Page 1 of 1

It was a cold and overcast day in February

PostPosted: Fri Feb 14, 2003 12:18 pm
by k3xf
It was 7AM on a cold and overcast day in February. While sipping my first cup of bean juice I marveled at the simplicity and function of a Pentium III heat sink assembly. The smoothness and precision of the heat collector/fan assembly were attributes I couldn?t relegate to the junque box. What do you do with an obsolete piece of computer? Why you make a class e single FET transmitter, that?s what . . .
I raided the cache of FQA11N90 FETs that were going to be a six FET transmitter when they grow up. ?only needed one . . . That one was bolted to the CPU heatsink after drilling a hole in it to accommodate a long 4-40 screw and nut assy. This aggregation was bolted into a 4X6 aluminum chassis with a convenient SO239 connector. It was only 8:30AM and the transmitter had begun to take form. The input network was next. Out came MFJs # 269 analyzer and a big ferrite ?bead? from the junque box. The bead is about one inch in diameter and one inch tall with a half-inch hole in the middle. I bet one of the Radio Shack RF suppressors that clip around a phone wire would work too. Six turns of # 18 Belden enameled wire for the primary and 1.5 turns of #14 and the input transformer was complete. The transformer output was soldered to the gate. One side of the primary and secondary were grounded. The FET source lead was held firmly under another 4-40 screw to make sure it was grounded well. I carefully connected the remaining transformer primary to the center of the SO-239, fired up the 269, and the match was equally poor from 1.7MHz to 8MHz. It wasn?t going to work this way. Too much C. Perhaps some L in series to balance that C. 12 turns of # 14 enameled on a mag light (about 1.5? diameter) was squished together and taped into a nice inductor. In it went, right in front of the transformer and surprise, surprise the broadband match was better and had a resonance SWR dip at 3885 KHz of all places. SWR was still too high. Hmmm it was resonant where it needed to be but needed more R. A 3dB pi attenuator preceding the L (R1=R3=300Ohms, R2=17.5Ohms) got the SWR to 3:1. Close enough.
Next the transmitter needed a DC power input. I visited the junque box and came up with a feed through capacitor of unknown value, drilled a hole and mounted it on top of the chassis. 0.1, 0.01 uF capacitors were connected to the bottom side of the feed through and ground bottom side of the chassis. 20 turns of #14 on a large yellow core toroid to keep the RF out of the PSU and the first connection to the drain was ready for soldering. A 330pF 500V dipped mica shunt capacitor was connected from the drain to ground
As long as we were working on the drain a piece of RG174 was included to connect to the output circuit. A hole and grommet later the RG174 appeared on top of the chassis. It was 9:45AM and time for the output network. Some research at Steves WA1QIX site yielded starting values. The 470pF series coupling capacitor connected to the output network coil connection point. Another 1850pF distributed among three mica caps and a three-gang variable capacitor (estimated about 990pF max. capacity) were installed on the output end of the chassis along with an RF pigtail. On to the final coil. Five turns was just too little L. Out came another large yellow toroid and 14 turns of #14 wire got me in the ball park again. After compressing the turns a little the reverse SWR read with the 269 looked right. It was noon and the transmitter was almost complete. Food, family, and other good stuff were the right thing to do through the rest of the day but the transmitter was sitting on the bench quietly calling to me for that all important smoke test.
About 7PM the call became too great and I returned to what was now known as the single FET 80M transmitter. Almost matter of factly I connected the inputs and output to the T/R switch, modulator, power supply, and swr/dummy load. Switches flipped into the T position, and 5W of RF applied to the gozinta. ZOIDS, there?s output. About 30W was coming out of the transmitter. A bit of tuning of the output cap and some compression of the output coil and output was 50W. It was meeting carrier specification. I hooked up the antenna and retweaked the final tune for max output, applied modulated DC from the QIX class H modulator and out came 50W carrier and 200W PEP. After a few yaalows, test, tests, and scope observations on 3827 Tom, W2KBW near Buffalo came back and gave me a good audio report. The TX worked. It was simple to build, fun to assemble, and put out real RF. What a deal.
It turned from a cold and overcast day into a dark and stormy night. I looked at the large heat sink on the workbench. It called to me . . . . . but that?s a story for another time.

PostPosted: Sat Feb 15, 2003 11:07 am
by Dan Brown

Great story and writing style.

And it even worked!

I thought of the CPU heatsink/fan combo as well, but have not made any heat yet on 75. Running a NorCal 40 Class E amp as in QST a few years ago.



PostPosted: Thu Feb 20, 2003 8:04 am
by k3xf
Thanks Dan. I enjoy writing as much as building. 'looking forward to QSOing with you on 40M. . .Norcal 40 to Norcal 40. . on AM. Undoubtedly a unique event. . .