Welcome to the ninth newsletter on all things THAT (THE ANALOG THING).
When the circuit boards for batch 1.5 of THE ANALOG THING arrived at our office, we soon noticed during testing (a very thorough process in our company) that something was not right. Power consumption was too high and the behavior of the integrators and summers fell short of our expectations.
Since all THATs we tested behaved similarly, it was clear that this was not due to some random faulty component but a systematic fault.
Eventually, it turned out that the operational amplifiers labeled TL074H - the correct part number - did not behave as they should. (Note that the TL074H is a new operational amplifier whose label only shares the "074" with that of its predecessor, the TL074C.)
Could these be counterfeit parts? How could we be sure? The only way to say for sure was to take actual microphotographs of the dies. Thus we set out to decap a real TL074H as well as some of the suspected counterfeit chips.
Decapping integrated circuits is a tedious process. There are several recipes out there in the wild to get rid of the plastic enclosure. We decided to dissolve the plastic in hot nitric acid (65%). DO NOT TRY THIS AT HOME!
The chip was put in a beaker with about 10 ml of nitric acid and placed on a heater set to 110 degrees Celsius.
⚠️ Note that the fumes resulting from this process are dangerous and should not be inhaled! Work like this must be conducted with appropriate ventilation, i.e., either outside or under a suitable fume hood! ⚠️
A friend of ours and an expert in microphotography (as well as many other exciting things :-) ), Dr. Ralf Nötzel, was kind enough to not only take pictures but perform some decapping work after a sample he obtained from our decapping process was shattered during postal transport, so he had to do it all over again.
His setup is shown in the following picture: A Keyence microscope for metallurgical and other research requiring incident light.
It turned out that the suspected counterfeit chips were indeed not what their label suggested. This is how a real TL074H looks:
A very modern design on a really small die (less than 1 mm side length). In contrast to this, the operational amplifiers that were actually soldered onto the BASE boards look like this:
Clearly, this is a much simpler and older design. The markings on the chip seem to imply that this might be a classic TL074C (instead of the modern TL074H).
Although THE ANALOG THINGs with these chips basically work, we - of course - would not send these out as they are not up to our standards. Accordingly, we are now manually replacing 1,500 counterfeit chips with genuine TL074H and retesting the entire batch again.
Unfortunately, this will delay the shipment of batch 1.5 by a few weeks as we first had to get a sufficient number of replacement operational amplifiers (and make sure they are the real deal). The manual chip replacement is a tedious process (hats off to Max Peschke, who takes care of most of the rework for his incredible dedication!).
About 30% of batch 1.5 have now been sent out with all their components conforming to our specification. The rest will be shipped during July.
Episodes of this kind are quite annoying for both our customers and us, and they cause real a loss of money (this is one of the risks of running a small company developing and selling electronic devices).
What we find difficult to understand is that someone even bothers to counterfeit rather inexpensive chips like a TL074H. The amount of work required seems quite substantial: Removing TL074C from reels, grinding them down a bit to get rid of the original labeling, relabeling, reloading back onto reels, and so on. All this at the risk of being found out and of losing customers for good. Who in their right mind would bother to do something like this?
The world of electronic parts procurement has become a strange place indeed.
Having encountered multiple problems with outsourcing manufacturing overseas, we decided to find local suppliers and PCB manufacturers here in Europe for the next batch of THATs. Of course, this increases cost, but it has the valuable benefit that we can observe the manufacturing process more directly, take samples from a production run and test these immediately without having to wait for a complete batch to arrive transcontinental shipping only to realize that extensive follow-up work is required to get the product to our customers at the appropriate level of quality.
Furthermore, we have decided to buy our parts only from distributors of the manufacturers to avoid problems like this in the future.
We are currently negotiating terms with a company in Hesse (Germany) and, if all goes according to plan, expect to start the production of batch 2.0 within the coming few weeks and to start shipping in mid-August 2022.
On June 30th the Breaking Lab had a two and a half hour long live stream which also featured THE ANALOG THING (in German):
anabrid GmbH and THE ANALOG THING will participate in the Festival der Zukunft at July 22 in Munich, Germany. There will be a short talk and a workshop with THE ANALOG THINGs. We would love to meet you there! Here are a few impressions from the conference in 2021 (where we did not yet take part):
anabrid GmbH was nominated at the 2022 Berlin Deep Tech Award in the category social and sustainable technology. You can view a recorded lifestream of the event.
At the event, we showed our new video featuring THE ANALOG THING for the first time to a wider audience:
The last Analog Consultation Hour two weeks ago had to be canceled since Bernd was unavailable. The next Analog Consultation Hour will take place on Sunday, July 3rd, 2022 at 6 PM Berlin time. Zoom link: https://anabrid.zoom.us/j/94750770887
Happy analog computing!
As always, please do not hesitate to contact us at that@anabrid.com if you have questions regarding THE ANALOG THING.
You can share this newsletter issue with the link https://the-analog-thing.org/newsletter/9/.