March 31, 1992
You've received this article either electronically or by surface mail because you've contacted us regarding the article entitled "Inexpensive Multi-Megabaud Microwave Data Link"which appeared in the December 1989 issue of Ham Radio Magazine and in the ARRL Handbook. Our intent with this update is to answer as many questions as we can, aid in construction and facilitate exchange of information between interested persons. You'll find attached a mailing list of all recipients of this article. Please accept our apologies for being so slow in responding. This letter has been written and awaiting mailing for well over a year.
The majority of the queries we have received can be divided thus:
To our knowledge, none is commercially available yet. John Conner, WD0FHG did quite a bit of work towards that goal. He made schematics as well as pc board layouts for both boards. We met with John at the 9th ARRL/CRRL computer networking conference in 1990. At that time the designs were nearly complete but he thought that he wasn't likely to have time for further work soon and had given all the files to Bdale Garbee, N3EUA. After returning from the conference and contacting Bdale we discovered that John had retrieved the files and soon finished them. At that point, Jon Bloom KE3Z at the ARRL lab was seeking to get some links operational and had some dialogue with John. Jon got a copy of the layout and because it was not compatible with his printed circuit board processes had both boards re-laid-out and a few sets of boards fabricated. These boards did not have thru-hole process and had minor errors but were satisfactory. Jon sent 2 sets of these boards to N6GN who populated one set with components, identified the errors and except for the audio amplifier, verified operation of the completed set. As of this time it isn't clear whether anyone is going to be able to offer a board kit but Jon indicated that he thought the League would at least be able to make the artwork available for those that would like to fabricate their own boards. Contact Jon at the ARRL for details.
Some working models have been built elsewhere, presumably hand wired. In recent months Bdale has indicated that four transceivers are operating in Colorado. In addition, amateurs in Belgium (I'm afraid I've lost their names and addresses) reported a functioning link. Is anyone else working on PCBs? If so, please contact Jon Bloom at the ARRL Headquarters.
The driver software is available over the Internet on suntan.Tandem.com under the "hamradio" directory if you've access to the Internet, if not, send Kevin, N6RCE a diskette and he'll return it. with code. Obtaining the cards has been hit or miss. Sytek will sell a new one, but for a lot of money. We would suggest checking your local college or university for surplus. Does anyone else have a source of these things? The supply which Kevin had is now depleted. When the project was first started, we didn't intend to make use of the PCLANA cards, they were adapted at the last minute, when the impossibility of slowing down 10 Mbps Ethernet cards was discovered. As part of our efforts toward moderate speed, wide area amateur network development we have produced cards which can serve as a replacement for the PCLANA. A design effort by N6RCE and N6PLO, MIO is nearly ready to be deployed at remote sites, along with 256 Kbps radios that N6GN has built, as part of a prototype network in Northern California. A reasonable quantity of commercially fabricated boards exist but this initial supply is designated for "Hubmaster" and high speed backbone network development and no boards are yet available for purchase. The 1991, 10th annual ARRL Computer Networking Conference proceedings have details of these cards, radios and protocol development. Has anyone yet attempted to use the Gracilis interface cards with the microwave hardware? At the 9th Computer Networking Conference Don, N4PCR (of Gracilis) indicated that he had tested it on wire lines at 2 Mbps and showed the microwave hardware as an important part of the Illinois area network which he and others are working on. At that time data clock recovery and TTL <-> ECL interface circuitry seemed to still be in need of design and construction.
The TOKO inductors available from Digikey are fine. Make sure to use the shielded versions. Digikey TK2705 and TK2709 appear to work. The NEC Transceivers are best obtained on the Surplus market in Akihabara, downtown Tokyo. CEL, NEC's North American arm as well as some of the representatives seem to have been giving mixed signals about the availability. It appears that the transceivers are a low volume item but quotes have been made of $40+ in single quantities. Has anyone actually purchased any from NEC stateside? Potential substitutes for the NEC transceivers are advertised by:
SHF Microwave Parts Company 7102 W. 500 S. La Porte, IN 46350
The MA/Comm gunnplexers are also very viable though more expensive alternatives. In the past,Chuck Swedblom, WA6EXV, has successfully built his own oscillators using packaged Gunn diodes which seem to be available, at least on the West Coast. We don't know if he has ever mounted a mixer diode for receive as well or not. If you have MaComm units which you intend to use, contact N6GN for modifications to allow use of the electronic tuning input. These units appear to support full Ethernet speeds and should be able to operate with unmodified Ethernet interface cards. In addition, there have been some inexpensive ($25) transceivers advertised in the "Ham Ads" section of QST. We have not yet obtained any of these for testing but N6GN has seen them and they look almost identical to the NEC units. They may in fact be made by NEC. As far as we know, dishes are still available from "The Antenna Center" as indicated in the article. (The proprieter was pleased with the free advertisement from the article). If you plan on using antennas with different F/D ratios you may want to change the horn feed to be more optimum. Small TVRO dishes (4-6 feet diameter) will probably work pretty well and typically have an F/D = .375. Use of the feed from the article will result in slight underillumination and loss of gain but is a fine starting place. If someone wants design info for such feeds, contact N6GN.
The Ham Radio Magazine article was written and submitted for publication June of 1989, so for us, it's been almost three years since that phase of our projects ended. As well as providing a fun project, the goal of the article was to encorage the revolution of amateur radio networking by demonstrating that information rates much higher than current amateur practice need not be neither difficult or expensive. As the need for high speed backbones is driven by higher speed user radios we plan to readress the microwave radio design. It is likely that future designs will use discrete transistor receive and transmit amplifiers. 100 mW output power and 4 dB system noise figure which could result in a system improvement of around 20 dB is obtainable with standard microstrip design and inexpensive components. It is even likely that such microwave hardware can be built for a price similar to that of the NEC transceivers. Since the article, we have been busy trying to produce new digital hardware, protocols and radios for 900MHz and 1200 MHz. See the 8th,9th and 10th Computer Networking Conference proceedings for several articles describing the justification for our approach, a protocol to support it and details of what we are doing.
The 1991 ARRL Radio Amateur Handbook has a description of the radio hardware. N6GN used this schematic as a reference in loading the boards which Jon KE3Z provided and found the following corrections:
We hope that this additional information along with the address list may help some of you cooperate and successfully build higher speed radios. Please contact us if you have information which would be valuable to others or if you feel another letter like this could be helpful. We are busy but desire to see amateur radio networking improve. If you do desire more information an SASE or remuneration for postage would be appreciated.
Glenn Elmore n6gn 550 Willowside Rd. Santa Rosa, CA USA 95401, email@example.com
Kevin Rowett n6rce 1134 Steeplechase Ln. Cupertino, CA USA 95014 firstname.lastname@example.org