If you frequent our E2E forums they can be like detective mysteries. They are brain teasers that exercise and build your troubleshooting skills, solving analog problems with limited information. But not all these mysteries can be solved, given the information provided in the original posting.

“My amplifier circuit doesn’t work. Can you tell me what is wrong?” Seriously, that is nearly the extent of the information provided in some of our E2E forum postings. No schematic. No description of the problem.

If you are reading this blog, it’s a sure bet that you wouldn’t post a question like that example. Still, there is an opportunity to tweak the quality of information you submit on our forums to speed you to an accurate and complete solution. Here’s a checklist for possible items in your posting. Of course, most can omit some or many of these but consider the possible value that each may have in your inquiry:

Schematic— Word descriptions of circuits can be tricky, even if your circuit is very simple. There’s always a possibility of misinterpretation or confusion so it’s best to include a schematic. And the details you provide in your schematic can be crucial. Check out this example:

Power supply voltages— Your schematic may not show the power supply voltages and this can be important in solving many problems.

Oscilloscope captures— A picture is worth a thousand words. Be sure to provide scale factors and settings that could be important if they are not captured on screen.

Oscillations— If you are reporting a problem with an oscillation, its frequency and magnitude are valuable information in diagnosing the cause.

Noise— Merely reporting that you see excessive or unexpected noise is too vague. What is the magnitude? Describe it. Is it random thermal noise? Or is it correlated in some manner suggesting an oscillation or some external source? Again, a scope shot is best.

Product model numbers— Be sure to provide accurate product model number of all ICs in the circuit. You might need to use an IC from a competitor (rarely, I’m sure). We need to know.

Expected performance— Occasionally the reported results seem to fit the behavior of the circuit provided. Tell us what you expected the circuit to do.

Measured DC voltages— If your circuit seems totally out of whack, provide measured DC voltages to help us debug.

The load— Load conditions such as low impedance, high capacitance or long cables frequently influence circuits. Provide these details if it is outside the ordinary.

Bypass capacitors— If your schematic does not show bypass capacitors on the power supply, tell us that you have included them in your circuit.

Your troubleshooting efforts— Tell us whether you tested more than one IC? What else have you tried.

The TI engineers and other forum contributors who troll the forums are great at what they do. They’re knowledgeable and amazing at solving problems without ever physically touching your circuits. But they all flunked their Clairvoyant Troubleshooting class. Help them out with some quality information and they will speed you to a solution. And maybe you would like test and build your troubleshooting skills?  Join us on the precision amplifiers forum. You may be the hero who helps solve a case. More tips on posting.

Did I forget something important on this list? Think about it and add it in a comment below. And, by the way, what’s the fatal flaw in the example circuit? (No TI peeps allowed.)

Thanks for reading,

Bruce           email:  thesignal@list.ti.com

          Index to all The Signal blogs.

Anonymous
  • While not being an analog guy in the first place, I don't think the proposed solutions are good ones. Yes, the problem has been identified (V+ vs. U1 gain), however, all solutions introduce new problems.

    A missing info is the voltage swing on U1+. SInce it is a mic input, it _does_ have a swing (or else the whole circuit is useless). How much is it?

    A coupling capacitor on U1+ would eliminate the DC offset. However, the OpAmp only has a positive supply, so an AC coupled input signal would bring it down to the GND rail (if it comes down that far at all - there is no output load specified and I didn't check the datasheet - has it R-R-outputs at all?) for half of the input wave. Since U1 does not have a negative supply to support a plain AC signal.

    Same for the proposed capacitor between R2 and GND. Not to mention that both solutions have an influence on the frequency response.

    Aleksander was on the right track, but looked into the wrong direction: U2 provides a virtual GND (from low-pass filtered U1 output signal), so U1 output swings around U2 output. That's fine. But since VCC is way too low for the input voltage (it surely wold work with a +12V supply), the input offset needs to be reduced. Reduced, not eliminated. Best way would be to add U2 output to U1- input. Attenuated to a level where it doesn't eliminate offset but lowers it enough to keep U1 output below 5V and above GND (includign th eunknown voltage swing). A rough estimation would give a ~1:6 attenuation. So a 4.7k or larger resistor between Vout- and U1- shoudl bring the circuit into a working configuration.

    BTW: I like the list of required info. I should set-up a similar list for the MSP430 microcontroller forum. While the info required is different, it is usually missing to the same extent when someone asks for help there.

  • 1. Gain of U1 is 1+R3/R2= 1+2.2=3.2. So, 2.1V*3.2= 6.7V. Power supply is only 5V?

    2. gain of U1 is 3.2, gain of U2 is R6/R5=1, so, output is not fiff.?

  • Thanks for your comments, Leendert. You have identified my preferred solution to the non-functioning circuit shown in the blog.  I also appreciate your suggestion to include circuit board layout or photo in posting questions. Though unneeded for most questions, when more subtle problems arise this can be helpful. Layout related problems, however, are among the most difficult for us to solve remotely. It generally requires hands-on troubleshooting to diagnose.

    Regards-- Bruce.

  • Hello Bruce,

    ? add to the list ?

    A layout, or better, a photo of the build circuit may be handy, as the layout influences the behavior of the circuit, even if it is a 'simple one'.

  • Hallo Bruce,

    I just printed the example schematic and I'll put it on the 'coffee-table' for my colleagues to have a look at it.

    I first estimated the DC-levels at U1. This op-amp will saturate to the the power rail.

    Inserting a 10u cap between R2 and GND will get the output back at 2.1 Volts, preventing the DC-current flow through resistors R2 and R3. (comparable to situation around U2)

    The list looks complete to me. I'll have a second look at it later on.