Nice work Charlie. I found better rejection by grounding all the crystals. (About 90 dB)The Spectrum analyzer BW of 300Hz is just below the CW filter bandwidth, so you might get some better results going to 100Hz BW. The Nano VNA maybe sweeping frequency too fast to get a typical response.
Thanks Kevin. I'm very new to the spectrum analyser. Still leaning. The grounding is interesting. I did that on a previous filter and had similar results to you. This time was different for some reason. I'll try it again tonight.
@@CharlieMorrisZL2CTM I have spent 45 years or so with analyzers. First repairing them, and the last 35 years or so, using them almost daily. The value for money these days is exceptional.. The new price of analyzers back in the 70s one could buy a house with!!
I’ve built a number of crystal ladder filters and came to the conclusion that the G3UUR’s QER design is the easiest and best performance to use. I have a few comments: 1. I use the nanoVNA-saver software with my nanoVNA. It scans, displays an reports results much better than just using a NanoVNA. You can have a much larger number of scan points (than 101) by increasing the number of segments; the SW breaks down the task and control the NanoVNA to do it. It also analyzes band pass filters. 2. Use NanoVNA to get key crystal data (F0 serie resonance frequency, 3-dB bandwidth, minimum loss, package capacitance). Derive corresponding crystal parameters Lm, Cm, ESR from the measure crystal data using established formulas (or use the G3UUR’s 3-dB method.) 3. Use the Dishal DJ6EV SW to design your QER filter (enter the number of crystals, desired BW, crystal parameters Lm, F0, ESR to get the characteristics impedance and the value of the shunt capacitors). Look in the IEEE QEX year 2009 archive “Steder-Hardcastle”. 4. Design some adaptive circuits to match the filter impedance to that of your system (usually 50 ohm). I prefer using a common base BJT amp circuit at the input and a common collector amp at the output. Alternatively, one can use L-shape filters using L and C. 5. Use SMA connectors and SMA cables to connect filters to your VNA. Pig-tails cables with alligator clips (and sometimes unknown cable impedance) are suspectful in showing non-truthful results. My results lately have been very good. I can get +/- 0.5 dB ripple and steep side skirts and BW that is within 100’s Hertz of target.
Hi Charlie, I have read somewhere that using two crystals of the same value on the input and output improves the performance, see the HF signals site where the BitX does this on his crystal filter. From memory where I first came across was in the 4SQRP site. Look for kits SAXVO (retired kit) where they use 3 crystals in parallel to stabilise the VFO.
Charlie, I had something similar with the VNA and I think it was something to do with the number of samples, changing that gave a smoother trace. There was another odd thing when I had to force it to do 'thru', not an 'SWR' or 'Z' measurement. It's been a while, so I don't recall the detail, sorry. Thanks for sharing.
Hi Charlie, nice work on the video there. I have a question regarding filters. I want to try and make a CW filter for a specific rig(the Kenwood TS520) and am wondering what frequency crystal I need to pick for the job. 73
I have a question about the dimensions of the RFC. Sometimes in a Amplifier instead of an Load Resistor a Inductor is used. How large should its inductance be and on what does it depend? Thanks for helping out
I used a $100,000 Rohde & Schwarz multiport VNA (ZNBT8) to look at my crystal filter. It just didn't work! I think the Q is just too high and VNAs aren't set up for such narrow bandwidth measurements. Now that I think about it, though, I didn't play around very much with the sweep parameters. Maybe it's still doable.
How did you arrive at the 170-ohm value? Or rather, what changes with different resistors there -- is it the shape of the passband? I'd like to know how to determine the correct values for my own filters!
@@CharlieMorrisZL2CTM do you use the value on both sides of the filter? When you change the resistor to a transformer, does the insertion loss improve?
Check measurement points, these cheap nano VNAs suffer on this very critical aspect, i.e. they have very low number. I use tinyVNA, more expensive but you greatly eliminate measurement uncertainties, hope it helps, 73
@@CharlieMorrisZL2CTM Not related to QER, however, this is one of the best crystal filter design tutorials I have ever seen ruclips.net/video/LyAQqqIsBOM/видео.html
Nice work Charlie. I found better rejection by grounding all the crystals. (About 90 dB)The Spectrum analyzer BW of 300Hz is just below the CW filter bandwidth, so you might get some better results going to 100Hz BW. The Nano VNA maybe sweeping frequency too fast to get a typical response.
Thanks Kevin. I'm very new to the spectrum analyser. Still leaning. The grounding is interesting. I did that on a previous filter and had similar results to you. This time was different for some reason. I'll try it again tonight.
@@CharlieMorrisZL2CTM I have spent 45 years or so with analyzers. First repairing them, and the last 35 years or so, using them almost daily. The value for money these days is exceptional.. The new price of analyzers back in the 70s one could buy a house with!!
Hi Kevin. I grounded both filters and gained another 3dB of attenuation. I'm happy with it. Thanks again.
I’ve built a number of crystal ladder filters and came to the conclusion that the G3UUR’s QER design is the easiest and best performance to use. I have a few comments:
1. I use the nanoVNA-saver software with my nanoVNA. It scans, displays an reports results much better than just using a NanoVNA. You can have a much larger number of scan points (than 101) by increasing the number of segments; the SW breaks down the task and control the NanoVNA to do it. It also analyzes band pass filters.
2. Use NanoVNA to get key crystal data (F0 serie resonance frequency, 3-dB bandwidth, minimum loss, package capacitance). Derive corresponding crystal parameters Lm, Cm, ESR from the measure crystal data using established formulas (or use the G3UUR’s 3-dB method.)
3. Use the Dishal DJ6EV SW to design your QER filter (enter the number of crystals, desired BW, crystal parameters Lm, F0, ESR to get the characteristics impedance and the value of the shunt capacitors). Look in the IEEE QEX year 2009 archive “Steder-Hardcastle”.
4. Design some adaptive circuits to match the filter impedance to that of your system (usually 50 ohm). I prefer using a common base BJT amp circuit at the input and a common collector amp at the output. Alternatively, one can use L-shape filters using L and C.
5. Use SMA connectors and SMA cables to connect filters to your VNA. Pig-tails cables with alligator clips (and sometimes unknown cable impedance) are suspectful in showing non-truthful results.
My results lately have been very good. I can get +/- 0.5 dB ripple and steep side skirts and BW that is within 100’s Hertz of target.
Hi Charlie, I have read somewhere that using two crystals of the same value on the input and output improves the performance, see the HF signals site where the BitX does this on his crystal filter. From memory where I first came across was in the 4SQRP site. Look for kits SAXVO (retired kit) where they use 3 crystals in parallel to stabilise the VFO.
Yet another great video, nice to see it all moving forward.
Glad you enjoyed it.
Charlie, I had something similar with the VNA and I think it was something to do with the number of samples, changing that gave a smoother trace. There was another odd thing when I had to force it to do 'thru', not an 'SWR' or 'Z' measurement. It's been a while, so I don't recall the detail, sorry. Thanks for sharing.
Hi Charlie, nice work on the video there. I have a question regarding filters. I want to try and make a CW filter for a specific rig(the Kenwood TS520) and am wondering what frequency crystal I need to pick for the job. 73
I have a question about the dimensions of the RFC. Sometimes in a Amplifier instead of an Load Resistor a Inductor is used. How large should its inductance be and on what does it depend? Thanks for helping out
What exactly are the issues with using those low profile crystals, as opposed to the regular sized ones? Why might they give more problems?
tks for post...nice work Charlie....
Glad you enjoyed it. 73.
What frequency is the crystal used? From ADI YB7UUI
I used a $100,000 Rohde & Schwarz multiport VNA (ZNBT8) to look at my crystal filter. It just didn't work! I think the Q is just too high and VNAs aren't set up for such narrow bandwidth measurements. Now that I think about it, though, I didn't play around very much with the sweep parameters. Maybe it's still doable.
How did you arrive at the 170-ohm value? Or rather, what changes with different resistors there -- is it the shape of the passband? I'd like to know how to determine the correct values for my own filters!
I varied the terminating resistance until i had a passband with acceptable ripple. It's mainly a tradeoff between ripple and insertion loss.
@@CharlieMorrisZL2CTM do you use the value on both sides of the filter? When you change the resistor to a transformer, does the insertion loss improve?
Check measurement points, these cheap nano VNAs suffer on this very critical aspect, i.e. they have very low number. I use tinyVNA, more expensive but you greatly eliminate measurement uncertainties, hope it helps, 73
Познавательно. Зачетно.
please schematic, crystal frequencies and capacitor values, thanks
You making one?
@@ShainAndrews I want to build my own version of Bitx20 in 40meters with 9 MHz filter
Sure thing. I'll put it up on the blog.
Look at a QER to improve the ripple
Yes, the QER was better. I will look to experiment more with that configuration. Thanks Garry.
@@CharlieMorrisZL2CTM Not related to QER, however, this is one of the best crystal filter design tutorials I have ever seen ruclips.net/video/LyAQqqIsBOM/видео.html
"Quick video" [not equal to] 16 minutes. 🙂