Pixie Ham Radio Information
IntroThis is one amazing little ham radio project. With it's exceptionally low price tag, it is well worth it, and a good beginner project, especially for those who have been told that ham radio is an expensive hobby. It took me, a good experienced assembler with decent tools, a little under 1 Hour to build, worst case might be 2 hours for anyone. All through-hole parts so don't need a microscope.
If you buy it from Vadien.com, then you will get 2 crystals, a set of earbuds that work well with the radio, and a power plug and 9V battery clip, so all you need then is a CW key with an 1/8" plug on it, an antenna with a BNC connector on it, and a good 9V battery. The two crystals are 7.023, which is an Extra-only CW band, and 7.030 which is a Tech and everyone else CW band, and one of the QRP calling frequencies. Another interesting thing is that the tuning range (pulling range) of the 2 crystals is different. The 7.023 will tune about 850 Hz, the 7.030 one will tune 1800 Hz, more than twice as far. But remember, the transmitter is always at the low end of the RX tuning range.
The quality of this kit is very good. The PC board is doublesided copper with plated though holes, and has good ground plane on both sides of the board. Green solder mask looks nice and tough. The top side has a white silk screen to identify the name and position of all the parts. Parts go on this side, in case you've never built a kit before. The back side has no white printing, that's the side you do all the soldering.
One bad aspect of the board is the size of the pads for the parts being very small, and the ground plane clearance around all the pads. If you insert parts the usual way, bending the leads over on the back to hold the part while you solder, you could bend a lead right into the ground plane and cause a short if the solder mask cracks. If you are careful and only bend part leads over to a 45 degree angle, this shouldn't be a problem. Another problem is any pad that connects to ground has no thermal relief, the copper of those pads and the ground plane are one piece of copper. This means you need quite a bit of heat to solder wires in these holes. If you have a temp controlled iron, crank it up to 800 for these holes. If you have a wimpy 25 Watt Home Depot thing, you either need to sit on these holes for a long time, or get your Weller 8200 soldering gun to do these holes.
Performance ExpectationsThis is a very simple direct conversion transceiver. Looking at the simple schematic, you will be amazed that it even works. I couldn't believe I actually heard many CW signals on it when I first connected it to an antenna. But there are still some drawbacks with this simple radio, typical of a simple direct conversion receiver without much filtering. First, the audio output is very sensitive to anything. Any noise on the power supply comes right through. If you use a cheap wall wart to power it, you will probably get a lot of hum. If you use a lab power supply, it sounds pretty good, but when you turn the voltage adjust knob, you will tell exactly how dirty the pot is in your power supply, noise will come blasting out. It will also receive audio signals from the antenna. If there is AC noise on the antenna, that will come through. This is why dipole antennas work better than a long wire and ground, there may be lots of low frequency noise between a piece of metal in the sky and earth ground, but a dipole is balanced and this kind of noise is likely to cancel. You may also pick up strong local AM broadcast stations, but it works.
Connecting to a full size 40m dipole up about 30 feet and tuned to 7050 KHz, I sent some IDs out from the Boston area and looked at the reverse beacon network. First few late evening tries I didn't see anything, but finally a try at 4 PM had it show up at 2 locations, one by Rochester, NY, and one just north of Pittsburgh, PA, both with 14 dB SNR. Tried my regular rig with 80W CW, showed up 5 places. Gives you an idea of how QRP works, those first few tries might have been when there was poor propagation or too many signals nearby. Have to be watchful of when you get on, or just check the reverse beacon network, a pretty amazing tool.
Build StyleIf you have built lots of things before, just skip all this instructions stuff and build it - Stick all the parts in according to the Bill Of Materials (BOM) and solder it all and you're done. Otherwise, there are two levels of advice here. The next section is for noobies, the one after that is my opinion on how to build it the easiest (like put L1 in backwards). But as for me (a 30 year experienced radio engineer), I just threw it all together and it worked fine. Then I blew it up holding the key down too long (8 seconds). See operation below.
Beginner Kit Builders:You need a soldering iron and some solder. Ideally you would shell out $125 for the Weller WESD51 "digital" temperature controlled iron with a nice clean new tip, and some relatively small diameter (.032") 63/37 alloy rosin core solder. If you have some old beat up thing with one tiny spot that still holds solder (like I did for the first 40 years of my life), you can get away with that, but it will be harder. If you only have a big soldering gun, if you're real careful you can use that too, but it will be hard not to cook something. Do not even consider using a propane torch. OK, now someone out there is going to prove you can do it, but it's a risk. The other tool you need is a good sharp pair of flush angle cutters. There are very nice ones of these at Michael's Arts & Crafts, or at places like JoAnn Fabrics, for like $7. Big old electrician cutters will do a poor job. OK, torch man, prove me wrong on that, too. One time I couldn't find my cutters and I just wiggled the wires back and forth and broke them off, that works, if you're into self punishment.
When you install parts, bend the leads a little to get them into the holes on the board, then push the part down, wiggling the part a little, until there is maybe 1/8-3/16" of lead between the part and the board. Then bend the leads on the back side to a 45 degree angle to hold the part in position for when you turn the board upside down to solder it. Parts that stand up (all resistors, inductors, and diodes) have one lead that will have zero length between board and part, and that's OK as long as it solders OK. I would put in a small group of parts, sodler them, and cut their leads off, before putting in more. The groups below worked well for me.
When soldering, put the iron/gun/torch so that it is touching both the wire lead and the pad, wet it with a little solder until you see it start to flow around, then add enough solder to fill the hole. Holes that are not separated from the ground areas will take a while to get hot enough for solder to flow down the hole. Make sure that you don't leave solder just on the wire, and not flowed onto the pad. It has to flow smoothly onto both.
If you put a part in wrong and have to remove it, this is not easy. It's very easy to wreck the board trying to unsolder something. I don't recommend trying to remove the solder first from a 2-leaded part. Rather, heat it up and pull on the part. Then suck the solder out of the hole (or heat it up and bang it on the table to get all the solder out). Multilead parts are harder, almost sure to wreck something unless you're real good at this. I don't have any simple advice for you. If you're the torch guy, you'll find it's real easy to remove parts by just aiming the torch at the bottom of the board, but 1 second too long and the whole thing burns up.
Building AdviceI would put the parts in the following order, with reasons and hints given:
1) Install 3 of the connectors, leaving off the BNC connector. These 3 connectors will protect the other parts from bending over when you turn the board over to solder. Having the BNC in gets in the way of putting in some of the parts, so leave that off for now. Align the connectors with their outline before soldering the first pin, if you want it to look very nice. The Phones connector didn't want to do this, the holes aren't laid out right, I had to push on it while soldering to get it nice and square.
2) Install the 11 disc caps. You don't need to push these down close to the board, this is only 7 MHz, but make the top of the caps about as high as the 3 connectors. I always orient them so you can read the value after it's installed, makes it easier to check later. 3) Install all resistors. All resistors are installed vertically, so you take one end and make a nice Ubend to get the wire to go down the side of the part.
4) Install the 3 diodes. All 3 are installed vertically with the cathode end up, so bend the wire nearest the band on the part in a Ubend. All diodes have their band up.
5) Install the 3 inductors. I would install L1 backwards from the way the silkscreen shows, with the body of the inductor close to Q2. 2 reasons for this the hot RF signal is not on a wire coming up off the board, the RFquiet 12V is. Also, I was probing around on the board with a scope and the ground of the scope hit that wire loop sticking up, which blew up the inductor, instead of shorting the power supply. Ya, the insuctor acted like a fuse, but then I'll bet you don't have a 22uH inductor laying around the shack to replace this, whereas I could replace a blown wire to the battery, or the power supply would go into current limit. Radio works with L1 either way.
6) Install the 2 transistors. The silkscreen doesn't quite match the outline of the parts, but make sure the flat side of the transistor is on the side of the silkscreen that is the longer flat line.
7) Install the 4 polarized electrolytic caps. The polarity matters, and the board is nicely marked with white lines on the negative side. Just line up the white area on the cap with the white side on the board. Or put the longer lead (+) in the nonwhite side. These won't push down all the way to the board, probably leave 1/16" under these.
8) Install the potentiometer (W1). I wonder where W came from?
9) Install the IC socket (without the IC in it). Make sure to put it in with the little half-circle notch lined up with same on the silkscreen. If you put it in backwards and realize it before you blow everything up, you just have to remember to put the IC in backwards too.
10) Install the BNC connector. This takes a ton of heat on the two big ground pins
11) Install the crystal. Do so with a space between the bottom of the xtal and the board, or else use some insulator between the two (black tape with the leads poked through). The crystal can may short to the pads if you don't. Polarity doesn't matter. Use either the 7.023 if you want to be only in the Extra Class band, or the 7.030 if you want to be in the everybody band. Crystal freq is printed on the top. I suppose you could put two pin sockets in the board and then plug in the crystal, that should work.
12) Plug in the 8pin IC. First, use a flat surface to bend the 4 leads on each side of the IC to be 90 degrees from the part. Then it will be easy to plug in. Make sure it's in the right way - there's a little half-circle indent on one end of the IC, make sure it's lined up with the half circle printed on the PC board.
You're done! You might notice that you have a few parts left over, mainly disc caps. This varies from kit to kit. One kit I built came with a 51 ohm 1 Watt resistor, to use for testing, that was nice.
WARNINGS:1) Note that there is a reverse protection diode on the DC input of this radio, whereby it dead shorts the power input if it's backwards. If you hook this up backwards to a big battery (gel cell, or LiPo), you will see smoke somewhere. Maybe you blow up the protection diode, then the whole board gets cooked.
2) Don't key it without a load on it. Plugging in a key with the power on may key it, though the stereo ring terminal of the key is not connected, only the tip, so you might get lucky.
3) Even with a load, don't hold the key down. 3 seconds MAX! 8 seconds will burn up the output transistor on 13.7 volts. It's a TO92 plastic part trying to put out over 1/2W, give it a break. Data sheet says it's a 1 Watt part, but that's if you keep it at 25C somehow. Hard to heatsink a TO92.
Operating Instructions:No tune up for this baby, it just works, or not. If not, it should be easy to find the problem, see below. First test I would do is power it up with a 9V battery (so if there are shorts you don't blow things up), with nothing plugged into it (no key, no antenna, no earphones). Listen on a "real" radio for a signal at the crystal frequency +500Hz, it should be there. This thing puts out -10 dBm (100 microwatts, a huge signal) to the antenna at that frequency IN RECEIVE! So you'll hear it, as will your neighbor ham down the street. Tune the frequency control pot of this radio up and down, and you should hear the frequency move (change pitch) in a USB/CW receiver. This pretty much proves out about 85% of the radio. The power supply current should be around 15 mA at 9V, 20 mA at 13.7V. Next, put a load on the antenna jack with a power meter of some kind (SWR meter?), plug in earphones or a speaker (walkman style stereo earbuds of the 32 ohm variety work fine), plug in a key (connected to tip and frame of an 1/8 inch plug, float the ring if it's a stereo plug). If you're using an antenna as a load, you should hear signals, if there's someone on. Key the transmitter for no more than 1 second. Note the power output, should be 1/4 to 1/2 Watt out (16 volts peak-to-peak on the antenna connector with a 50 ohm load, if you have a scope) Note that the "tuning" pot only affects receive, so it's really RIT, not radio tuning. And it only lets you tune the RX ABOVE the transmit frequency. Tuning range is about 850 Hz for the 7.023 crystal, 1800 Hz for the 7.030 crystal.
Some SpecsPower: 9-12V
RX: 15mA@9V, 20mA@13.7V
TX: 120mA@9V, 190mA@13.7V
TX Power out: 650 mW @ 13.7V, 350 mW @ 9V into 50 ohms, 2532% efficiency
Tuning: RX tuning only, upward 0850Hz from TX frequency (depends on xtal)
Q1: S9018H, 30V 50mA 400mW, "H" is high beta (97146), Ft 1 GHz, "VHF tuner", $0.065@100 Mouser
Q2: S8050D, 40V 1.5A 1 Watt, "D" is high beta (160300), Ft 100MHz, "Audio amp", $0.067@100 Mouser
How it WorksLooking at the schematic, you might ask "How does this thing receive??" The trick is that Q2, which is the RF power amp in transmit, is biased way low during receive so that it acts like a mixer diode. The received signal is coming in the collector, the local oscillator to the base, and the audio is coming from the emitter. During receive, there are 2 volts peaktopeak of 7 MHz on the base of Q2, but at a DC point the same as collector and emitter. This makes it a good switch, even though running in reverse.
The rest of the circuit is pretty straightforward, an Edwin H. Colpitts xtal oscillator with emitter output coupled hard into the base of the PA, 3 pole pinetwork low pass filter on the output. keying is grounding the emitter of the PA transistor.
The LM386 audio amp (all the receiver gain) is set to max gain (x200, 46 dB). It is powered through a 1K resistor from the battery, which drops the voltage down to around half (pulls around 4 mA). This is a problem when running on 9V, it is very close to quitting due to low voltage. Any loud noise does shut it down. Diode D3 kills the power to the audio amp on keydown, dumping 12V across the 1K (R3), which is why you don't want to make that much less than 1K. The LM386 doesn't seem to be unstable with lighter loading than 8 ohms. The output stability snubber probably takes care of this (R7, C11).
A 1N4001 diode is used as a varactor to push the crystal. DC bias from the frequency set pot drives the varactor, and also goes to zero on transmit.
Pretty nifty little circuit.
TroubleshootingNot a lot that can go wrong in such a simple device. If it doesn't work, it's probably worth it to re-check all the component values and polarities. Check all the soldering for good connections and unwanted shorts. If the oscillator doesn't run when you put 9V on the thing, there are only 6 parts that could be wrong. You can check the DC voltage on the emitter of Q1, it should be something like 6 volts. If it's zero, the transistor is blown or R1 is open, or there's no power on anything. If there's DC on that emitter, and you don't hear a signal around the crystal frequency, then the crystal is broken, or D2 is open.
If the oscillator can be heard on a receiver, yet you don't hear any signals or noise in the earphones on receive, the LM386 is probably plugged in backwards. Or the key is stuck on. Measure pin 6 of the IC (power, the pin nearst C5), it should be 4-5 volts. If it's zero, R3 is the wrong value, or CP4 is in backwards.