Vendor-Tech

Operational Excellence with Technology

Bar Code Basics

Bar Code Basics

"What is black and white and read all over?" An old riddle from my grade school days that couldn't be more appropriate today. While the answer back then was a newspaper, today it could just as easily be the subject of this series of articles--barcodes.

There are literally billions of barcodes being used every day. Just go to the local supermarket or drug store and virtually every product on the shelf has a bar code on it. When you think of how many boxes of corn flakes produced every year, you can see how completely integrated barcoding has become in society. And once you have been through a scanner equipped check-out counter then gone to a non-scanner equipped "convenience" store, you will also quickly realize why barcoding is so important.

Why barcode? The two biggest reasons are speed and accuracy. When compared to a reasonably good typist entering data, barcode data entry is approximately 5 to 20 times faster and anywhere from 50 to 1 million times more accurate. Take a factory blue collar worker or minimum wage check-out clerk and the speed improvement is even greater. Accuracy improvements depend more on the bar code implementation than on the capabilities of the person using them.

Bar coding has long since moved beyond the realm of the check out counter. Industries as diverse as General Motors or your local blood bank are using bar coding for rapid, accurate data entry for applications such as just in time manufacturing control, to inventory management, to sales order entry. And in the video rental industry the difference between profit and loss is often whether the store is using computers with bar coding.

Not All Bar Codes Are Created Equal

Unfortunately, as bar codes were being developed and integrated by different industries, several different bar codes developed. In fact there are now over 75 different bar code symbologies being used today. Each has different applications, strengths, and weaknesses. Thus an important step in any barcode application is the proper selection of your bar code symbology and then picking equipment compatible with that symbology. It's a lot like picking your software first and then finding hardware to run it on. You wouldn't want to go out and get a DataFlex based system only to find out it doesn't run on the IBM System 36 you bought for $50,000 last week.

Fortunately, you don't have to look at all 75 symbologies before picking the one right for you. Over the years, a handful have evolved into the dominant symbologies. And selecting the appropriate one from the list is usually determined as much by the kind of information you want to record as anything else (Although your choice of bar code symbology is often driven by existing barcodes provided by your vendors). We'll look at seven of the most common symbologies in use today. Figure 1 shows the same piece of data in each of the seven.

Two terms that need to be introduced before you evaluate bar code symbologies are First Read Rate (FRR) and Substitution Error Rate (SER). Along with density considerations, which determines the size necessary to encode the information you want, and the symbologies code set, which determines what information can be encoded, the FRR and SER are important criterion. For example, you usually want a First Read Rate, that percentage of the time you successfully scan the first time, to be greater than 85%. Anything less and your users become frustrated by re-tries and end up abandoning using the bar code reader. You requirements for accuracy determine the SER level required. For instance, if you are a blood bank, you have extremely low SER level requirements and wouldn't want to use 2 of 5 code, which can accept partial scans.

Obviously the most prevalent bar code symbology in terms of number of bar codes in use is the ubiquitous UPC, or Universal Product Code. They are literally everywhere. The UPC was originally developed in the early 1970's as part of a grocery store automation project sponsored by the grocery industry. The goal was to reduce the cost of check-out labor, which at the time was costing just one chain, Safeway, over $1 billion a year. The UPC is made up of two parts: a vendor identification number, six digits long, and a product identification number, also six digits long. UPC also features a check digit to improve the accuracy of the scanned information. One of the key features of the size and shape specifications of UPC is its ability to be read at a wide variety of angles relative to the scanner. Most bar coding systems require relatively precise orientation to have high first read rates. But using the relatively new holographic check-out scanners, you don't even need the UPC code directly facing the scanner. You will most likely pick UPC coding if you are dealing in goods that will ultimately be sold retail since the UPC will already be a requirement or already be present on the goods.

Becoming the most common industrial bar code symbology, Code 39 was developed in 1975. Code 39 had been selected by the Department of Defense for LOGMARS, by the General Services Administration, by the Automotive Industry Action Group, and by many other groups as their standard bar coding symbology. There is even an ANSI standard for Code 39. Code 39 has two important advantages over UPC and many other coding symbologies. First, it allows the encoding of not only numbers but also characters (upper case only). Second, it is a variable length code. You can have bar codes with only one character or with 50 characters, although there is a practical limit to how many can be scanned with desirable first read rates. And most bar code readers will let you mix different lengths on successive scans (some bar code symbologies require the length to be fixed for all reads to maintain data security). Finally, Code 39 includes a high level of data security, providing SER rates of anywhere from one in 1 million to one in 100 million. Generally, if you don't have some other requirement dictating another symbology, you should seriously consider Code 39 as your first choice.

If your requirements are strictly numeric, then perhaps you should consider one of the oldest symbologies, 2 of 5. 2 of 5 was developed in 1968. It's a simple black and white encoding of the binary ones and zeros of the code with wide bars representing ones and narrow bars representing zeros. There is also a parity bit to assure integrity within a digit. Unfortunately there is no information within the white spaces in the original 2 of 5 code. Thus in 1972, a second variant of 2 of 5, called interleaved 2 of 5 was introduced. Interleaved 2 of 5 is simply 2 of 5 coding with every other digit encoded in the black stripes, and every other digit encoded in the white spaces in between. As a result, the coding density roughly doubles and is much the much preferred version of 2 of 5 coding today. There is a significant problem with 2 of 5 coding however. Partial scans, that is reading only part of barcode, can go undetected. Thus, it is common for particular bar code implementations to require a fixed length 2 of 5 bar code to allow additional integrity checking.

Another bar code symbology most business people have seen is Codabar. That's because Federal Express uses Codabar on each of their airbills. And they process over a half a million a day! Codabar is also used by photofinishers and by many blood banks. One of Codabar's features/benefits is its high data integrity and very high first read rate. However, Codabar suffers from peculiarities in its use of several different bar and space widths and appears to be declining in popularity.

A relative late comer to bar codes, Code 128 was introduced in 1981. The primary benefit of Code 128 is its ability to represent the entire 128 characters of the ASCII character set. Thus, you have upper and lower case, numerics and a wide selection of symbols. Given its late entry to barcoding, it can represent complex alphanumeric messages and is suitable for a wide range applications. Code 128 can boast being the highest density of all alphanumeric codes. This is achieved by using variable width bars and spaces. Thus a given bar/space may be anywhere from one to four units wide. There is also a "double density" coding for strictly numeric coding. Although there is a derivative of code 39 called code 93 that is denser is strictly uppercase alphanumeric comparisons, code 128 offers both lower case and very high density in numeric only applications.

Getting It in "Writing"

There are several alternatives getting a readable bar code. Each has its advantages and needs to be evaluated depending on your application.

By far the least changeable, and most economical and with the highest density, is offset printing. In other words, incorporating bar codes as part of the packaging. Obviously this is extremely common in consumer goods which include the UPC on the package. This can be done blatantly, as in most packages, or very subtly, as in a bottle of wine I saw that included its bar code as part of the design (the bar code formed the base of a group of rushes or wheat that made up part of the label). One of the greatest advantages of offset printing of barcodes is they can be precisely controlled relative to other printing technologies. For instance, it is possible to hold resolutions of 1000 dots per inch, or more, during offset printing. Why else would printing companies invest in phototypesetters with 2700 dot per inch resolution? The disadvantage of offset printing is you can't change the code as often as with other techniques. Thus if you are putting a single code on a lot of packages, Offset printing is the choice.

An alternative to offset printing when you need the high density of offset quality printing, but with variable labels, is to go with pre-printed labels. Several vendors will provide you with labels, either totally "stock" (no customization), or semi-custom (with your particular format and additional information). The advantage of this method is you can achieve significant density improvements, and sometimes lower your costs, at the risk of needing more or less of a particular label than you have ordered. Still there are a lot of applications where pre-printed bar code labels are appropriate.

Finally, if your demands are completely unpredictable, then you need what's called on-demand printing. There are many applications that either require, or can make use of the convenience of, on-demand printing of bar codes. On-demand printing of bar codes means, simply, that the user can generate the needed bar codes whenever they need them, thus the term on-demand. There are three interesting ways to produce on-demand bar-codes: Dot-matrix, Laser printers, and specialize Thermal printers.

Dot matrix printing offers the least expensive entry into bar code printing. Anyone with a $200 dot matrix printer and, potentially $10 worth of software, can be in the bar code producing business. Interestingly, a lot of bar codes are produced on dot matrix printers. Not the inexpensive ones, but on Printronix "line-printers"--and at relatively high speed. The primary disadvantage of dot matrix production of bar code is the limited resolution of that style printer limits the ultimate density of the barcodes produced. Thus you can't get extremely high resolution bar codes out of most dot matrix printers. Another potential problem with dot matrix printers is they aren't compatible with some infrared scanners because there isn't enough carbon in normal dot matrix printer ink. You can get special ribbons, but in general you are better off not mixing dot matrix bar codes and infrared readers.

An interesting cousin of dot matrix bar code printing is the use of ink-jet printers. In those cases, the speed of application to apply bar codes, along with the flexibility of application surfaces and lack of contact, more than make up for the limited resolution. In fact, spraying very low density bar codes onto cardboard cartons is very common in distribution applications.

A recent introduction into on demand production of bar codes is the use of low cost laser printers to produce on-demand labels. Producing several labels per minute at reasonably high resolution (300 dot per inch), laser printers appear to be destined to become the on-demand printer of the masses. The primary disadvantage to laser printing of on demand labels is you have to produce a full page, several labels, at once.

Finally there are a number of specialized thermal printers designed specifically for on-demand bar code label printing. They are fast, quiet, and produce high quality bar code labels. Primary disadvantages of thermal bar code labels is they deteriorate in higher temperatures (and in some cases when exposed to sunlight). Plus, because of the materials used, are incompatible with infrared readers unless care is take to select specially designed thermal label stock.

Different Applications/Different Readers

Bar code readers are characterized in several ways. They can be contact or non-contact. They can be wands, fixed or moving beam laser, or even array sensing. They can be visible or infrared. They can be batch (data collection) or on-line. Which is the most appropriate for you depends on your application needs and budget.

Contact bar code readers are just as the name implies--they must contact the bar code during the reading operation. The most common contact bar code readers are bar code wands, often in a wedge reader configuration. In contrast, non-contact readers need not be in contact with the bar code during the reading operation. In fact they are usually from several inches to several feet away. The most common non-contact bar code readers are the hand held laser scanners and the laser scanners found in many grocery store check-out counters. In dirty environments, or in applications where the bar code is read many times, non-contact readers are preferred. They can't be contaminated by dirt on the bar code label, and don't wear out the bar code as most contact readers would. However, their high cost and bulk relative to the inexpensive contact wands limits their universal acceptance.

One bar code reader classification you should pay particular attention to is the frequency of the light used by the reader. As I mentioned before, not all scanners are compatible with all printing techniques. Readers use either visible light or infrared light to scan the bar code. Visible light readers are compatible with almost all printing technologies, including thermal and dot matrix printing. Infrared readers are less accommodating, requiring a high carbon content in the ink. That requirement must be taken into account when having labels printed and restricts many dot matrix and thermal printers. Then why are infrared readers even used? There are two distinct reasons. First, infrared readers cost less than their visible counterparts. Second, the very insensitivity to many inks can be used to an advantage in applications where security is important. The bar code can be printed in a high carbon ink, then obscured by over printing with a second infrared "invisible" ink. Thus the reader can read the bar code, but casual observers can't.

The final classification of bar code readers is their usage mode. That can be either on-line or batch, depending on the application. For instance, a video rental store or grocery store check-out counter require on-line access to the store's database of merchandise. In those cases, the bar code read by the reader is immediately transformed into a rental agreement or price on the customer's bill. On the other hand, many applications are strictly data collection where the data will be processed at a later time. For instance, taking an inventory doesn't require on-line data access. The inventory part numbers and counts could be collected using a hand held laser scanner and uploaded into the computer later for reconciliation. Similarly, many job costing applications, where bar code is an ideal data entry method, have no on-line requirements. The data often must be matched with other data not available until a later date before it is useful. Sometimes a hybrid approach is possible using the programmable data collection terminals that are available.

Conclusion

Bar coding is a fast, accurate way to get data into a computer, especially where the operator isn't a trained clerical person. Because of the efficiencies made possible by bar coding, bar codes are finding their way into every aspect of our lives--from the grocery store, to the mail system, to sorting our bags at the airport. It's hard to imagine an application that can't be adapted to, and benefit from, bar coding.

Members: