General requirements for the programming workplace for mass chip programming include maximum throughput for a reasonable price as well as reliability, i.e. ability to work without a fault for the longest possible time. With the increasing intensity of innovation of programmable chips as well as of end products, the availability and price of an upgrade for a programming workplace for the programming of other chips – also this chips, that were not considered to be used or simply did not exist at the time of purchasing of a multiple programmer – is becoming increasingly important.
One of the basic problems of desktop multiprogrammers – which are often used for programming medium-size batches of chips – is the price/number of supported chips ratio. Such a multiprogrammer is a good choice in the case of a small range of programmed chips and longer chip programming times only, e.g. for programming high capacity memories. But with the ever increasing range of programmable chips, the cost of a programmer’s hardware upgrade is growing rapidly, and it can often happen that simply no upgrade for programming other chips is available in an acceptable time.
The number of necessary programming sockets of a multiple programmer depends particularly on the time of programming of individual chips and on the chip handling time. That handling time includes the removal of a programmed chip from the programmer’s socket, attachment of an identification sticker, putting the chip aside, insertion of an unprogrammed chip into a socket of the programmer and start of a new programming cycle. The optimal number of sockets of a multiple programmer is equal to the ratio between the time of programming of a single chip and the chip handling time. In the case of a programming time shorter than the handling time, the use of a multi-socket programmer is a waste of money. At a 20-second programming time the optimum number of programming sockets is 3 to 4. If the multiprogrammer have more sockets as necessary, it is also wasting of money.
Let me also mention that most of the desktop multiprogrammers works as the parallel (gang) programmers, where all the chip are programmed at the same time. It mean, the whole multiprogrammer is not working during the time, when the operator "reload" the ZIF sockets. In contrast, the much powerful solution is a concurrent multiprogramming system, where each programmer works independently - therefore only currently reloaded programmer is not working.
Such conflicting requirements are fully met by a programming workplace consisting of a cascade of powerful versatile single-socket programmers BeeProg2 or BeeProg+. The connection of more BeeProg programmers to the same PC through an USB interface creates a powerful multiprogramming system with the same number of supported devices as that of BeeProg (currently 21022) and with a truly high programming speed. A multiprogramming system designed in this way allows to flexibly choose the number of programming sockets as needed. In contrast with standard multiprogrammers, each programmer works independently (concurrent multiprogramming system), i.e. a programmed chip can be replaced in one programmer while other programmers work, which increases even more the productivity of the programming workplace.
The performance of the BeeProg2/BeeProg+ programmer is a guarantee of low programming times and therefore low number of necessary programming sites of an optimal programming workplace. The programmer's robust hardware is a guarantee of a low error-rate of the programming workplace. Moreover, there is a none chance that all programmers break down, in other words, in the case of a fault of one programmer, the whole programming workplace continues to work with just decreased performance. If you have one spare programmer stored (which is with difficulty financially viable in the case of large multiple programmers), just replace the faulty programmer and the programming workplace will continue working at its full capacity.
In conclusion, the return of investment. Provided 25 % of time in the operation of a workplace is lost in waiting until chips are programmed, the return of investment into another BeeProg2/BeeProg+ programmer added to a programming workplace is couple of weeks, in the case of very low labour price, couple of months.
Therefore we think, the optimal solution for multiprogramming is a cascade of BeeProg2/BeeProg+ programmers.