First IF amp and the Xtal filter is build. IF amp is perfect and works exactly as designed. The Xtal filter is a bit rough as you can see from the plots below. It needs to be a bit narrower, closer to 500hz would be nice. Its a bit ripply which is probably due to imperfect terminations and because the xtals are not very well matched, being more like + or – 300hz not 50hz. I have a better frequency counter and a couple hundred more xtals coming slow boat from China, so I should be able to improve things quite a lot when I can measure down to 10’s of hz.
I made a few minor mods to the homebrew bending brake and have improved the quality of the bends, with corner radius much tighter and improved the accuracy and repeatability of the bends. All in all its a much better unit now. I have added a set of sheet metal bending pliers and should now be able to make much more intricate boxes for my projects. The pictures below are something i knocked up rather quick and rough. Its 80x60x30mm and turned out ok. Nice tight fits.
Bandpass filter as designed in Elsie. In practice was a little low in frequency when built. I changed the end caps to to make them a little lower to bring the frequency up. The bandpass is rather ripply and if i remember latter I will Bode Plot it on the red pitaya.
Here is the schematic. It works but does not have the smoothest bandpass. I will see how it goes for now and if its not good enough, redesign it.
So i have started to build the actual project now rather than build and test different blocks. Its about time that I actually built something from beginning to end. So lets start with the schematic.
A blank canvas all ready to go.
The first stage built, but not working in this photo. The J310’s i have, have a different pinout to the data sheet and i put a cap where a resistor should be 🙂
Got things fixed and working in this photo.
Maximum gain from the dual gate rf amp.
Minimum gain from the RF Amp.
So far things are going swimmingly and next job is to wind an output transformer for this stage and then start on the bandpass filter.
So after building the IF Filter haphazard looking like a spastic, I connected up the test gear to run a Bode Plot to see the shape of the filter. I then proceeded to sit here wondering what i did wrong for the next hour as my 500hz CW Filter is almost 4K wide and probably ok as an AM Filter. LOL. Well, slow as slow can be it dawned on me, I grabbed 470 capacitors out of the box, thinking they were 470pf not 471 capacitors which are 470 pf. So in my haste and not thinking i have 47pf capacitors in my filter not what i originally planned.
On a positive note, the shape is not all that bad and not to much ripple even though the input to the filter is not correctly terminated. That is another headache, trying to get accurate measurements of in and output impedances and matching right when you have not built the preceeding stages. So I am going to switch to Termination Insensitive Amps for the IF and make my life easier.
So i had another swing at the MPSH10 in LT Spice and what it says should happen and what actually is happening are 2 very different things. I got it sorted and it is working fine. But I am not sure if i run with this or revert back to the 2n3904. For now I will run with this and see how it goes, I just have to measure the input and output impedances. Which as of next week or so, will be easier when i have a vna. The circuit is a little different to what was in the schematic in the previous post. I will update the schematic if I end up running with these.
I build this a little different, I had these squares sitting there i made so I used them.
Oscilloscope output, green trace is the signal in, yellow signal out.
Output from the spectrum analyzer.
So i sat here this morning with a bag of 50 crystals playing a game of what is the resonant frequency. Thoroughly boring I can assure you and I am not sure how the ladies at uBitx Labs in India can sort through thousands to get sets of 8. But alas I have done enough for now to get 2 sets that are matched at a guess to within 50hz, being that my frequency counter is only good to 100hz.
One thing that I would not have predicted is the distribution in the resonant frequencies. Its like buck shot, all over the place and the other thing I would not have predicted is the number of dead crystals or ones with so low Q they might as well be dead being 8, that seems like a very high death rate to me. I still have 2 more bags to sort, but thats not happening today, i have 16 matched crystals I need for 2 filters and am moving on to the next thing.