More Bits and Pieces
In our last post we described the Band Pass Filter simulation with LT Spice and the measurement of the filter performance. There is no one more astonished than me that the O Scope plots seem to track very well what was shown in the simulation plots. Needless to say it is a lot more productive to spend time simulating the hardware and eventually honing the design to a point where one is satisfied and to only have to solder in the components --one time! I also want to highlight the use of the older publication SSDRA by W7ZOI in conjunction with the LT Spice.
By way of review we have built the Audio Amp Stage, the Product Detector, the LO and the Band Pass Filter. This has resulted in a Direct Conversion Receiver. All of these pieces will be used in the Superhet version of the receiver albeit with a modification of the LO frequency range and the addition of new modules. Items to be added include: Beat Frequency Oscillator, Intermediate Frequency Amplifier stages, a Crystal Filter, a DBM Mixer stage and an RF amplifier stage. For those would like to follow along our block diagram is as shown below. The darkened blocks are those that were built for the Direct Conversion Receiver and those in white will be built for the Suprhet version.
Noteworthy, as this came up as a question, virtually all interfaces are at 50 Ohms so there are matching transformers, where required, to make the match to 50 Ohms. Once again the use of the J310's are used in the remaining blocks. The BFO is a J310 and the RF Amp and IF Amp use the J310's in the cascode configuration to simulate a Dual Gate MOSET. The RF and IF amplifiers are virtually identical (and similar to the Product Detector). The differences are that the RF amplifier is designed for use at 7 - 7.3 MHz and the IF amp Stage design is centered on 12.1 MHz which is the IF design frequency. Later we will cover in detail the matching to the modules
Essentially the incoming frequency is up-converted to the IF frequency using the LO operating in the 5 MHz range. While an LCD would show 7 -7.3 MHz the actual LO RF being produced is in the 5 MHz range.
This is by design since there are those homebrewer's who do not like the use of modern technology such as the Arduino and AD9850. Some homebrewer's prefer an "absolute homebrew project" where everything and I mean everything must be scratch built from discrete components. There is much to be said for such an approach as it does lend itself to understanding what each and every component is contributing to the circuit and in detail learning how the circuit works. For myself I tend to build the first one with discrete components and then move on to the "black boxes". I know how the circuit works; but now want to take advantage of the expanded capabilities of the modern technologies. Both result in a working radio -- it is all a matter of choice.
Thus in an attempt to provide choice we will suggest that a VXO operating in/around 5 MHz as well as a conventional 5.0 MHz LC VFO or even a varactor tuned oscillator such as you might have in a self excited NE602 or SA612 will work. The choice is left to the builder. I will present a 5.0 MHz crystal switched VXO design that will give small segments of 40 Meters for those who feel technically or emotionally uncomfortable with the digital technology. Fortunately there are cheap computer crystals available to make this possible. The ranges cover portions of the SSB frequencies and only a limited amount of the CW frequency range. Mind you that is only because of the use of stock catalog crystals.
My plan is to build the Simpleceiver on a Bread Board just like in the old days and my build philosophy is somewhat different than others. A recent article I read on homebrewing started by the author suggesting that the builder start with the most difficult circuit and get that working first. Well I guess if you know in advance what will be the most difficult circuit then I suppose that would work. But if you don't know which is the most difficult, therefore some other process must be effected.
I always start at the back end of the project and work my way forward which implicitly suggests that as you build forward the radio itself becomes part of the test system. It also provides way points of proven performance. Heathkit pioneered that method which I think is sound.
So I started with a Bread Board and installed the audio amp on the board. When I built the Product Detector I left room on the board for the BFO as it should be as a close as possible to the detector. The following photos show the build and you can see the O Scope showing the waveform coming from the BFO -- it is about 4 volts Peak to Peak. We may add a small trimmer pot on the board to fine tune the BFO input to the Product Detector but that can be done later. In passing note the frequency on the O Scope. That will be used in the Arduino Code.
The very first bit and piece I would like to present is the BFO Circuit which is shown below. I did not simulate this in LT Spice but simply grabbed one of my stock circuits. The 50 PF trimmer cap is used to "wiggle" the 12.096 MHz crystal to place the signal on the appropriate frequency slope of the crystal filter. Since we are using the 5.0 MHz Frequency for the LO there is not a sideband inversion thus for LSB the BFO frequency must be ABOVE the filter center frequency in this case nominally 12.097500 MHz. Now this actual number becomes critical if you are using the Arduino + AD9850 as that then becomes the "Frequency Offset" used in the code. By having these numbers correct --when the display says 7.203 MHz -- it really is 7.203 MHz. This circuit gives a good account of itself and should be easy to replicate.
Now a few words about impedance interfaces where we attempt to make everything match 50 Ohms
- Product Detector Output to Audio Amp. On the Output side we matched to 10K which essentially is a 10K Pot so we can match to the audio amplifier input. On the input side the Gate #1 likes to see about 2.2K (SSDRA ~ Hayward). But we want to have this at 50 Ohms. So we need a step up from 50 Ohms to 2.2 K --that match is a 1: 44 (2.2K/50). If we have a 3 turn primary to a 20 turn secondary --we are awful close. The impedance transformation is based on the turns ratio squared. Thus 3^2 = 9 and 20^2 = 400 and following 400/9 = 44.44 --close enough. Now you ask how did you know where to start --well that is Tribal Knowledge! I have wound enough matching transformers to just know what gets you close. BUT at one time I did create an excel spreadsheet and then I had a catalog of values. By the way get a large stock of FT-37-43 cores as they are used through out the build.
- The 2nd IF Amp will feed the product detector and its output is already low impedance so a small 10 NF coupling capacitor can connect the output of the IF amp to the input of the Product Detector. On the input side it uses the same input as the Product Detectors so you have the same 3 Turn to 20 Turn Transformer.
- The Crystal Filter was designed using a canned program on the internet. The Z in Z out is 300 Ohms. So we a match from 300 Ohms to 50 Ohms which is a 6 to 1 transformation. So if we used a 8 Turn to 19 Turn transformer we would be close. 8^2 = 64 and 19^2 = 381. Thus 381/64 = 5.953. Good Enough!!!! the same transformer would be used on the input side only reversed 50:300 in to 300:50 out.
- The 1st IF amp is like the second only with a twist -- Hayward has suggested terminating a Post Mixer Amp with a resistive pad to provide a constant impedance to the amp and to prevent distortion. I simply used a 2dB Pi Resistive Pad Calculator on the Internet and came up with stock values of 470 Ohms and 12 Ohms. The In/Out is still 50 Ohms. So the output from the Pad is simply connected to the 50 Ohm Port on the crystal filter matching transformer. The input transformer is the same as the 2nd IF Amp and the Product Detector.
- The SBL-1 is already at 50 Ohms In/Out and so Pins 3&4 are connected to the input of the 1st IF Amp.
- The Receiver RF amp uses the same design as the IF amps only the frequency moved down to 7-7.3 MHz. The out is 50 Ohms and so a 10 NF between the output port of the amp and the RF in Port on the SBL-1 is all you need. The input side of the RF amp --yep just like the IF amps and the Product Detector is already set to 50 Ohms.
- Thus it should become obvious we are using a standard design albeit with some slight modifications that is used through out the receiver. This same J310 DGM template will find itself in the transmitter stages. Thus a lot of design work is already done by simply creating the very first model.