This transmitter was adopted from a MOPA broadcast transmitter described in a 1926 NRI course textbook. It wasn't designed to be a practical example, but a schematic used to explain the thoery of a MOPA transmitter. In 1926, all that was available were triode tubes but I wanted to try the Pentode #47 that was released in 1928-29. I finished the transmitter on 12-2-2001.
Here are pictures of the completed transmitter:
The front panel view of the MOPA transmitter.
Back view. Thats a #45 on the left, #47 on the right.
The transmitter consists of a Colpitts oscillator on the left driving a #47 on the right. The oscillator is tuned by 2 fixed mica's in series and grounded in the middle between plate and grid. The Cardwell condenser is in parallel with the grid condenser and will tune the entire 80 meter band with reasonable bandspread. The coil has 13 turns of 1/4 inch copper tubing mounted on standoffs. Grid leak consists of a 50k resistor (1-100k's in parallel) and a 200pf condenser.. The stages are condenser coupled together with a 50pf ceramic condenser. The transmitter works best when the grid end of the oscillator tank is coupled to the amplifier. When its coupled to the plate, there was a huge amount of grid hum. It totally disappeared when I took output from the grid. Might have been a feedback problem, but no matter, it works taking power from the grid circuit, so thats how I left it.
The amplifier is keyed along with the oscillator. The final is self biased by the drive from the oscillator. The grid leak consists of the 50pf cap and an 18k resistor in series with a 2.5 mH choke from grid to ground. Under keydown conditions, I measure about -40 volts of grid bias, but I did not take a grid current measurement. The plate circuit consists of an RFC (unknown value) that I had in the junkbox, then that is coupled to the tank via a 6800pf mica block. The tank is similar to the oscillator tank except it has 1 fewer turns. 13 turns on the oscillator, 12 turns for the output.. The link is a 5 turn link in series with a 3 section variable condenser for loading.
This thing runs off the same power pack as does the Hartley found on the Hartley page. The filaments are series connected with about 0.6 ohms of resistance in series with the 6.3 volt filament winding. Plate voltage is about 250 volts which is the maximum rating for the '47. The screen voltage is connected to the 250 volt supply thru a 22k resistor and is grounded with a 0.01 uF condenser.
This transmitter was used during the first weekend of the 1929 QSO Party on the 1st and 2nd. I did not operate on the 1st but worked the 'test for a couple hours Sunday afternoon and worked about a dozen or so stations with this transmitter.
Overall transmitter current runs at about 50ma, the oscillator drawing
about 10ma, although it really wasnt measured. Output power ran about
4 watts so the 10W input report I gave may have been a little high.
Upcoming improvements are to include provisions for spotting the oscillator,
also some kind of metering is necessary. A panel meter mounted on the
front panel will be the next step.
This transmitter is VERY stable and is workable at any time. The chirp is very slight and the drift is also very slight. It is a fun little rig to operate. It is a very easy rig to put on the air. No hand capacitance because the frame of the Cardwell is grounded.
More construction details will follow, as well as pictures taken during its development.
This page is still under construction.
Original schematic taken from the '26 NRI textbook.
This is NOT the actual diagram, my version has a pentode output tube.
The oscillator circuit is relatively unchanged.
Oscillator built, starting the amplifier.
halfway thru the amplifier.
closeup of the oscillator section.
almost finished.
Current schematic of the MOPA transmitter.