Topics

Bandpass filter measurement anomalies - problem identified

Steven Dick
 

Problem has been isolated on my Type II PHSNA. The ERA-3+ is causing the harmonic distortion. I first tried reducing the amplitude of the input signal to ther ERA-3+ by crude tests - jumpering the output of the first attenuator at R3 with various resistors. I realize it was mismatched, but ok for this crude test. The amplitude of harmonics went down a bit but they were still there.  Then, I eliminated the ERA-3+ in  the signal path by the following crude method: I soldered a jumper across R2 to increase the amplitude, removed the jumper at pins 2-3 on P102 just after the anti-alias low pass filter, and took the output directly from pins 1 and 3 to a coax to BNC jumper to drive my filters.  This gave me about -7dbM of output power, albeit mismatched all over the place. The response was beautiful!! - no more jiggles due to harmonics on the left slop of the bandpass filter responses. 

The harmonics don't show up if you are testing lowpass filters.  The harmonics don't show up if you are testing narrowband crystal filters. The harmonics are quite obvious when you are testing LC bandpass filters on their left edge.

For the Type II PHSNA, if you remove the first attenuator and drive the antialias filter directly from the push-pull transformer, you have 6dB more power output than the Type 1 as it is no longer doubly terminate.  This is permissible per the AD9851 data sheet.  My recommended fix is to eliminate the first attenuator, take the output from the antialias filter, and use it to drive the final 6dB output attenuator from the output of the antialias filter after removing the components in and out of the ERA-3+.  I'll do a write-up on this AFTER field day.

Regards,
"Digital Steve", K1RF

 

Steven Dick
 

I updated file "Bandpass filter measurement anomalies R1.pdf" in the K1RF filter showing improvement of LC bandpass filter responses on the left edge using the PHSNA Type II with and without the ERA-3+ output amplifier.

Regards,
"Digital Steve", K1RF

Steven Dick
 

I updated the file to rev2: "Bandpass filter measurement anomalies R2.pdf" in the K1RF folder in the files section showing improvement of LC bandpass filter responses on the left edge using the PHSNA Type II with and without the ERA-3+ output amplifier. I included circuit mods to eliminate the ERA-3+.  Don't do the change if your PHSNA application is working fine for you.  The mod if useful if you need to do accurate measurement on relatively wideband LC bandpass filters on the left side of the filter response if you see unexplained "jaggedies".

Regards,
"Digital Steve", K1RF