Crystal Ladder Filters
Carbon_Rod, Sun Sept 10 2017, 03:17AMI tried building a 4 Crystal Ladder Filter, but can't seem to get a pass band working properly.
The circuit is similar to the schematic shown here, and the values using Lloyd's method are similar:
However, the surface mount crystals I tried don't seem to work in this application.
I used a 14dBm source, but the signal is almost completely attenuated at the first stage.
Does anyone know of a working band pass filter example that uses compact surface mount crystal packages?
Re: Crystal Ladder Filters
Hazmatt_(The Underdog), Sun Sept 10 2017, 04:42AM
I don't know of an example... maybe the ARRL handbook.
But you should not need to be driving your filter with 14 dBm.
A good filter passband will attenuate 1 to 2 dB, so 0dBm input drive is more than sufficient.
What frequency and do you have a scalar analyzer? (I sure don't)
The board can interact and shift your center frequency too if you're not careful.
I tried doing filters here at home too from the cookbook, but I couldn't seem to get any to work properly. It's kinda a shame too, I have access to all kinds of toys at work.
Hazmatt_(The Underdog), Sun Sept 10 2017, 04:42AM
I don't know of an example... maybe the ARRL handbook.
But you should not need to be driving your filter with 14 dBm.
A good filter passband will attenuate 1 to 2 dB, so 0dBm input drive is more than sufficient.
What frequency and do you have a scalar analyzer? (I sure don't)
The board can interact and shift your center frequency too if you're not careful.
I tried doing filters here at home too from the cookbook, but I couldn't seem to get any to work properly. It's kinda a shame too, I have access to all kinds of toys at work.
Re: Crystal Ladder Filters
Sulaiman, Sun Sept 10 2017, 07:54AM
I built a four crystal Cohn filter like the second diagram in figure 2 - it was not 'flat' or selective enough for my ssb requirements,
so I built an eight crystal filter like in the bottom diagram using the closest matched 8 of 20 crystals that I bought & measured;
- and never got it to work properly,
As I do not have a sweep frequency generator or spectrum analyser, VNA etc., setting up Cohn filters is MUCH too tedious.
A few Cathodeon crystal filters (two lsb and one usb) came up on eBay so I bought them - now I have good crystal filters :)
(my diy filter was built with 8.86xx MHz (2x PAL colour burst freq.) but the purchased filters were for 1.4 MHz so a little re-design was required
I see no reason why smd crystals would not work - but the crystals need to be very closely matched
It IS important to feed and load Cohn filters with the correct resistances
(I forget, but somewhere over 1 kOhm for the 8-Xtal Cohn that I built)
EDIT: +1 on the not driving at +14 dBm ... 0 dBm or less seems appropriate.
Sulaiman, Sun Sept 10 2017, 07:54AM
I built a four crystal Cohn filter like the second diagram in figure 2 - it was not 'flat' or selective enough for my ssb requirements,
so I built an eight crystal filter like in the bottom diagram using the closest matched 8 of 20 crystals that I bought & measured;
- and never got it to work properly,
As I do not have a sweep frequency generator or spectrum analyser, VNA etc., setting up Cohn filters is MUCH too tedious.
A few Cathodeon crystal filters (two lsb and one usb) came up on eBay so I bought them - now I have good crystal filters :)
(my diy filter was built with 8.86xx MHz (2x PAL colour burst freq.) but the purchased filters were for 1.4 MHz so a little re-design was required
I see no reason why smd crystals would not work - but the crystals need to be very closely matched
It IS important to feed and load Cohn filters with the correct resistances
(I forget, but somewhere over 1 kOhm for the 8-Xtal Cohn that I built)
EDIT: +1 on the not driving at +14 dBm ... 0 dBm or less seems appropriate.
Re: Crystal Ladder Filters
Carbon_Rod, Sun Sept 10 2017, 04:36PM
With the SMD crystals:
I tried a OCXO at 9.99985MHz reference source for testing the passband, but could not get anything to calibrate against at any of the stages. After seeing your guys feedback, I became more suspicious of the SMDs... and tried a cheap Tracking Noise Source (a cheap zener+amp ebay unit). The insertion loss was almost >12dB at the initial stage irrespective of the tuning, and thus not behaving correctly given there are no shorts or bad MLCC etc.
After replacing the SMD parts with some HC49 packages in the identical setup, it now at least is tunable.
Note, the filter currently seems to inject +-253kHz overtones, but does clean up the 2% asymmetry of the test reference source.
As a side note, I usually recommend a ZFDC-20-5-N with replaceable SMA adapters for VNAs.
Even those cheap USB based AD based kit will give fairly repeatable results.
I did post the limeSDR VNA tracking source link in the SDR thread.
Unlike other chips it can lock both its front end Tx/Rx Mixers in a low phase-noise synchronized configuration (I didn't even notice that feature until after the campaign ended). However, there will be projects based on the system emerging next year if I get time, given it should theoretically replace most of my VHF/UHF kit.
Cheers,
Rod
Carbon_Rod, Sun Sept 10 2017, 04:36PM
With the SMD crystals:
I tried a OCXO at 9.99985MHz reference source for testing the passband, but could not get anything to calibrate against at any of the stages. After seeing your guys feedback, I became more suspicious of the SMDs... and tried a cheap Tracking Noise Source (a cheap zener+amp ebay unit). The insertion loss was almost >12dB at the initial stage irrespective of the tuning, and thus not behaving correctly given there are no shorts or bad MLCC etc.
After replacing the SMD parts with some HC49 packages in the identical setup, it now at least is tunable.
Note, the filter currently seems to inject +-253kHz overtones, but does clean up the 2% asymmetry of the test reference source.
As a side note, I usually recommend a ZFDC-20-5-N with replaceable SMA adapters for VNAs.
Even those cheap USB based AD based kit will give fairly repeatable results.
I did post the limeSDR VNA tracking source link in the SDR thread.
Unlike other chips it can lock both its front end Tx/Rx Mixers in a low phase-noise synchronized configuration (I didn't even notice that feature until after the campaign ended). However, there will be projects based on the system emerging next year if I get time, given it should theoretically replace most of my VHF/UHF kit.
Cheers,
Rod
Re: Crystal Ladder Filters
Conundrum, Wed Sept 13 2017, 08:52PM
Thanks for the tips. I believe that the problem with SMDs is that they are constructed differently to save costs, essentially a crystal is a quartz disk with wires held in a hermetic can whereas SMD ones are mechanically more rigid so need more energy to run. Not issue for most applications but can be for this one.
Conundrum, Wed Sept 13 2017, 08:52PM
Thanks for the tips. I believe that the problem with SMDs is that they are constructed differently to save costs, essentially a crystal is a quartz disk with wires held in a hermetic can whereas SMD ones are mechanically more rigid so need more energy to run. Not issue for most applications but can be for this one.
Print this page