First true NC milling machine ??

Several post on NC and paper tape machines leave me wondering. I want to ask the experts here. Can anyone say definately, what was the first NC metal working machine, or specifically a milling machine in some practical size with a real computer, an electronic computer, reading a program from some medium- card, tape or whatever. And, let me qualify futher by asking that this be a product available to public not a one-off machine. I would also think to be the first true nc machine requires position sensors, not cutting blind and ability to do complex shapes. I mean a real deal working, sucessful NC machine- who was first?

Can an example of that specific model be seen anywhere?

I dont know about computer controled machines, but punch card looms for weaving "orental type" carpets were developed centuarys ago. That the root of the technology.

The first "true" NC machine was a converted Cincinnati Hydrotel mill at the Servomechanisms laboratory at MIT in 1955.

The machine read commands in a numerical format from punched paper tape.

Depending on your drfinition of computer, this machine was or was not computer controlled. The control was electronic but could not store a program within it. But in all senses of the word the control was an analog computing device.

The information is coded in a binary coded decimal format, It was not converted into 1's and 0's but used as is by the controllers counters and registers.

The tecnical name for the NC controller was "director".

A machine controller is CNC if it has a stored program digital computer (CPU) within it. The information is processed digitally but in some cases converted to analog signals for the servo drivers.

Early NC machines used various commercial scientific or data processing computers as a "front end" when programming shapes that needed contouring, the simultaneous movement of two or more axes.

In the case of contouring NC controls, the machine had somewhat of a memory so that it could read the tape by "looking ahead" for all the axes infromation while the machine was in motion. That memory was called the buffer.

Most point-to-point work was programmed manually on a tape punchiing Flexowriter typewriter. Computer time in those days was frightfully expensive.

What we call "G code" was a word address format, developed so that the tapes could be punched by the simplest methods.

Each of the letters, N,G,M,F,S,T,X,Y,Z etc. is a "word" and it is used to address the corresponding registers in the controller. G code remains the internal machine language of the CNC's today.

A photo of the MIT machine plus more infomation on John Parsons..the "Father of NC" was included in a thread I started in the CNC forum about 3 years ago.

www.practicalmachinist.com/ubb/ultimatebb.php?ubb=get_topic;f=13;t=000008#000000

Three years ago is an eternity on the net.

I sat down and wrote my post right off the top of my little baldy head.

Yours truly

Jimmy K, the machine Jedi

Jim, if you didn't know it already, you'd be interested in that reportedly the first person to bring NC "down to earth" for use by the common job shop was a guy in the printing press business...Adrian Holmberg. He worked for Goss-Dexter in Chicago.

Paraphrased from the book Metalworking-Yesterday and Tomorrow below-

By 1955 he was already using 22 pieces of numerically controlled equipment and had 20,000 tapes made ready for active parts.

Because the printing press business consists mostly of small runs, NC was ideal for his operations.

He bought two NC positioning tables at the '55 IMTS Chicago show and paid $10,000 each, but he found the savings so great he was soon urging most machine tool builders to make NC versions of their machines.

Holmberg was open in talking about his accomplishments and wrote a few articles in American Machinist magazine back then. In one he said "It's out of this world what you can save. In the tooling of three new products in our various divisions we have saved aprox. $450,000 just on drill jigs and fixtures !"

($450,000 in the 1950's !!!!)

Others began to get the message. Relatively low cost modifications of radial and upright drills with NC positioning tables provided the first NC experiences for many plants and proved the practicality for small lots.

No, Don, I didn't know about Adrian Holmberg.

Goss built newspaper presses and that is out of my line. However, all printing presses consist of flat side frames with the cylinder bearing bores in them. Many of the frames also have the major ink roller bearing bores in them also.

If anything would be speeded up by simple point to point NC, that job would be.

The Goss Community web offset newspaper press is a small press and it is one of their most successful products. The little weeklies that are in about every town are printed on these presses.

All of the prats are small, I can pick up a side frame, although my show-off days are over.

The frames are planed on one side and then off to the boring operations. A single "press" might have five printing units or more, so think of how many side frames needed to be worked. And, oh yes, each press had a folder with two side frames and each unit has it's own roll stand.

Just a simple thing like a positioning table would constitute a revolution in that shop.

The NC lathe didn't fare so well in the printing press industry. There is so much straight turning involved with cylinders and roller shafts that the NC controls couldn't justify their expense and the troubles that the early controllers caused.

For the bearing boxes and other short turned pieces, the common turret lathe was difficult to supplant.

It is the saddest irony that all the wonders of the modern CNC's couldn't save our domestic press manufacutrers. One by one they came apart due to union labor problems, corporate wrangling and their failure to keep ahead in mechanical design.

I believe that you are in possesion of "When the Machine Stopped", the story of Burgmaster. In that book, they point out that Burgmaster put the NC machining center in the ordinary jobbing machine shop. It is said that the comapany gave the machining center it's name.

As a side topic, the idea of having part programs stored on individual tapes still appeals to me. In the printing press industry, parts don't change much if at all. The semi finished castings and the tape to run the job would come to the machine at the same time. No screw ups, no genius getting into the program and messing it up.

When the job is done, just wind the tape up and put it back in the can with the part number marked on it, ready for the next time.

Jeez! Even I could run a machine like that!

Hey, wow!

My new computer will do the icons.

"G" code = Gerber code.

Gerber files are still used in the printed
circuit board industry. They go way back.

Jim

Jim:

Did they spoon feed the pupils at the training courses?

Were the canned cycles really in little bottles with metal caps?

I am sure the course material was easy to digest.

And yes, they have been around for generations.

When I worked at Westinghouse Steam Turbine Division starting in 1978 we had a Moline NC drillpress that was 42 rail mounted spindles, IIRC. They brought an identical Moline from Lester PA after Marley bought the division in 1980 or so.

I think the Molines were built in the 60's and were used to drill condenser end plates and tube supports. Thousands of 1-1/8" holes in a very precise pattern. The operator sat and read the newspaper and just kept an eye out for a yellow light indicating a dulling drill. He was to avoid a red light which indicated a broken drill.

We also had a gundrill that was NC that was used to do the plates for feedwater heaters. I can't remember the name of the gundrill even though we referred to it by the name of the maker.

That was my first exposure to NC machines.

Les

And, let me qualify futher by asking that this be a product available to public not a one-off machine.

Click to expand...

There actually was an NC lathe before the MIT machine. Made by Arma Corp and manufactured in 1950, it used a punched piano roll to provide digital data for step turning. I've got a picture of it and it appears to be a standard South Bend with a motor on the cross slide. Probably only one or two made though.

The Milwaukee-Matic, introduced in 1958, was the first commercially sucessful NC machine however. It was also the first sucessful "mutifunction machine" or machining center, with tool changer.

The 20 page brochure is interesting in many ways, one of which is that it's bascially teaching the prospective buyer about NC and the potential fixture and cost savings.

Cost with optional pallet shuttle system was aprox $140,000. There is a depreciation chart in the brochure showing how funds become available to "pay off" the machine in "less than 28 months"

Note in the last catalog scan, dated 1958, the control is so early that it doesn't even have readouts of any sort. Just some buttons and tape readers.

Not sure what year the smaller Milwaukee-Matic model 100 that Ries showed earlier...looks mid 1960's in design. The "Matic" below predated that one by 7 years I suspect.







Spindle drive was via hydraulic controlled motor with speeds of 100 to 4,000 RPM. Hydraulic servo motor for axis drives. "standard multimeter and oscilloscope are only instruments required for maintenance and testing"

[ 10-28-2005, 01:29 AM: Message edited by: D. Thomas ]

Don, et.al.:

Notice in the second pic, the rails and pallet set up. There was a pallet changer option with that machine, it was the most popular configuration. I think the one with the table alone would be a rare specimen.

Nothing new under the sun.

Wow
I ran KT 200 and 800 back in the day, amazing the early ones they put the magazine where the operator could get right at it, ours it was way round the other side of the machine. The first one we had still had the tool id rings on the straight shank tool holders, there was a reader with pins in it that "felt" which rings were on a tool. That machine eventually got a new control.

Our KT's used F0 for rapid instead of G00

Bill

Regarding the lack of readout on the Milwaukee-Matic, this gets me to wondering what was the very first DRO ? And was it in seperate "add on" DRO form to replace those "optic scale systems", or was it integral to an NC control ?

Re early electronic *scales*, in the "Matic" brochure, there is a diagram that mentions "fine positioning linear scale, electrical resolution .0001 inch", "electronic position monitor", and "coarse positioning rotary transducers" but that's all the details they give.

But what was the first electronic numerical display on a machine tool ?

I don't know the answer to this, but by looking thru old magazine ads I might can find out.

Anyone have any idea ?

Don:

Some of the (maybe) second generation controls like the GE Maek Century has a numerical position read out. I think that some were Nixie tubes. Most of the controls had one position display, you had to select the axis that you wanted to show.

This set up was certainly not like the modern DRO's, it did, however, verify the machine's position. You used them when you dialed in an offset because they didn't show the command, they showed where the machine actually was.

The early machine manufacturers stayed away from linear scales for the most part, probably because of the dirt and scuzz factor. The Kearfott rotary resolver was the state of the art then,

Most makers put the resolver right on the end of the screws. The screws were as good as they could make them, the resolver system was not usually compensated for lead error.

This was in the days before integrated circuits All the TTL (transistor transistor logic) was on little circuit cards. A read out feature was terribly expensive by today's standards, in fact the controllers usually cost more than the iron in the machine.

This was in the days before integrated circuits All the TTL (transistor transistor logic) was on little circuit cards.

Click to expand...

Those old controls used some vacuum tubes too. I briefly owned an NC Burgmaster that was from the 1960's with Hughes control, and even as "modern" as the mid 60's this control had some tubes in it !

Man, I could sure kick myself for never even taking a picture of that thing. I traded it for some parts I subbed out on the tapping machines I used to manufacture...maybe the guy still has it...hmmmm....

Don:

I belive that the tubes you refer to were in the power supplies or other high voltage parts of the control. Tube type flip flops were out in the mid fifties.

Tubes defied the mean time befor failure calculations, they actually ran pretty well. It was the heat and current draw that could not be tolerated.

Simplicity was the order of the day. Pratt & Whitney made their own controllers, their boxes used only three different types of circuit cards.

I had watched IBM make their main frames in Endicott and I watched P&W making their controllers. Both companies used the little circuit cards (IBM called them SMS cards) and the internal wiring in the main frame was all connected from one point to another by a robotic wire wrap machine.

IBM went to integrated circuits with their System 360 machines. Even the integrated curcuits were somewhat discrete, Not too many functions were on a chip and the chips were still mounted on similar cards to the SMS cards. VLSI (Very Large Scale Integration) was still a long way off.

GE and P&W and the others were not in a position to go over to the new technology after spending large amounts of capital getting to where they were then. There was also the attitude that computing was done in the conputer room and not out on the shop floor.

One reason for that was that every computer of every manufacturer used it's own programming, unique to it. APT was a standard program but it had to have a compiler for every different computer that it was run on. APT also required the big main frames to run.

AD/APT came out and would run on the smaller "scientific" computers that were coming out. IBM's small 1170 series could run it.

We all complain about Bill Gates - The licensing and the computer time to run APT was horrendously expensive. Only the Big Boys could afford it. Bill's carrying on pales in comaprison.

We also complain about the lower priced imports that we think did damage to out machine tool industry. The ventures into the world of electronic controls did very much damage to many machine tool manufacturers. I am sure that NC hastened several of them into insolvency.

They got in way over their heads.

As I know of the first digital display for a machine tool was an add on to the Mark Century 100S. It was a separate box that used nixie tubes, Jim is correct. The Kearfott resolvers or Singer Kearfott resolvers were petty advance for there day and small. The were smaller than a C battery. Linear scales were around too, not optical they were magnetic, they were call "precision pin" Looked kind a like a hair comb but the pins had .001 spacing and the pins were .125 as I can recall, VERY expensive stuff. My dad serviced this stuff back in the day, even went to K&T school for the EB. I in the last few years thru away loads of parts and books for these machines, I had new resolvers, tape readers, precision pin sets, sets of books. The books could fill a draw of a file cabinet.

The name Milwaukee Matic became a general term, its just a NC/CNC K&T. The machine with the sales cat is the original called Milwaukee Matic and was in sizes #1, #2 ect. The machine with the tools on top is an "EB" later became a Gemini.

The EBs were hydraulic movements, the NC control drove a "servo valve" that could move an exact amount of oil, the valves were made by Moog...

The interesting thing is the tool changer. The story is fuzzy in my mind, who stole the design and wound up in a big mess?

Tom 1:

A man named Brainard, possibly Wallace E. invented the swing arm tool changer while at Hughs Tool Company. He then went to K&T and that company started using the design for their machines.

There was a Bru ha ha between the two outfits. How it got resolved, I am not sure.

The blow up is covered in the book "When The Machine Stopped" which is actually the story of Burgmaster.

I don't have my copy anymore.

Don? If you still have your copy can you tell us how the thing was resolved. That design is standard for almost everybody's tool changer now.