frequency to voltage circuit with LM2907

Discussion in 'General Electronics Chat' started by Tamir, Dec 13, 2018.

  1. Tamir

    Thread Starter New Member

    Jul 11, 2018
    10
    0
    Hello Everyone

    I'm trying to build simulation model in LMSpice with the LM2907 but without any success.

    my input is 45V@8Khz and I need Vout=0-4.5Vdc
    I'm using the LM2907 model (see attached files) for the simulation,
    with help from the datasheet, I calculated the following parameters:
    C1=1.5nF
    R1=25Kohm
    I'm not sure what my C2 need to be because it depends on Vripple, so I tried many options,
    the biggest problem is the output voltage is always in uV no matter what the C2 will be.


    where is my problem?
     
  2. ericgibbs

    Moderator

    Jan 29, 2010
    8,520
    1,711
    hi Tamir,
    Welcome to AAC.
    Post your LTS asc circuit file, so that we can check.
    E
     
  3. eetech00

    Senior Member

    Jun 8, 2013
    1,616
    332
    The output will vary with frequency.
    What is the Frequency range at the input?
    And what is the shape of the input waveform?

    eT
     
  4. ericgibbs

    Moderator

    Jan 29, 2010
    8,520
    1,711
    hi Tamir,
    A couple of obvious points.
    The LM2907 is powered from 15V, you have a 50Vppk sine wave input signal, centred at +25v and its DC coupled to the LM2907 input.
    E
     
  5. eetech00

    Senior Member

    Jun 8, 2013
    1,616
    332
    You want C1 as small as practical.

    For 4.5v FS@8khz:
    If C1=1.0n and R1=40.1k (should be a pot) then for 1% ripple C2=24.8uf-> use 27uF

    You state 8Khz, but in the simulation you used 400hz. Which is it?

    Also, you need to drop the input signal to a useable level.

    eT
     
  6. MisterBill2

    Distinguished Member

    Jan 23, 2018
    3,682
    723
    45 volts as a signal is way more than you need with the LM2907nor LM2917. 2 volts is plenty enough for the input circuit. But beware of simulators because they can lie! The application notes will tell you about what signal you should have. I have built several tachometer circuits using the LM2917, which includes a built in voltage regulator. AND, very important, both the supply voltages and the input signal must be within the ranges described in the data sheet or you can not expect any useful functioning.
     
  7. eetech00

    Senior Member

    Jun 8, 2013
    1,616
    332
    heh...I disagree with that statement. Simulators don't have intent to deceive, but the simulation results can be inaccurate.

    eT
     
    ericgibbs likes this.
  8. danadak

    Distinguished Member

    Mar 10, 2018
    3,468
    778
    What is the accuracy you want and the resolution you are seeking
    for the conversion ?

    Regards, Dana.
     
  9. MisterBill2

    Distinguished Member

    Jan 23, 2018
    3,682
    723
    I was quoting the honorable Bob Pease, who never trusted simulators because the models they use are seldom accurate. So possibly "lie" is to strong a term. "Simulators may present results inconsistent with reality," does sound better.
     
  10. Tamir

    Thread Starter New Member

    Jul 11, 2018
    10
    0
    Hello

    firstly, thank you for all your help.

    the attached zip file is only the model I used for the circuit, it's not my circuit.
    I need to build a circuit that the input level is from external tachometer.
    the input level is pulse wave 45V@8Khz and I need the downgrade to 0-5V level so I can read it with MC.

    I attached my .asc file
     
  11. ericgibbs

    Moderator

    Jan 29, 2010
    8,520
    1,711
    hi Tamir,
    Look at this option.
    BTW: I have not optimised the comp values.
    E
     
  12. Alec_t

    Expert

    Sep 17, 2013
    10,243
    2,504
    The IC will only tolerate input voltages up to ±28V, so you will need to have an input attenuator or clamp.
     
  13. MisterBill2

    Distinguished Member

    Jan 23, 2018
    3,682
    723
    That circuit and the traces look reasonable, and with the attenuator at the input it does indeed function. But I am a little bit confused, in post #10, about "downgrading" a pulse signal. Converting a pulse train frequency to an analog voltage proportional to frequency is a bit different. And there is still a bit of ripple, which is unavoidable with the principle of the 2907 family. So at this point you need to decide between fast response and the uncertainty of reading a waveform with ripple. The faster and higher accuracy approach is a counter scheme that measures the time between peaks and updates every cycle. It is fast, accurate, and very computation intensive. So the question is how much resolution does the application require, and how fast of a response time does it need.
     
  14. danadak

    Distinguished Member

    Mar 10, 2018
    3,468
    778
    If you need more accuracy a UP doing a period measurement and converting that into a
    PWM duty cycle value (followed by LPF to get DC) can approach 12 - 16 bit accuracy fairly easily.

    Regards, Dana.
     
  15. Tamir

    Thread Starter New Member

    Jul 11, 2018
    10
    0
    now it work good, and I can see Vout changes as the Vin and/or the Vcc modified.
    Thank you.


    Now I need to do some reverse engineering, I need to modify this circuit so it will measure the output of torque sensor,
    the output of the sensor is phase displacement.

    how can I simulate phase displacement in LTSpice?
     
  16. ericgibbs

    Moderator

    Jan 29, 2010
    8,520
    1,711
    hi Tamir,
    For what sort of phase displacement range are you asking.
    E
     
  17. Tamir

    Thread Starter New Member

    Jul 11, 2018
    10
    0
    It probably will be around 8Ghz but I don't know the phase shift itself.
    thats what I want to simulate
     
  18. ericgibbs

    Moderator

    Jan 29, 2010
    8,520
    1,711
    hi,
    Is that meant to be 8GHz.?:eek:
    E

    EDIT:
    This one option [ using 1omeg for test]


    Added a 8gigaHz plot.

    AA1 18-Dec-18 10.07.gif
     
    Last edited: Dec 18, 2018
  19. Alec_t

    Expert

    Sep 17, 2013
    10,243
    2,504
    The LM2907 is typically used for frequencies in the low kHz range at most. I can't find any upper limit in the datasheet, but I can't see it working for anything much higher. Incidentally, that IC is now obsolete.
     
  20. Tamir

    Thread Starter New Member

    Jul 11, 2018
    10
    0
    it good option for 10 different signals, but the sensor output is one signal with changed phase.
    I tried the SFFM option and it looks good, it give me an output similar to the sensor and with changed modulation index it change the phase.

    I will try couple more option to figure out how make phase modulation input
    Thank you
     
Loading...