Early Numerical Control Machines

Long time Lurker, Infrequent Poster. I have been to quite a few antique machinery museums and shows, and I have never seen any early NC lathes / mills out there. I was wondering if anyone out there still has a NC Lathe or Mill still out there operating. I guess technically the NC age started in 1949 with John Parsons and a Cincinnati Hydro-Tel Milling machine so I think this thread would be appropriate for this forum.

I am probably one of the younger guys collecting old machines, and have hoarded a bunch of antique machines. I was going through some paperwork at my grandmothers house, and I found my grandfathers NC training certificate from the 60's. It makes me wonder with all the commotion about robots "taking over jobs" lately that a huge part of our history in regards to Numerical Control is sort of forgotten. My grandfather would be close to 100 years old, I wonder how many guys are out there that ran these machines?

About 5 years ago I visited a 1 man shop in Minneapolis MN and he has some Pratt & Wittney (think that what they were) and he had a teletype next to the machine he plotted and then typed his code to run into the huge computer. He said they still ran good. I'll have to see if he is still there when I get back home hopefully next week (stuck in snowstorm). I was think Sunstrand were one of the first ones to build them.

I ran a Cinc. retrofitted Hydrotel back in 1973. I had to lay out my moves on a graph paper and then type into the teletype. Bourn & Koch now owns Sunstrand and I will ask them if they know anyone who has them or if there is a museum. Rich

My first 'controlled' milling machine was a Moog Hydrapoint 1000 using an early Bridgeport varispeed as a starting point it had all hydraulic servo drive to all axis. (No lead screws, just Rams) Used a series of plates with holes through that slid against each other giving a vernier style accuracy. Target point was set up on a remote console as a 24 bit word on pneumatic switches. Console was connected to the machine via 1/8 tubes and the tube drove pneumatic to hydraulic actuators, hundreds of them !

It would also read or punch 8 hole paper tape, again pneumatically by blowing air through the holes!

Mr Moog of the NC machine was the brother of the chap who made the Moog Synthesiser

I did manage to interface the machine to a very early IBM PC shifting data out on one bit of the parallel port into a register which I arranged to drive a bunch of pneumatic actuators. Amazingly it worked!

The oldest N.C. machine I worked on was a " Richards " Hor bore a few years back that had been one of the first large N.C. machines ever built over here. The drawings were all dated 1963.

It had a tape reader as big as old fashioned juke box. The DRO reading heads were about 5" cubed ! The scales were made out of a stiff stainless steel ribbon graduated very much like the later glass scales that came out.
Each axis had it's own individual hydraulic clamping system. It was a point to point type control.
How it worked when it was new I don't know. My job was to remove all the old N.C. stuff and fit a new modern DRO set up so the machine could be run as a conventional machine.

Regards Tyrone.

I had a mid-sixties H-40 Cincinnati. Acramatic 220 control, point-to-point but you could do contouring if you were clever. It was 2 1/2 axis and very reliable. More reliable than my newer stuff, in fact. Big round linear ways, variable speed spindle drive, and nixie tube readout Kwik-Switch spindle and an interesting set of geared resolvers to tell you where the table was. You did not have to home it because it could not lose its position. The corner of the table was 0,0 und by gott, that was what you programmed from. Manual feedrate control with a potentiometer. Had an eight-position depth stop on Z axis. Mechanical tape reader. Four position rotary table.

Could be wrong but I sort of remember that it had rotary dial inputs, where you could enter a position then push "go" and the machine would go there. You could mdi run it, after a fashion. But that might be faulty brain cells

It was pretty nice to run, actually. I don't see much advance in controls over the past fifty years. Sure, new stuff is faster but it's all the same ideas.

Also ran an early Series II Bridgeport - but this was not the later one. I've never seen another like it. This had an air Spindle Wizard and was entirely nc with a Bridgeport control of some sort. Preset tooling. Frieden Flexowriter (those were worse to use than a Teletype.) Did its job but it was a Bridgeport, what can you say. Would have been about 1970. The Cincy was nicer.

Used to do timesharing. Both Sundstrand and Westinghouse offerred it. That was kind of a kick. Basically, you prepared a part model then sent that over the telephone, and the remote service sent back the program. Then the Trash-80 came out and a bunch of places offered software for doing your own geometry, so time-sharing dried up.

A lot of those old tools would still be faster than what people love now. Graphical interfaces kind of suck. I see people struggling like hell to make G76 work when the job would be a picnic with G33, but by god they are going to do it "the easy way" if it takes them all year ! Damned if I know. For a lot of work I'd take an Acc220 on an old horizontal in a heartbeat. Paying $10,000 to have someone bend me over for a new board on a modern control, well heck. I could have a pretty good time in Dallas on that money instead.

The first post shows what appears to be a Milwaukee mill.

if so, I did a layout where the customer removed (2) of those, and replaced them
with Mazak Integrex machines.

I don't recall if the old machines had run just before being removed.

As far as Parsons being the first, I always thought that as well, however I posted
awhile back on a find of a numerically controlled flame cutter, predating that.

it had drums with tick marks, one for each axis, and an electric eye to read those marks for speed and direction (not a more modern line tracer).

So the verdict is still out on this machine.

Don't forget pegboards.

gbent said:

Don't forget pegboards.

Click to expand...

I always thought a pegboard machine was much like a W&S auto, not really computor
controlling the axis's, more like commanding a turret to feed in, to a stop.

digger doug said:

As far as Parsons being the first, I always thought that as well, however I posted
awhile back on a find of a numerically controlled flame cutter, predating that.

it had drums with tick marks, one for each axis, and an electric eye to read those marks for speed and direction (not a more modern line tracer).

Click to expand...

That's not nc tho. With nc you enter a number and the machine goes there. That was the breakthrough.

SeaMoss said:

That's not nc tho. With nc you enter a number and the machine goes there. That was the breakthrough.

Click to expand...

You enter a number somewhere (a Friden or other typewriter), and it get's punched on tape or holerith cards No ?

No matter how you get it into the control, the breakthrough was that you tell the control to go to X12000 or similar, not put a line on a drum or set a relay stop or change pegs or whatever. That made all the difference.

Now anything you can describe with numbers is possible, basically no setup -- templates, stops, lines, cams, none of that. NC changed machining from a physical thing to an abstract thing.

That's pretty revolutionary, when you think of it.

SeaMoss said:

No matter how you get it into the control, the breakthrough was that you tell the control to go to X12000 or similar, not put a line on a drum or set a relay stop or change pegs or whatever. That made all the difference.

Now anything you can describe with numbers is possible, basically no setup -- templates, stops, lines, cams, none of that. It went from a physical thing to an abstract thing.

Click to expand...

And the torch I showed, was programmed in the office with a machine that took
the numbers, and converted them to the tick marks on the drums.

If you look at the pictures, the machine is cutting an arc.

Just like punching a number on a Friden, and then it punches the paper tape.

So the maker of the torch didn't have any of that available at that time,
(paper tape & teletype) so had to design it all from scratch,
and used drums with tick marks (the drums are covered with paper, and are
exchanged with others stored in the office) just like paper tape.

digger doug said:

And the torch I showed, was programmed in the office with a machine that took
the numbers, and converted them to the tick marks on the drums.

If you look at the pictures, the machine is cutting an arc.

Just like punching a number on a Friden, and then it punches the paper tape.

So the maker of the torch didn't have any of that available at that time,
(paper tape & teletype) so had to design it all from scratch,
and used drums with tick marks (the drums are covered with paper, and are
exchanged with others stored in the office) just like paper tape.

Click to expand...

Not really. The tape carries the numbers themselves. The tick marks are placed at a distance on a drum, just like lofting a frame for a wooden boat. If a tape had tick marks at a certain distance determined by numbers they would be comparable, but that isn't how it worked. This flame thing is a tracer with an additional device that can convert numbers into a template on a drum. An NC control takes the numbers themselves and uses them.

You could possibly say that the thing that made the drums was an nc machine, but it wasn't a very useful one

Oops, read the page more carefully. No, they are not numbers converted directly to tick marks. They are "laid out by a stylus or a scribing machine." This is more like those electric eye things that follow a line, except that instead of a line it's x and y locations drawn on a drum. They still lay out the locations by hand first. There's no abstraction happening. This is an electric pantograph or automated lofting machine.

Pretty cool but it's not nc.

Here’s an old K&T video that was post a while back.
The Numerically Controlled Machining Center 195s Educational Documentary WDTVLIVE42 - The Best Docu - YouTube

SeaMoss said:

No matter how you get it into the control, the breakthrough was that you tell the control to go to X12000 or similar, not put a line on a drum or set a relay stop or change pegs or whatever. That made all the difference.

Now anything you can describe with numbers is possible, basically no setup -- templates, stops, lines, cams, none of that. NC changed machining from a physical thing to an abstract thing.

That's pretty revolutionary, when you think of it.

Click to expand...

I think we can extend your definition to any configuration that uses a binary representation of the numbers.

Previous to this everything was analog in nature, gears and cams. Even the point to point controls using rods were analog.

Why I bring this up is that there was quite an electronic revolution going on that didn't change the foundational mathematics but definitely changed the technology in constructing the control. Relays to RTL to Logic Gates to Microprocessors and the memory going from some pretty weird schemes to Core memory to Ram.

Bell Labs developing the transistor was the real game changer in being able to actually construct a control that was practical.

Here's a picture of a Sundstrand Tape Checker that was in a tool box that I picked up years ago.
I imagine that this was used in conjunction with the older machines/technology mentioned here.
Back in the late 60's when we needed to "customize" a part on a car we were building we would go to a friend's father's shop where he had a few machines that used similar technology.

The OP asked about early NC machines in museums and shows. Here’s one on display at the National Museum of Scotland in Edinburgh. It’s a Ferranti/Kearney & Trecker-CVA NC milling machine.

The museum give a date of c.1960, but Ferranti were making NC machines in the late 1950s. An article in The Engineer in 1959 described their latest machines, including a milling machine which was up to Mk IV. Article link below. You need to sign up to view, but it’s free and safe.

https://www.gracesguide.co.uk/Special:MemberUsers?file=images/4/40/Er19590828.pdf

NC machines at that time were straddling the gap between simple olde world mechanics and new-fangled black boxes, so even I can understand some of the workings. The article mentions the use of stainless steel strip reflection gratings rather than glass, with the advantage of more compatible thermal expansion.

The milling machine’s electronics was transistor-based. Input was either by magnetic tape or punched paper tape, with the facility for direct positioning by dial. It's mentioned that with the earlier Mk III version the magnetic tape ran at 15” per second, needing 3600 ft of tape for 20 minutes runtime! With the Mk IV version they managed to get more pulses per second on the tape, and could get 190 mins from 3600 ft of tape. As long as it didn't break.

The article also described another, simpler machine which used Ferranti magnetic tape control for two-dimensional profiling to produce cams and templates. It drove two axes of a ‘Contourmatic’ milling machine made by High Precision Engineering Ltd. The milling head was driven by a 3 HP hydraulic motor.

Also described was a co-ordinate measuring machine. It only resolved to 0.001” (although there were lights to show whether it was 0.0005” over or under the displayed number). No mention of how the measurement probe worked.

One way to abstract the question is this:

An "NC" machine depends on interpreting commands from an input medium, such that the size or form of the medium wouldn't change the program. It implies the machine ingests a code, and then interprets that, rather than "homing to" an analog value.

So a paper tape machine could in general be driven by mag tape, or by a computer emulating the tape reader and pretending to be a tape (that was surely done.)

This does suggest that Jacquard Looms might count, since they seem to have been long binary command words, each hole/non-hole pulls/does-not-pull a "shed" - and the mechanism could provide the same behavoir from any input mechanism that could present these long control words.

I've got some basic literature on the old control system I'm mentioned earlier, I'll see if I can dig it out. That " juke box " in the photo is very similar to the one I'm referring to.

Regards Tyrone.

Fanuc produced a "Tape Checker" - still have one tucked away somewhere. It was to ensure that tape "reperforators" (or punches to you and me) punched tape inspec. It was just a flat bit of plastic with a grid etched into it to ensure that the holes ended up where they should. I used to carry one in my tool kit for setting up Fanuc 4070 tape reader / punches back in the early 1970's