For the proper functioning of this website we recommend to turn on JavaScript.

Elnec programmers T51prog/T51prog2, PIKprog+/PIKprog2, MEMprog/MEMprog2 and SmartProg/SmartProg2. Selftest error: "Pindriver (TTL) test ... ERROR" reason.

In this application note I would like to provide my final statement on a relatively high error rate of the T51prog, partly T51prog2, PIKprog+/PIKprog2, MEMprog/MEMprog2 and SmartProg/SmartProg2 programmers, specifically on the error, indicated in the programmers' self-test as a "Pindriver (TTL) test ... ERROR".

Elnec makes a great effort, with maximum consistency, to make sure our programmers serve their customers reliably. The hardware of the programmers is designed with a substantial reserve for permanent operation and our universal programmers LabProg/BeeProg/BeeProg+/JetProg are additionally equipped with effective protective circuits to protect them against ESD. Similar care is given to the production process and, of course, each programmer is burnt in.

Our universal programmers have therefore recorded the maximum long-term error rate of 1 % and specialized programmers have recorded the maximum error rate of 2 %.

This is why were have been seriously concerned about the increased error rate of the T51prog and PIKprog+ and partly of the T51prog2, PIKprog2, MEMprog/MEMprog2 and SmartProg/SmartProg2 programmers, which manifests itself by the "Pindriver (TTL) test ... ERROR". In 99 % of all cases this error means a damaged Xilinx FPGA Spartan XCS05 chip, which is the core of the programmers and which also works as a TTL pin driver. The damage to the chip is specific and the error is exactly the same in all damaged chips.

On an example of the T51prog programmer, I would like to clarify what we mean by the term "increased error rate". By the end of 4/2005, totally 6% of T51prog programmers were claimed, while 5.5 % (!) of the cases were due to the damaged XCS05 circuit. Therefore, with some proudness, we record just a 0.5% error rate of Elnec's T51prog programmer.

Since here at Elnec we do not tend to pass the blame to others, we have made a number of technical and statistical analyses of the damages to the programmers and to the XCS05 chips with the aim to find and eliminate the source(s) of the error. We let also the programmers' circuits be analyzed by an independent developer and, in the end, by an application engineer of Xilinx.

The circuitry analysis has not found anything wrong. The statistical analysis, however, has produced some interesting information:
1) Damaged were ALWAYS those pins of XCS05 that were connected to the ZIF socket.
2) Atmel has bought 100 pcs of T51prog and NONE of them have been damaged (verified in Atmel company).
3) The error occurred several times at some customers (!).
4) In winter and spring months we registered an increased number of claims.
5) The same chip is used in the same place (as a TTL pindriver) also in BeeProg, but NOT A SINGE XCS05 in BeeProg has been damaged.

We started to have some suspicion, but not certainty. We therefore agreed with Xilinx that they would take a few circuits for detailed testing. After some demanding work (chemical de-capsulation = removal of the plastic package, grinding off and analyzing individual layers of damaged circuits under the electron microscope) we received a final statement for all chips we had sent them. We can unveil only abstract of the final report:

You can see, our suspicion was confirmed: the circuits were damaged by electrostatic discharge (ESD).

The interpretation of the statistical findings is now simple:
1) The user touches pins of the ZIF socket and of the ISP connector when working with chips being programmed.
2) At companies that have introduced (and observe) proper rules for the working with chips sensitive to ESD, the error does not occur.
3) Customers, that ignore the proper rules for the working with chips sensitive to ESD, damage the programmer repeatedly.
4) When the air humidity is lower, the accumulated charge tends to be bigger and an electrostatically charged object tends to keep electrostatic charge longer.
5) BeeProg (as well as other programmers of higher price category, i.e. LabProg+, JetProg, BeeProg+ and multiprogrammers based on the BeeProg/BeeProg+) has ESD protection on all ZIF socket pins and on all LPT socket pins, which ensures safety at least for discharges up to 15 kV (human body model). Therefore none of this programmer has been damaged by ESD.

Just for more information, the above-mentioned Xilinx Spartan XCS05 chip has a standard (2kV) ESD endurance. By walking on a carpet in winter months, a body may be charged to more than 30 kV.


Damage of the programmer, caused by the failure to observe the rules for handling the device, is considered to be the result of improper handling of the device. Our warranty does not cover faults caused by improper handling. The warranty therefore does not apply to T51prog/T51prog2, PIKprog+/PIKprog2, MEMprog/MEMprog2 and SmartProg/SmartProg2 programmers whose XCS05 chips have been damaged by ESD.

In conclusion, I would like to point your attention to Elnec's application note Few notes about handling of electronic devices or directly to the Electronic Discharge Association website, where you can find more information about the rules for proper handling of electronic devices.

Yours faithfully
Jan Puobis, Elnec

Presov, May 2005, revised November 2008
Author: Jan Puobis, Elnec

Return to Application notes