• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 47 subscribers
  • Views 1540 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

THS4131 Active Filter Cutoff Question / Issue

jbuckf50
Prodigy 80 points

Other Parts Discussed in Thread: THS4131

Hello.  I'm using the THS4131 as a single-ended to differential converter at the output of an amplifier circuit.  I've pasted a portion of the circuit schematic relevant to this discussion below.

The configuration is similar to Figure 38 of the datasheet, with a capacitor in parallel with each Rf to get a single-pole low-pass filter.  

Using a 220pF capacitor and Rf=4.99k, I calculate a cutoff frequency around 140kHz (fc=1/(2*PI*Rf*Cf).  I confirmed the cutoff in simulation using LTSpice prior to PCB fab.  

The problem is this - in test I measure a cutoff frequency in the 90kHz range (equivalent to a 320pF capacitor).  The PCB has 10 channels and they are all consistent.  I verified the proper components were purchased and installed.  I also confirmed the stage preceding the THS4131 has a wider bandwidth - i.e. the THS4131 is the bandwidth-limiting stage.  

Regarding the layout, I'm using a 10-layer board, 0.063" thick.  I followed the layout guidelines in the datasheet except I left the ground plane underneath the part pins, and added a ground pad underneath the part body for thermal reasons.  I can see how this would add 10s of pF, but not 100pF (unless I'm missing something).  The PCB is straight from the assembly house so it's reasonably clean.  

I can experimentally figure out the new capacitance needed to get a bandwidth of 140kHz, but I'm not quite clear why these values result in a bandwidth much lower than expected.  Any ideas?

It's worth mentioning the load during bandwidth measurements was a high-impedance oscope channel, with coax length <6ft.  

Thanks!

Josh

  • 0
    TI__Expert 4950 points

    Hello Josh,

    I believe you are correct in your assessment.  To have the -3dB at 90kHz instead of 140kHz as calculated would indicate that you have 1.3dB loss in either the previous stages, as the loss would accumulate, or in the measurement.

    Have you tried placing your source directly on R8? On using the scope, the cable may account for some loss due to the added capacitance on the 1Mohm input impedance.  1 1.3dB loss roughly corresponds to 140kHz bandwidth which is equal to 1.2pF in parallel with the 1Mohm input impedance of the oscilloscope.