This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

  • Resolved

THP210: Issue Fully differential Amplifier

Prodigy 90 points

Replies: 17

Views: 354

Part Number: THP210

HI,

I’m trying to use the THP210 amplifier to get a fully differential amplifier. I followed the designed made by FilterPro Desktop by TI.

Here is the provided schematic.

Here is the implemented schematic in altium designer.

and here is the implemented layout

In the previous image plane +5V and plane GndA are hidden.

I connected gnd to In- and a sine wave between 0.5V and 4.5V, with mean voltage equal to 2.5V and 50Hz of frequency as shown in this picture.

Here are the signals at the inputs of the THP210

and here are the output

It seems that the opamp is stacked.

What’s wrong in my design?

Thank you in advance for your suggestions.

  • Hello Claudio,

    I will be reviewing your design using SPICE/TINA simulation,  I expect I will get back to you later today, or by tomorrow at the latest.

    Many Thanks,

    Kind Regards,

    Luis

  • In reply to Luis Chioye:

    HI Claudio,

    I simulated the circuit on TINA and the AC Frequency response and Transient response of the filter, and the circuit at first view appears to work fine on AC small signal and Transient simulation.

    Is the issue occurring on both stages of the THP210 (U5 and U6)? 

    Is this issue occurring in one PCB or several PCBs?

    The PD pin 7 is usually pulled-up to the supply (connected high); this is how PD pin is setup on the EVM; and appears to be the recommendation on page 24 of the datasheet.  Can you verify/check the voltage with an external meter on PD is high on the PCB board?  I will need to verify with design to see if PD needs to always be pulled-up to the supply or it can be left floating.

    Thank you and Regards,

    Luis 

  • In reply to Luis Chioye:

    HI Luis

    Thanks for fast reply, issues is both stage U5 and U6 and it's this in more PCB, so there is a problem in layout,  the outputs have been reversed.

    Best Regards,

    Claudio

  • In reply to Claudio Maria Lamacchia:

    Hi Claudio,

    If you decide to change the board layout, one recommendation would be to place a 10kOhm pull-up resistor between the power down PD pin (7) and the +5VA supply.

    Many Thanks,

    Kind Regards,

    Luis

  • In reply to Luis Chioye:

    HI Claudio,

    If you are concerned with distortion performance of the circuit, a general recommendation is to use NP0/C0G ceramic or Polypropylene film capacitors in the signal path, and thin film surface mount resistors for best linearity and lower distortion.  Among ceramic capacitors, the COG/NPO grade offer the lowest voltage coefficient dielectric.   You may find 0805 footprint COG/NPO capacitors available up to 20nF.  Alternatively, you may choose polypropylene film.

    A suggested modified filter is below. Essentially, using the same design in FilterPro, I scaled down the capacitors by typing 20nF as the largest value in the circuit. The TINA AC frequency response of the modified circuit looks clean. I also performed the AC small-signal open loop analysis on both stages, to verify the circuit is stable.

    Please see suggested filter below,

    Thank you and Regards,

    Luis

  • In reply to Luis Chioye:

    Hi Luis,

    thank you for you answer.

    I checked that I wrong the electrical symbol of the ampiifier and I switched inverting and non-inverting output resulting in two positive feedback path that caused output saturation.

    Now I restored the correct schematic in altium as in Tina

    THe inverting input is a fixed voltage equal to 2.5V and the non -inverting input is a sine wave @50Hz between 0.5V and 4.5V.

    THis is the voltage at the output

    CH1 and CH2 are the outputs while MATH track is the difference between them.

    I can see that the common mode is 2.5V but it seems that the signals are not balanced. Why?

    The mean value of difference is not 0V.

  • In reply to Claudio Maria Lamacchia:

    Morning Claudio, 

    Yes, your measured output appear to be sitting a large differential DC offset. 

    1. Ground both inputs, and measure the final output DC voltages - be careful in these scope and DVM measurement to isolate a probe capacitance in making the measurement - I usually tack a 100ohm resistor on the output pins and probe on the other side of that

    2. Not sure why you are spending so much on the Vcm bias? Levels of simplification include

    a. Share the output of the 2.5V reference on each FDA

    b. Develop the 2.5V as a simple R divider from the supply with decoupling - and again you can share that on the two Vcm inputs

    c. Let the default internal mid supply default set up the output CM level, be sure to include a decoupling on those pins for noise

    Normally the output CM voltage can be a little less accurate as the next stage usually rejects it anyway (ADC?)

    Michael Steffes

  • In reply to Michael Steffes:

    Hey Claudio and Louis, 

    I went off to run the target design in my MFB design tool and did not like what I was seeing, fixed some things and did an example design here, just FYI, 

    500Hz 4th order Butterworth active filter with the THP210.docx

    here is the resulting TINA V7 file, 

    THP210 4th order 500Hz Butterworth.TSC

    Michael Steffes

  • In reply to Michael Steffes:

    Michael Steffes

    Hey Claudio and Louis, 

    I went off to run the target design in my MFB design tool and did not like what I was seeing, fixed some things and did an example design here, just FYI, 

    (Please visit the site to view this file)

    here is the resulting TINA V7 file, 

    (Please visit the site to view this file)

    Thanks a lot Michael 

  • In reply to Claudio Maria Lamacchia:

    You bet Claudio, I went on just now to verify the <$0.1 2% C0G C's in 47nF, it was actually Kemet automotive that had those, I usually assume at least 2% C's and 0.5% R's for active filters in E24 and E96 steps 

    And then here is the integrated noise comparison, this is just mainly the lower R's using the higher max C value. 

    Michael Steffes

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.