This page started on Feb 2, 2016.
This page last updated on Mar 30, 2016.
This page will contain all the information to build a qrp transceiver. It is assumed that you have about 1 to 2 hours per week to set aside for this project. This is not a race and this is not a required course in college. But, there will be homework. :-)
Like all buildathons, the leader gets to pick the project. I chose the 1W-20 for several reasons. 20m is good year after year. It has some down time, but it is not like 15m and 10m, which are very dependent upon the phase of the sunspot cycle. 20m also requires less real estate for an antenna.
I chose the 20m version from all the bands available as it is the only one, right now, with the V3 PCB. This one has the two 4.7M (4M7, or 4.7 Megohm bypass resistors in the muting circuit). I am 7 states away from WAS on 20m with my 20m 1W and I need you to get yours going to help me out. :-) Especially if you are in DE. K7TQ needs you in NE and MT.
I am pacing this course for several other reasons. In the March time frame a new kit company will start up. There are a couple of kits that I am excited about that will aid you now for this build in later life for all kinds of RF projects. You will need to have some sections unbuilt in order to take full advantage of what they will do for you. So, be patient Grasshopper.
I bought serial number 501 from Diz. I get nothing for this project. I am doing this to help you. After we are done, I ask that you set up a local group to build the same transceiver and you lead them. It is important to get more hams excited about building and operating. We need new blood or this hobby will die.
Some research material to save and read.
Building and Measurement Steps.
I know at this point you are chomping at the bit to do some more. But, I want everyone on the same page. Please. Take a break and study some theory. Where would you go next, since I assume you do not have a lot of test equipment? Think about what you would like to do about putting a finished transceiver in an enclosure. I will get to that issue early on so that you have time to get your solution done before you get the transceiver completed.
Click on the thumbnails to get the full sized image.
1W bags of parts
Mono and SMD caps plus -43 toroids
The first photo is of my workbench. It is usually not this clean.
The bags that come with the kit.
This bag is important. The two black toroids are -43 material. The round toroid is for L4, which is installed almost at the end of construction. Put it off on the side somewhere safe so that you do not get it mixed up the -61 toroids. Please. The binocular toroid (the rectangular with two holes) is used for T5 and will be used what I call Phase 4 of construction. I suggest you use the green mono caps for the 100nF caps. Put the SMD caps in your parts supply. Also, do not install the mono caps all at once as per Diz' instructions on the web page. I will install them in order they are needed.
Mono caps and -43 toroids
Header + D8 + C49 Installed
I separate parts out and put them in 1oz cups that I get in the plastic cups department at Wal*Mart. $2.79 for 50 with lids.
I am using a 12V gel cell for this build.
Here is a 0.10" header for power connection. D8, 1N5817, diode and C49, 47uF electrolytic capacitor installed.
Photo of Measurement at C7.
Look at the schematic. Look at the lower right hand section where J6 is located. The LM-386N-4 audio amp is connected to the +12V input through pin 6 (U6-6). The voltage measurement is to make sure that we have power going to the IC when we install it next.
Why is the voltage not the same at C7 and U6 as it is at the battery? Well, there is a diode in the path. The voltage drop across the diode will be 0.10V.
Leave the power supply hooked up for a bit, say 15 seconds or so, and then remove the battery and hook it back up in reverse. That is, the positive terminal to ground and the negative terminal to the positive terminal of J3 (pin 2). Now measure the voltage at C7. You should see the voltage start at some voltage between 0 and the voltage you read previously. You will note that it is decreasing in value. What is happening is that you are discharging the voltage across the battery left from the previous connection. With the reverse voltage at J3, you will see the voltage go to zero and then go negative. Do not leave the battery connected for long after this happens. You could possibly damage the 45uF electrolytic capacitor.
Before building the next section, let's read up on just where we are going and what we are doing. You don't learn by just willy nilly building something. Unless that is what your goal is. Mine is to learn something new every time I build something.
First look at the block diagram of the transceiver. We are going to build the receiver section first. Without any test equipment, you can hear what is going on with just your ears.
The signal comes in from the antenna and then goes through the following sections before being heard in the speaker or headphones. Here is the path. Print off the block diagram and highlight the path.
ANT -> T/R Switch -> HF Mixer -> Crystal Filter -> BFO Mixer -> Audio Amp -> Speaker
Without any test equpment, other than a DMM, let's see just how far we can get before we might need something else. OK?
For now, let's assume that we are listening to a carrier on 14.059MHz. If it is a S9 signal level, then the voltage level will be 50uV RMS at the antenna terminal. It would take a high dollar piece of equipment to look directly at that signal level. The receiver will take that signal and pass it through several stages of amplification and conversion to get a tone at the speaker that we can hear. As I progress through this building of the receiver in stages, we will discuss what is going on.
The only thing that we can detect without electronic test equipment is sound. Let's use that to our advantage. We'll start at the end of the receiver chain, the audio amp, and build it first.
Let me be the first to warn you. Your ears are precious. You only get one set issued to you for your entire life. Please. Take care of them. I will remind you of this from time to time. NEVER put on headphones and then turn on a rig. NEVER. Especially when building something. You do not know what the volume level is set at and you certainly don't know if there is a loud signal going to come from the audio amp. I turn on a rig and then put on my earphones/earbuds. I never listen to CW with a speaker. I hate it and that is a book unto itself. Onward and upward.
The datasheet has some typical audio application circuits, but all of them are what is called single-ended input. One of the differential inputs (the + and - in) is grounded. That is not the case for the 1W. Both inputs are fed without grounding from the differential output of the NE602A mixer U4. Just ignore the muting JFETs for now. In this mode, the noise or hiss background is significantly reduced. That is why the receiver is quiet.
I find that the typical CW operator, for some psychological reason, turns the audio gain up on a receiver so that the atmospheric noise level can be heard. I prefer to turn down the audio level as low as possible until the signal that I am working is barely readable. It saves my ears and it makes me concentrate. CW can not be hammered into the brain. Period. It has to be worked for. IMHO.
This is the first section we are going to build, test and measure. But, we are going to build it in two sections. First starting with C37 and C38 and everything to the right. You have to decide how you are going to connect the transceiver to the volume control and headphone jack. Either using the connector kit (optional) from Diz or use your own wiring. While building you do not want to mess with wires getting in the way of putting parts in and turning board over to solder. I like to use the 0.1" headers and connectors. But I have used a short pair of wires to the 3.5mm jack and just be careful in moving things around.
At this point, when you put on the headphones or use the computer speakers, you should hear a hissing sound. This is the thermal noise in the LM-386N-4 IC. Remember, we are not using the volume control pot, thus the volume is at maximum value for this test. That is why you should not put on headphones upon power up. If you are using powered speakers, then have the volume set to a minimum at first and then turn it up until you hear the hiss.
Take a 15cm length of INSULATED wire with 1mm or so stripped off one end. Touch the left pad of C37 and then the left pad of C38. You should hear an A/C hum in your speaker(s) or headphones. This proves the audio amp section is working.
Here I am going to repeat the testing sequence above, but this time output the audio from the LM-386N-4 circuit to a graphical analysis program on a computer. Now, there are a lot of programs available. I am an avid linux user, thus I will be using linux programs for a lot of my stuff. I just don't have the time to research and learn Windows 10 software. If you want to contribute, feel free. I will put your results with the student photos, but I must be able to use your name and call with the material to give credit where credit is due. From this you will probably get some emails asking just how you did it. Be prepared to become a teacher too.
First a baseline spectrum audio display with just the DSP audio in the computer with no input. I have not tampered with the input gain control.
Next comes just the Phase 2 Part 1 build. This is just the audio amp, LM-386N-4 running wide open to the input of the computer.
Now, repeat the step whereby I touch the left pad of C37 with the 15cm wire. Note the 60Hz signal, if you can, on the far left of the graph. There is some blank space between the start of the noise and the peak. The divisions (of which not all are shown in the screen capture) are 1KHz wide on the horizontal scale.
With the addition of stages in the receiver I will measure the audio output to see how much background noise is added by each stage and if the spectrum response changes significantly, which it will.
Here is a sound snippet of this stage of building. First 5 seconds or so is just the sound card. Then power up the audio amp. Again for about 5 seconds and then touch the left pad of C37 with a 15cm piece of insulated #24 wire with 1mm stripped from the end used for touching the pad. Try to be very careful when running this tests. I am not responsible for accidents.
Double check your work. Power up the xcvr and listen for any changes. Now, using the 15cm test wire, touch pins 4 and 5 of the U5 socket. This will inject AC hum into the audio circuit. You should get results similar to doing the AC hum test at C37 and C38 before.
Here is a picture of the assembly up to this point.
This concludes the first Phase 1 and 2 build. You now have a working audio section that we will use to double check the rest of our work on building the receiver and even in building the transmitter section. Since I am taking a week between sections, I ask that you either build, borrow or buy a crystal checker. It will help double checking crystals for the IF filter later. The Chinese are on holiday for their New Year celebration (Feb 9th at the day of this writing), so ordering will be delayed one week for ebay items. Don't spend more that $8 USD for a crystal oscillator and freq counter combo like this one.
I showed you how to build this kit as a first exercise at:
Your due date to complete the above Phase 2 work is Feb 15, 2016. I'd be honored to receive photos to put on this page. Be proud. Show off your handy work. Prove to the world that I am not the only one doing this. Please.
Here is the schematic of the audio detector mixer or BFO of the receiver.
This section takes, for illustrative purposes, a 6.001MHz input into the NE602A from the IF filter (X3. X4 amd X5 plus the caps C29, C30, C31 and C32). The 6.001MHz signal is combined (mixed) with the internal oscillator signal generated with X6 (6.000MHz) and results in a 1KHz signal and some other stuff that we can't hear. This 1KHz tone is propagated through the muting circuit and into the audio amp and amplified into a large enough 1KHz audio signal to drive the headphones or a speaker. We'll learn how the IF filter works in the next section. So, for now, let's build this puppy..
The first step is to install a 6.000MHz crystal (20m version of the 1W). There are five such crystals. If you have a crystal tester, then first measure all 5 crystals and pick 3 of the crystals that are closest in frequency. If you can not do this, then just insert a crystal at random. By matching crystals, you can get a better IF filter and a more narrow filter by 10's of Hz in bandwidth. Probably not enough to be heard by the human ear.
OK. The moment of truth once again. Clean off a spot on the bench. Get ready to power up again. Double check everything. Connect headphones or speaker to audio jack and then power up the transceiver. Should hear the same old boring hiss.
Now take your 15cm piece of wire with 1mm stripped end and place the end at the upper pad of C32. This is where the signal will be coming into T5 and then into the BFO. You should hear a different noise than in previous tests. There may be a faint tone of about 1KHz and if you are lucky you may hear some voice in the noise level. This is caused by local broadcast station(s) demodulation from all kinds of mixing arrangements in your environment.
If you have a crystal oscillator. Take one of the remaining 6.000MHz crystals (not the 8.064MHz crystal) and put it into the test oscillator and turn it on. You should now hear a tone in your 1W at this time. Frequency can be anything from a few hundred Hz to more than 1KHz. Try to estimate what it is and write it in your notebook. You are keeping notes, aren't you?
Here is an audio spectrogram of the above scenario at a desk other than the building workbench. Just to get close to this computer. Note the several tones shown as straight vertical lines in the waterfall display.
This concludes the build and testing of the audio detector mixer or demodulator or BFO section.
Due date for this build section is Feb 22nd, 2016.
This section narrows the response bandwidth of the receiver to about 500Hz or so.
Here are 3 photos taken during the build of this section.
IF Filter Section
IF Filter Section
IF Filter Section
The only test we can do at this stage is to again power up the transceiver and using our 15cm test wire, touch pins 1 and 2 of U4. You will hear a quiet click when touched or removed. Not much noise at all. This is because the U4 mixer has no local oscillator input. In this case we have not built the VXO circuit. So, there are no appreciable signal levels coming out of the output pin U4-5 (pin 5).
After we finish the receiver, we'll come back and measure the response of the filter using some computer software to determine the characteristics of the crystal filter.
In this section we will build the heart of the transceiver, the variable frequency oscillator (VFO) using a variable crystal oscillator (VXO). We use a crystal oscillator with a variable capacitor in the form of a varicap (voltage controlled capacitor), MV209. This allows us to move our receiving and transmitting over a range of frequencies instead of being fixed at one frequency using the crystal. The range we can cover is band dependent depending upon the frequency of the crystal and just how much we can change it.
OK, now time to test our build. Setup as previously for power up and testing.
When powered up you should now hear your usual noise in the form of a hiss, but maybe quieter. Now, taking your 15cm test wire probe touch either pin 1 or pin 2 of U4, the receiver mixer. You will hear a much lower range of frequencies for the background noise. In fact, the range is going to be less than 1kHz for the maximum amplitudes. Here are two screenshots using baudline to show the background noise with no input to U4 and then with wire touching pin 1 of U4. Note the vertical lines, if they show up on your display, are 1kHz apart and the peak is below the first 1kHz marker. And if you look the IF filter looks to be about 600Hz or so. We'll make more detailed measurements later to determine the exact response curve of the filter. We are almost done with the receiver section.
No input to receiver.
Input at pin 1 of U4.
Hopefully, at this point you should have a working receive signal chain from the first RF mixer to the audio PA section. Congratulations.
This section narrows the response bandwidth of the receiver to about 500Hz or so.
For this section we are going to wind T4 and install C26, the variable trimmer cap.
The test for this section is to fire up the transceiver again. Touch the 15cm test wire to the left pad of L3. This will inject a signal at the top of the primary of transformer T4. You should hear atmopsheric noise. It is difficult to say just how loud this should be. In my transceiver this is not much volume into the powered PC speaker system, but I can hear the atmospheric noise and from time to time a CW signal can be heard, when 20m is open. We do not have far to go to get the entire receiver working now.
You should be able to peak the noise level by carefully adjusting C26, the trimmer cap. It should peak in two positions. I will explain this in detail later. This will be adjusted again later.
If you are going to use a digital display, such as the one shown in the next video, then this is a good point to test it. The display shown here is from a new kit manufacturer that is about to be announced in early March 2016. The display costs $15 USD with $3 shipping for US domestic postal service.
I have put a 200pF mono cap from my parts bin into the spot marked C_s. This is the sampling coupling capacitor that provides the VXO frequency to the display plugged into J7. I would start with 200pF and then lower the value until you no longer get a steady realiable reading and go with one of the lowest values. This decreases the load on the VXO oscillator. I found that there was no change in the VXO range with even 200pF, which I consider a high value typically used for coupling caps in these situations. You can see the frequency in the video is steady and I checked it with a DDS signal generator. The display is only off by 100Hz and can be adjusted with the trimmer cap (green in the upper left of the PCB).
I will come back later to see if there are some more steps we can do to give you more information.
The tone you hear is from a DDS signal generator on the workbench.
The transmit/receive switch (TR) section consists of only the L3, D! and D2 and C5 and C51 components. It also serves as a tuned band pass filter (PBF) at the same time due to the series resonance created by the capacitors and inductor being in series. You can neglect the effect of the D1 and D2 pair of diodes in the receive mode.
I am concerned about the level of understanding of an individual, when they post on the Internet or in emails and postings online, that a diode does not conduct until the applied voltage reaches 0.7V in the forward direction for a silicon diode. BS. It conducts at any non-zero applied voltage. It is the effective resistance of the diode and the current level that matters. For very small voltages in the microvolt range, the current is very very small. Almost to the level of counting the number of electrons that move. :-) With small levels, in the receive mode, we can consider the resistance of the two diodes to be very high and neglible for this discussion.
So now we have eliminated signals below the 20m band with C5 (C51 is not used on 20m) and L3. If you carefully go back to the advanced test in phase 7 and view the video, you can hear a signal that occurs around 8.0597MHz for the VXO, which corresponds to 14.059uMHz AND 2.0597MHz on the input from the wire (pseudo antenna). Since this is the difference between the VXO and the IF frequencies. You do the math. Not difficult.
Let's build the TR/BPF section and test.
Now hook up transceiver and power up again. Put pseudo antenna at the lower pad of C4, which is not installed at this time. You should hear noise level increase at this time. Tune C5 for maximum noise level.
If you have a tuning pot setup, then tune the entire band and you might hear a signal, if 20m is open and the antenna is long enough. Note the abscence of of the prior tone, if you had one like I did, at the lower end of the band spread.
OK. Screw up on my part. Installed C5 and L3 as per my instructions. I hooked the system up and played with it. Works as instructed. I just did not get a video. Bummer. You do the experiment any way. Enjoy.
The low pass filter section, LPF, has a cutoff frequency just above the 20m band so that any signal(s) generated by the transmitter are not allowed to the antenna. This is for FCC regulations and requirements to prevent interference from out operating at any time. And a number of other historical reasons. Ever note that most of the radio amateur radio bands are harmonics of each other? I knew you did.
The assembly of the LPF section requires the installation of the following parts.
Now would you believe that you have now completed assembly of the complete receiver in the 1W? Congratulations.
Now connect up and power up the transceiver and connect an antenna to the ANT input pad on the top left hand side of the board. Now readjust C26 and C5 for maximum background noise and signal levels. If you have the tune pot setup, you can now tune up and down the band and listen to CW signals from around the world when the band is open. I hope all is working for you. Here is a video of what I got at this time.
Note the absence of the tone at 8.0579MHz dial reading. I think that it was the multiplexing of the display by the uP. The signal was present in a previous video when I had the receiver powered up without an antenna. This is how you find things out.
OK. Let me explain what the receiver is doing and how it operates and why I did things the way I did to simplfy our lives.
The following discussion is not into a lot of detail and I've used some leeway in the discussion, so don't get retentive on me. This is an approximation in some areas to keep from having to write a book. Let's look at how the receiver gets to the signal we want.
First, let's assume that there are signals very 2KHz from 1MHz up to 30MHz. Here is what the spectrum would look like at the antenna. We will neglect selection effects of the antenna for this discussion. I am also assuming that there is a signal of interest at 14.059MHz, say a NM CW operator.
The first thing the signals do is pass through the low pass filter (LPF) as shown in the following diagram. Sub the parts values for the band of your choice. Here we are assuming 20m and the appropriate values. The LPF will remove signals above the cut off frequency. Let's assume 15.0MHz for this discussion.
Now the signal spectrum will look like.
As it turns out, the LPF doesn't do anything for us for 20m. The receiver can only copy signals at 8.059MHz (the VXO freq when listening for 14.059MHz), 6.000MHz the IF frequency, 14.059MHz the sum of the VXO and IF and 2.059MHz for the difference between the VXO and IF freqencies. All of these are below 15MHz already.
Now the spectrum has to pass through the TR switch.
The circuit components are C5 and L3 in the transmit/receiver switch (TR), which you can locate in the main schematic. D1 and D2 we can neglect, as the signal is so small that the D1 and D2 effectively have a very high resistance/impedance. C5 and L3 form a tuned series circuit and thus become a band pass filter (BPF). I am going to assume that C5 and L3 are peaked for the middle of the bandspread, so let's say 14.059MHz. Let's also assume a Q of 200 for the circuit. SWAG it. With a Q of 200, the bandwidth will be 75KHz, so let's allow 37.5KHz on either side of 14.059MHz to get through. Just for illustration purposes.
Now the spectrum will now look like, with frequency band expanded in the graph....
Now this merry band of signals gets fed to the input of the first receiver mixer through a transformer to pins 1 and 2 of the NE602, U4. The VXO frequency is, say, 8.058MHz, which will be subtracted from the input frequencies and then the output sent to the IF filter from the output pin 5 of U4.
The spectrum will look like the following. Note the shifted frequencies.
Now the signals go through the crystal IF filter. Let's say the filter is 700Hz wide and it is centered on 6.0005MHz. This removes all the signals except the one we want. Now the spectrum looks like the spectrum graph following the link to the IF filter schematic.
Now, at 6.001MHz, we are unable to hear this signal. We need to convert it to an audio frequency we can hear. So the beat frequency oscillator circuit, shown in the next schematic section, subtracts a 6.000MHz signal from the input signal, thus resulting in an audio signal. This gets passed on to the audio amp through the muting circuit, which is disabled during receiving periods of time.
As you can see, we wind up with a nice clean 700Hz tone, assuming that all the frequency offsets are tuned just right. Of course, I did skip some gory details and I did not do the arithmetic exactly right on the plots and offsets, but you get the general picture. I will come back and clean up as time permits. But for now, let's get on with the build.
This page has gotten long enough. Let's now go on to building the transmitter and keyer sections and for that I will start a new web page. Go to it and bookmark it for immediate access.
My email address is the usual hiding from Internet bots.chuck dot adams dot k7qo at gmail dot com