NoWire, a 1985 Microwave LAN Experiment

20 Jun 2024

An Experimental WiFi LAN I designed and tested in 1985 under an FCC Special License

Working on building up a unit

So in 1985, I was kind of adrift. I had a living by doing what I’ll loosely call “consulting” but many would call computer support. Much of it was a kind of freelance piece-by-piece existence. One day I was at my major client’s offices on some project or another and he observes that it would be superior if the computers didn’t have connecting wires. This was not an untoward comment, the outfit provided a networking software product for TeleVideo Computers machines, 8-bit Z80 products. I guess it was boredom and potential, but I piped up that I could do that.

I had done a few projects for these folks, and I had a measure of credibility. Next thing up I am going to design a wireless LAN on the cheap. Now at this time to the best of my knowledge, no one else had a wireless LAN. My life has been interesting, and I had experience as a microwave maintenance petty officer for a US Navy communications station as well as homebrew experiments with surplus klystrons and tin cans from the top of Stone Mountain Georgia, so I figured microwaves would be the trick. Modern microwaves at that time were Gunn diodes; low power and relatively cheap.

Think about that for just a moment. Here we have this addlepated n’er do well who has gained the absolute confidence of a funder to work some nifty magic. The least I could do was deliver, right? ☺

My idea was to build up a proof of concept system linking a pair of machines. To do this I used a functional block microwave unit as the transceiver. These units were typically used as speed detectors in railroad marshaling yards to determine railcar weight by virtue of their acceleration downhill after being pushed over a hump.. Or police speed-detecting radars. Both of those concepts used the Doppler shift of reflected radio waves from moving objects. My idea was to use them as both my receiver and as a CW transmitter. So I got to design some nifty circuitry doing this. For one thing, the Gunn diode was voltage critical, a change in voltage gave a change in frequency. If you turned it on slowly its frequency kind of smeared on its way to quiescent, so I needed something that was very fast and stable. Another consideration was the output from the microwave module as a receiver. In this case, it was a train of high-frequency RF pulses at 4Megabits rep rate. I fed the resulting pulse train to my IF and used a biased Schottky diode bridge “detector” to remove the high-frequency component. I designed an interesting IF amplifier that is fully differential and that uses trifilar wound ferrite toroids as IF “transformers” to obtain an IF bandwidth of 8MHz at a 6MHz operating frequency. RCA Semiconductors gifted me 10 or so of their then-new RF integrated circuits. These were initially proposed for use as for example RF mixers, and of great import to me was a differential pair that was well matched. Thanks, folks!

There is no schematic for any of the circuitry. At the time I would just build things out of my head with no previous written schema. In my view the actual realization instantiated, the actual “as built”, that is the ultimate “schematic” or “logic diagram”. I also had this kind of weird thing where I would visualize the circuits with their actual components and connections before I built it, sometimes like an animation or tutorial movie about how it worked.

For this concept and my work to have more legitimacy (and to be legal. I was using commercial frequencies for my experiment to lower expenses), I decided to get an FCC license to cover this effort. That also established a demonstration of when and by whom this work was done. An official scratch on the wall of time, markings on the wall of the cave. I have been unable to locate the Documentation regarding the license or I would have included it here in the article.

In the end, the units worked fairly well. I had a functioning wifi LAN in my house as a result of this work. I got a phone call from some fellow from Microwave Associates asking me about the project. I described it in a sentence or two and he replied “Oh that won’t work.” My reply was well fly your ass on down here and you can see it working right here. Guess he didn’t believe me.

Meanwhile, what happened? The fellow I reported to, the one with the remark about the wires, told me I should try to curtail my desires because the owner had recently lost a good bit of money financing some scheme of his nephew’s. I did not feel like it was up to me to subsidize or compensate for his bad choices or his nephew’s mistakes. Then the owner himself said, “Give it to me, I know what it is worth and who to give it to.” Interesting idea, maybe he even had the very best of intent. But there were no contracts or organizations that I had a part in, so disheartened once again, I just quit doing anything further with it. I did not do this work to give it away. Such is life.

This next is a photo of one of the test units,

One of the LAN's test units

A view of the encoder decoder circuitry. Simple Manchester encoding. The unit had its own integral power supply.

There is a power cord and an i/o connector that mated the computer.

The enclosure is fabricated with copper-clad PC board laminate. This is the enclosure rear panel.

This is a view of the circuitry internal to the enclosure.

Another view shows a closer look at the transmitter driver circuitry.

Here is a final shot of the enclosure interior.

This is the first page of the CA3028B Integrated Circuit datasheet and you can see the internal circuit here. Makes a dandy diff amp. You can see that Harris Semiconductor took over the RCA operation as shown on the datasheet.

The microwave module was manufactured by Microwave_Associates and is apparently no longer available. New, suitable replacements could be found.

Hopefully, you found this article entertaining and informative. As always comments, criticisms, and suggestions are welcome.

God bless all.