Creating templates for moulder knives

Thoughts on producing templates for cutting moulder knives. June 14, 2000

I'd like to know what the consensus is on the quickest and best way to make templates for moulding knives. We're a high-end mill shop with a new Wadkin K-23. We typically run 500 to 1,500 linear feet.

I'm assuming that metal templates take a bit too long. We use plastic. We have a miniature CNC mill that we can use to machine them, but I'm not so sure this is the answer because it takes a while to draw the moulding shape in a program like AutoCad.

Once we've got the shape, we're minutes away from a template -- but I can generally knock out a plastic template by hand twice as fast as anyone seems to be able to draw it in AutoCad. I guess we could scan a drawing, but programs to scan drawings, which actually give you something AutoCad can use, seem to also take time.

Forum Responses
I usually start by scanning a tracing or scanning a fax someone has sent. I can have a working drawing in 5 to 15 minutes, depending on how complex it is. I send my file via e-mail to a template shop and get my template the next day, $25 a pop; I can't make them for that.

I just started using this service when I found out I could get them done for this price. Prior to that, I had always made metal templates (for 16 years). If I'm pushed for time I make a metal template by pasting the drawing on the template steel, then cutting and filing. The key to making metal templates is using the proper steel. Wadkin's is the right hardness, but a bit pricey. With a little research, there are other sources, and cheaper. I've had my K-23 for about four months now. It replaced a GA. I also have an older NX grinder.
I do a lot of one-off work in addition to the regular stuff. Average runs are 1,500-5,000 feet.

What software are you using to scan your drawings? My problem is getting that "working drawing" you mention into the computer. If you don't mind my asking, which CAD program do you use to clean things up?

Strange as it may sound, I use a paperport scanner and save it to tiff format. Open it with the scan utility in TurboCad (this is all I use TurboCad for), save it as a dxf, and open it in VisualCADD to do any editing, and save it as a dxf for the template shop. (VisualCADD has a scan utility by Corel, but I haven't taken time to learn it.)

The template shop normally charges more for drawings or actual wood samples, but because I've done some of the work he does it cheaper. If I get my file e-mailed to him by 4 p.m. it goes out that day. He sends it regular UPS, but because we are both in NC I usually get it by lunch the next day.

VisualCADD is a modern-day version of generic CADD. VisualCADD has most of its commands available via two-letter keyboard commands. It has a very intutive interface, any new user can be using it productively in a few hours, as opposed to years with AutoCad. VisualCADD has a fully functional demo of the program available at -the company's website, just serach on the product name.

The scanner utility included with TurboCad is by Trix Systems. I think they have a website. Their stand-alone program is a PITA to move to from one computer to another.

As I mentioned, I use a tracing of a piece of mould to start a lot of times. This saves me time over starting a profile from scratch with the CAD program. I don't worry about being exact because if it's oversize from the tracing I just dimension it and scale it to the proper size.

I suggest Lexan templates made on a CNC router of some sort. To get the informatoin to the router I use AutoCad. The profile can be entered from a scanner, or more accurately an optical comparator.

Once I have the Lexan template, I duplicate it to metal for long run profiles. I do this on the profile grinder using a technique taught in the Training Center, and this has been in one of the past issues of "The Profiler" newsletter from the Grindermans' Association.

It should be noted that a scanned image can change size some from the original. In most cases, this is not a problem. When I scan an image in, I cut a thin slice of the profile and set it on the scanner. If I have a drawing, I simply scan the image in, but the true transfer seems to be off about 3 percent (on average).
Dave Rankin, forum moderator

As I noted, when scanning a tracing of a pattern I dimension it, measure the actual mould size and scale the one on the computer to those dimensions. I then print it out at full scale and place the mould sample on it for comparison. It's not rocket science, but it works 99 percent of the time. It's fast and easy!

The size of the scanned mould is not off because of the scanner, but because you have traced on the outside of the mould profile, making it slightly larger than the original piece.

From the original questioner:
First of all, thanks for your replies. They're helping me come to a new understanding of this beast. Second, Dave, do you use a raster-to-vector conversion program of some sort after you scan? If so, what is it?

When I scan a profile I accept any variation that I may get. When I need high accuracy templates I use AutoCad to draw the profile, then transfer the file to the template machine.
Dave, forum moderator

The comments below were added after this Forum discussion was archived as a Knowledge Base article (add your comment).

Comment from contributor A:
I usually use 1/16 inch spring steel to make templates with. Very cheap and easy to work with. A CAD program sounds interesting. Paper cutouts? Sounds like you're doing the same job twice. Why not just etch your engraving on the steel? Sometimes technology seems to work against us. For example, the CAD program - how much time are you wasting on the computer trying to make a drawing, then transferring it to an engraver? Sounds kind of time consuming when on average it takes about 30-1.5 hours to make one from scratch depending on difficulty. Besides, knife grinding is a craft.

Comment from contributor B:
I agree, knife grinding is a craft. I've only been at it for a few years now, but I have found that making templates from steel does only require 1/2-1 hour (dependent on the profile), and the longevity is much greater than plastic. As well, it is easier to fix or modify a steel template that was made in-house.

Comment from contributor C:
I have made many knives by using a scanned drawing, then, using a raster to vector program, I convert it to a dxf. format and dimension it in Autocad. The dxf. drawing is then e-mailed to my grinder and I have the knives in a few days. This system has worked well for me... approximately 180-200 knives.

Comment from contributor J:
I use a CAD program called Cutting Shop by Arbor Works. It is a lot easier than TurboCAD/AutoCAD. You can convert raster to vector drawings with their program tools more efficiently. I tell it the template size I need, import or scan or draw on top of the temp layout, then convert to a DXF format. My desktop mill converts it to a CNC program. The program is easy to use.

Comment from contributor T:
I also use metal templates. I etch the profile onto the template by tracing a sample or paper cutout. Then rough grind on a pedestal grinder using old grinding wheels that have too small a diameter to use on the grinder (Foley grinder). I finish the template using various files. I rarely spend more than 30 minutes creating a template. I've done over 400 to date this way. CNC is nice, but expensive.

Comment from contributor S:
My argument would be that by drawing in CAD and using a CNC router would be a time and money saver. I really have never run into a template that I could not complete both the drawing and finished product in less than an hour. The beauty of drawing the template and having a record of it would be if the template is ever worn out or lost it can be instantly recreated. There is also the degree to which the tolerance of a CAD design and CNC machined template can be made verse a hand formed template. You simply cannot reproduce smooth curves when done by hand. Also I am a big proponent of acrylic templates as opposed to metal because of the wear of the stylus from metal on metal contact.