Beware of Fake Chips: A Hard Lesson in Electronics

Post Stastics

  • This post has 625 words.
  • Estimated read time is 2.98 minute(s).

If you’ve been following my blog for a while, you know that I love tinkering with electronics. Recently, I encountered a situation that has driven home the importance of sourcing components carefully, especially when dealing with critical parts like EEPROMs.

A couple of weeks ago, I decided to purchase ten AT28C64B-150PU EEPROM chips from Amazon for a project I’m working on. The chips arrived about ten days later in a familiar MidZoo box—a small white box with black lettering. Having bought several items from this seller over the years without any issues, I trusted the quality of their products.

However, when I began using the chips, I encountered a frustrating problem. I placed three different chips in my programmer, and none of them would program. My first thought was that there might be an issue with my programmer, so I built a small Arduino programmer for the 28CXXX family that I found on GitHub. Unfortunately, the chips wouldn’t program using this circuit either.

Due to other responsibilities, a few days passed before I could revisit the issue. Today, I decided to manually program the EEPROM using a breadboard setup. For the AT28XXX family, programming is initiated by an edge of the programming pulse, making it safe to program this way. Following the example set by Ben Eater, I breadboarded a simple circuit using DIP switches for the data and address lines, buffers (7406s), and LEDs to monitor the data lines.

I inserted one of the suspect AT28C64B-150PU chips and applied 5VDC to the board. I read the first 16 cells, which all showed 0xFF—a usual value for an empty chip, though sometimes new chips contain random values. I then attempted to program the first position at address 0x0000, and this is where the problems started. Despite checking the wiring and DIP switch positions several times, I couldn’t program any new data values into the chip. I even attempted to write the unlock sequence to the new chips and then reprogram them, but this too did not work. The chips simply wouldn’t accept new data.

Finally, I dug through my old parts and found an X28C64 chip that I had purchased from Marvac Dow back in 1990 (as evidenced by the dated price tag still on the Teknopak card). I placed this chip in the manual programmer circuit, and it worked perfectly. It also worked in my other two programmers, confirming that my circuits and programmers were functioning correctly.

With this confirmation, I tested all ten AT28C64B-150PU chips I had received from Amazon. None of them would accept new data, even though they read as empty (0xFF). It was clear that these chips were faulty, and my initial suspicions were correct. Notably, all my chips contained a date code of 2132. I have never received ten chips, all with the same date code in a single order, unless that order was from the manufacturer. This alone should have been enough to suspect the new chips.

The lesson here is clear: buying chips from what seem to be reputable dealers, especially those based in China, can be risky. Despite the convenience and often lower prices, you may end up with counterfeit or defective components. I wasted several days troubleshooting what I thought were issues with my circuit and programmers. Without that old, reliable chip in my scrap box, I might still be struggling to diagnose the problem.

To all my fellow electronics enthusiasts: be forewarned. Always scrutinize your component sources, and if possible, stick to well-established suppliers known for their reliability. It may save you from a lot of headaches and wasted time in the long run.

Stay vigilant and happy tinkering!

Leave a Reply

Your email address will not be published. Required fields are marked *