Objective and implementation:
Building a simple guitar pass through LCD audio visualizer, from new MCU based guitar pedal PCB. At this point I had bread boarded, programmed the code, and tested the concept, so I next wanted to transfer this into a stand alone device. To build this LCD visualizer, my plan was to leverage one of my more advanced MCU based guitar pedal designs that has yet to be fully vetted and build a “do not stuff” (DNS) version that omits all the components not needed for the objective. The DNS’d boards that I created would have a 9V input/output throughput buffer and a 5V level shifting amplifier for the MCU input. The original fully populated “Stuffed” board has all the parts above with multiple op-amp stages, mixers, and foot switch with LED.
Issue:
Audio visualizer not displaying guitar notes on visualizer of both DNS’d PCBs and the throughput buffer is not working on one.
The differences:
First, I tested and re soldered all the junctions that looked questionable on both the non-operational boards. I quickly found a solder glob between the source and gate on one of the buffers. Which fixed the buffer issue, but the LCD was still not responding as well as the operational Stuffed board. I noticed that when I would change to the DC offset control the screen would react nominally, so this pointed me straight to the level shifting amplifier as the issue. Since both the DNS’d boards were not working and had the same issue, I just went after one of them.
From here I beat up the schematic as much as possible, tested all the mechanical parts/values, confirmed operation voltages, swapped the socketed TL072CP op-amps. I finally pulled out the oscilloscope and read the outputs of the Stuffed and DNS’d board. This narrowed the issue specifically on the level shifting amplifier and omitted the 5V buffer out of the equation. Now, why was this amplifier not working.
I went through the datasheet confirming operational voltage of the op-amp, but quickly determined the op amp to not be the issue as the TL072CP’s were brand new* and at this time I already tried a couple of them along with a random LM358 that didn’t work. At the same time the Stuffed board had a TL072 op amp that was in and working. This pointed me back to the schematic. After not finding anything that should be an issue I went back to the board to measure all the passive components. Now I really had to start using my imagination on what the issue could be. Because this should all work… (3 hours)
I went back to the schematic and started adding in parts of the circuit that I assumed wouldn’t be an issue, but wanted to confirm.
- Isolated the output buffer on the DNS’d,
- Added in all components the components that touch the signal splitting node on the DNS’d,
- I isolated the same circuit on the Stuffed board and pulled out all additional OP-amps
- Added in unnecessary VREF voltage divider on the DNS’d
- Lastly, I placed all passive components on the DNS’d board
Still bloody nothing, all those potential solutions were a complete stretch, and still not working. (3 hours)
Going round and round troubleshooting other projects… assuming more dead ends I moved the TL072 op-amp from the Stuffed board onto the DNS’d board and the scope reading show amplification. I rubbed my eyes in disbelief and did a swap back.. and it didn’t work, then I did a swap back and it worked.. I noticed the TL072 of the Stuffed board was a TL072IP and the TL072 of the DNS’d board was TL072CP. So, case-closed right? No, what is the difference between the TL072IP vs. TL072CP. After going through the datasheet, there is no difference other spec production tolerances, temperature range, and material used to make them; All which should be negligible in this application. So, I tried the specific datasheet that was supplied by distributor, still nothing. (1 hour) Did I get a bad batch of op-amps?
Testing the op-amps, the TL072CP is definitely not working as expected. The only thing this could be is mixed up of pins on the TL072CP, I actually found a TL072CP from an old batch of operational op amps that did work. Which makes this plot even thicker, and points to a bad batch of op-amps. I did a final side by side comparison on voltage readings of the old TL072CP op-amps that worked and the brand new* versions that did not work. This was evidence enough for me me.
I reached out to the distributor of these, who is a well known distributor in the SF Bay Area, that I used to will call parts to when I was in college. The email thread went a little further than below but this was this gist. The below is no excuse, they should not resell parts unless that are validated. In the full email thread they gave me a ton of excuses and said it was normal. In my opinion there is no excuse for this. You buy parts, confirm they are what you order, and resell them. As simple as that.
Their reply: “Thank you for the photo. Please accept our apologies, these chips are not authentic, that date code has now been purged from the bin and our sales person can now give you a replacement going forward.”
My Question: Is this common?
Their reply: “In 2011, a congressional report cited widespread problems with counterfeit electronic parts and American military systems. 67% of counterfeits are discontinued products, 19% are of OEM production parts. More than 1.4 million purchase parts were involved in suspect counterfeit and high risk transaction during each year for the past decade. The proliferation of counterfeit electronic parts into the supply chain cost the US government and its contractors billions. Electronic contract manufacturers lose $100 billion a year due to this problem.Yes, it’s widespread. No distributor or supply chain is immune.”
The email thread got a little more involved. DM in Instagram for distributor, and full email thread.
