Printer Friendly

Anatomy of a machine tool.

Part 2: Controls, servos, and other electromechanical mysteries

Once you've made the major decisions about the mechanical aspects of the machine tool you're going to purchase (see the February 1993 issue of Tooling & Production), the next set of questions you have to answer concerns the selection of a control package. This will normally consist of a CNC, a set of servo drives and associated electronics, perhaps a programmable logic controller, and some sort of interface to make them all work together.

Depending on exactly which machine you've settled on, the choice may be as simple as accepting the builder's proprietary control package--about which you have NO option; or as complex as putting a package together yourself from a menu of available components. The first choice, while very common today, promises to be less so in the future, while the second shows signs of becoming much more widely available.

Here again, there is a strong, and understandable, tendency on the part of machine tool suppliers to sell you what they have, even though it may or may not really fit with what you need. In the case of proprietary controls, it's often a "take it or leave it" proposition. But never lose sight of the fact that it's your money being spent, and, for good or ill, you're going to live with your decision for a long, long time.

Bits, bytes, & bulldroppings

Before getting into the details, it may be useful to take a quick look at what goes on inside a computer. You've probably heard of "32-bit" controls. They're the newest thing on the market and are quickly replacing "16-bit" controls as the de facto standard.

To understand the difference, you have to know what a "bit" is, and why it's important to a computer. Once you clear away all the mystery, a computer is nothing more than a huge bank of extremely fast on/off switches, and a "bit" is simply the status of one of those switches.

It's either "on" or it's "off." To the computer, "on" is "1" and "off" is "0." So the computer "thinks" in the Binary number system, in which the "bit" is the basic unit of information.

In the very early days, processors handled data four bits at a time and were called "4-bit" machines. The next step was to double the capacity to eight, then again to 16, and then, more recently, to 32. Theoretically, the next step will be to 64, with a possible intermediate stop at 48.

So, if you think of the data as a long line of 0's and 1's, the 4-bit processor handles four of them at a time, and the 32-bit processor handles 32 of them at a time. As you might expect, you'll get to the end of the line sooner with the 32-bit processor.

In most applications, the 32-bit processor also handles the data faster, because it works at a higher "clock rate." Essentially, the "clock" tells the processor how often to grab a new set of bits.

In the Intel family of chips used in IBM PCs and their clones, the original "clock rate" was around 5 MHz in the first PCs. Some of today's chips operate at nearly 70 MHz. So, not only do they process more data at a time, they do it almost 14 times faster.

The net result is that the newer processors can handle more data, faster. In the computer on your desk, that may mean your spreadsheet recalculates quicker. In your CNC, it may mean you can perform more sophisticated machining operations or simply run faster to produce more parts in less time.

Theoretically, there's not much a 32-bit processor can do that a 16-bit can't. The 32-bit machine just does it faster. In practice, however, that speed edge can make a difference by allowing the processor to complete complex tasks quickly enough to be useful in a machining operation--if the rest of your system can keep up with the control.

Don't lose sight of the fact that a control that processes data and issues commands faster than your machine can execute them isn't doing you any good at all. It's highly probable, in fact, that such an unbalanced system will actually be less productive than one in which all the components are complementary.

So, if you have a 16-bit control, and it's doing the job for you, it's probably not worth your while to upgrade to a 32-bit model unless you want to replace your servos, ballscrews, spindle, and much of your tooling at the same time. If you're buying a new control, and you have a choice, try both versions on your parts, and make the most cost-effective decision.

Specialization is the key

The CNC is nothing more than a microcomputer which has been optimized for the kind of computational functions required to run a machine tool. It's a really fast number cruncher with sophisticated, but highly specialized, input/output (I/O) capabilities.

Yes, the control your salesman recommends probably has special chip sets, esoteric busses, and the latest 32-bit processing technology along with EIA standard axis coding, RAM, ROM, RS-232, and a bunch of other things you probably don't understand, and really don't need to. The fact is, virtually any control you can buy today has the technical capability to run your machine tool, unless your machine tool is very unusual, indeed.

Things like jog, cutter compensation, automatic tool offsets, and various interpolation capabilities are all essentially standard features of virtually every control on the market today. So are the ability to accept either absolute or incremental positioning commands. CNC reliability and serviceability are two other non-issues about which much is said which doesn't need to be, a holdover from the days when such things really were matters of concern.

The market is sufficiently competitive, and the technology sufficiently accessible, that nobody keeps a technological advantage for very long. Remember, regardless of what the control builders would like you to believe, we're talking about something just slightly more sophisticated than a personal computer in a shop-proof box. In fact, with the right software and a few I/O boards, the PC on your desk probably can run your machine tool.

The real difference between CNCs and PCs is not power or sophistication, it's specialization. The CNC does two things very well:

1. It makes it easy to translate dimensional information into cutter paths required to machine a part, and

2. It is very good at managing the activities of servos and other electromechanical devices.

That's it, just two things. Everything else either makes those two tasks more efficient, or it's window dressing. That narrows the real selection criteria down to just two questions.

How do I tell it what I want it to do?

Once upon a time, numerically controlled machine tools got their instructions from paper tape with coded holes punched in it. That's not very common anymore, but the tradition lingers in the practice of defining program memory in a CNC in terms of feet of tape.

Here's a useful bit of information you can use to get even with the next person who attempts to bury you in computerese. The original punched tape carried 10 characters per inch, or 120 characters per foot. So, 1 kb (1 Kilobyte or 1024 bytes) of program memory is equivalent to just slightly over 8.5 feet of tape, or 1 mb (1 Megabyte or 1,024,000 bytes) is just over 8,500 feet.

Today, of course, things are different. Most programs are created off-line and loaded into the CNC on a floppy disk or directly through a network. Make sure your CAD system can talk to the CNC you're considering, but beyond that there isn't much to be concerned about.

Don't buy MDI for the wrong reasons

The alternative to off-line programming is Manual Data Input (MDI). It's something you want a control to be able to do, but not for the reasons that are usually given. Unless you're careful, MDI can be a real two-edged sword.

The lure of MDI is that it all sounds very simple. You just stand in front of the control with a print in your hands and punch in your program. The ability to directly translate your machinist's skills into a workable program can be very comforting, especially during a transition from manual to CNC operation.

But if you intend to stay in business, you won't be transitioning for very long. That's why this particular benefit of MDI is not worth paying a great deal of money for, no matter how hard the salesman pushes it.

What is worth paying for is MDI's ability to quickly make simple parts, prototypes, or small changes or corrections to programs written off-line. These are the right reasons for insisting on an MDI capability in your control.

Human nature being what it is, the big danger of MDI is the fact that while you're punching in data, the machine tool is most likely sitting there patiently waiting to do something productive. Even if the control is able to cut one part while you're punching in data for another, in practice the odds are good that it won't be.

In that case, the more complicated the program is, the more money you lose writing it. That is not a very good bargain.

Don't let the apparent simplicity of MDI blind you to the potential cost. The whole point of buying a CNC machine is to improve your productivity, and an excessive reliance on MDI works against that goal.

Will it remember in the morning?

Program storage, on the other hand, is something you should be concerned with. Within limits, more is definitely better. And as memory costs continue to decline, more isn't all that expensive either.

Program memory comes in two flavors, volatile and non-volatile. The difference is quite simply that volatile storage is lost anytime the power goes off while non-volatile isn't. Non-volatile memory is worth the money in terms of the peace of mind it offers.

Chances are the control you're looking at will offer a floppy disk drive for program input and perhaps an internal hard disk for storage. Both are excellent ideas which offer a lot in terms of flexibility and reliability. Network compatibility also falls into this category. You may not have a network today, but chances are you will during the life of the machine tool.

What you see is what you get

You'll also want to look at the display. Your choices are screen size and black & white or color. In screens, the current standard is 9", but bigger is better if the cost isn't excessive. You'll not only be able to display more information, you'll also be able to see it better, which means fewer mistakes.

The question of color isn't purely aesthetic. Color displays are much better at showing part geometry and cutter paths than monochrome displays. They're also more expensive.

If possible, look at the same part on both before you make a decision. And remember, as hardware and software become more and more capable, you're going to see a lot more detail on the screen in the future.

It's not what you say, it's how you say it

There's another reason to consider a big, color display. In the future, CNCs will almost certainly follow the current PC trend toward mouse-driven, icon-based interfaces, the infamous Graphical User Interface (GUI) pioneered by Apple and cloned as Microsoft Windows. AGUI simply works better on a big color screen.

Today's norm is the menu-driven text-based interface where you select the operation you want to perform from a list of choices displayed on the screen. There's nothing wrong with this approach, but you can count on the Windows steamroller to make it obsolete within a few years. Take a look at the options available, and pick the one you're most comfortable with.

Why not just buy the builder's control?

There are a number of machine tool builders who don't give you a choice of controls. You either buy their "captive" unit or go somewhere else for your machine tool. Like everything else, this position has its pros and cons.

On the plus side, the builder knows his machine better than anyone else, so you can assume his control is as perfectly tailored to his machine as humanly possible. You can also assume that any special functions needed to make the machine work most productively will be supplied. Chances are the "captive" control will do what needs to be done.

On the other hand, the average builder sells only a few thousand machines a year at most. He has to amortize the cost of control development and production over that number of machines. As a result, you're less likely to see cutting edge technology in a "captive" control than in one that must compete in the open market where sales volumes can be much, much higher.

What's the answer? If you simply must have that particular builder's machine, there's no choice. Or if the "captive" does everything you need, and the price is right, there's no reason not to buy it. Only in cases where the "captive" can't do something you need done, and there's a good alternative machine, does it make sense to base your purchase decision on the control alone.

If it feels good, buy it

The whole topic of user interfaces, which is at the heart of what we've been examining to this point, is a very subjective one. If you're already using several controls from one builder, it makes sense to specify the same controls on your new machines. After all, you've already made the investment in training. There really is no point in re-inventing the wheel.

If this is your first purchase, then the control that is easiest for you to use is probably the one that's best for you. If you're buying a reputable brand name control with the capabilities you need, it's pretty difficult to make a bad choice.

One word of caution is in order. Today's controls can do some really remarkable things, but unless you can use these advanced features to improve your productivity, there's no point in buying them. Be careful to stay focused on what's really important to you and the parts you make.

And just around the corner...

Like all businesses related to computer technology, the machine tool control industry is changing rapidly. One of the most important trends is the narrowing of the gap between CNCs and general purpose PCs.

Right now, the major difference is speed, and here the purpose-built CNC wins out. The difference between updating a servo in 12 milliseconds or 8 milliseconds is significant. Four milliseconds would be even better, and two better yet. All else being equal, faster response means more precision, less scrap, and greater productivity, all of which show up on the bottom line.

The fact is that you can now buy a PC in a shop-proof box with software that will run a machine tool. But those systems won't give you the speed of a CNC--yet. That will certainly change.

We may be witnessing the first stages of a major change in the control industry, in which today's hardware vendors will focus more and more on the software aspects of the business. Already the industry buzzword is "open systems," where off-the-shelf hardware and software can be combined in a virtually unlimited number of configurations.

The situation is very similar to what happened in the personal computer industry during the 80's, where hundreds of "clone" producers combined standardized components to produce custom products. If that experience is any guide, however, the promise of "open" CNCs will not be fully realized in the immediate future. Still, it's clearly the direction in which the industry must move, if only because it's the direction firmly established for the underlying technology.

Like all computer-based technologies, machine tool controls will become increasingly software dependent. The confluence of more generic CNCs with extremely flexible "software configurable" servos will make software the real variable in machine tool performance.

Today, the standards required to make that feasible simply don't exist. But they will in the future, either through industry cooperation or as a fiat accompli by one of the major players or groups. You can count on it.

A short course in servo technology

A servo is basically an electric motor teamed with a device which reports back on what the motor has done and a controller to turn the motor off and on. The motor is usually coupled to an axis (although servo-type spindle drives are also used), the position sensor may be attached to the motor or the table, slide, or other driven device, and the controller is either a separate "box" or an integral part of the CNC.

Servo motors are designed to minimize rotor inertia and maximize torque. The first examples were DC powered and suffered from relatively short brush life, which gave them a reputation for maintenance problems in many machine tool applications.

This situation has changed with continued improvements in brush materials, and today's DC servos, while still not as maintenance-free as AC units, are, nevertheless, an excellent choice for many machine tool applications. The major advantage of DC servos is a substantial cost savings compared to their AC counterparts.

AC servos don't use brushes, so in that regard they're more durable and reliable than DC types. On the other hand, they require considerably more complex electronics to perform effectively. Given the reliability of today's solid state devices, this is not a negative factor, but it does make AC systems more expensive than DC systems of similar capacity.

Give me some feedback

The thing that makes a servo system different from a simple motor drive is the fact that the position of the driven device is constantly being monitored. It is the combination of servo technology and microprocessors that makes today's CNC machine tools possible. The microprocessor does the "thinking" and makes the necessary decisions. The servo carries out the decisions and verifies the result. There are several ways to do this.

First, you can attach a sensor to the driven device itself. In this category fall optical and magnetic linear and rotary scales and some sophisticated laser-based devices.

The more common approach is to attach the sensor to the servo motor's shaft and simply count the rotations. In this category fall optical and magnetic encoders and resolvers.

Servo builders prefer the latter approach because it gives them complete control over the marriage of motor and monitoring device. It also improves their bottom line, because they get to sell you both pieces of the system.

Scale builders point out that because their devices are attached to the moving machine elements, they measure the real motions produced by the motor, not just shaft rotations, and are, therefore, potentially more accurate. The reducto ad absurdum for this position is the case of a broken ballscrew, where the shaft-mounted servo could report motion even though the mechanical system was totally disabled.

In practical terms, if you stay away from absurd situations, you'll get perfectly acceptable performance from either method. Unless you're retrofitting an existing machine tool, you probably won't have a choice anyway.

Encoder vs resolver/absolute vs incremental

The difference between encoders and resolvers lies in the way the device senses motion. A resolver acts like a rotary transformer producing an output voltage based on a reference voltage as modified by the rotation of its elements. Encoders count the number of times a signal is interrupted by the motion.

They are roughly comparable in terms of precision and reliability, but resolvers are slightly easier to handle from an electronic point of view. All else being equal, either will do the job for you.

A difference that may be of concern, however, is the difference between absolute and incremental encoders. In very basic terms, an absolute encoder attached to a servo motor's shaft counts the number of rotations since the system was last zeroed, while an incremental type counts the number of rotations since the system was last sampled. The absolute encoder stores this data in some form of non-volatile memory, while the incremental encoder typically updates the status of a volatile memory location in the control.

If power is lost, the absolute encoder permits you to begin operation without re-zeroing the axis. Theoretically, you could shut an absolute encoder equipped machine tool down on Friday and start it again Monday with no downtime to re-zero and no loss of accuracy.

As a theory that may be true, but very few people of my acquaintance would try it in practice.

In an intermittent power failure, however, the absolute encoder could possibly offer a time savings. Whether or not that justifies the increased cost and complexity of absolute encoder installations depends on your operating circumstances.

Putting it all together

Servos are general purpose devices, and their exact response characteristics have to be "tuned" to match the mechanical and electronic characteristics of the system they're used in. Within limits, the more "tunable" a servo is, the better your machine will perform.

Think of it this way: The machine tool is really three basic sub-systems which must work together efficiently. There is the mechanical system of ballscrews, ways, spindle drives, etc; the control system inside the CNC; and the servo system which links the other two together.

For purposes of example, lets look at a hypothetical machine tool that is mechanically capable of feeding a cutting tool at 1"/min at a given speed and depth of cut. To accomplish that motion in that time, assume that the CNC has to perform 100 calculations and the servo system has to receive 50 inputs and generate 50 outputs.

Now, if you were to replace the CNC with one capable of performing 150 calculations per minute, would you be able to get more out of the machine tool? No. The servo system still could handle only 50 inputs and generate 50 outputs in a minute, and the mechanical system could still only handle the stresses of a 1"/min cut. You would have to upgrade all three systems to realize a productivity gain.

But in reality things are never that clear cut. Mechanical capabilities vary widely, and recent advances in cutting tool geometries can squeeze extra capacity out of previously limited machines. CNCs are available in a broad range of speeds and computational horsepower. And connecting the mechanism and the CNC, as noted above, is the servo system, which is a general purpose device with a range of response capabilities.

If in this example the servo could handle a range from 25 input or output signals per minute to 75, you could then "tune" its response to match the capabilities of both CNCs by simply speeding it up or slowing it down. Yes, you'd still be limited by the ultimate capabilities of the mechanical system and the tooling. But, in principle, you should be able to optimize the interaction of the three sub-systems to squeeze the most productivity out of the machine. That, in an extremely simplified way, is what "tuning" is all about.

Until recently, "tuning" a servo meant adjusting a set of potentiometers in the control circuitry to change the system's response characteristics, an operation that required specialized skills and, often, a great deal of time. Those factors are the major reason behind the recent proliferation of so-called "software configurable" servos.

In these systems, the parameters that used to require physical adjustments can simply be specified in software. Since the adjustments are no longer electromechanical, there is no mechanical or electronic "drift" in the system. That translates into better accuracy and repeatability, but at the price of somewhat higher initial cost.

All in all, it's probably a good investment.
COPYRIGHT 1993 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:includes related article; part 2
Author:Morningstar, David
Publication:Tooling & Production
Date:Jun 1, 1993
Previous Article:US equipment helps pack pills.
Next Article:Dial in for high production machining, accuracy.

Related Articles
NC software - how it stacks up.
CAD/CAM drives NC software.
CAM for moldmaking.
Faster machining.
Medical manufacturing shop gets 'hip' to better production: the Irish subsidiary of a Kalamazoo-based company improved throughput and reduced...
From bar to box in half the time: turning centers cut diesel part production time in half.
The keys to calibration: machine tool 3D volumetric calibration replaces 1D linear calibration.
Getting by with a little help from friends: when faced with an increase in demand, a machine-tool manufacturer bought another brand to fill a need...
Get behind the wheel: efficient tool research makes the job easier, faster.
Four-spindle technology drives lathe including pallet changer duties.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters