SSB Single-conversion superhet transmitter

 

Juggled some wires from the receiver I made yesterday to get a transmitter.  The video shows me picking up the signal on my G90 (I'm transmitting into a dummy load).

My signal path is going microphone to the input of the first Si5351 (where the antenna was coming in on the receiver).  I switched the wires coming out of the Si5351 (so I didn't have to reprogram anything for now), so the ~9MHz is going into the first mixer and the ~5MHz is going into the second (opposite of the receiver).

I'm using this Adafruit board that has an electret microphone and amplifier together.  I made a little board that has a blocking cap (22uF - what I had handy) to block the DC bias the board has.  In the configuration, it is doing 60db gain and is enough to get 80mv P-P to the dummy load (~ -18dbm).  Will probably need to use another amplifier to drive the QRP Labs 10W linear amp.

SSB Single-conversion superhet receiver

 

I strung together a bunch of parts to make an SSB receiver.

Details (mostly for my later reference)

• Currently set up for 20m
• Bandpass filter is the QRP Labs design, just with parts I had laying around.  Don't have as nice of capacitors as his kit (they are no-name trimmer caps) and the bandwidth is more like 2MHz vs his 1.4MHz, but the insertion loss was about the same (-1.75dB) and does the job.
• I have one RF amplifier in - an AliExpress "low noise 30db" thing.  (I need to remeasure what gain I'm actually getting at 9V.)  Will probably replace for a MostlyDIYRF IF amp that has AGC.
• Both mixers are AD831 boards off of Aliexpress.  Want to see about swapping them out for some of the DBMs I built when playing with the SolderSmoke DCR.
• The audio amplifier is one from SparkFun, a class D amp based on the TPA2005D1 chip.
• The speaker is a Jameco Valuepro "4" 8 Ohm 5W 800Hz-10kHz Paper 97 dB"
• The IF bandpass filter is the Mostlydiyrf one that I posted about before.  I made little stand-alone matching transformers for it.  I'll probably swap it out for a similar crystal filter I made years ago.
• The VFO/BFO is from a MostlyDIYRF "UDVBM-1 Universal Digital VFO/BFO".  Will replace with my own arduino/si5351/display setup if I actually box this up, as this board is really nice for mocking things up and would like to keep it out for that.  Also, the code he has for this has a few things I'd change (the VFO shows the actual frequency, not the frequency+BFO, and the "click to change what digit you are changing" doesn't go high enough).  There are a lot of libraries out there or might roll my own.

No matter how many times I think about it, figuring out what side of the IF filter to put the VFO/BFO always does my head in, and of course I had it on the wrong side to start.  I like using FT8 to test, as you always know there are going to be strong signals there, and if my iPhone decodes it, we're probably good to go.

I think next step is that I want to make it a transmitter.  Really just need to make an amplifier for the microphone, use the QRP Labs 10W amp (must remember to take out the 40M bpf it has in it).  Use one of my various RF amps to get the power up enough to drive that.  I have various LPFs I've made over time sitting around.

Then see about making it a transceiver.  I have some relay modules that are ready-to-use with microcontrollers (have the diodes in for the current spike, etc.)  I think will be mostly figuring out how I want to attach different parts, updating the code to handle switching, etc.  The whole "boxing and socketry" thing is the part I'm least good at, so might take quite a while to get to it.

Anyway was well pleased 

Playing with a new crystal filter

 I recently put together this board - I wanted a crystal filter that looked a bit neater than the one I built previously.  Here is the board and it's frequency response directly.

That is a lot of ripple, as the filter does not want to "see" 50 ohms.  The email from the vendor said that this has an approximately "native" impedance of 150 ohms.  So I made input/output transformers to get to approximately that.  Was better, but still a lot more ripple than I wanted.

I found this post talking about matching the impedance of the filter to fix the filter's response.  It didn't click at first exactly the procedure, but eventually got it.  Inserting a resistor in series with the filter increases the impedance it "sees", and the filter response adjusts accordingly.  So I made to little fixtures of sma connectors with an in-line potentiometer.  The only ones I had were 10k, so it was touchier than I'd have liked.  But found the impedance looked more like the 300-400 ohm range.  Note that the impedance the filter "sees" is the amount of inline resistance + 50 ohms (from the nanovna).

I made a new transformer that had a 7:20 turn ratio, and got a much better result.  (So far have only made one on the input, still using the lower-impedance 5:9 transformer on the output.)

The bandwidth is 3.8KHz at the moment.  I need to have a play with the other capacitors provided to get the value down to something better.  (This is why the caps are on there the way they are at the moment - to easily remove them to try others.)  

Quick Pi Pico TX test

 Changed the hardcoded frequency to the WSPR frequency, threw on an AliExpress 2W amp and a QRP labs LPF, and used the transmit feature of the iOS "WSPR Watch" software (just holding my earbud up to microphone.  

Pi Pico SSB transmitter

I threw together the Pi Pico-based SSB transmitter from here.

I haven't had a chance to do much with it yet, but got immediate results and was very pleased to hear myself over my G90 from it.

Soldersmoke Direct Conversion Receiver Challenge

 

Preliminary success with new devices

 A nice chap that goes by Stretchyman put me onto some devices I was not aware of.  Specifically, the UCC27614DR gate driver and the GAN111-650WSBQ GaN FET.  I had a go this afternoon with throwing them together in a similar layout as my last try with the IRF510.

So far I'm getting good results.  Out of the gate I was getting 10W out with decent efficiency.  I haven't let it run long yet.  

As an aside, I noticed that my probe was having an effect when I put it on the dummy load.  Switched it to 10x and that fixed that.

A lot of progress over the weekend

 I haven't been writing this down, so it is hard to remember all I got done over this weekend.

I spent some time trying to build this HF PA.  

I was not getting at all the output that I was expecting.  And also found that the drive waveform looked weird.  The guy who designed it was super nice and responded to my emails really quickly, but I couldn't quite figure out what was up.  But I did observe a few things

• The drive waveform was "weird" when using my signal generator.  But it looked way better when I used the SI5351.  I'd like to investigate this more later.
• The drive still wasn't as square as I'd like.  My theory is it is due to my construction technique introducing stray capacitance/inductance.  Want to look into rebuilding (maybe a pcb)
• Even when I increased the power to the drain, the output power didn't increase.  Don't know what was up there.  One theory is that this isn't a good IRF510 - it was one from Radio Shack years ago, and I couldn't find my new ones anywhere.  Want to order new ones and try it.
• I built a little adapter board where I could use 3 BS170s in place of the IRF510.  It does look like increasing the the drain voltage increased the power out here, which points to the IRF510 being bad.  But also want to investigate to make sure I didn't do anything dumb with the output transformer.

I also wanted to try the NCP81074A again.  This time I decided I would try with an actual PCB.

First time I've etched a PCB this century.  I have tried a few different toner transfer techniques (acetone/alcohol no-heat, various papers for heat transfer).  I ended up trying this paper, and it worked pretty well.  I don't think Brother's are supposed to be best for this, and my toner is running low, but it turned out ok anyway.  I used an iron on high for two minutes.  The edges weren't great and there were a few lines broken, so I filled in with Sharpie.  I used ferric chloride with a sponge (so I barely used any, and it was done quickly).  I checked with meter, and there was only one broken trace (that I fixed with solder).

And here it is populated.

Got nice results - looks like 85% efficiency.  I tried some different voltages, could get 10+ watts out.  I ended up putting a heatsink on, but turned it back down to 5W.  I can't quite tell how how the driver gets, and at 14MHz it seemed pretty toasty.  Will play more with different drive voltages, etc.

Overall, a nice result for the weekend!

10W amplifier - success!

 Woke up early this morning so was able to swap out the BS170s before work.  The hardest part was figuring out where they were :)

Well, also desoldering.  The wick works good in general, but isn't good at clearing holes.  But I managed to use the soldapullt on the other side of the board from the iron and it worked a treat.  Would have been good to have a third hand or a helper, but made it work.

On the plus side, I have output!  And gain!  But not as much gain as I should have.  

With a 13.8V power supply and 9v p-p from the signal generator, I was getting about 4.6W out (15.2 vrms).  I am unclear on if that is too much power input.  There are discussions on the forums about input impedance not being 50 ohms, so I'm not sure if that is 200 mW (at 50 ohms) or 25mV (400 ohms).  Should try to get a proper power input measurement...

But based on things in the forums, I increased the power, and was able to get 10W at 17v.  Can experiment more with the input power.  But, regardless, I fixed it!  :)

Another go at a QRP Labs 10w Linear amp

 I had a great time at the Minnesota Amateur Radio Technical Society meeting this past Saturday.  Great conversation, and free stuff to boot!  A beloved member had recently become a silent key (at almost 100), and his family had sent a bunch of his workbench stuff in to go to a good home.

Most interestingly, there was a 10W amp!

This is a QRP Labs 10W linear amplifier.  Which I know about because I've built one, but never got it quite working.  (I can't find my notes about the issues I was having).  So this is exciting.

First of all, Mert did a way better job of boxing things up than I ever have.  

On the inside, I found that there was a low pass filter on the output.  Measuring with the vna, it looks like it is for 7MHz.  Also, the transmitter powers the bias when the /tx pin is low.  He has a phono jack for this, and made a little shorted plug to keep it on.  Nice!

I hooked it up to a dummy load and powered it on with no input, and it looks like the bias is already set correctly.

Unfortunately, when I applied signal, nothing came out the other side.

Poking around, the bias at the irf510s looks right, but there was no signal.

The input BS170s have power and signal, but no output.  So, those look blown.  It got late tonight, so I'll have to wait to swap them out for another day.

Towards a working class E amplifier

 I've had a few attempts in the past of building a class E amp, but never quite got the right results.

I dug out an old board where I tried to replicate the NS-40 and decided to try a different drive circuit.  I took out the oscillator and drive circuitry and replaced it with a NCP81074A gate driver.

It somewhat worked - I was able to get some power out.  But it isn't as efficient as it should be because the gate drive is ugly.  Need to look into it more.

The circuit

The ugly gate drive