Crystal Filter Bilateral Amplifier Board

Finally finished drafting up the board and etched it this afternoon.  It's interesting, spoken sarcastically that is, as I think I'm getting a handle on using KiCad and then problems present that were not an issue with previous boards. In this case, connecting the grounds to the ground plane.  The work around was easy enough though somewhat messy.  Looking at the board if you see something strange, its likely where I did my connections as I couldn't be bothered wasting more time to figure it out. It all became time consuming in the end as I went through the board several times applying ground connections I would discover something that needed to be changed.  This in turn would make me delete the connections after the new changes, reapply the fill fore the ground plane and redo the grounds.

Another problem I'm having with KiCad is the modifying or developing a new component or footprint and saving it.  Life is becoming to short to figure this one out.  Maybe next week I'll change my mind.

Anyway I'm getting better at developing boards as far as what will work and what areas to be careful with.  Overriding the Electrical Rules Check (ERC) function has being a blessing, but I'm sure it may have been causing me headaches in other areas. As mentioned above.

Here's my latest board and I should be able to get around to populating later this week.

Xtal Filters

Spent the last couple of days off and on in the shack matching crystals and sweeping crystal filters. Results from crystal matching yields a wealth of data most of which I'm not sure yet what to do with.  Frequency is the easy one and then things can get a little complicated.  Data I have access to is as follows:

 Xtl                  Fs               -3dB BW  Rs          Cm          Lm          Q

x4        4,913,470 95 9 0.0184 56.96   195,438 x9        4,913,508 100 11.14 0.0182 57.51   159,442 x14        4,913,536 100 11.14 0.0182 57.51   159,443 x15        4,913,536 96 9.11 0.0186 56.56   191,570 x5        4,913,545 96 9.11 0.0186 56.56   191,570 x8        4,913,547 97 11.14 0.0177 59.29   164,374 x1        4,913,587 92 9 0.0178 58.81   201,816 x10        4,913,600 85 8.76 0.0166 63.22   222,727
I took out 7 other crystals as the Q for them was below 150,000.  I read that number somewhere as the defining line, but I suppose open for debate.
So far anything I've read on the subject, I'm not clear as to how to use all this data in making choices for crystal selection.  Testing so far I've used a breadboard, not the best I know, and a board I etched to help with testing crystal filters.  Results while interesting so far, are very time consuming in obtaining.  I can see the effects of changing capacitors and a much more detailed and a better test jig is in order before being able to have something useful to go with.
One surprise was the minimal effect on the ripple when substituting a good crystal with one that has a poor numbers or a Fs much further away. What I can say for sure is that I can see that the crystal filter as a complete circuit may be the way to go when testing such circuits.  The capacitors and matching transformers have a large effect on the filters response and when ever I get around to doing this testing again, I will incorporate the torroids as well. Not just the caps and crystals.
Now back to building the LBS project and hopefully not get too distracted again.
73, Peter

Xtal Filter Bilateral Amlifier part 3

Found a little time for the shack today.  Repaired the R5 vertical again.  Blew over this morning and broke the radials I repaired last week.

Re-checked the test amplifier board and found an error.  Made some changes and all's well.  Checked the circuit on the xtal filter board and of course found a couple of wiring errors there as well. So no time to fix that and test but at least I found something and get to repairs next week.

73, Peter

Xtall Filter Bilateral Amplifier - part 2

Quickly put together an amplifier to test the output circuit of the xtal filter thereby bypassing the last stage amplifier and see if things clear up. Well, sort of.  The signal clears up alright, but the amplification does not.  Below is a picture of the setup for reference. Maybe I've missed something, but I can't think of what? If I've made a mistake in the circuit, I've done it twice and I'd like to think that's unlikely.

I'll need to take a step back for a few days with this circuit as I need to fix my antenna and get ready for my tower/antenna installation this Friday.  Hopefully others will have some data soon with their boards to report on.

70mv is the input to the first stage amplifier. B trace is the input to the final amplifier at 800mv and the A trace at 300mv is the output.

I'm getting pretty good at building complicated attenuators.

73, Peter

Xtall Filter Bilateral Amplifier

Seemed to take the better part of yesterday for the testing and trouble shooting of this board.  So, here are the results and observations.  Be as it may. :-)

1, I'm not completely confident on the phase relationships of the traces. I saw numerous differences as I made the measurements. Not sure why.

2, Reference is the schematic by KK6FT and N6QW, fig 7,  page 28 of the April 2015  QRP Quarterly.  I've uploaded a picture of the schematic here for reference to go along with my scope screen shots. Incidentally, QRP Quarterly is a great magazine for the homebrewer.  I highly recommend  subscribing to it.

3, Frequency input is 4,914 mhz.

4, 6vdc fed to both circuits and both n/c positions of where RY2 will be are grounded.

First screen shot shows A trace on the bottom and B on top. A is the filter output and B is the output of the toroid T2. No issue here as the voltage out is about 1/2 of what is fed into it.  

2nd screen shot shows the input to the filter on A and filter out on B.  As other screen shots will show the signal is somewhat distorted going in but gets cleaned up by the filter.

Here I'm looking at the output of the first toroid T1 to the 39pf cap on A and B is looking at the input of the xtal. Again the distortion is cleaned up a little through the cap.

 

The 4th picture shows the input to the board from the signal generator.  This was a surprise as I expected a nice looking sine wave.

Next up is the signal generator output, but not connected to the board.  I wanted to make sure how the signal looked in case it was the board having some effect.

 Finally a shot of the board's output depicting how it was processing the signal. Looking at it I'm not sure I really am happy with the results.

Up for discussion:

1, Why is the input signal being distorted by the circuit?  Is it due to an impedance mismatch? I did try various levels with no change in behavior.

2, The signal is cleaned up completely after the first crystal. I'm glad it happened but I did not expect the filter would have to do this.  I thought it's only function in life was to reject all but the band pass frequencies. 

3, The signal looks fine coming out of T2 and at the base of the 2N3904.  After that its a mess and you can see the output after the 100nf cap.  

4, Construction of the board is me-squares style etched into a board. I suppose its a bit of a mess and is this contributing factor? i don't think so but since I was treating this a prototype board I was not all that concerned. There was not too much thought put into the layout, just follow the schematic and solder.  So there would be room for improvement in a finial version.

5, I did play a little with the supply voltage.  I varied it between 5 to about 6.2 or so while watching the output signal. My thinking was where in the characteristic curve of the transistor was the signal operating? I was curious to see if the quality of the signal improved or not.  It did not.

6, Input from the signal generator varied from about 50mv up to around 300 to see the effect. Nothing to really note as it was done quickly to see if that was an issue.  The screen shots were taken with the sig gen set at 100mv.  A more detailed investigation will be done here once I know what the output of the previous stage in the LBS project is known.

7, I think I need to investigate further the 2nd amplifier.  I'm not confident its working correctly. I am not getting a signal I expected on the collector of the 2N3904.  1volt signal, clean, is on the base and 40mv with hash on the collector.

8, Even though testing of the xtall filter showed a bandwidth of  of about 1,500hz, I noticed when adjusting the frequency of the input, it appeared to be passing far more.  In fact a great deal more. At least a meg.  Now it was quick changes to see the effect and I didn't take any measurements, yet.  But while varying the signal in 4 steps, 4.913, 914, 915, and 916, the output signal did not vary much if at all with 4.914 or 4.916.  Again crude adjustments done quickly and a more detailed look is needed.  

9, I covered a lot of ground and discovered I left out 2 caps, installed connectors and jumpers, bad dc jack, forgetting an important step in the setting up the signal generator, and more than a few scope issues.  Long day at the bench and could have overlooked something easily. So be gentile if there are any comments, (hihi) hope there are, and look forward to any suggestions or ideas.

For those of you who know, beware,  this is only a preamble to the thesis. hihi.

73, Peter 

Antenna Plot

I ran a return loss sweep with the sna and exported the data into Excel. The resulting graph below.  There is no vswr data in the exported data file so I can't display it here.  It is displayed on the sna screen though.

It's an interesting plot, that seems to work very well. I did a sweep from 7 mhz to 29mhz. It was a little chilly this morning and I didn't want to mess around with several runs focusing on the individual bands.  Maybe this afternoon. What I don't know unfortunately, is how much effect the tree has on the results with the antenna leaning up against it? I hope a lot. hihi. 

I do remember years ago trying to tune up a ground mounted vertical I had then. I recall very well the dramatic effect I had when I stood near the antenna.  I can't recall how far away I had to stand to minimize the effect, but I think I was limited by the amount of coax I had and was forced to make due. I seem to remember trying to look at a very small screen from a distance so I reduced the effect my presence was having. Don't have an issue in that regard any longer.

I don't really have time to set up a temp. structure to mount the antenna for further testing so I'll have to hope for the best once its up in the air and re-run the sweep for a comparison.  That alone should be interesting regardless of the results.  If its really out of wack, then I'll make due for the winter and haul it down next spring and set up something.  I have a 10' section of tower that I could use and just guy it off, but that maybe easier said than done.  Anyway I have a 5 band antenna and I see 5 peaks in the plot.  Good enough for now.

73, Peter