20m BPF Elsie Design vs homebrew - need help?

Discussion in 'General Technical Questions and Answers' started by S21RC, Jun 12, 2021.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: abrind-2
ad: Left-3
ad: l-gcopper
ad: l-BCInc
ad: Subscribe
ad: Left-2
  1. S21RC

    S21RC Ham Member QRZ Page

    Hello All,
    I have made a 20m BPF based on the idea of K4VX published in QST 1988 (https://www.arrl.org/files/file/Technology/tis/info/pdf/8809017.pdf).

    I have used ELSIE to design the 3rd order shunt input bandpass butterworth where the center f is 14.2Mhz and Bandwidth is 6Mhz.
    elsie 2.JPG

    Now, I have used LC meter to check from a lot of capacitor and parallels few to get within +/- 2pf of the elsie design. Then used a miniVNA to adjust the inductor to make each L/C to resonance at 14.2Mhz. Core is T80-6 wrapped with plumber Teflon tape. used #21 SWG magnet wire.


    Problem is it shows 1.5db insertion loss (compared to 0.2db at elsie design) at 14.2Mhz. Also the impedance showing not 50 ohm.

    Is it ok to have 1.5db loss for bandpass filter? If not how to improve the 1.5db loss here?
    (I am intended to use such 20m and 40m BPF for two 100w stations operating from same location in a upcoming IOTA).

    The plot of the Elsie and VNA is below:


    Thanks in advance, 73 de S21RC
    Last edited: Jun 12, 2021
  2. K6BSU

    K6BSU Ham Member QRZ Page

    I don't recognize those "grey" toroids. At 20M, I only use "mix #6 (Amidon)" That is a yellow core, iron opowder.

    Good luck winding toroids inductance to 5 decimal places! (In my engineering courses I was taught to ignore even the 2nd decimal place)
    N1OOQ likes this.
  3. VK2TIL

    VK2TIL Ham Member QRZ Page

    Elsie assumes rather high values of component Q, not always achieved in "real-world" components.

    Your #6 iron-powder is a good choice for that frequency but your capacitors may not be ideal.

    Use higher-Q capacitors such as NP0 or C0G ceramics for QRP or receive-only; 500v mica types for higher power.

    Polypropylene capacitors are not usually thought-of as RF-suitable but might be a low-cost alternative to mica types; I have used them as RF coupling capacitors and had no problems..
  4. S21RC

    S21RC Ham Member QRZ Page

    Thanks for your feedback, those are yellow mix#6 but looks gray as I have put some teflon tape on it :)
    There is no way I am measuring the inductor to all those decimal points, after getting in the ballpark I used a miniVNA to tweak the inductor to get in the resonance frequency.

  5. S21RC

    S21RC Ham Member QRZ Page

    Thanks Kerry for pointing out the capacitor Q issue. Now after reading some about it I get the point. Previously I only thought about the voltage of the cap, so ordered the high voltage ceramics (1-2kv) [the 47pf disc capacitor in middle was for test only] :(

    I will buy some silver mica capacitor, hopefully that will reduce the loss and make the plot sharp. Will update again.

    Thanks 73 de S21RC
  6. S21RC

    S21RC Ham Member QRZ Page

    Opps, the post title has an unintended "?" mark and now I can't edit it. My apology (happens when you spend hours to tune inductor ;) .
    K4PP likes this.
  7. VK2TIL

    VK2TIL Ham Member QRZ Page

    We sometimes worry, in filter design & construction, about the Q of inductors but forget about the Q of capacitors; that doyen of experimenters, W7ZOI, wrote a very good paper (attached) on the subject of Q.


    Attached Files:

    S21RC and K4PP like this.
  8. SM0AOM

    SM0AOM Ham Member QRZ Page

    Unless you have a measurement error somewhere, the Q values of your components appear to be very low.

    Elsie, and other simulators, assume ideal or lossless components unless told to include the "physical model" in their calculations.

    I made a script, using the filter tables in my old "Taschenbuch der Hochfrequenztechnik", that modelled the filter insertion and return losses with inductor Q as parameter.

    In order to arrive at your figures, it would take an inductor Q in the order of 50 at 14 MHz, which is very low.

    According to the toroid core manufacturer,

    the inductor Q should be in the 250 or so range which would provide an insertion loss of around 0.3 dB.

    Suggest that you enlist the help of someone with a Q-meter in order to find out what is going on. An LC-meter does not show the losses at radio frequencies for typical inductors.

    Last edited: Jun 14, 2021
    S21RC likes this.
  9. S21RC

    S21RC Ham Member QRZ Page

    Thanks Karl, As suggested by Kerry/VK2TIL above I will assume now that the "very low Q" is the cheap ceramic/disc capacitors in the circuit. I have ordered some assorted Silver Mica capacitor from China (Mica cap locally not available). If after putting the silver Mica the plot do not improve then I guess the culprit is the Yellow T80-6 core which is also brought from China (these cores are not available locally as well) - and bringing from USA the shipping is often prohibitive.

    You may have noticed my call sign, I won't dream to find a Q meter here in Bangladesh :)

    I will measure the inductor only and take some graph data using the miniVNA to see what it shows.

    73 de S21RC
  10. SM0AOM

    SM0AOM Ham Member QRZ Page

    The Q-meter is one of the most versatile instruments available for the practicing RF engineer.

    It is for this reason that the Boonton 190A and 260AP have prominent places in the lab, taking "places of honour" alongside the Rohde&Schwarz VNA.

    S21RC likes this.

Share This Page