recomendations for table top milling setups?

    thinking of doing some small (and I do mean _small_) milling, to make a set of molds for casting type for hand setting printing (aka "Letter Press").     Not sure it is a cost effective method of getting a font, but this would be a hobby, not a business. Unless maybe I get into the type casting foundry bizz.
    So, what I am looking for is something which can give good precision at the .005 inch level. (I'm sure, finding the mills is going to be another issue, but one thing at a time.)

  Quoted Text Here  
Freehand carving on a milling machine is very difficult, even a rectangular panel meter cutout can go astray from backlash. Smooth curves and straight diagonals are nearly impossible.
If you can define a line by an equation you can have a spreadsheet print a list of x,y points along it and then plunge cut them. I would first rough cut them somewhat shallow with a fairly wide spacing, then finish at final depth and tighter spacing to minimize tool deflection or backlash jumps. Define the spacing between points in a cell so you can change it.
My antique Clausing mill is nice for milling small details because the table handles have very low initial static friction and enough rotational inertia that they don't jump from a light push. I can divide 0.001" of travel into 4 or 5 steps. Currently manufactured small mills I've tried were the opposite, sticky and jumpy.
I spent a year drawing the fonts for a new printer, dot by dot, the full IBM character sets, in regular, bold and italic.
https://en.wikipedia.org/wiki/B%C3%A9zier_curve

On Sunday, June 16, 2019 at 2:25:13 PM UTC-7, Jim Wilkins wrote:
  Quoted Text Here  
Yow! It only took me an afternoon to create a lowercase font for the old Versatec plotter; the V- 80 was a good low-cost producer, with terrible hardware design and first-class software. It came with a connect-the-dots font editing utility, but only one uppercase font.
Did those IBM character sets include APL fonts?
Electronic design trivia: the V-80 power supply used screw-terminal capacitors, fastened directly to a thick (3mm) printed wiring board, which was then wave-soldered (solder over screws). Some diodes were under the capacitor, and when heated (and lifted) in the solder process they melted through the insulation jacket on the electrolytic cans, which shorted completely after some time in the field. Melted metal sprayed at each event...
Repairing it, you then had to loosen the screws, with the solder-blobbed heads.

  Quoted Text Here  
No, it was an office printer, the Howtek color inkjet of 1986, developed by the Centronics engineers. https://en.wikipedia.org/wiki/Talk%3ASolid_ink
During an Army field exercise I designed a simple font in about 2 hours, that displayed in the punched hole patterns of Teletype paper tape. The regular printout appeared to be random garbage that satisfied the requirement to exercise the system somehow. A full, credible Times Roman or Bodoni font isn't nearly so quick or easy.

On Sunday, June 16, 2019 at 3:02:46 PM UTC-4, pyotr filipivich wrote:
  Quoted Text Here  
http://www.taigtools.com/ nice CNC setup that should work fine for your purpose. There is plenty of free software that can convert drawings to gcode to run it.

  Quoted Text Here  
How much does the entry level model give up, compared to the ballscrew one?
Which 2D (or 3D) CAD program would you use?

On Monday, June 17, 2019 at 12:26:31 PM UTC-4, Jim Wilkins wrote:
  Quoted Text Here  
I am using the leadscrew model, and once the backlash is adjusted (there ar e adjustable nuts) and you've applied backlash adjustment in the control so ftware, pretty much what you're giving up is rapid speed.
For cad/cam I use (mostly) Fusion 360 which is free for mere mortals, and v ery well supported. Additionally, I have been engraving logos into a black plastic box at a depth of 0.0005" and filling the engraving with white pain t. The result is beautiful.
As far as precision goes, I've been isolation milling printed circuit board s with excellent results. This involves .006" traces with .006" spacing, an d once I got the machine dialed in, the results are excellent. The software for this is also free - Kicad for the design and Flatcam to convert the ge rbers to gcode. In between, there's Autoleveller which probes the board sur face and modifies the gcode appropriately. Engraving with v bits requires t ight control in the z axis.
One big change that I did make is replacing the spindle with a 24K rpm wat er cooled motor with VFD. I did this because my early engraving tests appea red to be affected by vibration from the stock motor. In the Taig group on facebook, there is a lot of talk about motor upgrades, with many people usi ng a consew motor and controller (from a sewing machine). There are also pe ople selling upgraded frames (way more rigid) and even an upgraded head wit h R8 holder and power drawbar.
For the work you're describing, though, you wouldn't need to do any upgrade s.

On Sunday, June 16, 2019 at 3:02:46 PM UTC-4, pyotr filipivich wrote:
  Quoted Text Here  
Just wondering out loud here, but with your small feature sizes and modest 0.005" accuracy requirement would a CNC router be a good choice? Should give you a higher speed spindle which would be useful with tiny cutters.

On Monday, June 17, 2019 at 7:16:21 PM UTC-4, Carl wrote:
  Quoted Text Here  
It wouldn't be a bad way to go, and probably considerably less expensive to get started, but compared to the Taig, you're giving up a lot of Z capacity and rigidity. You're not going to cut steel with a router.

-0400 typed in rec.crafts.metalworking the following:
  Quoted Text Here  
    That is what I am wondering. I realized after some of the first responses, that I really should be looking at what I "need" for dimensions, and work up a spec from that. As much as I might want 4 pt type, it might not be doable. Let me rephrase that, "it might not be doable _on my budget_." (I mean, IBM made letters out xeon atoms, but a) that's _too small_ and B) I don't think I can afford the power bills for a scanning electron microscope and accessories.)

on Mon, 17 Jun 2019 19:15:34
  Quoted Text Here  
On the font design system I used the horizontal pitch resolution was 1/480 inch, 240DPI dots shifted half a space. Moving a dot 0.002" could make a noticeable improvement in the smoothness of curves from normal reading distance.

-0400 typed in rec.crafts.metalworking the following:
  Quoted Text Here  
    As I said, I'm at the start of the project. I need to figure out what sort of detail / repeatability I will want.
  Quoted Text Here  

on Tue, 18 Jun 2019 15:49:21
  Quoted Text Here  
The Centronics printer engineers had specified that resolution to achieve commercially acceptable visual quality at maximum printing speed. Laserjet font dot spacing at that time was 300 DPI for both X and Y but toner spatter blurred the dots together and hid the jaggies.
Those are two expert opinions of the minimum necessary resolution/accuracy for fonts, limited in part by 1980's microprocessor and ASIC speed since vector and .pdf images need to be rasterized. Later Laserjets went to 600DPI. https://www.formula.co.uk/tech5.htm
In other words you don't need 0.001" accuracy for type but 0.005" (from backlash) may be objectional, depending on your budget, the application and the intended customer. https://www.commarts.com/columns/know-if-a-font-sucks
The mills I've used or examined typically had 0.001" resolution on the hand crank dials, 0.0005" or better on a DRO (CNC and manual Bridgeports). Achieving that accuracy in practice may require careful adjustment. I tried to square up an RF-31 mill-drill whose column wasn't quite perpendicular to the table, and the tee slots weren't parallel to the X axis motion. It was good for maybe 0.005" in X and Y as received though the dial graduations were 0.001". Z was more like 0.020" because locking the quill shifted it.
If you can't clean your machine up to meet your needs you've wasted your money. In that instance it was good enough to machine electronic control panels as intended but not to make unexpected repair parts for IC handlers and testers, so I roughed them out on it and trimmed them to size at home.
0.005" error could be too much for the sideplate bearing holes of a geared robot transmission. Having a milling machine empowers your creativity.

-0400 typed in rec.crafts.metalworking the following:
  Quoted Text Here  
    I will keep that in mind.
    I will also keep in mind that "saving money" in the short term, is not always cost effective. (Anybody want a Harbor Freight router table - with router? Runs fine, but adjustment in Z are ... "problematic". It is on the "replace next" list. Will see what I can find at the pawn shops.)
  Quoted Text Here  

on Wed, 19 Jun 2019 09:36:53
  Quoted Text Here  
If you had machine tools you might be able to improve it. I've made plenty of metal replacements for worn, broken or overly 'value-engineered' plastic parts, when the machine was otherwise worth saving and a purchased part would be just as frail, or in many cases unavailable. This reclining office chair came with arms and a low back. I milled replacement brackets that raise the back to support my head and let it roll under the keyboard shelf.
My machine tools have never and likely will never demonstrably pay back their investments. Instead they gave me abilities and a bag of sample work that may have gotten me good electronic + optical + mechanical R&D jobs.
Perhaps the decision reduces to whether or not you see better ways to do things.

-0400 typed in rec.crafts.metalworking the following:
  Quoted Text Here  
    I have found that there are alot of things, which the only way I can afford to do them, would require a large lottery jackpot. At which point, with that much money, why fix what I could replace? (Aside from the challenge of it.)

on Wed, 19 Jun 2019 13:12:10
  Quoted Text Here  
I learned to evaluate the build / buy decision early in my custom equipment building apprenticeship. I had to build several things myself at least once (on company time) to understand the time and effort involved. For example I can braze a refrigeration system but am much better off contracting it out, likewise I can make strong flat welds in steel but not vertical or overhead, or any in aluminum. I can weld a steel patch into a car rust hole to my own satisfaction but maybe not anyone elses's. That paid for the MIG machine with the first use, the body shop quote for one fender was $800 before painting..
I've liked creating stuff since I was little and am good enough at it that my machines and electronics usually work with little or no modification, for instance a homebrew computer, several parts of a satellite network controller, a hydraulic bucket loader for my tractor and the bandsaw lumber mill mentioned in the Setscrew thread.
Yes, metalworking machine tools are expensive, much more than woodworking equipment because the cutting forces are much higher, and their value to a hobbyist can be hard to justify. It takes less force to push a bandsaw through 20" oak than 1" steel. I stopped at a 10" collet lathe and small vertical knee mill, plus a 30" combo sheet metal machine, all fine for a hobbist or prototyper but inadequate for a money-making shop. I haven't bought a CNC mill or lathe because my manual machines have proven capable of everything I seriously wanted to do.
If I had more artistic talent I might acquire a small CNC mill and blacksmith forge and maybe a 3D printer. I've dabbled with them but the results weren't worth the cost. The CAD system I learned for circuit board design can draw only straight lines and circular or elliptical arcs, which along with my manual machine tools have proven to be enough for everything I've wanted to make.

On Tuesday, June 18, 2019 at 2:03:40 PM UTC-4, pyotr filipivich wrote:
  Quoted Text Here  
st
  Quoted Text Here  
That atom-moving video is pretty cool
https://youtu.be/oSCX78-8-q0

https://youtu.be/oSCX78-8-q0
for thos e who haven't seen it.
I'm not so sure that four-point type is out of reach of reasonably priced m achinery. You don't have high cutting forces, so you can use spring-loaded anti-backlash nuts without a problem, and both mach3 and linuxcnc (to which I am strongly leaning) have the capability of compensating for irregularit ies in motion (lead (or ball)screw mapping). As I said, I am regularly gett ing better than 0.001" repeatability on a leadscrew taig.
Also, on the subject of (free) software, have a look at www.scorchworks.com . F-engrave may be useful to you in converting fonts to gcode. I have used it to make aluminum tags to label refrigeration valves. He's got some othe r cool stuff too.

On 6/17/2019 9:56 AM, rangerssuck wrote:
  Quoted Text Here  
Does the Mach3 software support G41/G42 cutter compensation?
David

On Monday, June 17, 2019 at 8:26:32 PM UTC-4, David R. Birch wrote:
  Quoted Text Here  
Yes, with some limitations https://www.cnczone.com/forums/tormach-personal-cnc-mill/241162-tormach.html
I am using mach3, but I'm really thinking about switching to linuxcnc (formerly emc2).