Green Bobin Filter

Design And Reality

For even complex TTC, I come to software instead calculation by myself (Even for a DTC -:).

It's quite simple design process with help of EMRFD.  I choose 47p Series End C.  default bandwidth 250 Khz.  Such a inductor might have a Q  factor range from 80 to 100, I suppose.

Unfortunately,  filter according this calculation perform badly, insert loss 20db. Something going so wrong. So first thing is checking the Coil Q  Factor.

Adjustment

On 7Mhz, the Q of  this Coil is only 50 ! So redesign with unloaded Q=50, and must broaden the  bandwidth to 500 kHz.   Design given a insert loss about 4.3 dB.

Circuits testing still had large insertion loss. After several round change and test,  I found two key factor affect it so much.  

The most important one is input transformation capacitor.  Use 150pf:220pf instead of 68p:53p, the insert loss reduce  from 20 dB to 6dB !  The side tuned tank with higher frequency, and the middle one resonated at a lower frequency.  

Reduce the insertion loss to minimal is almost identical to adjust the RL to biggest one. Too badly S11 leading to high insertion loss.

And then tune the two side coil slugs could reduce the insert loss further, to 4.5dB. Or slightly shift the band pass up/down.

The final result, I got a band pass filter just by tuning and experimental:

• 6.85Mhz - 7.25 Mhz,  400 khz BW

• insert loss 4.5 dB

• Return Loss  -20dB

All of this  indicate the design parameters selected for this coil is not the actual value.  The main parameter is the Q and series capacitor. 

To enable slightly shift the frequency and  given enough tuneable margin, this Green Bobbin Coil better to use following design parameter:

• nominal inductance 2.85uH

• Q=60

• reduce the transformer ratio given by design

By this way, the most suspicious problem is what the filter exactly is .

Make All Correct

Since all these experiment, something learned:

• The middle resonant tank dominating most of insert loss. If tuned central resonator could not reduce IL to around -10dB or less,  there gonna something wrong.

• Two side resonator could optimized the RL, or slightly shift frequency response.

• Need checking 2 to 3 Mhz wide band width response to make sure the filter is perform well. There should not be bumpy.

Then, use following parameters for this green bobbin coil, the design and actual circuits match very well.

• nominal inductance 2.9uH

• Q=65  (60 ~ 70)

Design with 7Mhz, 2.9uH, bandwidth 400 Khz, estimation IL 4.7 dB,  actual circuits Insert Loss 4.6 dB.  Actually band width 420 Khz, center frequency 6.95Mhz (6.75 - 7.17)