That's a very good looking 4 pole SSB filter. Your 31 minutes of filter design went much faster than than 5 minutes of my US history classes ever did. And I love the simple methods you use and the Cohen filters. Since I have around a thousand 12 MHz xtals but not a thousand varieties of capacitor values, this looks like an ideal way to go. Thank you.
Thanks Nick. I do watch Charlie's videos on this also but it is nice to hear someone else go through it again. I believe you pick up things you may have missed from the first person. good job!
I find that the 'classic' textbooks that you and others usually mention, start at a level a little beyond where I feel comfortable. Your explanations and demonstrations bridge the gap. Thanks again. Ps. I watched in one go. More entertaining for me than whatever is on the Beeb on a Saturday night. 73 2E0FXZ
You can probably fine tune the shape with small trimmer caps across the fixed ones. This will tune out the small variations between caps and crystals. The one small disadvantage of this type of filter is that one side drops off faster than the other, making it more suitable for one sideband than the other. Something to think about when selecting injection frequencies.
Thanks, great explanation. I've been researching crystal filters so I can put one on the input frequency of our 10m repeater on 29.580 MHz. Finding the right crystal is a challenge, plus designing the thing isn't something I've done before.
That was a fantastic video! You have explained very clearly all the points that people get confused with when attempting to build these filters. Just by that one video alone, you have earned a subscription. When I build these, I like to get a visual on the passband and this can be difficult if you don`t have a vna or speccy. I used to use a white noise source into the filter, hetrodyne the output down to around 10khz and at this point you can view the passband on a standard computer sound card using the free spectrum laboratory software. It`s also interesting to use variable caps so each can be tuned individually while observing the shape. You can get it very flat where you want it this way and then change out the variable caps for NP0s. I will be directing quite a few chaps to this video, as I really think, it`s one of the best (clearest), explanations on the internet! Complements from the lads of 1977 UK, yes sometimes we use ssb as well as am.
Hi Chris! Thank you very much for your kind words. I'm pleased you thought the video worth watching and even recommending! I'm certainly no expert in the field but I've built a few filters and made more than a few mistakes! Making mistakes is fine too - it just gives me more opportunities to learn how to do it better. If I can give some encouragement to others to give it a go then that's enough for me. Thanks for the info on your methods. I particularly like the idea of variable caps to determine the desired value. I might try that in future builds. 73 Nick
Great video, Nick. An easy way to keep track of the frequency of the crystals as you measure them is to use a permanent marker and mark each one with a number, measure them, and put them in a spreadsheet. Then you can sort them by freq and get the ones closest to one another.
Hi Stan, thanks very much. Yes, that's a very good idea. To be honest, if I'd had a permanent marker to hand I might well have done this. Where it is really useful is when you do measure the two crystal frequencies in the G3UUR oscillator and then calculate the motional parameters. To be able to store this info in a spreadsheet and then cross reference it to the particular crystal would be very useful indeed. Thanks again. 73 Nick (M0NTV)
Thank you for a great tutorial. You described a very compicated subject in such an easy to understand way. I'm very tempted to buy a stack of crystals and have a play - even though I don't do radio any more.
Thanks very much! I've actually just built another crystal filter (an 8 pole Cohn type) which works even better but is just as simple to build. It'll be in the next video which is coming soon. Thanks again. 73 Nick
I just watched this again, and I'm sad to think of my mentor, now a silent key, who cannot also watch this video. He loved to build QRP gear, but he advised me years ago to avoid any attempt to build a crystal filter because it was nearly impossible to build a good one. (Of course, that didn't keep him from collecting boxes of crystals, all of which he bequeathed to me.) He had strong opinions about things, but he would have been very pleased to be proven wrong about this one. And that is exactly what you did.
Thanks Doug. Yes, I've tackled 'Winding Toroids for the Terrified' before! Once you get your head round it it's not too difficult. It is pretty essential for most RF design and particularly crystal filters - where you want as flat a passband as possible. Thanks for sticking it out to the end! 73 Nick
Fantastic video. I'll be using this for building a filter just for the fun of it! I've got a load of 11.0592's from Amazon that took *forever* to arrive. My method for determining the frequency was to listen to the oscillator on a radio and use Spectroid on my phone to view the frequency of the tone. It worked well enough, but my oscillator (built on a breadboard) left a lot to be desired because bumping anything changed capacitance. I need to build it like you built yours, but the filter I built (a BITX40 12 MHz > 11.0592 MHz replacement) works well enough. 73!
First off, very appreciative of you doing this video and showing your workflow. I do wonder (as I need to make one) what further refinements would be necessary for a CW filter (500Hz or lower). Would you simply have to increase all the capacitors on the Cohn circuit? Also, did you ever get this to the point we could hear it? Thanks & 73...
Thanks very much. I've never built one for CW but as the bandwidth is determined by the capacitor value I imagine it is just a matter of getting the right value. I've built several of these now and you'll certainly be able to hear the latest one when I string together the complete IF section (very soon!). Thanks again. 73, Nick
Hello Nick, Thanks for your very informative videos. I've begun building modules for Eamon Skelton's Transceiver. In his book, Photo 1 is of the exciter module and shows a commercial crystal filter on the right side. This filter does not show up in any of the schematics and is only mentioned in the testing paragraph; " the 10.7 MHz crystal filter used for the tests is shown on the right". Does this transceiver need a crystal filter in the exciter? if so, would you recommend commercial or homebrew as per your video? Also, Please provide your source for variable capacitors. The ones visible in your videos look excellent.
Hi Gregory, what Eamon calls an 'exciter' I would break down into smaller modules i.e. a mic amp, oscillator, balanced modulator and crystal filter. If you are building a Superhet kind of design then yes, you'll need a crystal filter to filter out the opposite sideband. The variable caps are just ones I've found on eBay, Amazon etc. 73, Nick
Very good video ..... And Charlie does some great work over on his channel as you said , I've been watching him and Peter VK3YE for quite some years now ... Regards de 'Gw0wvl' 👍
Thanks for the video. It's really great. I am still trying to understand the impedance matching part. I have watched many different videos including Charlie Morris', but still a bit confused. So you attach one trim pot on the input side and another on the output side. You turn both of them independently until you get a the most flat response. Then you measure each (and hopefully they are close?) Then you add 50ohm to that number as filter side impedance? So if I read 150 ohm then you assume 200ohm filter impedance and match 50:200 in ohm? (or 1:2 in turns ratio). Am I getting this right? So then the filter impedance can never be less than 50 ohm?? Some other sources suggest trying it parallel and series so I am getting confused.
Hi there! Yes, I think you've understood me correctly. If you build the crystal filter in the way I did then you'll probably get a total impedance (measured value + 50Ω) of around 100-200Ω. Most of mine are closer to 100Ω. Thanks for watching and commenting. Incidentally, my next video will deep dive on a different method of impedance matching that you might also be interested in. 73, Nick
"Epic" in every sense of the word (I can just see it now in CinemaScope or even Cinerama"!) - well explained and in a logical sequence - keep up the good work! 72/73 G4FUI
Enjoyed the Manhattan building style, suspect a breadboard /might/ not hurt for this one, to proto on... thanks for the interesting notes on parts selection
Your oscillator can measure the xtal frequency with the crystal going directly to ground or through a capacitor to measure different parameters . Why did not you do the switching to measure these different values?
Hi! In short - because I didn't need to. All those motional parameters are necessary if you build a crystal filter in the conventional way. However, a crystal Cohn filter does not require this. All the capacitors are the same value. It is a simpler method which I've found to be very effective. Thanks. 73 Nick
Great video, I have a question on your schematic, on the lower left transformer, isn’t it a 5:7 ? Turns ratio. The schematic shows 7:5. 50 ohms up to 100 ohms, then on the right 100 back to 50…
Hi there! Yes, you are right it is a 5:7 on the left and a 7:5 on the right but the schematic hopefully makes it clear that the 17.17uH end goes in to the crystal filter and the 8.75uH end faces out to the rest of the radio. I just referred to both the transformers as 7:5 because they are identical - it's just that they need to be fitted symmetrically. Thanks for watching and for your observation. 73, Nick
Hi Nick! Just sent you a note on the GQRP forum. Great post as always. Silly question: where does the crystal filter go in the circuit? Is it in lieu of the pass band filter? Or is it after the front end? Jack, NG2E
Hi Jack! Thanks very much. There's no such thing as a silly question! You would normally use a crystal ladder filter in your IF section. Because it is so narrow (mine is 2.6 kHz) it's job is to filter out the other sideband (on transmit) and anything else lying in the other sideband slot on receive. So in a typical SuperHet receiver you might have: Antenna -> Band Pass Filter -> Mixer -> IF Amp -> CRYSTAL FILTER -> IF Amp -> Product Detector -> Audio Preamp -> Audio Power Amp -> Loudspeaker Basically it is the heart of a SuperHet configuration. Your question makes me think I should maybe tackle the subject of how all these filters and things connect together - so thanks for that inspiration. 73 Nick (M0NTV)
Hi Nick, very well explained video. A question. How close does the actual center IF frequency have to be to the resonant peak of the filter and still be reasonably effective? I suppose as soon as you move more than 1KHz, you will be mostly out of the filter pass-band. I'm just wondering if you can move the resonant peak with changes in capacitance and inductance. If you know any tips to do so, I'd be very interested or any links where others have done this. With the declining number of crystals in the world any methods to better utilized what we can get will be helpful. My need is to build roofing filters for receivers already in my collection. 73 es cheers Glenn WA4AOS
Hi Glenn, thanks very much. I think the question of how effective the filter is depends on the bandwidth. Clearly the wider the filter is the more tolerance there will be for a signal not on the filter's centre frequency - but the downside is it will let other signals through as well! When it comes to the design of a Cohn (Minimum Loss) filter the bandwidth is proportionate to the capacitance. Each capacitor is the same value and the lower the value the wider the filter. I'm not sure this has much effect on the centre frequency of the filter though. I don't know how you'd be able to easily shift the frequency itself. These kind of filters are fine for IF stages where the precise frequency isn't critical because you build everything else around your crystal filter. Your proposed use case is somewhat different and might require a different design or even a whole new approach. Sorry I can't be of more help with this but I wish you every success! 73 Nick
Nice filter, I am curios what the impedance match is. My experience is that getting a reasonable passband is the first step, but getting a good match the next. Did you try the dishal software to calculate the filter? I have build some 8 xtall filters and it is "fun" to do. Add a hycas if and you have a great receiver
Hi Paul, I actually go into the impedance matching in the video. Mine had a Z of about 100 Ohms so I made a 2:1 matching transformer (root 2 for turns ratio). I didn't model it first just built it - added the trim pots - tweaked for flattest passband - removed trimpots - measured them + 50 (= 100 ohms). It's all in the video. The thing with these Cohn filters is that unlike the regular ones you don't necessarily get better performance by simply adding more crystals. It does make them easier to build but you won't get the skirts as steep as a regular filter. Thanks for watching. 73 Nick
@@M0NTVHomebrewing What I was refering to is in the video you show a very good passband S21 graph, but you don't show the s11 graph (reflection). Because I don't understand the assumption to add 50 to the input impedance. In the Dishal software you can calculate a Cohn filter based on the motion parameter of your xtals. But the input impedance calculated differs very much per chosen xtal. But I have to say I like the video very much. Building xtal filters is fun to do and a learning experience.
Hi Paul, sorry - I'm with you now! Yes, the adding 50 puzzled me for a while. It is Charlie Morris' technique. It goes something like this. So we know the impedance (Z) of the NanoVNA is 50 ohms but we don't know the Z of our filter. What we also know is that a matched impedance will result in a flat passband. So we add resistance via the temporary trim pots until the passband is a flat as possible. We can now assume that the Z is as matched as possible. So we remove the trim pots and measure them. We know that this figure + 50 ohms of the NanoVNA = Z of our filter. In my case it turned out to be 50 (NanoVNA) + 50 (trim pot) = 100 ohms. So I knew that my filter had twice as much Z as my preceding stage. So I built 2:1 matching transformers. Hope this helps! 73 Nick
The input and output values ought to be much more similar than that. You'll always get some difference depending on the components you use (NP0 caps etc.) and how well-matched your crystals are. You might want to try again with better frequency matching of the crystals and good caps. Or you could just stick with what you have and build different impedance matching transformers for the IN and OUT. Once you've done it a NanoVNA will soon show you how well you've done ... or not. Keep at it! 73 Nick
Every homebrew crystal ladder filter I've ever seen is the same. Am I missing something??? My old Icom 761 double conversion transceiver used 9 mhz and 455 kc if stages. Why would you just get a hand full of random filters and build a 13.3 mhz crystal filter? Wouldn't you have to have the frequency equal to one of the if sections? I'm just a self taught electronic hobbyist and don't have an engineering degree. I realize there are large gaps in my electronics education. Are you doing this just as a theoretical exercise to demonstrate how one would build a ladder filter??
Hi there! No it is certainly not just a theoretical exercise: the last two homebrew SSB transceivers I've built have had 13.3 MHz IF stages and the one before that used a 16 MHz IF. I seem to remember confessing in the video that I thought I'd bought 100 x 40 MHz crystals but I hadn't realised they were 3rd overtone. There is nothing set in stone about 9 MHz - it is just that there were a lot of commercial filters produced for that frequency (and 455 MHz was in many of the old AM radios). Basically you can build a filter for whatever frequency you can get the crystals for. But you need to calculate how that will mix with your incoming RF signal. Also low IF frequencies were preferable when people were just using analogue VFOs because it was easier to get a stable signal. These days with magic modules like the Si5351 that is not so much a concern. Thanks for your question. 73, Nick
Thanks for the reply Nick. Thanks for making things clear for me. I've been confused for years about Wes Haywards article on crystal filters. He never mentions he is building a crystal filter for a specific if frequency. He just says he has a bunch of crystals laying around and starts building a ladder filter. I guess Wes Hayward assumed you knew but I don't have an EE degree. I'm just an experimenter and hobbyist. It never made since to me until now. Thanks again.@@M0NTVHomebrewing
Please verify your calibrations. i seen this quite a lot in your videos, you say you've calibrated it but looking at the screen @ 28:37 that's not calibrated. On the left side of the screen cDRSTX means un-calibrated, when calibrated it should look like this CDRSTX. its all about the (C) if its in calibration, its uppercase (C), out of calibration its lowercase (c).
Well spotted! Perhaps when I made some of the recordings I was uncalibrated. I do however try to calibrate every time I make a measurement (as opposed to just showing a result on the camera). I'm now using a later version of the NanoVNA with a bigger screen - so it is easier to notice this kind of stuff! Thanks for watching so closely. 73, Nick
That's a very good looking 4 pole SSB filter. Your 31 minutes of filter design went much faster than than 5 minutes of my US history classes ever did. And I love the simple methods you use and the Cohen filters. Since I have around a thousand 12 MHz xtals but not a thousand varieties of capacitor values, this looks like an ideal way to go. Thank you.
Thanks very much John. 73 Nick
Thanks Nick. I do watch Charlie's videos on this also but it is nice to hear someone else go through it again. I believe you pick up things you may have missed from the first person. good job!
Thanks very much Curt. 73 Nick
I find that the 'classic' textbooks that you and others usually mention, start at a level a little beyond where I feel comfortable. Your explanations and demonstrations bridge the gap. Thanks again.
Ps. I watched in one go. More entertaining for me than whatever is on the Beeb on a Saturday night.
73 2E0FXZ
Thanks very much Stephen :)
You can probably fine tune the shape with small trimmer caps across the fixed ones. This will tune out the small variations between caps and crystals. The one small disadvantage of this type of filter is that one side drops off faster than the other, making it more suitable for one sideband than the other. Something to think about when selecting injection frequencies.
Thanks Kenneth.
great video, i do gain a lot of info from them. Thanks
Thanks. Glad to hear it. 73, Nick
Thanks, great explanation. I've been researching crystal filters so I can put one on the input frequency of our 10m repeater on 29.580 MHz. Finding the right crystal is a challenge, plus designing the thing isn't something I've done before.
Thanks very much. All the best with your own project! 73, Nick
Very comprehensive video and no-no nonsense like this Texan's first car, a Triumph GT6.
Thanks very much! Glad it was of interest. 73 Nick
That was a fantastic video! You have explained very clearly all the points that people get confused with when attempting to build these filters. Just by that one video alone, you have earned a subscription. When I build these, I like to get a visual on the passband and this can be difficult if you don`t have a vna or speccy. I used to use a white noise source into the filter, hetrodyne the output down to around 10khz and at this point you can view the passband on a standard computer sound card using the free spectrum laboratory software. It`s also interesting to use variable caps so each can be tuned individually while observing the shape. You can get it very flat where you want it this way and then change out the variable caps for NP0s. I will be directing quite a few chaps to this video, as I really think, it`s one of the best (clearest), explanations on the internet! Complements from the lads of 1977 UK, yes sometimes we use ssb as well as am.
Hi Chris! Thank you very much for your kind words. I'm pleased you thought the video worth watching and even recommending! I'm certainly no expert in the field but I've built a few filters and made more than a few mistakes! Making mistakes is fine too - it just gives me more opportunities to learn how to do it better. If I can give some encouragement to others to give it a go then that's enough for me. Thanks for the info on your methods. I particularly like the idea of variable caps to determine the desired value. I might try that in future builds. 73 Nick
Best video on this topic I ever found. And I follow the fantastic Charlie Morris too.
Thank you very much indeed. Glad you found it helpful. 73 Nick
Great video, Nick. An easy way to keep track of the frequency of the crystals as you measure them is to use a permanent marker and mark each one with a number, measure them, and put them in a spreadsheet. Then you can sort them by freq and get the ones closest to one another.
Hi Stan, thanks very much. Yes, that's a very good idea. To be honest, if I'd had a permanent marker to hand I might well have done this. Where it is really useful is when you do measure the two crystal frequencies in the G3UUR oscillator and then calculate the motional parameters. To be able to store this info in a spreadsheet and then cross reference it to the particular crystal would be very useful indeed. Thanks again. 73 Nick (M0NTV)
Great video Nick, another nice piece of work!
Cheers Alan!
First time I understood how to impedance match. Thank you.
Thanks very much Doug. Glad it was of some help. 73 Nick
Really good video Nick and your explanation and demonstration is very clear. Has helped me lots and I shall hunt down those crystals in my stock!
Thanks very much Steve. Have fun with those crystals! 73 Nick (M0NTV)
Thank you for a great tutorial. You described a very compicated subject in such an easy to understand way. I'm very tempted to buy a stack of crystals and have a play - even though I don't do radio any more.
Thanks very much! I've actually just built another crystal filter (an 8 pole Cohn type) which works even better but is just as simple to build. It'll be in the next video which is coming soon. Thanks again. 73 Nick
Nice job! Keep up the good work. Thanks for the video
Thanks very much! 73 Nick
great video! Thanks a lot!
Thanks very much. Glad it was of interest. 73 Nick
I’m certainly no expert but I did watch the whole thing.
Thanks Bruce. Hope it was of some help. 73 Nick
I just watched this again, and I'm sad to think of my mentor, now a silent key, who cannot also watch this video. He loved to build QRP gear, but he advised me years ago to avoid any attempt to build a crystal filter because it was nearly impossible to build a good one. (Of course, that didn't keep him from collecting boxes of crystals, all of which he bequeathed to me.) He had strong opinions about things, but he would have been very pleased to be proven wrong about this one. And that is exactly what you did.
Thanks very much John. 73 Nick
Kind of deep with the impedance matching technique but I did watch to the end. Thanks.
Thanks Doug. Yes, I've tackled 'Winding Toroids for the Terrified' before! Once you get your head round it it's not too difficult. It is pretty essential for most RF design and particularly crystal filters - where you want as flat a passband as possible. Thanks for sticking it out to the end! 73 Nick
Fantastic video. I'll be using this for building a filter just for the fun of it! I've got a load of 11.0592's from Amazon that took *forever* to arrive. My method for determining the frequency was to listen to the oscillator on a radio and use Spectroid on my phone to view the frequency of the tone. It worked well enough, but my oscillator (built on a breadboard) left a lot to be desired because bumping anything changed capacitance. I need to build it like you built yours, but the filter I built (a BITX40 12 MHz > 11.0592 MHz replacement) works well enough. 73!
Thanks Ryan! All the best with your homebrewing. 73 Nick
Great Video!!!
Thanks very much Mauro! 73 Nick
very useful research sir, low past filter
Thanks very much.
First off, very appreciative of you doing this video and showing your workflow.
I do wonder (as I need to make one) what further refinements would be necessary for a CW filter (500Hz or lower).
Would you simply have to increase all the capacitors on the Cohn circuit?
Also, did you ever get this to the point we could hear it?
Thanks & 73...
Thanks very much. I've never built one for CW but as the bandwidth is determined by the capacitor value I imagine it is just a matter of getting the right value. I've built several of these now and you'll certainly be able to hear the latest one when I string together the complete IF section (very soon!). Thanks again. 73, Nick
Hello Nick,
Thanks for your very informative videos. I've begun building modules for Eamon Skelton's Transceiver. In his book, Photo 1 is of the exciter module and shows a commercial crystal filter on the right side. This filter does not show up in any of the schematics and is only mentioned in the testing paragraph; " the 10.7 MHz crystal filter used for the tests is shown on the right". Does this transceiver need a crystal filter in the exciter? if so, would you recommend commercial or homebrew as per your video?
Also, Please provide your source for variable capacitors. The ones visible in your videos look excellent.
Hi Gregory, what Eamon calls an 'exciter' I would break down into smaller modules i.e. a mic amp, oscillator, balanced modulator and crystal filter. If you are building a Superhet kind of design then yes, you'll need a crystal filter to filter out the opposite sideband. The variable caps are just ones I've found on eBay, Amazon etc. 73, Nick
Very good video ..... And Charlie does some great work over on his channel as you said , I've been watching him and Peter VK3YE for quite some years now ... Regards de 'Gw0wvl' 👍
Thanks very much! 73 Nick
very informative, I need a filter, and will make it. Subscribed
Thanks very much.
Thanks for the video. It's really great. I am still trying to understand the impedance matching part. I have watched many different videos including Charlie Morris', but still a bit confused. So you attach one trim pot on the input side and another on the output side. You turn both of them independently until you get a the most flat response. Then you measure each (and hopefully they are close?) Then you add 50ohm to that number as filter side impedance? So if I read 150 ohm then you assume 200ohm filter impedance and match 50:200 in ohm? (or 1:2 in turns ratio). Am I getting this right? So then the filter impedance can never be less than 50 ohm?? Some other sources suggest trying it parallel and series so I am getting confused.
Hi there! Yes, I think you've understood me correctly. If you build the crystal filter in the way I did then you'll probably get a total impedance (measured value + 50Ω) of around 100-200Ω. Most of mine are closer to 100Ω. Thanks for watching and commenting. Incidentally, my next video will deep dive on a different method of impedance matching that you might also be interested in. 73, Nick
Very helpful thanks
Thanks Karl. Glad it was of use. 73 Nick
"Epic" in every sense of the word (I can just see it now in CinemaScope or even Cinerama"!) - well explained and in a logical sequence - keep up the good work!
72/73 G4FUI
Thanks very much Martin. I'm humming the Pearl & Dean music!
Enjoyed the Manhattan building style, suspect a breadboard /might/ not hurt for this one, to proto on... thanks for the interesting notes on parts selection
Thanks very much.
Your oscillator can measure the xtal frequency with the crystal going directly to ground or through a capacitor to measure different parameters . Why did not you do the switching to measure these different values?
Hi! In short - because I didn't need to. All those motional parameters are necessary if you build a crystal filter in the conventional way. However, a crystal Cohn filter does not require this. All the capacitors are the same value. It is a simpler method which I've found to be very effective. Thanks. 73 Nick
I love it good
Great video, I have a question on your schematic, on the lower left transformer, isn’t it a 5:7 ? Turns ratio. The schematic shows 7:5. 50 ohms up to 100 ohms, then on the right 100 back to 50…
Hi there! Yes, you are right it is a 5:7 on the left and a 7:5 on the right but the schematic hopefully makes it clear that the 17.17uH end goes in to the crystal filter and the 8.75uH end faces out to the rest of the radio. I just referred to both the transformers as 7:5 because they are identical - it's just that they need to be fitted symmetrically. Thanks for watching and for your observation. 73, Nick
Thanks I figured it was just a clerical thing. It was clear from the discussion and calculations so I was just curious if I was missing something
Hi Nick! Just sent you a note on the GQRP forum. Great post as always. Silly question: where does the crystal filter go in the circuit? Is it in lieu of the pass band filter? Or is it after the front end? Jack, NG2E
Hi Jack! Thanks very much. There's no such thing as a silly question! You would normally use a crystal ladder filter in your IF section. Because it is so narrow (mine is 2.6 kHz) it's job is to filter out the other sideband (on transmit) and anything else lying in the other sideband slot on receive. So in a typical SuperHet receiver you might have:
Antenna -> Band Pass Filter -> Mixer -> IF Amp -> CRYSTAL FILTER -> IF Amp -> Product Detector -> Audio Preamp -> Audio Power Amp -> Loudspeaker
Basically it is the heart of a SuperHet configuration. Your question makes me think I should maybe tackle the subject of how all these filters and things connect together - so thanks for that inspiration. 73 Nick (M0NTV)
@@M0NTVHomebrewing Yes, please tackle how these things all fit together - especially how the impedances of the stages must match. Kevin M0XYM
@@M0XYM OK Kevin. Thanks - I'll bear that in mind for a future video :)
Hi Nick, very well explained video. A question. How close does the actual center IF frequency have to be to the resonant peak of the filter and still be reasonably effective? I suppose as soon as you move more than 1KHz, you will be mostly out of the filter pass-band. I'm just wondering if you can move the resonant peak with changes in capacitance and inductance. If you know any tips to do so, I'd be very interested or any links where others have done this. With the declining number of crystals in the world any methods to better utilized what we can get will be helpful. My need is to build roofing filters for receivers already in my collection. 73 es cheers Glenn WA4AOS
Hi Glenn, thanks very much. I think the question of how effective the filter is depends on the bandwidth. Clearly the wider the filter is the more tolerance there will be for a signal not on the filter's centre frequency - but the downside is it will let other signals through as well!
When it comes to the design of a Cohn (Minimum Loss) filter the bandwidth is proportionate to the capacitance. Each capacitor is the same value and the lower the value the wider the filter. I'm not sure this has much effect on the centre frequency of the filter though. I don't know how you'd be able to easily shift the frequency itself. These kind of filters are fine for IF stages where the precise frequency isn't critical because you build everything else around your crystal filter. Your proposed use case is somewhat different and might require a different design or even a whole new approach. Sorry I can't be of more help with this but I wish you every success! 73 Nick
Nice filter, I am curios what the impedance match is. My experience is that getting a reasonable passband is the first step, but getting a good match the next. Did you try the dishal software to calculate the filter? I have build some 8 xtall filters and it is "fun" to do. Add a hycas if and you have a great receiver
Hi Paul, I actually go into the impedance matching in the video. Mine had a Z of about 100 Ohms so I made a 2:1 matching transformer (root 2 for turns ratio). I didn't model it first just built it - added the trim pots - tweaked for flattest passband - removed trimpots - measured them + 50 (= 100 ohms). It's all in the video. The thing with these Cohn filters is that unlike the regular ones you don't necessarily get better performance by simply adding more crystals. It does make them easier to build but you won't get the skirts as steep as a regular filter. Thanks for watching. 73 Nick
@@M0NTVHomebrewing What I was refering to is in the video you show a very good passband S21 graph, but you don't show the s11 graph (reflection). Because I don't understand the assumption to add 50 to the input impedance. In the Dishal software you can calculate a Cohn filter based on the motion parameter of your xtals. But the input impedance calculated differs very much per chosen xtal. But I have to say I like the video very much. Building xtal filters is fun to do and a learning experience.
Hi Paul, sorry - I'm with you now! Yes, the adding 50 puzzled me for a while. It is Charlie Morris' technique.
It goes something like this. So we know the impedance (Z) of the NanoVNA is 50 ohms but we don't know the Z of our filter. What we also know is that a matched impedance will result in a flat passband. So we add resistance via the temporary trim pots until the passband is a flat as possible. We can now assume that the Z is as matched as possible. So we remove the trim pots and measure them. We know that this figure + 50 ohms of the NanoVNA = Z of our filter. In my case it turned out to be 50 (NanoVNA) + 50 (trim pot) = 100 ohms. So I knew that my filter had twice as much Z as my preceding stage. So I built 2:1 matching transformers. Hope this helps! 73 Nick
@@M0NTVHomebrewing Yes, that is a practical technique. Thanks for the explanation.
I put Trin por in serie, i have nice flat filter but in the in 600 ohms AND in the out 200 ohms i dont understand
The input and output values ought to be much more similar than that. You'll always get some difference depending on the components you use (NP0 caps etc.) and how well-matched your crystals are. You might want to try again with better frequency matching of the crystals and good caps. Or you could just stick with what you have and build different impedance matching transformers for the IN and OUT. Once you've done it a NanoVNA will soon show you how well you've done ... or not. Keep at it! 73 Nick
Did you experiment with the QER topology?
No I didn't. It's a simple Cohn (Minimum Loss) crystal ladder filter.
Every homebrew crystal ladder filter I've ever seen is the same. Am I missing something??? My old Icom 761 double conversion transceiver used 9 mhz and 455 kc if stages. Why would you just get a hand full of random filters and build a 13.3 mhz crystal filter? Wouldn't you have to have the frequency equal to one of the if sections? I'm just a self taught electronic hobbyist and don't have an engineering degree. I realize there are large gaps in my electronics education. Are you doing this just as a theoretical exercise to demonstrate how one would build a ladder filter??
Hi there! No it is certainly not just a theoretical exercise: the last two homebrew SSB transceivers I've built have had 13.3 MHz IF stages and the one before that used a 16 MHz IF. I seem to remember confessing in the video that I thought I'd bought 100 x 40 MHz crystals but I hadn't realised they were 3rd overtone. There is nothing set in stone about 9 MHz - it is just that there were a lot of commercial filters produced for that frequency (and 455 MHz was in many of the old AM radios). Basically you can build a filter for whatever frequency you can get the crystals for. But you need to calculate how that will mix with your incoming RF signal. Also low IF frequencies were preferable when people were just using analogue VFOs because it was easier to get a stable signal. These days with magic modules like the Si5351 that is not so much a concern. Thanks for your question. 73, Nick
Thanks for the reply Nick. Thanks for making things clear for me. I've been confused for years about Wes Haywards article on crystal filters. He never mentions he is building a crystal filter for a specific if frequency. He just says he has a bunch of crystals laying around and starts building a ladder filter. I guess Wes Hayward assumed you knew but I don't have an EE degree. I'm just an experimenter and hobbyist. It never made since to me until now. Thanks again.@@M0NTVHomebrewing
No worries Douglas! We are all learning all the time. Keep on experimenting and have fun. 73, Nick
Please verify your calibrations. i seen this quite a lot in your videos, you say you've calibrated it but looking at the screen @ 28:37 that's not calibrated. On the left side of the screen cDRSTX means un-calibrated, when calibrated it should look like this CDRSTX. its all about the (C) if its in calibration, its uppercase (C), out of calibration its lowercase (c).
Well spotted! Perhaps when I made some of the recordings I was uncalibrated. I do however try to calibrate every time I make a measurement (as opposed to just showing a result on the camera). I'm now using a later version of the NanoVNA with a bigger screen - so it is easier to notice this kind of stuff! Thanks for watching so closely. 73, Nick