EMRFD Message Archive 13215
Message Date From Subject 13215 2016-10-22 09:14:57 k1rf_digital_stev... QRP order 3 Chebychev bandpass filter Hi all. I just posted a .pdf to the Files section entitled "QRP Order 3 BPF with bifilar-wound L1 and L3.pdf. The filter can be shared between receive and transmit functions in homebrew transceivers, providing their transmit output stage is push-pull to help suppress the second harmonic. These filters are based on the two part article in May and June 1998 QST articles entitled "Clean up your signals with bandpass filters". I made a QRP version of these using T50 toroids and 500V silvered mica caps. The pc board for these is 0.7 inches wide by 2.5 inches long and are compatible with the AQRP lowpass and bandpass filter footprint. You need a return loss bridge to tune these. They are not designed for "wind the toroids and forget". They will not work if you don't tune them. But if you are patient with tuning, you will end up with high performance filters that can be tuned in an orderly fashion. I shared the pc board at: OSH Park ~ K1RF BPF.zip
"Digital Steve:, K1RF
13216 2016-10-25 15:16:56 artus1947 Re: QRP order 3 Chebychev bandpass filter Steve,Thanks for doing this -- I actually have looked at the original writeup thinking it might be a good solution for the Hermes-Lite project I have been dabbling with. Your solution of bifiliar winding on t50s is very interesting. I have ordered some boards and am anxious to play with the design (I may breadboard one later this week to play.) Have you calculated any other bands? I ran the 40 meter filter in ELSIE to see your work and will plan on using it to calculate other bands if you haven't already done them. I finished a Scotty's Spectrum Analyzer last spring that should be useful in tweaking the toroids. Thanks again,Jack (W0FNQ) 13217 2016-10-25 15:39:37 k1rf_digital_stev... Re: QRP order 3 Chebychev bandpass filter Hi Jack. Let me know how they work for you. A few tips: I found that number 22 wire is as large as you want to use for L1 and L3 as it becomes difficult to wind bifilar windings for larger sizes. These turns stay in place fairly well during tuning. For L2 which will have more turns, I used number 28 wire for the ones I built. These turns tend to be slippery on the toroid and it is somewhat difficult to maintain tuning as they try to spread out after you let go of them. I found that if I use some polyurathane varnish and let dry or even hot glue on one end you can control it a bit better, and some wet varnish on the rest will help hold it in place during adjustment. This for me was the most difficult part of tuning. Also, for the bifilar windings, I made the mistake of spreading turns of the bifilar winding for fine tuning. Don't do this. You will get the right inductance, but you will have a slight impedance mismatch for the 50 ohm to 200 ohm impedance transformation. It won't affect the filter shape much, but the return loss will degrade a bit.
An easy way to use ELSIE for the bifilar wound toroids L1 and L3 and corresponding capacitors is to temporarily set the impedance to 200 ohms in ELSIE. I computed a few more bands but don't have a solid set of values and didn't test other bands yet. You can start with the original QST article as a starting point for parameters.
For my source of silvered mica caps, there are many sources, and I used justradios.com as they have a large set of available values at reasonable prices. But silvered mica caps are not inexpensive. Most of the ones I tested were raated 5 percent but were measured at 2.5 percent or better.
"Digital Steve", K1RF
13221 2016-10-26 18:57:04 artus1947 Re: QRP order 3 Chebychev bandpass filter Steve,Playing with ELSIE, I am thinking I may not have set it up quite right. I looked over the orginal Wetherhold articles as well as the July/Aug QEX article he wrote on "Receiver Band-Pass Filters Having Maximum Attentuation in Adjacent Bands." In the later article, he shows his ELSIE setup and the transformations he did before running. As you mentioned 200 ohms for bifilar winding is the right starting point for impedance and L2 must be increased by a factor of 4 and C21/C22 decreased by 4. I've posted in W0FNQ Folder of the Files section, the 40m ELSIE schematic and the insertion/return loss plots. My question is about the predicted values, the return loss and insertion losses are much higher than expected. On your runs, are you seeing better results. Are my topology and input values consistent with yours?Thanks for your advice73,Jack (W0FNQ) 13224 2016-10-27 05:02:54 k1rf_digital_stev... Re: QRP order 3 Chebychev bandpass filter Jack, Elsie cannot be set up to represent this type of filter with tapped (bifilar-wound) windings on the input and output inductors of the order Chebychev filters. ELSIE should be used to simulate 50 ohm filters throughout. The input and output parallel tuned circuits are tapped halfway down. The tap is at the 50 ohm impedance point. The whole trick of this filter is that by winding the input and output coils in a bifilar fashion, the number of turns can be twice as many as a monofilar-wound inductor designed for 50 ohms. There are double the number of turns in the bifilar winding, which gives four times the inductance. Similarly, the capacitance only has to be one fourth that used with the monofilar-wound 50 ohm design. You effectively are increasing the Q of the input and output inductors, because their total resistance increased by a factor of two, due to two windings in series while their inductance increased by a factor of four (total number of turns squared). ELSIE is useful for showing the response of an ideal 50 ohm filter. I use LT spice to simulate the bifilar-would input and output inductors. I'll post an LT-Spice model in the files section for you for the 40 meter filter.
"Digital Steve", K1RF