Op amp simulation for timegrapher

Discussion in 'Programmer's Corner' started by Rafał Kotrys, May 7, 2016.

  1. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    Hi Folks,

    I'm writing to ask for ciruit validation. I'm very begginger to electronic and my knowledge was based on some polish books and some of texas instruments manuals I could find in the internet. I'm trying to build a timegrapher and I "pre-designed" the circuit to gather sound signal from the watch. According to my assumptions cirucit should consist of:

    -stabilized power supply +12 V
    -op-amp to gain microphone signal
    -high-pass filter
    -inverting comparator which should give output signal to uC

    I have some concerns about if I'm doing everything properly. I list them below in description of each part. I'd appreciate if You could help me in solving it:

    -Stabilized power supply:

    Designing this part was based on internet tutorials. As a V1 I used 18V/1A stabilized power supply. Then I made gretz bridge and split it to receive two receive two types of voltages:
    +12V - to supply op amp which required 6-16V voltage
    +5V - to supply uC
    In both sources of supply I used stablizer series 78xx respectively to voltage I wanted to receive. Before and after stabilizer Iput capacitors 0,1uF to filter high peak voltage noise.
    To pre-stabilize voltage in both cases I used 1000uF tantalum capacitor. I read that easiest way to find good capacitor to filter is to use a rule which states that for each 1A of supply current I should use 1000uF filtering capacitor.

    1) Question: In this case current splits equally 500mA each wire. Shall I use 500uF instead?
    upload_2016-5-7_20-3-3.png

    When I made simulation everything looks correct: Voltage measured close to R1are around 5V DC and same right after 12 V stabilizer. Below screenshot:

    upload_2016-5-7_20-13-32.png

    Next is op-amp. Here I used ready to work kit from some polish online shop. It is dedicated to electret microphones and according to specification has 100x gain. Power supply should be from range between 6-16V. Below I upload the circuit
    upload_2016-5-7_20-23-15.png

    I made a simulation of whole cirucit. Below I'm uploading whole ciruit from LT Spice so You can quickly checked against above picture if anything is wrong.

    upload_2016-5-7_20-35-40.png
    Resostpr R8 represents rest of ciruit.
    V2 = 0,01V/1kHz represents signal from microphone.

    I made simulation to see gain of an op amp and if everything is right. Below are the plot of
    -non inverting input signal (+) - (Red line)
    -output signal of op-amp measured right before C11 - (Green line)
    -output signal measured between R6 & R8 - (blue line)

    Plots look respectively:
    upload_2016-5-7_20-45-28.png

    2) Why signal till around 0,35 looks so strange?
    3) What causes and why such big drop of voltage between green and blue signal? There is no gain between input (0,01V of microphone singal) and blue one?

    Filter design:
    I made high pass filter based on the texas instruments manual "design of filter in 30 seconds:

    http://www.ti.com/lit/an/sloa093/sloa093.pdf

    I'd like to connect it to the circuit (in R8 place):
    cut off frequency = 10 kHz
    C1=C2=3,3nF
    R1=R2=3,4kOhm
    Cout=3,3uF
    upload_2016-5-7_20-50-46.png
    4) Will that designed filter work with op amp?
    5) Can I use simple TL072 Op amp or shall I use different solution?

    Comparator:
    Do You have any data how should I could calculate resisor/capacitor values for comparator?

    Thanks for Your help and validation.
    Regards
    Kuba
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,647
    2,346
    Hello,

    The TL072 has not enough power to drive a 10 Ohms load (R8 in your simulation).
    Try to repeat the simulation with an 2K resistor as load.
    See this graph for the effect of the load resistance on the output voltage:

    TL072_max_output_voltage_versus_load_resistance.png

    Bertus
     
  3. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    upload_2016-5-7_21-13-17.png

    Bertus,

    Thanks for Your hint!

    I made once again simulation and right now plot of R8 is red. It's still on a level of 0,5V so rough calcualtion will give:

    0,5V/0,01=50x so right now it seems to be good. I guess that I still would need additional 10x to use this signal as an input for comparator, am I right?
     
  4. bertus

    Administrator

    Apr 5, 2008
    15,647
    2,346
    Hello,

    I have edited my post.
    The datasheet recommends a 2K load resistor.

    Bertus
     
  5. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,779
    1,103
    In answer to your questions:
    1) Since your circuit shown uses only a few mA, 470uF should be enough but it won't hurt to have more.
    2) C10 has to charge via R7 and R5.
    3) R8 is too low a value and effectively shorts the output to ground.
    4) It should, but wouldn't a band-pass filter tuned to the frequency of interest be better?
    5) TL072 should work.
    6) Why do you need a comparator? Can't you feed the filter output directly to the micro?
     
  6. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    Bertus Thanks for Your reply!
    Does that mean that before I add filter as a next element I should add load resistor before to enable correct work of op - amp?

    Alec_t thanks for Your input!

    4) It should, but wouldn't a band-pass filter tuned to the frequency of interest be better? - Add the beginning I was thinking about band-pass, but frankly speaking I made a simple check in audiacity and measure an impulse of watch with iphone. Results seemed preety good (picture below) therefore I stayed with high-pass.
    upload_2016-5-7_21-45-47.png - non filtered signal
    upload_2016-5-7_21-45-57.png - 10kHz fhigh pass filter

    6) Why do you need a comparator? Can't you feed the filter output directly to the micro?

    I don't know if it's a good answer, but what I was thinking was that microphone signal which I would need to measure is going to look like below:

    When You are measuring mechanical signal from watch You are interested in three top signals visible below. In fact there are more signals in behind which may be a false input for measurement. To ensure it is corect I'd like to setup kind of treshold which would let me filter only required signals (picture below). I wanted to put on Vin V supply ( I planned to supply it with 5volts) connected via potentiometer which would let me fluently regulate treshold. As a Vref I planned to put microphone signal.

    Here I have concerns. My concern is if such big frequency will be send to uC as a huge number of signals related to frequency, not only three (I hope You catched my point). I was wondering If I shouldn't send instead af raw signal only the signal envelope. If so, how to do it cause I have no idea.
    upload_2016-5-7_22-0-30.png
    upload_2016-5-7_21-54-50.png

    Regards,
    Kuba
     
  7. bertus

    Administrator

    Apr 5, 2008
    15,647
    2,346
    Hello,

    Between stages of opamps there is no load resistor needed.
    Have a look at the attached PDF for a lot of opamp information.

    Bertus
     
  8. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    Thanks for manual. I'll take a look at it. Is it possible You could also provide me some hint about above solution I proposed?

    I'd like those three peak of watch to be measured kind of this and send signal directly to uC

    upload_2016-5-8_10-40-4.png
     
  9. bertus

    Administrator

    Apr 5, 2008
    15,647
    2,346
    Hello,

    If you want to detect signals above a certain level as a block as you show, I would use a envelope detector with a comparator at the wanted level.

    An envelope detector will follow the shape of a signal:
    https://en.wikipedia.org/wiki/Envelope_detector

    Bertus
     
    Last edited: May 8, 2016
  10. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    Hello. It's me back again. I got to ask for Your further help in my sensor design. I added next part of circuit to the siumulation, which is 10kHz filter. I made simulation once again to check how the voltage signal will behave and what frequencies will be. To my surprise I observed big voltage drop. before and after op amp Can anybody explain me the reason of it?

    Refering to the circuit:

    upload_2016-5-9_21-7-56.png

    Red line is voltage after measured after C11 capacitor
    Green line is voltage measured after C12 capacitor
    Blue line is measured for filter output.

    Voltage drop is significant + my signal amplitude dropped a lot. My filter was designed to filter out frequency above below 10kHz. Frequency of the microphone signal was 20kHz and amplitude (blue) at outpus is so low.
    upload_2016-5-9_21-7-27.png

    I'm very rookie to the topic and I'm really nuged by time here so I'd really appriciete Your hints how to sort out problem. I'll owe You a beer!
     
  11. bertus

    Administrator

    Apr 5, 2008
    15,647
    2,346
    Hello,

    The opamp U4 has no DC bias.
    You could make a 1:1 voltage divider and lift the ground of R8.
    Use a decoupling capacitor on the lower resistor in the voltage divider.

    Bertus
     
  12. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    I cut the screenshot of the circuit therefore You didn't see the bias of op amp U4. I'd add capacitor but I don't have bloody Idea what value should I pick. I used the same value as two other capacitors: Results were the same as previously.

    upload_2016-5-9_22-24-1.png

    And here the ciruit without supply as we already talked it's ok
    upload_2016-5-9_22-24-23.png
     
  13. Rafał Kotrys

    Thread Starter New Member

    Apr 22, 2016
    12
    0
    Hi Guys,

    Can anyone help me on above? I really need to complete this sensor Today and do not have a clue what is wrong:(

    Good Will people, I'm looking for You:) :)

    Regards,
    Kuba
     
  14. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,779
    1,103
    Try this:
     
  15. dannyf

    Well-Known Member

    Sep 13, 2015
    1,779
    360
    maybe your filter is incorrectly designed?
     
Loading...