Advertisement

Blog

RC in Front of the Op-Amp: Why Did It Fail?

Its 3:55 p.m., and I'm just a few seconds away from turning off the O-scope, detaching my ESD strap, and hanging up the alligator clips. Suddenly, the door to the electronics lab opens, and who do I have the pleasure of seeing, standing in the only doorway in and out of the lab? It's the boss. Not just any boss, but a boss holding a drink in his hand getting ready to leave and to tell me to stay just a tad bit longer. A boss who is asking me to stay because someone needs just one more prototype built to show the customer the next morning.

Now, any electrical engineer knows that the typical boss with no knowledge of electronics believes creating a hand-built circuit board can be done with the wave of a soldering iron. Ha. How many times have we created the perfect circuit that still didn't work? One hour to everyone else is five hours in EE terms.

I flip the scope back on, heat up the soldering iron, and lay out the components. I hear all my coworkers leaving, lights being switched off, and my stomach growling as I get hungrier. I begin to build. This time, it isn't quite as bad as it is sometimes. This particular circuit has only 22 parts, the largest and most complicated being a quad op amp. The night is still young, and there is still hope for a nice dinner.

Well, as most of us EEs know (and as I previously hinted), nothing ever goes right the first time around. Remember, one hour to everyone else is five hours to an EE.

A couple of hours pass, and the last leg on the op amp is soldered down. All components are double checked, and the layout is compared to the schematic. All looks good. The three-foot power cord with alligator clips is connected, and power is applied. The circuit is working — 120Vac in, 12Vdc out. There's the 5V signal, and the lights are working. It's time to wrap it up.

Just as I'm getting up to stretch and clear the solder fumes from my head, something tragic happens. Remember the drink the boss had in his hand? Well, it's sitting on the top shelf of the lab bench. The drink gets bumped by the three-foot power cord, and, in what seems to be slow motion, I see the drink gracefully fall over and pour its contents on the only spot on the workbench where it's not welcome. It pours on the freshly built circuit board, which is still warm from the half roll of solder used.

After a few words that would not be recommended for saying during normal business hours, I grab the circuit and begin blowing off the drink. Now, in a twist of fate that shows the universe has a sense of humor, this drink is an RC soda. This is not the RC I prefer to have next to my op amp.

After several minutes of drying off the circuit, I notice that, though seemingly dry, it is still sticky. Worried that the sticky soda deposits may be conductive and damage the circuit, I apply a little water to a cloth and pat down the circuit. After drying the circuit with a fan, I hesitantly repower the circuit while holding my breath and crossing my fingers. Nothing blows up. All the voltages are just right. The sensor is working, and the lights are on. Tragedy has been sidestepped. Or has it?

The next morning, I immediately go to the lab to check on the circuit. I look it over and power it up again. Oh, no. It's not working. The power supply is working mostly. The 12Vdc is there, but no 5Vdc. What is going on?

After pondering a bit and testing the op amp, I determine that it's not functioning. It has small white deposits on the black epoxy housing and possibly some hard water stains on the leads. I replace it and catch a break. Everything seems to be back on track. It now works. So why did the op amp work last night after the tragic RC spilling, but it did not not work after sitting overnight? Is it possible the minerals in the dried water created a short? Is it possible mold grew in the tiny pores of the op amp device and short circuited it? What made the op amp go bad?

Here is the twist: It isn't long before I notice several of our sensors using the same op amp on other boards are either not working or working incorrectly. I didn't dump RC on every board we received, so that wasn't the problem. After a couple of weeks of testing and tracking date codes and lot numbers, and multiple overnight teleconferences and meetings, I come to the conclusion that we received a bad batch of op amps. They aren't just bad — they're counterfeit.

With my trusty $2 magnifying loupe, I find some slight font differences on the manufacturing markings between a known good op amp and the bad op amps. I take a trip to the local hospital to do an X-ray on the good and bad op amps, and I find they were constructed totally differently. The bad ones had several large voids inside.

Did the RC or the water cause a failure? Nope, that was just a coincidence. And we know that correlation does not imply causation. It was actually due to fake parts. Since it focused my attention, in the end, perhaps putting an RC in front of the op amp was a good thing. Ha.

35 comments on “RC in Front of the Op-Amp: Why Did It Fail?

  1. goafrit2
    January 23, 2013

    >> It was actually due to fake parts.

    I am very curious – how did you end up with fake parts? Was it bought in Asia or through the normal U.S. distributors? I have never had that experience and will like to avoid it.

  2. goafrit2
    January 23, 2013

    >> Since it focused my attention, in the end, perhaps putting an RC in front of the op amp was a good thing

    Sure, one must look at the load issues and also the current that the opamp can provide to drive the resistor in the RC network. There is no general statement that having RC in front of opamps is always the best thing. It has to do with the circuit and the system. In yours it worked which is good.

  3. Jason Bowden
    January 23, 2013

    Thanks for the reply. Well first, let me say, I have had my run with fake components, op amps, triacs and LEDs. It is absolutely amazing how an electronic component can be counterfeited. I have a totally different story for the 3 fake parts I just mentioned, but let me try to give some advice in general overall. All of my PCBs and components came from Asia. The particular product we manufactured had such a high demand that year that it caused our CM to drain the distributor stock. This distributor was authorized and where they had purchased for years. They had to make an Emergency buy from a different distributor who was not authorized. Low and behold the entire single lot was fake. Basically, here is what I implemented and enforced that would at least help a little for me to always know where my parts came from. I made my CM put together a monthly report of only the major components to show where they were purchased, how many, datecodes, lot numbers and the datecode, lot number and production line the parts were used at the CM. also, I made certain that the CM would not buy from any different source until I was notified of who they were. I tried putting other controls in place, but some were too difficult.

  4. eafpres
    January 23, 2013

    In the relatively early days of Satellite Radio commercial ramp up, there were a few parts used in the antenna amplifiers that were only made by certain companies in Japan, and they were often in short supply or on allocation.  This created a counterfeit market.  Since the antennas are roof mounted, these parts had to meet the broadest temp specs.  Often the counterfeit parts don't.  At one point I recall the design group testing a batch of ICs in the temp chamber to verify they were fake.

  5. goafrit2
    January 24, 2013

    >> All of my PCBs and components came from Asia.

    Now it makes sense. Largely, I do buy PCBs from Asia but never components. The reason is that they do not have companies that make components. It does not make a lot of sence to order PIC from Asia when Microchip Inc is here in USA. Yet, I have friends that do just that.

  6. goafrit2
    January 24, 2013

    >> Often the counterfeit parts don't.  At one point I recall the design group testing a batch of ICs in the temp chamber to verify they were fake.

    When I was starting in the industry, I never knew the importance of supply chain experts. That job goes beyond clicking to buy, it involves solving these fake products entering a supply chain.

  7. goafrit2
    January 24, 2013

    My colleague at work just read this and asked me to share that having a cartoon or a drawing would have made it more accessible to many people. Engineering concepts fly better when there are drawings.

  8. jkvasan
    January 25, 2013

    Hi JB,

    Interesting read with an easy style.

    I had several honeymoons with such fakes. In your case, you could tell a physical difference whereas I was left to feel the fake was looking more authentic than the original. Everything was same yet performance was horrible. From a simple LM358 to a complex LM2576 gave me dreadful experiences. I even had problems with PCF8574 which is a I2C I/O port expander.

     

  9. Jason Bowden
    January 25, 2013

    I think 80 percent of all components these days are made in Asia or CMs buy from distys or brokers or someone in Asia. I don't think many components are made in the US anymore.

  10. Jason Bowden
    January 25, 2013

    Now I did buy alot of parts from the US, but they were not made here.

  11. JohnaIrl
    January 25, 2013

    Jason, thanks for sharing this. I have heard so many stories about counterfeit chips. I admire your determination to get to the bottom of it. Well done!

  12. TheMeasurementBlues
    January 25, 2013

    Do engineers know how to solder or do that just use simulations now. IEEE is giving a soldering course in Boston in early March. Soldering for Engineers tells some tales about soldering.

  13. TheMeasurementBlues
    January 25, 2013

    Jason: Counterfeit connectors are also out there. It's not just ICs, but everything. Fakes often cheat on plating and housing are often not good fits.

  14. goafrit2
    January 26, 2013

    Yes, the components are made in Asia technically but most are outsourced plants of U.S. firms. That is different to when an Asia imitator makes a component. That TI has a plant say in China will not worry me if they run and manage it themselves. I do not think any product from that plant will turn out to be fake.

  15. goafrit2
    January 26, 2013

    That should be a very good one. When I was in graduate school, the lab had an assistant who was doing that. It was like that has been “outsourced” and it was not that a great skill to have. Yet, if you are a designer at transistor level, soldering may be rare for you in most of these bigger firms where they have clear division of labor.

  16. Jason Bowden
    January 26, 2013

    Yeah but once they leave the plant, noone can really control what they get mixed in with. It then becomes up to the disty or whoever to make sure they are not taking parts from unauthorized people.

  17. Jason Bowden
    January 26, 2013

    Yep,had an issue of connectors that looked exactly like my suppliers but they melted at low temp. Come to find out they really weren't V0 rated material. Thanks to a nice material analysis.

  18. eafpres
    January 27, 2013

    When I was with a company making antennas for WLAN high-end access points, we once had TNC connectors where the inner dielectric melted and deformed during the application of heat shrink tubing.  We found some of them were molded from low grade polyethylene.  

  19. DadOf3TeenieBoppers
    January 28, 2013

    I too have had a run-in with counterfeit parts. A board that had failed in the field after many long years of service, and was no longer in production, had to be repaired. Our technician easily figured out what had failed (a Uni-junction transistor that was no longer in production). But when he replaced the part, several times I might add, without getting it to work again, they gave it to engineering.

    I replaced the part one more time with the same dismal results. It was then that I noted the failed parts both failed identically, which was odd; they behaved like regular bipolar transistors, rather than a dead short or dead open like a failed part should. Checking with the technician confirmed this.

    I quickly figured out that we had received a batch of counterfeit parts remarked to look like the unijunction transistor. After alerting components engineering (who subsequently blacklisted the vendor we got them from) we found six parts in an old warehouse our company owned overseas. Two were shipped to me, and on the first attempt, the board worked again!

    As engineers we have trouble understanding the mindset of someone who will compromise their integrity for a few hundred bucks. It is disturbing that they are out there. Will we have to wait for the 2nd coming for them to recieve their just desserts?

  20. goafrit2
    January 28, 2013

    >> Yeah but once they leave the plant, noone can really control what they get mixed in with.

    It is indeed a supply chain issues which can be controlled if the distributors are serious. If you decide to buy from only authorized distributors, this problem will be solved.

  21. goafrit2
    January 28, 2013

    >> As engineers we have trouble understanding the mindset of someone who will compromise their integrity for a few hundred bucks.

    Sure but engineers in this age are seeing a lot of things. Who can swear with his CEO? These days guys are not even sure when their CEOs are speaking the truth because he will email the whole staff that there will not be layoffs only to cut 10% next week. That said, any engineer that does not understand that people can do anything for money is in trouble.

  22. goafrit2
    January 28, 2013

    >> We found some of them were molded from low grade polyethylene. 

    I have worked in the industry and noticed that companies sometimes have bad production. Say, they have a big order. They produce and quality is not met. They need to redo it. However, instead of throwing away the parts, they sell it to mass market. It happens. Sometimes the products are not fake, it is just that companies want to make money from compromized parts.

  23. eafpres
    January 28, 2013

    goafrit2 you raise a good point.  There are some variations.  In some cases, there are parts that have a range of performance due to tolerance variations and process variations.  Sometimes companies sort the parts and offer different grades.  Not all buyers are aware when lower grade parts (say, they might not meet the top grade temperature specs) are being offered to them.

    It has been a problem with contract manufacturing that “scrap” goes out the back door.  I would say in my experience this has decreased in the last 10 years in China, but it depends a lot on who you are dealing with.  Auditing is required to stop this.

    Another issue is factory buyers at CM or even company-owned sites in China etc. may do things they should not do to manage cost.  Often they are highly incentivized to “save” and will trade quality for price with suppliers.

    Some of the selling of out of spec parts is not condoned but happens due to poor audting control as well.  I would say that an out of spec part sold onto the broker market is equivalent to counterfeit.

  24. goafrit2
    February 14, 2013

    >>  Not all buyers are aware when lower grade parts

    Do you think we need to have more disclosures on this? I think it may be necessary in the industry.

  25. goafrit2
    February 14, 2013

    >> Some of the selling of out of spec parts is not condoned but happens due to poor audting control as well.

    This seems to be a well-structured business model for many U.S. firms. They simply name the product something different and sell it. However, they are very carefull to update the datasheets accordingly.

  26. eafpres
    February 14, 2013

    The situations I have seen in this regard are inexperienced buyers under pressure to find parts.  If a manufacturer is sorting for performance levels, they go to lengths to keep them straight.  This isn't the same as reject parts getting into the supply chain.  I think in this case it is up to the buyers to ask questions and keep their engineering engaged as well.

  27. eafpres
    February 14, 2013

    In my experience the US parent company did not condone, authorize, or organize “business models” around selling bad parts.  On the other hand, what you refer to is yet another scenario for buyers–factory generates part A out of spec; datasheet for relaxed specs for part B is created and parts go on offer.  Again, the buyers need to know what they are doing.  

    Has anyone had the experience where a cross-reference service has matched a part but there were specs that didn't match, and did it affect you?

  28. goafrit2
    February 21, 2013

    >>  I think in this case it is up to the buyers to ask questions and keep

    Most times, that is not possible. These companies may not reply if you are not a big customer. I think a startup has no chance.

  29. goafrit2
    February 21, 2013

    >>  Again, the buyers need to know what they are doing.  

    It is possible you are writing from the perspective of a big company where you can use size to extract data from the big vendors. Vendors categorize their customers based on their purchasing power. If you do not have the financial influence, you may not a lot of feedback.

  30. eafpres
    February 21, 2013

    @goafrit2–I agree with you; however, although today the fashion is software/internet startups with mainly youngsters with great ideas; in my career hardware startups often had very experienced people either on the team directly or on the board, and those people with 20+ years in the industry had lots of connections.  So I would say hardware startup is very different from software or services startup.  But you are right, it is hard for a small, unknown company to get a lot of support from semiconductor companies.

  31. Brad Albing
    March 28, 2013

    Quite right – unless they are rad-hard or mil-spec parts – then they are almost certainly from the US

  32. Brad Albing
    March 28, 2013

    That could turn into a serious problem – and a fire hazard.

  33. Brad Albing
    March 28, 2013

    So that would also muck up the impedance of the connector – more trouble.

  34. Brad Albing
    March 28, 2013

    So first, they need to teach the young whipersnappers which end of the iron to hold. And then, teach them to not hold the solder in their mouth like I used to do.

  35. TheMeasurementBlues
    March 28, 2013

    Brad,

    Someblody is teaching soldering, at least here in Boston. The local IEEE is offering a soldering course. We discussed this rather heavily in Soldering for Engineers. Some say that although soldering is never something employers look for on an engineering resume, it's a skill you should have. Others say everhting is done in simulation now.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.