Today, with a wide variety of programmable components available, you can easily store millions of bits of programs and data in FLASH/EPROM, or pack hundreds of traditional digital components into a fingernail-sized CPLD/FPGA; you can also turn an entire microcomputer into a single chip. As long as the R&D engineer has creativity, he can accomplish this seemingly ordinary task on his experimental table.

Looking back to the early days when microprocessors were just emerging, only a few companies had the capability to develop photomasks, create a MASK ROM or an ASIC. Besides the high cost, they also had to bear significant risks and deal with considerable inventory pressure.

With the continuous innovation of semiconductor manufacturers, various programmable components have emerged, such as non-volatile memories like PROM, EPROM, EEPROM, FLASH EPROM, etc., with capacities ranging from a few kilobits in the early days to 1 G bits that may be available in 2002. The digital circuits that could only be designed using TTL have gradually been replaced by PLDs, evolving from simple PALs to CPLDs/FPGAs with millions of gates. Coupled with the ever-changing tools provided by IC burners, the dream of "customizing ICs on a desk" has come true.

Classification of Burners

According to the specific requirements of usage, burners can be roughly classified into two types: those for research and development purposes and those for mass production. Different usage purposes have led to different design concepts and characteristics, which are described below:

1. For research and development purposes

Based on the types and functions of the processing components, the burners used for research and development can be further classified into single-function type and universal type. The single-function type burners, which provide simple functions for a single type of component, usually have the advantages of being simple to use and having a low price. However, they also have the problems of being unable to be expanded and having inconsistent quality. Due to the low technical threshold of such products, there are various options available in the market, including both professional and hobbyist suppliers. As long as the price is low, they can be placed on the shelves.

If your sole purpose is for a single development task and you only need simple components such as EPROM, MCS-51 or PIC, and you don't want instability or poor yield to delay your development schedule, then professional burner manufacturers or products authorized by IC original manufacturers might be a better choice for you.

The current mainstream product is the universal programmer. From the perspective of the R&D engineers, there are numerous programmable components that can be applied in product development. In the same design project, the use of single-chip, FLASH, EPROM, and CPLD components in combination is quite common.

Purchase a universal burner that can support thousands of types of ICs through software upgrades. Although the price is a bit high, it is a good choice in the long run.

For the manufacturers of burners, the main difficulties and challenges for universal burners come from the continuous innovation of semiconductor manufacturers. Currently, the number of burnable components on the market is approximately 4,000, and they are constantly changing at a rate of 300 to 500 per year. This does not include the number of old components that update the burn algorithm. Therefore, the manufacturers of universal burners must have a large R&D team to cope with the endless software updates. Not only must the speed be fast, but it must also be completely accurate. Otherwise, they will be eliminated quickly. Moreover, most IC original manufacturers will not allow the market to have a large number of unqualified burners, which would cause disputes that cannot be resolved. Therefore, they will only provide IC samples and burn algorithms to a few certified excellent manufacturers. Finally, new competitors entering the market must also face the task of supporting more than 4,000 types of components. The huge manpower and costs required for this are even more daunting.

From this, we can see that the entry barrier for this product is extremely high, and its sustainability is extremely difficult. The sweet and bitter moments involved are deeply engraved in the minds of the manufacturers. It's no wonder that major manufacturers from advanced countries such as Europe, the United States, and Japan have gradually withdrawn from the market over the years.With no major competitors emerging in the mainland and South Korea, and with the continuous efforts of Taiwanese manufacturers, another world record was achieved (with an estimated market share of over 50% of the total shipments).

II. Mass Production Version

This type of burners are usually classified as part of the production equipment. Cost is not the primary consideration. Yield, stability, production capacity and service are the important factors. There are three elements that affect the burn rate of components: the burner, the process changes of the IC, and the operator. Among them, the burner accounts for the largest proportion. As long as the design is good, fully complying with the IC manufacturer's specifications of the Algorithm, and promptly adapting to the changes of the IC manufacturer, providing users with the best burn method within the shortest time, combined with appropriate error-proofing design to avoid the operator's mistakes, this has resulted in an almost perfect burn rate.

Unlike an office, a production unit has more complex environmental factors. When using the recorder for 8 hours or even 24 hours a day, the stability of the recorder determines whether your production line will be disrupted. Moreover, if you are burning FLASH/EPROM or other large-capacity ICs, you need to consider the output rate of the recorder. Under the same IC specifications, the difference in speed can be several times. Considering your working hours and overall production capacity, the output rate is also a major factor. If the differences in the previous points are not significant, then the quick service and sustainable operation of the recorder manufacturer determine everything. After all, the production line cannot "stop and wait".

In my humble opinion, apart from the aforementioned points, a mass-produced burner that can operate independently without a computer connection, is easy to operate, and allows for the replacement of various modules, would be the perfect choice for you.

The early production-level burners were almost monopolized by foreign well-known brands. Their prices could reach tens of thousands or even millions of yuan, which was a heavy burden for domestic manufacturers. In recent years, thanks to the efforts and improvements of domestic suppliers, high-quality, cost-effective and reliable production-level burners have been widely available to meet the domestic market demand. This has also made a small contribution to Taiwan's PC industry becoming the world's number one.

Some hidden cost issues

When you are assessing the budget needed for purchasing a burner, there are some hidden costs and potential problems that you must also take into account:

The direct financial losses caused by the defect rate

For a well-designed programmer, the defect rate is typically around 0.15% to 0.3%, and it can even be calculated in terms of ppm. However, some programmers may go out of control and reach a rate of 2% - 3%. The difference here represents that for every 100 ICs processed, there are two or three defective ones. Regardless of whether the unit price of each IC is high or low, after some time, the amount you saved when purchasing the programmer may be completely recouped. Of course, this does not include the increase in labor costs, delays in delivery, etc. and all these factors.

The nightmare caused by data loss

When the IC that you burned with the recorder was judged as "PASS", soldered onto the PCB, the casing was locked, and finally delivered to the customer, only then did you realize that the machine couldn't be turned on and the functions were abnormal. After several twists and turns, it was discovered that the BIOS or the program and data within the single chip had been lost! You had to apologize to the customer, recall the defective products, deduct the payment, and also caused damage to the company's reputation! And the production unit and engineers had to hold a series of earth-shattering meetings for reflection, be in a state of chaos and rush to reprocess.

These situations do not only occur in nightmares, but are also common occurrences. It is always wiser to prevent problems before they happen. To avoid the realization of future nightmares, you must be more cautious when choosing a burner.

The SOCKET of the burner is a defective product.

All burners have a SOCKET (or called TEXTOOL TM) for inserting the components to be processed. Generally, the wear from contact and mechanical fatigue make the SOCKET a consumable with a limited lifespan. The SOCKET for DIP packaging is approximately 15,000 to 20,000 times, while for more precise SOCKETS like PLCC, TSOP or QFP, the lifespan is only 10,000 to 15,000 times. Once the usage exceeds this limit, the yield may decrease, or even the burning process may fail completely. Therefore, it is necessary to replace the SOCKET regularly or in fixed quantities as an expense.

The number of supported components for the burner and the assessment of the support energy for new components

If you are a R&D engineer, the types and quantities of components supported by the burner are the first condition you need to compare. However, numerous IC packages (such as DIP, PLCC, SOP, TSOP, VSOP, QFP, TQFP, μBGA, etc.) are also an expense for future expansion. Some burners, because they are not truly universal, have over a hundred modules just for PLCC packages. Calculated at an average price of 3,000 to 5,000 yuan per unit, it is indeed a considerable burden. Therefore, some manufacturers have launched truly universal burners. Each Pin Drive has Vcc, Vpp, GND, CLOCK and logic signals, which can support all pin types (from 20-84 Pin, SOP modules are from 8 to 44 Pin) in one PLCC module. Although they are more expensive than ordinary burners when purchased, in the long run, it is still a good deal.

In addition, low-power ICs are currently the mainstream products. 3.3V, 3V, and 2.5V are no longer special. 1.8V components are also coming one after another. FLASH EPROM, monolithic chips, and CPLD all have this power specification number. After all, the benefits brought by low voltage are not only energy saving, high speed, and no heat generation, but also an inevitable choice in the process of semiconductor manufacturing changes.

Therefore, the burner not only needs to be able to supply 1.8V power to the test IC, but if the logic signal level cannot be adjusted accordingly, latch-up problems may occur when handling the IC. (Note: When the input pin voltage of the IC is lower than GND or higher than VCC, a sudden large current may be generated, causing changes in the IC's characteristics or damage to it.)

Pre-check of the burned-in components

To prevent human errors, when the user inserts the IC into the SOCKET, the burner should be able to conduct a contact check on each Pin of the IC to rule out cases of the component being placed upside down, shifted, or having incorrect IC pin count; after the contact check, the ID Code of the component should be identified to confirm that the IC manufacturer and model are correct.

Only through rigorous pre-checking can the burner deliver power and signals to start processing the IC. This is particularly important for increasingly precise SMD components.

Conclusion

Faster speed, greater capacity, more precise packaging, more pins, lower operating voltage - the semiconductor industry is advancing at an incredible pace, bringing about rapid and significant changes. Besides constantly absorbing new information to develop better products and produce more valuable goods, does the IC programmer you have chosen also follow your footsteps and move forward vigorously, assisting you? Or does it delay your work, causing losses of time and money? We believe this is something that all IC programmer manufacturers and users, such as yourself, should carefully consider.