In various newsgroups Steve Quest has described an AM transmitter for mediumwave and shortwave that has produced great results and is quite inexpensive. Here is the verbal description he offered:
However, for now, just picture a standard Pierce oscillator, crystal controlled on your frequency of choice, with the output connected to a standard class C amplifier stage. The only modification to the standard class C amp stage is the insertion (series, so cut the trace and insert) of a modulation transformer between the power supply line and the isolation inductor on the collector. What works great is the $1.95 at my local audio shop, SPECO line matching transformer. Connect the transformer backwards! In other words, the input side goes to the class C amp, and the output side, normally the side connected to the speaker becomes input. Transformers work both ways you know.
Use the 500 ohm and common taps on the class C amp side, and the 8 ohm and common tap on the input side.
Now hook up a power audio amplifier to the (now) input of the transformer, and a mic or other audio source to the line input of the power audio amp. Adjusting the volume on the audio power amp will increase and decrease your MODULATION! Use a scope to set your modulation, just lay the probe near the antenna to see the AM waveform. Adjust to about 80% modulation, don't overmodulate, it sounds like crap AND is quite inefficient.
*IMPORTANT* Use a chebichev filter between the output of the class C and your antenna! If you don't your harmonics will be over a watt for sure!
This configuration, when powered by 12 volts will generally give you an input power of about 5 watts. That's based upon the class C amp giving you a times 10 gain, and the input power from the Pierce oscillator of half a watt. Sloppy design of the Pierce, and poor tuning on the class C will cause the power to go DOWN. However, if you're good, you can get 10 watts or more by tweaking the Pierce up to a 1 watt output into the x10 class C stage.
About a 4 to 5 watt input signal to a 150 foot random-wire antenna on the 41 meter band during the average sunny winter day resulted in a clear signal copied at greater than 200 miles and a weak signal copied at 1000 miles. This range example is reproduceable by several successful tests.
And here is the schematic:
R1 = 47k 1/4w
C1 = .001uF
C2 = 470pF
C3 = 470pF
C4 = 100pF
C5 = .01uF
C6 = 470pF
C7 = 820pF
Y1 = 7.3728 Mhz series type (off the shelf, CPU application)
Q1 = ECG or NTE282 or similar
Q2 = ECG or NTE235 or similar
T1 = Speco type T7010 70 volt line matching transformer
L1 = 28turns around 1/4" ferrite overwrapped with 9turn secondary
L2 = 28turns around 1/4" ferrite
L3 = 1.5uH choke
To make L1, tightly wrap 28 turns, then wrap 9 turns over that tomake the secondary. Space the windings out to evenly cover thefull length of the 28 turns. Use light gauge enamel wire for all.
To connect the modulation transformer, wire backwards using the normal secondary as the primary. Connect the red wire of theinput side to Vcc and the black wire to the class C amp choke. Connect the output side white wire to the amplifier speaker jack positive, and the black wire to the speaker jack negative. Use the amplifier volume control to adjust modulation level.
A few more details of this circuit:
... since the rig operates in the 41 meter band, I just used generic CB radio transistors. The Pierce oscillator uses a generic CB radio driver transistor, and the final is a CB radio final. Thesetransistors are cheap, readily available, and probably will never beoutlawed (like I suspect VHF RF components will eventually be). Since CB radio is 27 Mhz, running them at 41 meters (7 Mhz) is very easy on them.
Any type NPN, TO-220 package CB radio driver and final will workfor this application. There are probably _better_ transistors out there,but these are cheap, available at the local radio shop, and already instock for those of us who fix a few CB radios from time to time.
And here's some additional detail:
I get my forms from a distributor, and my inductors as well. However, I know there are ferrites that are "close enough" in the Radio Shack assortments. You should be able to find one fairly close.
The question this time is could one substitute say 5/16" core for L1 & L2 and adjust the windings to come out with the same results. If the answer is yes, then could you recommend a good starting point to make the alterations?
Sure, use a 5/16" core, and the same number of turns. Remember, I didn't specify the wire gauge, and I should have said "about 1/4" ferrite rod" instead of being exact. Use a heavy enough enameled wire that you don't vaporize it with current flow, I used the middle gauge out of the Radio Shack 3 pack for my windings. This fact should probably be added at the website. :) You may have to adjust secondary turns between the oscillator/driver and the final to tune for maximum power. Don't glue the windings and leave plenty of extra (not looped but curved up) so you can do that. Once you find the right amount, glue them down. I use wax usually, but have also used shellac, varnish, silicone, you name it.
On the inductor in the final, an extra few turns will not hurt, but don't take away! Same goes for the primary on the osc/driver. However if you add turns, you'll have to adjust the secondary turns and capacitors. One thing to remember with HF verses VHF, the windings are NOT as critical. 3 turns in an air wound VHF oscillator tank coil will get you, let's say 88 Mhz, and one more turn and you're down to 64 Mhz, and another and you are in the 40's. It isn't like that with HF, it's not so touchy.