Multistage Operational Amplifiers with gain and offset

Discussion in 'The Projects Forum' started by BastiaanMDC, Oct 28, 2015.

  1. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Dear Support, Designers,



    I am not so familiar with Opamps, but I would like to combine the following circuits with each other. and would greatly appreciate the help I can get in designing this. I have red some theories about opamps, books like opamps for everyone, simulated with Qucs and I aint getting the results I expected using the program, I tried configurations with Differential and selfbuild Operational amplifiers but to no avail, you are probably familiar with this.



    Information

    What I do know is that the input and the output are connected to a 50 Ohm BNC cable.

    0.15V TTL input unipolar and like to amplify it towards maximum 5V with an Potentiometer, and should have a variable offset also connected with a potentiometer.

    Configuration(required)

    AC to DC Adapter 12V 400mA

    LTC3265 Voltage Pump for stable 15+-V

    Low pass filter 6 order, gain, offset, voltage follower.

    5x TLV074 for the opamps



    Question:

    is the configuration in correct order? filter, gain, offset, voltage follower?
    I havent found that much on google about this.

    Does the configuration of the opamps need to be all inverted due to negative feedback?

    what do I need to keep in mind? as in where to start with, and the steps to follow?

    is there a general rule for configurations like this?

    as a last thought, is it advisable to place extra ELKO's between the 15+- to common? to prevent disturbances from common mode?



    sincerely yours



    Bastiaan
     
  2. ScottWang

    Moderator

    Aug 23, 2012
    4,853
    767
    What you have to is to draw a block diagram and upload to the forum.
     
  3. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Added a Block diagram, I excluded the LTC3265 for convienience.

    the caps on the V15+- are all 10uF

    IMG_0274_1024x768.JPG
     
  4. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,533
    1,251
    What is the overall application for this circuit? Is this a sinewave generator?

    Also, all of the stages are DC coupled. Have you worked out the maximum offset error for the circuit? Is maintaining the input DC level important to the output? If not, an AC coupled stage will relieve the accumulated offset error.

    ak
     
  5. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    The overall application is to amplify small AC signals (block pulse shaped/TTL)
    + Offset +-2.5V on top of the ground.

    No it is not important to have a offset in the beginning of the circuit,
    the reason for this was I was not sure what the circuit should do.( I do know that it will bring in more offset error, noise etc)

    For each of the OPA177 the max offset is 25uV
    the TL072 for the offset and amplification have an offset of 6mV (which is quite a lot)
    the Buf634 I didnt found out the max offset error, but its still the best on the market.
     
  6. dl324

    Distinguished Member

    Mar 30, 2015
    3,242
    619
    I'd move the filter ahead of any gain stage and replace the 10uF decoupling caps with 0.01uF ceramic. If the input and output to your circuit are 50Ω (not just the cable), you should match impedance so you have the correct voltage levels.

    Schematic critique:
    • It's nice to see care taken in drawing a schematic. Straight lines are visually pleasing...
    • Since you're placing decoupling caps on all of the opamps, omitting them and stating in a footnote that all opamps are decoupled would make the schematic less cluttered.
    • Same for the opamp supply connections.
    • Draw resistors/caps vertical whenever possible.
    • Avoid unnecessary wire jogs.
     
  7. ScottWang

    Moderator

    Aug 23, 2012
    4,853
    767
    What do you mean by TTL, is that Transistor Transistor Logic?
     
  8. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Thanks Dennis:),

    I know it was quick drawing because my pcb designer program killed itself on the moment. Could you tell me why I should filter first then amplify?, (I am working with low voltage signals 0.15 to 5V, 33/32 amplification).

    Ill replace the caps with 0.01uF MLCC ones and terminate the input and output with 50 Resistors.

    is it advisable to have a Trimpot at the TL072, for offset error comp? the environment is stable as in 22C°/71.6 Fahrenheit

    is there some point where I could optimise the schematic?
     
  9. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Well the scientists told me that is, the original selfmade device is in somewhere Europe, and I am on the otherside.
     
  10. dl324

    Distinguished Member

    Mar 30, 2015
    3,242
    619
    No filters are ideal, so your filter will pass low levels of out-of-band signals. If you amplify the out-of-band signals, the filter will have more to pass.
    You could also use 0.1uF. Whatever is handy; unless you need better high frequency decoupling.
    Depends on your accuracy requirements.
    That's a personal decision. Personally I tweak mine until I'm satisfied with my workmanship.
     
  11. ScottWang

    Moderator

    Aug 23, 2012
    4,853
    767
    What is the 0.15V signal looks like(sine wave, square wave ) and what is the meaning of 33/32 ?
    If the 0.15V is a normal TTL signal then it is only a low signal as ground, and how is the frequency?
     
  12. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    well Scott,
    the signal is used for measuring changes is a kind of a square wave, based on the info I got( plus Variable gain and offset).

    The 33/32x is for inverting and non inverting opamp configuration amplification.

    Work Frequency is under 1 KHz.
    ---------------------------------------------
    Dennis:)
    great, learning day by day:)

    last question: is it usual for small band amplification to put caps between the gain and the offset?

    Ill include a board if the design is finished.
     
  13. dl324

    Distinguished Member

    Mar 30, 2015
    3,242
    619
    Not sure what you're asking...
     
  14. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    based on Analogkid question, is it required to put caps between stages (In this case the gain and the offset stage)
     
  15. ScottWang

    Moderator

    Aug 23, 2012
    4,853
    767
    If that is square wave why you need so many stages of filter?
     
  16. dl324

    Distinguished Member

    Mar 30, 2015
    3,242
    619
    Still depends. Any DC component would be amplified by any gain stages. If that's a problem, AC coupling would remove it.
     
  17. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Ill put some headers in between just in case:)
     
  18. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    I Included the Gerber File, I got some comments from designer friends on the Output of the signal, because I have a 50 Ohm Load, at 5V max, I require a 100mA current.

    to do this I included a LM6181IM-8 Opamp used as a load driver, the LTC3265 isnt sufficient for driving large loads.

    -----
    one of the comments:

    You could start with a circuit like the one shown below. It uses eight TL074 op-amps (two quad packages). The following considerations apply:

    • It runs on + and - 15 V. To drive a 50 ohm load to five volts requires 100 mA, so make sure your power source (12 V adapter and LTC3265) can handle this. Bypass caps on the supply leads (not shown below) are necessary.
    • The TL074 op-amps are fine general-purpose FET input devices, good for DC through the low ultrasonic range (what frequency range do you require?) You'll need to check if their specs are suitable for your application. If they lack in any particular area, you could of course choose something with better specs. Naturally I would recommend ADI amps. I don't work for them, but I've found over many years that when I design in an AD part it just works.
    • Since your input and output are single-ended coax lines, there appears to be no need for a differential input. If such is actually needed, an instrumentation amp IC would be preferable to rolling your own IA with op-amps. The later method needs tightly matched resistors to preserve CM rejection.
    • The first two stages shown below are both inverting. Stage 1 has an adjustable linear gain of -20 to --30 V/V. Stage 2 has a gain of 4 with adjustable DC offset (bias), Overall gain is thus +80 -- 120 V/V. With a 0.15 V p-p input and a 50 ohm load, a 5V p-p output can be obtained at the load by adjusting the gain pot.
    • The third stage is a second-order filter. As shown it's a 10 KHz Butterworth lowpass, which you can easily re-design to your desired response using Analog Device's Filter Wizard at FILTER WIZARD | Analog Devices
    • The forth and last stage has five op-amps in parallel to handle a 50 ohm load, with current-sharing resistors. I recommend you substitute a dedicated 50 ohm line driver, as even five paralleled op-amps may have trouble with this load.
    I did run a simulation which worked fine, of course you'll need to carefully evaluate this or any circuit that you end up designing to be sure it meets your requirements,



    [​IMG]



    [​IMG][​IMG]
     
  19. dl324

    Distinguished Member

    Mar 30, 2015
    3,242
    619
    You're still putting your filter after your gain stage. If you have 10dB of gain in your first stage, you increase the amplitude of the out-of-band signals you want to filter. I wouldn't do that, but maybe it doesn't make a difference...

    To increase the output current drive, you could use an LH0002 or grow your own:
    upload_2015-10-29_7-44-24.png (From Nat Semi AN-4)
    To minimize crossover distortion, you can use the usual techniques to bias the output transistors so they're always conducting.
     
  20. BastiaanMDC

    Thread Starter New Member

    Oct 27, 2015
    23
    0
    Thanks for all the replies guys!

    First Ill start a new Topic, because I would like not to loose track.
    second I had to change the design for 1 application.
    still most of the designs here are not in vain and are a great help!

    new Thread:Multistage Opamp design 100KHz
     
Loading...