Sunday, June 18, 2017

Transmitter experiments

I've never managed to get more than 1W out of my Mighty Mike amplifier.  While I've tinkered a bit on things on it, thought I would try some other circuits.

A few weeks ago I tried was using a 74AC240 octal inverting buffer as the driver for an IRF510.  There are a lot of these circuits out there (here, here, here, here are a few, but there are lots more).  What I ended up building was pretty standard - one part of the inverter with the crystal to make the oscillator, then feeding that into four other inverters in parallel to increase the driving power, and putting that into an IRF510 (with a DC bias just below turn on).  It worked... ok.  Again was having trouble getting more than a watt, and the transistor got pretty hot.  

Some of it may have been the construction technique (I don't usually build RF on breadboard), but I didn't have a lot of time that night, and wanted to see something on my scope :)

Today I thought I would try my hand at building the circuit for the ns-40.  It is a Class E transmitter using a 2n7000 to drive the IRF510.  The actual kit is retired.  Its neat feature was that its inductors were part of the PCB (thus the 'None Simpler 40').  I've seen somewhere else where someone made this circuit using toroids, and I thought it would be a good morning project.

Here are all the parts

Not much there.  I used T50-2 toroids, since that is what I have most of.  Here is what I had two hours later (with my dummy load hanging off to the right):

And here is some output:

Pretty excited it worked the first try.  However, it isn't being a proper class E transmitter.  I was measuring 28.8Vpp while consuming 370mA from the 12v battery.  I'm pretty excited about the 2w result (double what I was getting from an IRF510 before), but that looks to be <50% efficiency, and the IRF510 was definitely getting warm after a while.  I'm not quite sure how to tune a Class E without being able to view the current waveform on the drain.  Also, the gate voltage on the IRF510 (yellow below) isn't pretty, not sure what is up there (though I did just notice that in that screen shot it was putting out 3w!)

Now what I'd like to do is not drive this with a crystal but rather an Si5351 or the Raspi.  I wanted to build it as is first (a) to get it working and (b) I was unsure what the drive requirements were.  I tried simulating the circuit in LTSpice first, but getting crystals to behave in that program is difficult.  Also, I tried driving the 74ac240 circuit with the Si5351 the other week and managed to smoke the chip... 

One other experiment I did over the past few weeks was buying a cheap "Forty-9er" cw transmitter on EBay.  It was pre-assembled and was suppose to put out 3W.

My plan was to convert it to be digital (per the QST article the other month).  Was disappointed to find it only put out 1W, and even then if you held the key down too long the power output went down.  Probably some dodgy parts or something in it.  Oh well.  First day I got it there was some CW contest that weekend - I could hear people, but didn't want to join in.  Tried calling CQ a few days later, but no takers (couldn't even hear myself on a local WebSDR).  Will probably shelve that for a bit.  

Monday, April 24, 2017

Poking at the amp

I've been poking at the amp tonight, trying to compare it to simulations in LTSpice.  I've noticed at least two places where I actually have different values than the original circuit (some DC blocking caps and the choke on the final), but putting those back into the simulation didn't seem to affect much.  Definitely getting different readings, but haven't found the cause yet.  Frustrating.

On the plus side, still making it really far on wspr :)

Saturday, April 22, 2017

Trimming the dipole

I took 20 inches off each end of my dipole (12 inches then 8 inches), and based on my measurements I think I got it to resonance around 6.9Mhz.  Ballparking, I think the SWR there is 1.6, going up to about 1.8 at 7.2mhz.  I was going to tinker with it more, but don't really have time today.

Also, got a new SLA battery (12v, 12 Ah), which means I'm not running through AA like mad.  However, the IRF510 in my amp is running really hot.  Maybe it is because I'm pumping in square waves rather than the sinusoidal I think this was originally designed for, maybe I just have other biases off.  Never did get the 5 watts I expected, so very likely I wired something wrong in this.

In any case, it has started thinking about maybe looking into class E amplifiers here (since I'm starting with a square wave anyway). 

Monday, April 17, 2017

Random reflections

As I was turning things off last night, I happened to run my scope across the outputs of the LPF.  I noticed it was doing like 44 Vpp, which was odd because it was doing more like 22 Vpp into the dummy load.  I realized that  this was indicative of a higher SWR than I was expecting (possibly 4:1, I didn't take exact measurements).  I ran a little experiment tonight, and I *think* my dipole has its peak resonance somewhere in the 6.5-6.6mhz range (where it is more like 1.6:1).

So this weekend I may try to shorten it a bit to get the resonant portion up to where I actually want to use it.

This also makes me want to finish that antenna analyzer I started messing around with last year...

Also, my batteries died part way through my experiment tonight.  I was running off of a back of AA batteries (same as when I was experimenting before), which don't exactly have the power to do this.  I tried with the random wall wart again, but it was noise city.  Am going to look into getting SLA and charger.

Sunday, April 16, 2017

Playing with WSPR

Having installed WsprryPi, I've been testing things out.

First issue is I didn't have any wires that had the pin connector end.  I ended up finding some some header connectors that I could cut some pins off of so that I could have single tall pin I could alligator clip onto.

I put that through a 1nf cap into the LPF I built previously, into my dummy load, and looked at the output on the scope.  Didn't like what I saw - definitely some harmonics in there.  Was getting about 2.3v p2p

Next I decided to see if my amp would do better.  I put the output of the raspi into the Mighty Mike amp, into the LPF, into the dummy load.  Much better!  The output looks pretty clean.  Doing 23.3 Vpp, which is about 1.3w/31dBm.

Ran this command:
sudo wspr --repeat --terminate 7 KC9DLM CM87 31 40m

Waited a few minutes, goofing around and what not.

And then I checked wsprnet.


This is so exciting I can't even explain.  Especially given I just threw this together in like an hour.


The IRF510 is probably a bit hotter than it should be (I don't think I even adjusted the bias level since I put that board in a box six months ago).  

I hopped over to a SF area web sdr ( and found my signal :)  

The farthest looks to be VE6JY at maybe 1970km away, followed by NO1D at about 1000km.  The closest seems to actually be someone in my town :)

So, that looks to maybe be 941 Miles/watt.  Not the most impressive, but given my antenna is not that high (nor have I ever verified the SWR on it) and it is is around 1pm on 40m, I'm pretty happy with that.

edit: tried again later (8:45pm) - the 40m band is maybe better at night :)

Installing WSPR on Wheezy

Got the urge to try out WSPR.  Installing WsprryPi on a Raspberry Pi I had laying around.

Didn't work because my version of gcc was too old (didn't recognize one of the flags).  So went to upgrade gcc

(Truth be told I probably should have just upgraded to a newer version of Rasbian full stop, but this was something I could easily do in the background while I was doing some other things.)

Much happier after that.

Sunday, July 31, 2016

Power supply woes?

I built another little preamp today (this time just a CE followed by a CC rather than the shunt thing I used before), just to see if additional drive was the issue.  Did not seem to help.

However, it occurred to me that it might be the power source I'm using.  I've been using alkaline AA batteries.  But calculating things out, for 12v to produce 5 watts, the batteries would need to be delivering around half an amp (not counting what is dissipated in the transistors).  I always assumed that the batteries could deliver more amperage for a shorter time (the amp-hour rating).  However, I forgot that that batteries have voltage curves.  I measured the voltage output when the amp was running, and it was closer to 9V.  So, I think I need to go find a better battery to use :)

(I actually tried a wall wart, but I think I was either getting feedback into it or it was just noisy, was getting a weird output when I tried to bring the power above what I had with the batteries.)