EMRFD Message Archive 13788

Message Date From Subject
13788 2017-04-10 05:00:30 Ken Chase BPF testing advice
Hi All

I have built a 40m and 20 m BPF. Not of my design but from EMRFD page 6.77 Figure 6.135.

I would like to test them but I don't have a SA or fancy equipment, just a scope.

My signal source is a SI5351 into a 40 m or 20 m LPF to get rid of the square waves, then into the BPF.

I assume I can start with say 6 MHZ, note the amplitude, increase the frequency, note the ampitude, etc.

Am I taking the correct approach?

I also want to test the input and output impedance of the BPF's.

I set the SI5351 to 7.15 MHZ and measure the no load amplitude. Then hook up the BPF and measure the amplitude. If the BPF has 50 ohm input the loaded amplitude will be half the unloaded amplitude.



13789 2017-04-10 05:51:37 Steve Dick Re: BPF testing advice
Hi Ken. Couple of comments.  I’m not sure what the LPF filter looks like after the SI5351 but you might want to add a resistive attenuator, say at least 3dB between the SI5351 LPF and your filter under test to provide a wideband impedance close to 50 ohms for the filter under test, more independent of any impedance variations of the SI5351 LPF.  The filter under test must also be terminated in 50 ohms.  Your approach will be fine for testing for 3dB bandwidth and in-band ripple. However, the scope will likely not have enough dynamic range if you want to test out of band attenuation far away from the filter passband.  For that you would need a higher dynamic range power meter such as the W7ZOI RF power meter based on the AD8307. See: http://www.qsl.net/sz1a/download/build%20an%20rf%20power%20meter.pdf  A PCB based on this design is available from the PHSNA yahoo group from time to time.  One more precaution: This approach will NOT work very well when looking at the low side of the BPF at frequencies below half the center frequency of the BPF because you will see energy of harmonics of the fundamental. For example, say you have a 7MHz bandpass filter under test.  If your SI5351 is set to 3.5MHz,  you will not be measuring the 7 MHz BPF attenuation at 3.5 MHz.  Instead, you will see a strong second harmonic of 3.5 MHz which will show up as a strong signal within the 7mhZ bpf filter passband, giving bogus results. So the approach can be used only if harmonics of the SI5351 fundamental fall considerably outside the passband of the SI5353 LPF.
-Steve K1RF
13790 2017-04-10 09:07:05 Ken Chase Re: BPF testing advice
Hi Steve

Thanks so much for your input. I don't have much experience in this area, so your thoughts are appreciated.

OK, so using an attenuator makes a lot sense making the SI5351/LPF more broadband for a 50 ohm output. I was planning on doing this for reading the levels with an Nano in the near future.

I do have the W7ZOI RF power meter, so I better get it finished.

So if I was measuring the characteristics of the 7 MHZ BPF, what frequency should I start testing from? 3.5 MHZ? 4 MHZ? 5 MHZ?



13791 2017-04-10 10:44:12 Steve Dick Re: BPF testing advice
Hi Ken. I would suggest putting this filter design into one of the available programs such as the student edition of ELSIE at http://www.tonnesoftware.com/elsie.html or AADE filter design at http://www.dxzone.com/dx17873/aade-filter-design.html  If you simulate the filter, the simulation will give you a pretty good idea of how the filter should perform.  The simulation is a good starting point but does not take into effect stray inductances or capacitances or inductances in grounds. That might give you an idea of how far you go in either direction around the in-band part of the response.  First, tune each section to give maximum output voltage at the center of the band.  That should be the geometric center.  For 40 meters, that would be the square root of )7.0 X 7.3) or about 7.15MHz. If you just measure the 3dB frequencies (half power points or 0.7 voltage points) and these agree with the simulation, the filter is working fine.  The simulation will predict within reason, how much attenuation you might expect with ideal components.  High Q components will give results close to ideal. Using a scope, you will not be able to measure very far down the response. 
-Steve K1RF
13792 2017-04-10 18:43:55 Ken Chase Re: BPF testing advice
Hi Steve

I have done a LTSpice model of the filters and it does show good response. Reason I wanted to measure the filter to see exactly what was going on.

Instead of using a scope to measure, could I use a 50 ohm load on the output with a diode detector and monitor the DC level?



13793 2017-04-12 07:18:48 Steve Dick Re: BPF testing advice
You’re better off sticking with the scope.  The diode detector will not work for very small signal amplitudes due to the diode not conducting blow its threshold voltage.  You can also add a good wideband amplifier and switchable attenuator between the terminated filter and the scope.
-Steve K1RF
13794 2017-04-12 12:35:51 k7tfc Re: BPF testing advice
Though not using the gear you have on hand, you might be interested in Bill Meara's (N2CQR) video showing the sweep of the band-pass response of an IF stage. The same method can be used for a filter (LP, HP, or BP).



Todd K7TFC
13795 2017-04-12 14:40:26 k1rf_digital_stev... Re: BPF testing advice
Thanks for the post - looks like a nice low cost signal generator if the frequency range suits your needs.
Steve K1RF
13796 2017-04-12 20:08:28 Ken Chase Re: BPF testing advice
Thanks Steve. This is new ground for me but nothing tried, nothing gained.

And check out Todd K7TFC link.