knives, the more money they are willing to spend. I was at a Sami market on saturday and an old man sold sami knives. Small and ugly . . . 400 USD, so there is a nice chunk of money to be made. You then have the reindeer herders who are not happy with factory made sami knives, and of course the tourists in summer and several markets throughout the year.

Up until now I have made knives by hand and people here are crazy about them, but I refuse to sit in a dustcloud with sanding paper for days on end for producing just one knife.
 
If you read my last post you may already have figured out that I'm playing with the idea of making a simple and specialized cnc myself. Linear bearings, shafts and lead screws plus stepper motors can be bought off ebay. I also have a friend that teaches machine work at the local high school. Tons of lathes, milling and cnc machines there. He could surely help out.
I just might be able to pull it off if I'm lucky.

This really needs to be posted over at "the zone" www. see and see zone dot com

I will ask, how about using a 3-d scanner to get your model ?
 
... but I refuse to sit in a dustcloud with sanding paper for days on end for producing just one knife.

That's kind of knife making. Most if not all of the fine knife makers still hand finish their knives and command top dollar as such. If you're just trying to stamp these out but also demand the hand made price you may be on the wrong path.
 
That's kind of knife making. Most if not all of the fine knife makers still hand finish their knives and command top dollar as such. If you're just trying to stamp these out but also demand the hand made price you may be on the wrong path.

I won't charge the price they charge for the most excellent handmade knives. They will be cheaper but also as nice as the better handmade ones.
I mentioned I was at a market this weekend. Quite ugly, non symmetrical and not very well made. Still the man wanted 400$ for them.
The factory made ones cost from 100 to 150$. I plan on charging about the same amount.

Edit:
A top notch handcrafted samiknife costs about 2000$. They had a few of those in a small shop in Finland. They sat on the shelf for a couple of years until the shop closed down. These knives usually end up in museums or as exibition items or get sold to very wealthy collectors.
The other alternatives are factory made knives that people in the reindeer husbandry are not happy with, or quite crude and not very pretty hand made knives with factory mass produced blades. I forge and harden and temper my own blades. Differentially tempered which gives a hard edge and a springy spine. My handles are made from reindeer antler and stacked birch bark which is very pretty. No shrinkage over time as with wooden handles that tend to come loose after a number of years. People can choose,-do they want an ugly and crude knife with a mass produced blade or a pretty and good looking knife with a hand forged blade and a "mass produced" handle?
If they pay approx. the same as for a factory made knife I think they get a decent deal.
-And who knows,-maybe those top notch super expensive knives are made with a dremel tool?
 
Last edited:
I'll look into the cost of one of those.

Your best bet is to go with someone who already has the specialized equipment and software to capture your knife design. People who do 3D scanning for a living are set up to get you good results for a fraction you would pay to get set up on your own. We're talking many thousands of dollars (I have around $4500 in my hardware alone) to get started and even when you have the equipment and software it takes a while to get good results with the proper lighting setups, indexing, and post processing.
 
Your best bet is to go with someone who already has the specialized equipment and software to capture your knife design. People who do 3D scanning for a living are set up to get you good results for a fraction you would pay to get set up on your own. We're talking many thousands of dollars (I have around $4500 in my hardware alone) to get started and even when you have the equipment and software it takes a while to get good results with the proper lighting setups, indexing, and post processing.

I've looked at prices. The cheap ones (500$) looks like useless toys and those who looked like serious bussiness cost from 5000$ and upwards. I'm not spending that for something that will only be used for one job. Your advice of letting some pro do the job seems like a sound idea.
 
Hi again emtor:
In post # 10 you described what you envision as your machining device.
If I understand it correctly it won't make what you want.
The reason is that your cam lobe is constant cross section but your handle is not.

Imagine an extreme case where the butt of the handle is much larger than the blade end of the handle.
If the small end were a proportionally scaled duplicate of the large end, the rise of the required cam would be different for the small end compared to the large end.
So if the cam rise is constant your small end shape will be wider from top to bottom and skinnier from side to side than the proportions of the big end.
Your cross section must vary constantly along the length of the handle, so you need to make a master no matter how you slice it.

Sure it's a pain, but I don't think you have a choice, so you may as well stop avoiding it and haul out a chunk of aluminum or brass; break out the belt sander, the files and the buffer and go to it.
You only have to do it once for each size of knife you make...embrace the pain and just do it.

And save yourself a lot of pain...get the little Deckel...it's gonna be SO much easier than bodging together the device you're talking about making from scratch, and I can pretty much guarantee it'll make WAY better handles than what you want to build unless you're an experienced machine builder with a generous budget.

I assume of course, that you want to get handles off your machine that are ready for finishing: a bit of hand sanding, a bit of polishing, and linseed oil or whatever treatment you intend to apply as your finish.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining
 
Going back to a copy lathe and the question about duplication. Take a look at a rifle stock. These were and probably still are made on a duplicating lathe. It does mean a physical model which you seem reluctant to make but after that it is all gravy.

Tom
 
Hi again emtor:
In post # 10 you described what you envision as your machining device.
If I understand it correctly it won't make what you want.
The reason is that your cam lobe is constant cross section but your handle is not.

Imagine an extreme case where the butt of the handle is much larger than the blade end of the handle.
If the small end were a proportionally scaled duplicate of the large end, the rise of the required cam would be different for the small end compared to the large end.
So if the cam rise is constant your small end shape will be wider from top to bottom and skinnier from side to side than the proportions of the big end.
Your cross section must vary constantly along the length of the handle, so you need to make a master no matter how you slice it.

Sure it's a pain, but I don't think you have a choice, so you may as well stop avoiding it and haul out a chunk of aluminum or brass; break out the belt sander, the files and the buffer and go to it.
You only have to do it once for each size of knife you make...embrace the pain and just do it.

And save yourself a lot of pain...get the little Deckel...it's gonna be SO much easier than bodging together the device you're talking about making from scratch, and I can pretty much guarantee it'll make WAY better handles than what you want to build unless you're an experienced machine builder with a generous budget.

I assume of course, that you want to get handles off your machine that are ready for finishing: a bit of hand sanding, a bit of polishing, and linseed oil or whatever treatment you intend to apply as your finish.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

I trust you on this. I've been wondering myself if this concept would really work for the same reasons you mention, but I assumed that as the assembly moved towards the butt the guide would force the router to get further and further from the workpiece and thus create a larger version of the cam lobe.
I've been thinking of making a very simple test assembly just to see what happened, but I never got around to do it.

Making a model may be easier if I make a lot of thin brass cross sections of the desired shape, let's say 1/16 of an inch thick,-every one a tiny bit larger than the previous one, drill centerholes, sticking a brass rod through them all and stack them together and then solder them. Then I would get away with less filing. A copy machine would then be much easier to make, sort of a beefed up fishing lure copy machine.
Sami knives come in different sizes,-the most popular being 6 inch, 7 inch and 9 inch blades. These all have the same size handles. Then there are the small knives traditionally made for women,-these are 3 to 4 inches, so I only need to make two models.

You guys have been a good help and have steered me around a few pitfalls.

And no,-I'm not an experienced machine builder. I've only made a power hacksaw and a belt sander so far.

You assume correctly,-I want the handles to come off the machine ready for hand sanding.

Deckel is a brand I've never seen here in Norway, but they are made in Germany so I could order one from there.

And lastly,-what is the greatest pain in the butt . . . spending lots of time making a machine that won't work, or spending less time making a model that WILL work?
 








 
Back
Top