By Peter Davie
A few years ago, I was asked by a client to reverse-engineer an industrial Pulse Generator/Power Supply. This was to enable us to carry out local repairs on the 50 or so of the units in his employer’s production-line. These PSUs are basically high power pulse generators, producing an output pulse stream, in this particular case, at up to 200V at an amp or two with a repetition rate in the hundreds of KHz range.
Most of the job was fairly uneventful - just digging through lots of potting compound and unraveling the circuitry found inside, documenting it and verifying part availability etc.
Then things got a bit more interesting… The heart of the pulse-monitoring circuit consisted of a chain of several plastic 8-pin DIL comparators set up as sequentially cascaded window detectors. These compared a divided-down and filtered version of the PSU pulse output under load, with locally produced references.
The PSU manufacturer had gone to some pains to protect his design from plagiarism; one step taken was to grind off the IC part numbers from the top of the comparator chips. In a lot of cases this wouldn’t have been a problem as any general-purpose comparator would have done the job. However in this case, I already knew that the entire chain had to resolve in a few 10’s of nanoseconds to give the required degree of accuracy to the resulting sampled voltage data.
So I had to find out exactly what the mystery ICs were.
There were three or four of them in the circuit and they were all entirely devoid of manufacturers’ markings and no other clues were visible to determine the relevant part numbers, or even who had made them. I was considering setting up a test scenario and developing a “virtual data-sheet” to compare performance spec’s with likely suspects when I had another idea.
I remembered seeing photomicrographs of IC dice in an old National Semiconductor data-book and noticing that the die part number was often etched into the circuit margin along with the familiar “NS” logo. I wondered if there was any possible way I could uncover the surface of the die in one of my mystery chips and read the information I needed from it.
Sounds unlikely, doesn’t it? Well, to cut a long story short, after destroying all but one of my sample ICs, I eventually managed to grind, chip and peel off the encapsulating plastic just right and reveal a clear view of the silicon die inside the mystery IC revealed that the chip was an Analog Devices AD9696 high-speed comparator. The Analog Devices Logo could be seen clearly with the”9696″ part number as plain as day in the lower left. Armed with this new information, I finished the circuit diagram and parts list, reported to my customer and went on to the next job - fixing a batch of failed EDM PSUs!
Peter Davie became interested in electronics in the 1960s at the age of 6 when one weekend his father, a college physics lecturer, bought home a Tektronix 545 oscilloscope and an AF signal generator for him to play with. He now runs his own electronics company, The Owston Grange Group Ltd, in rural Scotland. The company specializes in the reverse engineering, repair and servicing of industrial electronic equipment. Over the last 12 years Peter’s Lab has become a magnet for precision instruments amongst which, Peter considers his collection of 15 Tektronix oscilloscopes to be a fitting homage to his childhood introduction to the world of electronics.