Friday, November 27, 2015

Simpleceiver ~ Part 16

Hey? Nobody has asked about the Crystal Filters that are being used with the Simpleceiver Project!

Addendum: 11/29/2015 Another on the air video of the Simpleceiver
Addendum #2: A view of the completed Simpleceiver "al fresco"
Addendum #3: 12/01/2015 Data for a 20M RF Amp
Addendum #4: Signal Output Data for the AD9850
Addendum #5: Band Pass Filter Data for 20 Meters

The following is the schematic for the RF Amplifier stage. Please note about the Resistor "R" and how that is made.

Addendum #3: To use this RF Amp on 20 Meters (Query from WA7RHG) simply make L1 = 8 Turns on the FT-37-43 core and make C2 = 10NF.

Addendum #4: Based on an earlier input regarding why I used the AD9850 straight into the SBL-1 without a "booster amp". I guess the simple answer is I hooked it up and it worked --so keeping thing simple. But to get full advantage of the SBL-1 it probably would be a good idea for a booster amp. Now when you measure the output of the AD9850 + Booster Amp make that measurement with a 50 Ohm load and a scope. Do not do it connected to the SBL-1.

Here are three measurements:
  • With the AD9850 terminated into 50 Ohms V= 324 MV PTP
  • With the AD9850 No Load V = 800 MV PTP
  • With the AD9850 connected to the SBL-1 V = 440 MV PTP
The SBL-1 is a 7 dBM device and likes to see 1.414 Volts PTP and a homebrew DBM most likely will want to see 2 Volts PTP (or 10 dBm). So a homebrew DBM will need the booster amp!
 Addendum #5: Band Pass Filter Data for 20 Meters. It appears there is interest in putting the Simpleceiver on 20 Meters so here is the Band Pass Filter data. Please note if you are using the AD9850 you will need USB and so you need to take the LO plus the Offset to work 20M. Thus the VFO has to be in the 2.10 MHz range for 14.2 MHz. [NOTE in Part 17 you will see why this is not a good choice for a LO Frequency.]


The Crystal Filter

I guess this is a significant input to me as I have had no inquiries about the crystal filters being used with the radio. But just in case anyone was wondering here is some preliminary information to get you started.
First it is important to start by having three or four crystals (depending on which filter you build) to have the total frequency difference be no more than 50 Hertz across the units. To make that clear when you measure the frequencies the total difference from high to low of any of the crystals should be no more than 50 Hertz!!!! Typically I buy a dozen crystals  and from that batch will find that you can usually squeeze two filters with crystals that match that hurdle. Avoid buying four crystals and simply plugging them into the circuit. If by chance you would do this and the four crystal you purchased meet that criteria -- then stop wasting your time on homebrew projects and buy a batch of lottery tickets for you are one lucky dude!
This now gets to the problem of how to measure the frequencies to within 50 Hz. I have found that many hams new to homebrewing really lack some basic test equipment and there are few alternate paths beyond such a situation. A homebrewer needs something more than a rusty screwdriver, a beat up electric drill and analog VOM that is 10% accurate.

Eventually the serious homebrewer needs an O scope, a stable RF signal source , a frequency counter and a Digital Volt Meter (DVM). Most of the modern Digital Storage Oscilloscopes have a built in frequency counter. Thus one way of measuring each crystal  would be to build a test oscillator and measure the output from the oscillator and simply read the frequency on the scope. Another would be to have a frequency counter. (I happen to have both.) But something I have been recently using is my SDR Softrock transceiver along  with the Power SDR software.

Typically with the Softrock, I fire up the oscillator and have a "sniffer loop" (short chunk of wire connected to the SDR antenna port) from the Softrock brought near the output of the oscillator and simply read the frequency from the software display. The desirability of this approach is that I set the Power SDR parameters to CW with the narrowest filter and then look for a peak reading on the display. Then I can note down the frequency.
 But that only finds the three or four crystals that are close in frequency thus something more substantial is required to build a high quality crystal filter.There is a documented way to do this and I refer you to the following link Crystal Filter Construction from WA5BDU. Building a high quality filter can be done but must not be done in a haphazard manner. This tutorial is quite excellent and provides the formulas, equations and theory for filter construction.

Typically after I find the four crystals I sometimes think about purchasing a commercial filter. There is no reason that a 9.0 MHz commercial filter could not be installed in this circuit by modifying the 12 MHz amps to work on 9.0 MHz. With LT Spice that is not a difficult task. INRAD sells a 9.0 MHz a Four Pole filter with a Zin/out of 200 Ohms. (Model 315). 9.0 MHz filters are available also from the GQRP Club, Z in/out = 500 Ohmz. The beauty of the Arduino driving the AD9850 is that should you change the filter frequency, a few lines of code changes and you are there. The problem is somewhat more difficult using the LC VFO or a VXO. RF from the these LO's would have to be in the  2 or 16 MHz range for the 9.0 MHz IF. You get the idea.

Now for a not so high quality, not rigorously calculated, and not formally approved by the EMRFD and BITX reflectors, I usually  skip the tutorial and use some values that seem to work for me. Here is exactly what I did here. I assumed a Z in/out of 150 Ohms. My listening tests have shown that the values are not too far off. But then again my goals are a bit different as I am trying to define a template for a complete project and taking a shortcut on the filter may not be the best practice; but it does get the radio to a least the point of working.

But I encourage the readers to do the rigorous calculations so you can say I know how to do it and have done it! A rigorously designed filter will undoubtedly perform better and will satisfy the need to have everything precise and tidy!
Once you have the filter built and before soldering in the circuit it would really be a good idea to test your filter. (It will be impossible to test using only the rusty screwdriver, a beat up electric drill and the 10% accurate analog VOM. Thus you will need something more. )
One of the test procedures is to terminate the filter with a 150 Ohm resistor and with the AD9850 programmed to be a signal generator as the input and placing the scope across the 150 Ohm resistor "sweep" the filter at every 50 Hz starting at 5 KHz above and extending to 5 kHz below the center frequency. Record the amplitude points and then make a plot of the points versus the frequency and you will get a shape of the filter.
If you use my method, it will be a bit ugly (maybe more than a bit ugly) but if you use WA5BDU's method you should see plots similar to what is shown in his tutorial. Alternatively you can use the AD9850 and if you have a SDR receiver repeat the same procedure and use the S Meter numbers for the plot data.
Or if you have one of those snazzy SNA's (Scalar Network Analyzer) you can run an automated test and see the plot
I want to repeat my "good enough" filters will work BUT you should do the rigorous approach as outlined in the link. Once you do that then there will never be a question of how good is your filter.
The final chapter for this part of the project will cover the receiver RF amplifier stage which is very similar to the IF amplifiers stages previously covered.
Since yesterday was Thanksgiving here in the USA, I am still in a partial eating too much Turkey coma.
Addendum; On Air QSO K6XQ and KF6ZZ 11/29/2015
Pete N6QW


  1. Pete. I see your AD9850 output is feeding straight into the SBL-1. What Vp-p are you getting at the SBL-1? I too have a AD9850, but I only get around 600mVp-p when connected. I thought the SBL-1 needed at least 1Vp-p (ideally 1.4Vp-p) I also have a Si5351, and that manages 800mVp-p. I wonder if I should build a small Class A amp and boost it up to 1.4Vp-p?

    1. HI Charlie,

      Thanks for your email. Yes the AD9850 is feeding straight into the SBL-1. I have not measured the amount of signal but am planning on making a whole series of measurements which I will post on the blog. Just in case anyone else is reading this -- the AD9850 must be measured into a 50 Ohm Resistive load. It should not be measured while connected to the SBL-1 --the readings will be different. I have used a "booster amp" with the Si5351; but that was because of a homebrew DBM.

      I am trying to be careful here as frequently I will get an email like "YOU said this was the case and I didn't find the same result -- why are you wrong?" My original intent was to have a simple receiver -- now there is always a better mousetrap. Keep watching the blog for the update on numerical measurements.

      Yes for maximum signal conversion the SBL-1 is a 7 dBm device which translates to 1.414 Volts Peak to Peak. Actually a better lower drive device is the ADE1-L which is a 4 dBm device -- 1volt Peak to Peak. I have found that a homebrew DBM frequently will require almost 10 dBm (2 Volts Peak to Peak) and that meant a booster amp.

      Pete N6QW

  2. As an aside, I'm also building a 40m tramping (hiking) rig, but my SBL-1 output will go straight into a Teensy where I'll do the rest in software. Don't tell Bill or he'll have a fit!
    Charlie, ZL2CTM

    1. Hi Charlie,

      Wow please keep me posted -- I have no problem building radios with "black boxes".

      Thanks for your post.

      Pete N6QW

  3. Beautiful build as usual. I might use that architecture on my next radio. I like the fact that the J310 comes in a SMT package.

    1. Hi Dean,

      I picked the leaded version of the J310 because many homebrewer's shy away from Surface Mount. I know you built the TIA transceiver with surface Mount parts. Since you feel at home with SMD then move up to the BF991 from NXP Phillips. They cost about the same as two J310s but are good to 200 MHz and can tolerate up to 20 VDC. They work really well.



So How To Do CW on a Homebrew SSB Rig?

3/16/2019 ~ Remote Antenna Tuning 3/14/2019 ~ Calling All Junk Boxes???? Calling all Junk Boxes, N6QW is looking for a part! ...