3
\$\begingroup\$

I used this online calculator to generator a LC low pass filter with impedance transformation for interfacing a Variable Gain Amplifier (VGA) with an ADC. However the result is for a single ended input and output.

enter image description here

  • R1 200
  • C1 7.95775e-11
  • L1 0.000127324
  • C2 5.45612e-12
  • C3 2.54526e-12
  • R2 2000

Would it be possible to use this as a starting point to figure out what values I'd need to do this differentially?

VGA: AD8369 has differential output impedance of 200 Ohm.

ADC: MAX1438 has differential input impedance of 2k Ohm.

The low pass has a cutoff at 20 MHz and the matching is for 5 MHz.

Edit: There is an example circuit in VGA datasheet, but it is for an ADC with a 1k input impedance and I'd like to also add a low pass filter in front of the ADC.

enter image description here

\$\endgroup\$
5
  • 2
    \$\begingroup\$ At 20 MHz, you very likely don't need to do impedance matching. You normally want to drive the ADC with as low a source impedance as possible. Do you have any requirement to have the amp and ADC far apart from each other? What is your sampling rate? \$\endgroup\$
    – The Photon
    Commented Feb 7, 2012 at 22:47
  • \$\begingroup\$ I wasn't sure how to interface them if I weren't doing any impedance matching. The amp and the ADC sits within an inch of each other on the PCB. I can go as high as 65 MSps in sampling rate, but I'd probably be operating around 40 MSps. Thanks. \$\endgroup\$
    – jeep9911
    Commented Feb 7, 2012 at 22:56
  • \$\begingroup\$ I tried simulating the circuit from the AD datasheet, and I can't work out what they had in mind. It is a LPF with about 80 MHz cutoff; but the input impedance has dropped very low at 70 MHz making more work (not less) for the amplifier output stage, and output impedance is peaking slightly up at 70 MHz, which would cause potential ADC loading problems (though they did use an internally-bufferd ADC, so its probably not a problem for them). Maybe you'll get lucky and someone else can figure out what they're getting at. \$\endgroup\$
    – The Photon
    Commented Feb 8, 2012 at 3:03
  • \$\begingroup\$ Where is the sample & hold (S&H) function? If the input changes during ADC capture, this would cause large errors? \$\endgroup\$
    – D.A.S.
    Commented Apr 10, 2012 at 5:37
  • \$\begingroup\$ Please add units to your components' values. And expressing a 5pF capacitor to 10aF precise is silly at least. \$\endgroup\$
    – stevenvh
    Commented Jun 10, 2012 at 11:00

1 Answer 1

Reset to default
1
\$\begingroup\$

Any single ended passive filter can be made differential by mirroring it, in your case about the x axis. Taking care to double or half the LCR components should they end up in series or parallel.

This is only useful if your driver and load are both differential. Otherwise just ground on input (no current would flow in the other half of the filter).

Try a SPICE simulation to check your changes, e.g. LTSpice.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.