Monday, February 1, 2016

MSP430G2553 beacon oscillator

The next step to building the ham beacon is to create an oscillator.  I chose a simple Pierce oscillator similar to the ones you find in books.  This particular one is in Art of Electronics, Experimental Methods in RF Design, and a dozen or more other publications.  The idea is to use two caps to oscillate a transistor using a crystal to set the frequency, with a variable cap between crystal and ground to help "bend" the frequency a bit.  There are several features in my circuit that you might find helpful.



First, notice that I salvaged a 3-wire 0.1 pitch pin socket from some old equipment and I use that for a crystal holder.  You can go buy "crystal holders", or use a machine pin socket, or pull something out of the junk pile.  I like having the little wires to give my fat fingers enough wiggle room to change crystals easily.  At RF frequencies, you have to watch out for lots of long wires, but this little beacon at 7 MHz is not going to suffer from a couple inches of wire.

Another trick I like to use is a solder-style tube socket for the slightly larger, older style crystals.  Below is a 9-pin Chinese cheapo socket that I decided was too flimsy to put into a tube circuit.  Made a fine crystal holder, though.  I soldered half the pins together on one side, and the other half for the other.  A Sharpie makes is simple to see how to plug in the crystal.  This old perf board had strange imbalances in the oscillator, probably due to the lack of a ground plane.  Plastic. Blech.  It's in the junk box now.



The beacon circuit is being built on a small scrap of brass plate I bought from K&S Metals (via the local hardware store).  Some people swear by copper-clad, but I am offended by the prices.  It's just as easy to build small things like this on a simple piece of brass or copper sheet metal.  I marked spots for holes later so I can use 6-32 hardware to mount the thing in a project box.

The air variable cap is an antique I got from some eBay deal where an old guy filled a box with junk he had on the bench.  Dan's Small Parts has them, too, and probably many others.  This one claims to be 3.1 - 31.5 µF, but I measured it to be 3.6 - 35.2.  In any case, fully meshed I notice that the voltage developed by the oscillator is about 1.8V peak-to-peak with 11.9VDC as the supply voltage.  You don't need one, but I like building circuits with one of these in a strategic spot.  For the finished beacon I'll take it out, since I don't really need a variable beacon.  Instead, I'll adjust it to value that suits me, measure the result, and replace it with a fixed cap that matches.



As I turn the cap, I can raise the frequency of oscillation from 7.0394 MHz (fully meshed) all the way to 7.044 MHz (no mesh) - a full 5 kHz spread.  That's quite nice when you're otherwise rock-bound and need to nudge your frequency a bit.  The crystal is labeled 7.040.

Interestingly, as I "bend" the frequency up, the peak voltages go down until I hit 7.044 MHz, where the voltage is about half what we see here.  Taking the peak voltage here times .7071, squared, divided by 50 ohms, I get about 8 mW when the cap is fully meshed.  The transistor is a 2n3904.  That's a pretty strong signal for a 3904.  I usually get a result somewhere between 4 mW and 7 mW.  [For those of you who like to think in dBm, 8 mW power is around 9 dBm (decibel milliwatts) - for the rest of us, that's 10 raised to the 0.9 power].

The red silver mica cap is the 100 pF isolator that will connect the signal to the next stage, a buffer.  I went ahead and put it on there so I can connect the oscilloscope without touching any DC.  About half the parts you see are from old equipment I've dismantled for salvage.  That little silver mica cap came out of a very old piece of lab equipment that dated from the 50s.  I've used it in a half dozen circuits.  First thing I ever made with it was a grid-leak for a regenerative receiver.

Another detail you may notice are the little square soldering pads.  Those were included in a kit I bought, so I thought I'd give them a try.  They are called "MePads".  Frankly, I prefer ugly construction, just soldering components together as I go.  The little pads are supposed to be glued down, but I already notice them pulling up as I work on the circuit.  I intensely dislike circuits that are not rugged, so I probably won't use these again.  I'm also suspicious of gimmicky "helper" bits of flotsam that snorkel money away from the active device budget.  No offense to those who like these.  If you obsess over having "pretty" circuits, maybe they are for you.  I care whether it works and couldn't care less how tidy it looks.  I found the MePads to be fiddly and not very helpful.  They are certainly slower than ugly construction.  They aren't even self-adhesive.  Not for me.  Instead, I'll use high-value resistors as standoffs (1M or higher).  We'll see that in the next post, where I will add a buffer to the oscillator, and then the keying circuit.


No comments:

Post a Comment