Anyway, there are some other people that had also expressed some interest in participating in this discussion and kinda find out what the heck us crazy people were doing. I wonder if KE6VHM is on frequency. . .
TRANSCRIPT transcribed by Martin Crundall AC6RM
Microwave Data Project
First Datalink Net
March 20, 1997, approx 21:00 PST
WA6BFH: QST QST QST [this is] The discussion group that has been called for this evening to bring together a few of the people that are interested in building a high speed datalink network on the microwave band. The equipment that has been chosen, or that is otherwise available, is on the 10GHz band. This will be wideband FM modulation scheme, being talked about from the <> of various pieces of hardware that have already been assembled and have bits and pieces of more and more going on, and this first discussion -- this first semi-formal discussion -- will probably lead to having similar discussions on a simplex frequency that will also ultimately be used for coordinating activities with this system to tweak and tune and prune and get it up to speed, etc. And we've got a number of people that have already been engaged in this, as well as some others that I think want to become involved. I already have W6QS on frequency, KF6HQC and AC6RM. I wonder if WQ6S might be around . . . Steve, are you there?
WQ6S: I sure am, and I have my 807 with me too!
WA6BFH: Well, such a fine thing. I don't think we'll be using any 807s on ten gigahertz, but in the process of getting the gunnplexors going -- getting these gallium arcenide diodes going on frequency -- probably several 807s will be consumed! An 807 is a vacuum tube, for those that may not be privy to that little bit of humor. Kinda looks like a beer bottle!
Anyway, there are some other people that had also expressed some interest in participating in this discussion and kinda find out what the heck us crazy people were doing. I wonder if KE6VHM is on frequency. . .
KE6VHM: . . . Yes I'm here, John; pad and pencil in hand!
WA6BFH: Cool; very good. AC6SX; are you out there mike?
AC6SX: I'll be here for about another five minutes, unfortunately; I hope you're tape's rolling John. AC6SX.
WA6BFH: QSL, dude--I may not be able to get it to you tonight, but anyway, the plan is still afoot! WN6EKZ, are you there, Alvin?
<no response>
WA6BFH: Okay, nothing on Alvin so far. KE6OLH, are you out there?
<no response>
WA6BFH: Nothing heard on OLH. How about AC6UV?
<no response>
WA6BFH: Okay, nothing on Godie, so far. KE6BKA?
<no response>
WA6BFH: Well, sad to see that we have those drop-outs. Those are the folks that had expressed an interest on the last Tech Bench Elmer's Amateur Radio Society formal net, or those who had already been engaged--in some way shape or form, or fashion--in this project. I wonder if there might be anyone else on frequency who might be aware, or might be privy to what we are intending to do here, as far as a wide-band, high-speed digital linked network. This would work somewhat similar ... what we're hoping to build and get going here is something that would work somewhat similar to be able to carry the same kind of normal services that you might do on the Internet, except do it through a purely RF link system; no telephone lines involved. At least if we ever get to the point of possibly hooking it to the Internet through a normal network link. Otherwise what we're talking about doing is being able to carry the same kind of data; be that video data, audio data, keyboard data; anything that can be digitized to carry it back and forth amongst the various stations on a very high speed and, uh, either half duplex or possibly full duplex sort of arrangement. And W6QS will tell us about that eventually.
Anyway, anybody out there listing in to this that would like to be listed here to be called upon, or would like to interject points, or would like to make their presence known to possibly interject points later on in the discussion. WA6BFH standing by listening for call-signs at this time.
Responses: KE6ONZ, KD6RHD, AC6NI, KD6VRQ ("is here about half the time").
WA6BFH: Okay, very good. There was somebody just before VRQ; you were kind of weak and broken in to the machine. I wasn't able to pull out a call sign or modulation, or much intelligence out at all. So try it one more time ... after AC6NI, before VRQ.
<no response>
WA6BFH: Okay, well, he must have driven off the edge of the earth.
All right! Later on we'll look again for WN6EKZ, KD6OLH, AC6UV and KD6BKA, who are folks that expressed a specific interest at the last Tech Bench Elmer's net to possibly get involved with this and see what they might wanna do. In the meantime, let me give a general synopsis of what we've got going here, and then I am going to turn this over to KF6HQC for some commentary and just plain further the possible potential of what can happen.
KF6HQC: Okay, well, yeah--that's pretty much the idea here. I've got these gunnplexors pretty much available, but I don't really want to just hand them out to people with a passing interest; they'd be much better used for somebody who's definitely got a desire to get on line here with either digital, or---actually, my interest in this isn't in the digital realm like Martin or John or I think Mike, AC6SX, he also wants to go digital. I'm certainly willing to participate in the experiments, but my interest in this is going "mountain-topping" and "seeking the DX." But either way, if you do have an interest and a desire to build either the kit that hopefully we can provide--we can certainly provide schematics if you want to home-brew this stuff yourself--I'd be willing to hand a gunnplexor over to you, for the cause, and there's probably in the neighborhood of ten, twenty, may be more, that will be available in the near future, and I believe we've got one or two sitting on the bench that are--I'm not going to say "up-for-grabs", but they are certainly available for somebody who's got that interest. And you can contact me through a various number of ways; e-mail, or phone number if you've gotten my phone number on the repeater, or whatever.
WA6BFH: Nope. I think you covered that rather fully. We may talk later on about the idea of what the different parameters would be that would make the use of a gunnplexor more adaptable or more desirable for mountain-to-mountain communications versus this sort of a data system. But suffice it to say, even for a data system, you can always plug a microphone into this thing and use it to talk to people, whether that be from a mountain top or with a dish bolted to the chimney of your house, or something like that. And that's the system we envision, and that's certainly the availability of what is built upon this IF board. The IF--or intermediate frequency--generally speaking, controls the function of the radio both in terms of, primarily receive signal, but also as a package. Or within the constraints of that receive signal. That's how the transmitter's built to match up with those specs. And pursuant to some of the test equipment that was mentioned before, like the absorption wave meter, I know that there are at least two absorption wave meters floating around within relatively easy grasp now; maybe three. Also some of the people in the [San Bernardino] Microwave Society are also available to help with this project. They're not probably as wedded to our particular end for the project, mostly because they are interested in sideband on these same frequencies. But I am sure they would be quite enthusiastic and helpful in getting this going. So we have the test equipment available, we have the expertise available, and a lot of this is actually pretty easy to do. I mean, if you'd seen--it's too bad we don't have pictures available from Fred, myself and Martin, AC6RM, working over at Martin's house, me with hacksaw in hand, sawing loose a well-locked-down lock-nut on a tuning screw! The way these radios are tuned, the way these gunnplexors are tuned is by threading in an insertable stud, which is basically a bolt, into the gunnplexor, basically to change the shape of the cavity; actually to change its relative inductance and capacitance, that you might think about it as changing the shape of the cavity.
KF6HQC: Yeah. As far as the hacksaw technique is concerned, that's obsolete. I've found a nicer way to do it and it leaves that screw completely intact, along with it's locking lock-nut, or whatever, and then just--better than the silicone method that you came up with--a spot of Elmer's glue will keep it all in place. So no more hacksaw! Back to you. KF6HQC.
WA6BFH: Okay (laugh), well I'm not suggesting it as the OPTIMUM method, but anyway, just trying to illustrate a point about what can be done. You don't need to be gentile with these, necessarily. I mean there are certain things you don't wanna do: you don't wanna be jamming too much voltage or current through the diode, etc. and burning them up, but I was just trying to, mostly, illustrate the point that you can be a little bit mechanically extreme and, for the most part, so to speak, do no harm! So, anyway, it's a project that's easy enough to do, and there are some things to be learned, and I think that's the most fun part of it. We've got a bunch of good people around by way of those who have already sampled some of the delights and pleasures of playing with gunnplexors, as well as a bunch of people that have been doing this for decades in the San Bernardino Microwave Society. I think we're in good shape all the way around!
WQ6S: No emergency traffic, or anything, so thank you John. My name is Steve. Welcome and hello to those I don't know.
WA6BFH: Nothing I can think of. Since you mentioned the price in the way that you did it there with the board that you got from Chuck, and we should also mention probably that the ARR board was ALSO designed by Chuck: it's the same board; it's virtually the same board. He provided them his design previously. So any of the ARR systems that you might hear about are virtually the same thing. Since you mentioned the price on yours, and that's a good benchmark for me, in some of the quotations that I'm gonna be doing, 'cause when I know what basically the single-unit price is kinda doing it the hard way, that'll help me on a percentage of scale when I'm talking to some of the people for the board that I would like to see produced. To clarify my point, Steve had to drill all the holes to mount the components in the board that he got. Chuck etched the circuit traces on the board, and provided a board to him that was not drilled! That's not the way I would wanna do it; that's a pain in the butt. For the number of components, many of which are resistors with two leads, are capacitors with two leads, and some ICs with eight leads, or whatever, drilling a bunch of holes, is in my scheme of things, not the way to go.
AC6UV: Sorry about being tardy; just got in! AC6UV.
WA6BFH: Okay, no problem, good to hear you out there. KE6BKA, are you around?
W6QS: You wanna go through and look at Ethernet as basically the cards that are on there--the terminal adapters--[as], basically, high-speed modems that are set to be half-duplex. What they basically do, your normal Ethernet, or standard Ethernet (10-base-2 or 10-base-5) is running ten megabit per second, or ten million bits per second that'll go through and dump down the wire, and all the other terminal adapter boards in there are NIC [Network Interface Controller] cards (their proper name) are going through and listening to the signal that is coming down the wire and determining whether it's addressed to their particular card. If it is, they'll go ahead and pick it up and dump it in to a memory on board and then off to the processor. If it's not then they'll dump that particular packet.
WA6BFH: Nope; that covered most of what I wanted to cover in the basic scheme of what we would intend to be the ultimate goal in the modulation provided to the [gunnplexor]. Otherwise, like I said, you plug a microphone in to it and you talk in to it, but for surplus Ethernet cards/hardware: where are they available and at what kind of expense is typically found? Much like we're dredging up these gunnplexors, and they're "falling off the turnip truck," and if they would be come available, how available is the Ethernet hardware from the used market, etc.?
W6QS: I've gotten several of the 10-base-2 cards, the old "thin-net" which a lot of the companies are pulling out and replacing with the twisted-pair stuff. I've been able to find several of those just for relieving them of the storage space for them. I don't know how many more I can get at that price, but I've seen them out at the swap meets, Pomona Fairgrounds, when they have the swap meets out there. I've gone through and seen them in the $20 to $25 range in new boards, and used ones go on down from there; you're not gonna wanna pay a whole bunch past that point. If you're talking about a PCMCIA card, then you're talking a little bit more; I think I've gone through and seen some at around the hundred dollar range, or there about (if you're trying to do a set-up with your notebook on a mountaintop with it). But as far as the straight ones that plug in to a normal PC, that type of a range and cheaper. WA6BFH, W6QS.
WA6BFH: Okay, I think that'll cover that end of it for the time-being. Let me ... I'm thinking that an explanation of the whys and wherefores of twisting screws and building things to get these basic things going might be in order from anybody who has questions at this point in time. I can tell you that I built a "modulated power supply." Basically, all that a "modulated power supply" is is a three-pin regulator, like you can get out of radio shack very easily; the recommended regulator is an LM317 and that, and a hand-full of components, will yield you both the power supply to run the gunnplexor and also a way to--what I would describe as--wiggling the control voltage on it to provide the FM modulation. That's what an additional op-amp does, as well as a few other components, and initially a microphone input.
WA6BFH: Okay, I'm gonna pick 'em up in the order I heard 'em. So AC6RM, why don't you go ahead there first.
AC6RM: Okay, I'll do that. First a comment: The Ethernet cards; if you look in the junk boxes at the DeVry swap meet, I saw a couple, three of 'em in boxes. They're Novel cards; you'll see Novel on them. It's got a little 'N'-type connector on it, and then another multi-pin connector. That's an Ethernet card and they're like a buck or something; they're really cheap.
WA6BFH: Okay, I'm gonna come back to your point, Martin, later on to elicit comments from some of the people that may be more experienced with that particulate, Don W6QS, as far as thrashing that out. I'm gonna go to VHM right now, but I'll interject at this point in time, I think we're gonna have a certain amount of just basic testing and RF set-up between the assembled units. Obviously it's going to take different people different amounts of time to get their IF boards built. I mean when we drop a kit of parts in their lap, and they've got the board ...
<end of side one of tape A>
... parts put together in a couple of hours or something like that, and they're, like, up and ready to go. Others might take a couple of weeks. In fact if they take much longer than a couple of weeks, the rest of us will be prodding them and kicking them in the rear end saying "c'mon!" cuz we wanna get a field trip planned to go out all together to some football field or something or parking lot and start pointing our dishes and horns at each other and start talking to one another getting some of these "dialed in." So we'll get to THAT part before we get to the data link part. The data link will come after the speaking in to a microphone and going "Hello, can you hear me?" and having somebody else talk back to us. And I think all of that will be kinda the fun of it.
KE6VHM: My question is a little more basic; I am wondering about the dish antenna. There are a lot of these RCA dishes around. I am wondering if that is what you're edging toward, or is this a home-built dish? I am also wondering about these wireless cable TV dishes. They are more of a run-of-wires built into the form of a dish. What are you lookin' at here, John? WA6BFH and the group, KE6VHM.
WA6BFH: Okay good question. Both for "type of hardware" and also for "where do you get these kinda things." Probably the best desirable dish that is gonna fit this system very well, is the same kind of elliptical, solid metal, parabolic dish that you see that Radio Shack sells ... your digital satellite service dish, your DSS dish that you've seen being sold for the last year or two at Sears and every place else. That kind of egg-shaped, elliptical, parabolic dish antenna of solid metal. The wireless cable system's screen-grid pseudo-dishes; they're kinda shaped like a dish but they are wire rods welded together, that ain't gonna work! You don't want anything like that. But it'd be a solid metal dish that looks like just the normal round parabolic dish, like you see on ... much in the shape of the eight-foot dishes, like you used to see on the old earth satellite systems, the old TVRO--Television Receive Only--dishes, except it doesn't need to be that big. In fact you don't wannit that big. You wouldn't want a dish--even the round parabolic type--larger than about two foot in diameter, and something of the order of eighteen or twenty inches would be better, cuz you don't have to point it as critically. But if somebody put a three or four-foot dish in your lap, that'd be okay! You could use that. So that wouldn't be bad either. But the DSS satellite dishes, like Sony is putting on everybody's roof now, and RCA and Radio Shack and all these other places that are selling these digital satellite dishes, that's about the kind of dish you want.
The military type, the round dishes, are available surplus; you find them all over the place. Fred came up with one that was probably part of a "Z-TV" dish converter (some of you may remember the "Z-TV" back from ... when was that, 1975? 1976? Somewhere around in there ...) Any solid conductive, any solid surface parabolic dish of about fifteen, sixteen, eighteen, twenty-some-odd inches is a real desirable antenna.
So that's the basic haps on that. You find 'em wherever you can find 'em! The guys in the [San Bernardino] Microwave Society are telling me that the elliptical dishes are showing up surplus! Seems kinda funny to me because they've been selling them as brand new equipment for everybody that is selling digital satellite systems on their roofs and hanging them from the eaves of their house, and that kinda stuff. But I guess they're around; they've been building them by the ton, so I guess they're around. I'm probably going to be using a military type round eighteen inch dishes, or maybe something a little bit bigger. I'll do what ever I need to do. I've got a few different dish antennas I've been "glomming on to" from various surplus outlets for a while now.
KF6HQC: Okay, well, that's covered it pretty well as far as I can tell. It sounded pretty good. It's funny because that's exactly what my question or my comment was. VHM, I was going to tell you that at one of the microwave meetings I attended (I believe it was January) one of the guys there, one of the members, he had one of those eighteen-inch dishes there and he was saying that it was available for sale; you didn't have to buy all that other garbage. You could buy the eighteen-inch dish with the little arm that extends from the bottom and, basically puts you where you want to be with your gunnplexor, for about $30. So you're looking at roughly $30 for a dish. And that'll be either the RCA or the DSS, you know the little teeny ones that you see on the houses. And they were saying that that little dish yields about a 30db gain, which would take your ten milliwatts up to about a ten watt effective radiated power from the dish. So it's not too shabby. So basically that's what my comment was, and Frank I'm glad you asked that question. And Frank, was your call WA6VHM--break?
KE6VHM. No, Kilo Echo Six Victor Hotel Mike; we spoke the other day when I was up in the mountains. You had a fella who just got his license in your vehicle with you, and I lost you in the valley (laugh).
KF6HQC: Okay thank you for that. Hopefully that answers some, or all, or most of your questions. Back to WA6BFH, KF6HQC.
"WQ6S" <break>
WA6BFH: Yeah, go ahead Steve. WA6BFH.
WQ6S: Okay, for Frank and Fred, if you're interested MCM Electronics is a mail order house that sells those DDS dishes. I think the numbers come up closer to $35 but still that's an excellent price for a brand new antenna. They have a toll-free eight hundred number if you guys wanna just give 'em a call and ask them for a catalog: 1-800-543-4330. WA6BFH and the group, WQ6S.
WA6BFH: Okay, very good, and just to throw a couple other ideas in to the pot here, some stations may not even need the dishes. The dish is just merely a way to build up gain, like some people on two meters can get away with a "J" antenna nicely, while others are gonna need some sort of a beam. The Solfan gunnplexors come with a HORN antenna on them, that's probably about a 12 dB horn. It's approximately two inches wide and maybe an inch or so tall. And that may be enough for many people just the way it is, pointing that at the other station.
WA6BFH: Go ahead Steve.
WQ6S: Thank you John. I guess I'd like to direct this to Don. Not knowin' much about the digital end of this stuff, do we have to worry about bit error rates? Do these Ethernet cards have any kind of error correction modes in 'em--CRC stuff or anything? Just curious. Oh, and how much band width? Save me from looking it up and doing the math, if you happen to know, how much band width do we need to have to run ten megabits per second? W6QS, WQ6S (boy, that's tough!)
W6QS: (Yes, that is a little bit of a challenge there [referring to the exchange of closely matching call-signs]!) Part of that is according to how much you want to dev' it. A fairly narrow deviation is going to be about the same as your modulation signal, which if you've got your symbol rate of--well, with Manchester, with the bit-reversal--you can have up to ten megabits signal going down there if you're sending all "ones" or all "zeros." So I'm guessing it's going to be at least about one and a half times that; something around 12, 15 megahertz bandwidth would be necessary there.
WQ6S: Thanks Don. And was I correct: the prices you quoted, twenty, twenty-five bucks for a brand new card, was that correct? If so, what are some of the preferred brand names, or whatever. Could you run that by me again?
W6QS: Yeah, the ones that I saw, like $25 dollar ones, were like Novels in like five packs. So if we got a few of us together. But like Novel is a common one, 3Com, I think Future Domain, there's a whole pile of them. The main one that we'd have to take a look at for something like this would just be which drivers? You'd go through and easily find out, for your particular PC--if you're running Windows95(tm) you'd go through and see if you could find drivers for 'em. Or some of the other ones. The particular ones I have a bunch of old ones around here are a bunch of 3Com, but no magic there. Most of them are the same two or three chipsets. It's just a matter of who ever happened to stick them on the board, and wire up a power supply and a couple of those things through it. Go through and put together the driver code again, which, half the time, is supplied by the chip manufacturers, or most of it is. So fairly generic in there. Some of the fancier ones have bigger buffers on them, give it a little easier job on a slow PC, but, again, a '386, so you can run something like Windows(tm), so you can get the TCP/IP off of one of the servers from Microsoft, or whatever, or something that can run Windows95(tm); not a whole lot of overhead there; most of the fast signaling is done by a couple of chips on the Ethernet card. WQ6S, W6QS.
W6QS: Thanks Don--that's encouraging! Maybe I'll jump right in to this thing with both feet! Sounds like fun! WA6BFH and the group, WQ6S.
WA6BFH: Okay, and I was even thinking about, what I would like to do--and maybe Don, you would like to do this actually--I suspect that might yield a better chain of communication, or line of communication, if maybe YOU got a hold of TAPR, and explained to them what we're talking about doing and seeing what they have done already. What I have HEARD is that TAPR has done similar things. I think it'd be kinda nice to know where they're at, and what hints and kinks they've already ironed out, and even what direction they're going--even if we find that to be an undesirable direction, even if the path that they've taken leads them to do things that we go, "well, gees, that's kinda weird? Why are they sacrificing this or that to do whatever?" But it'd be nice to know what they've got in mind and how that might be going on, because I understand that they are doing pretty much exactly this, at least the same basic idea: high speed data on the ten-gigahertz band. Now, what they're doing, and what their modus operandi or plan of attack has been to do that, I don't know, but it might be interesting to find out.
WA6BFH: Okay, KE6VHM, go ahead.
KE6VHM: This Ethernet card you're referring to is simply a local area network--LAN--card for a computer system with a server and terminals? Is that the same thing you're talking about? KE6VHM.
WA6BFH: I'm gonna defer that to Don in a minute; let him give an explanation--or add to this explanation: I'll tell you what I know about Ethernet from the history of it. A bunch of guys from Xerox Park, which was--or IS--kind of the R&D wizz-bang guys at Xerox, where all the different workstation kind of communicating, mostly originally for their own purposes, so that one Xerox facility could talk to another Xerox facility could talk to another Xerox facility, they could trade their pool of knowledge and design talents for doing various things. And then they also developed this inter-communications system that--at least during the early eighties--was slick stuff. I mean, that was the really neat thing. Where it's gone from there, or how--relatively--it's gone from there, I dunno. Don, why don't you correct me here; the pictures that I have just painted, or tell us a little bit more about that. W6QS, WA6BFH.
W6QS: Yeah, that's basically their development, there. It's been out there for quite a long time at this point. And yeah, the Ethernet card, or just plain-ol' LAN card, the NIC (or Network Interface Controller) [depending] on who's talking about it, but yeah just generic "thin-net" or "cheaper-net" is what it was called coming out on the coax, although I've also been going through and looking at some ones that have been set up for "twisted pair" which are becoming a little bit more common now. It might go through and save us a hybrid on it. Nice [thing about] the coax type is you can have a couple computers at your site and easily tag both of them, or however many, plus the Ethernet going out on your local network. Or if you do twisted pair as opposed to the coax, it's basically the same signaling signals; it's just they separate the RX from the TX, right at the wires instead of having to build something in to separate it. WA6BFH, W6QS.
WA6BFH: Okay. I'm guessing that the "hot-tip" card to find in the surplus market will probably be the 50 ohm RF cable cards, and I can tell ya lots of times--crawling around attics crawl spaces and what-not, tracing down lines for gas sensors and doing things like that--Ethernet cables are pervasive. Generally speaking coax cable, fifty-ohm coax cable intended for Ethernet service, is either orange or yellow in color and says "Ethernet" on it, and it's a fifty-ohm cable, very--in outside appearance--very similar to Belden 9913 cable, except that it's either kinda yellow or orange, and it says "Ethernet" on it, and it's got "N" connectors at the end, and those "N" connectors go in to a card some place that's got a female "N" connector on it. So those might be the cards to look for, but Frank, did that kinda answer your question? KE6VHM, WA6BFH.
KE6VHM: Yes it did, John, and I might be able to come up with a source for those. The company I work for is currently changing out a lot of stuff and one of the things they're doing is getting away from the old LAN. I don't know what technology they're using, but they're putting something new in, and they are changing a bunch of stuff out at the control center down at the pipeline. I'll look in to that and see if there's a source there and let you know when I do find out. WA6BFH, KE6VHM.
WA6BFH: Cool. That would be great, because the OLD Ethernet systems are probably about twenty years old, I think--pretty close to that-- and the newer ones are probably at least ten years old; the type we are talking about. So it wouldn't surprise me if, in the idea for amortizing for best efficiency, that even ten year old hardware would be getting the boot. So it wouldn't surprise me at all that that stuff is available, and if it's available as handily to us as somebody in our midst that knows where it is, that'd be really cool.
WA6BFH: Yeah, ONZ, go ahead.
KE6ONZ: My question, on using the Ethernet card, is -- basically, we're looking at point-to-point with it; well, somebody somewhere along the line is going to have to be a relay to another station. How are we going to go about doing that? WA6BFH, KE6ONZ.
WA6BFH: That may relate to Martin's earlier question about running double-duplex systems or, as I understand Ethernet systems--and I'll refer this question to Don in a little bit, to see if I can get some feedback that way--basically, the Ethernet card, the transmission system, is what basically can be described as a half-duplex system. A gunnplexor is capable of being a FULL-duplex system. In fact, that's normally the way it works; a gunnplexor is a full-duplex device. If you're on mountain-top "A" and you wanna talk to your buddy and he's on mountain-top "B" while you're calling him--if my buddy is KF6HQC and I say "KF6HQC, KF6HQC, this is WA6BFH calling." well, I hear myself out of my speaker, out of my receiver, while I'm talking to him, while I'm calling him. And when he comes back I hear him and he hears him. So we might want to adapt the system, versus that hardware, to be able to do things like that; since we have that capability on the RF side, we may not want to "paint ourselves in to a corner" with half-duplex systems on the data side, but I'm really getting out of my realm here. Don, what do you have to say about that? W6QS, WA6BFH.
W6QS: I've been thinking about that type of thing. There are two types things that you're monitoring for when your listening: if either the signal is headed towards you for reception, or if the data link is busy, so you don't force a collision--jump on top of it when it's already transmitting. The thought I was looking at, if you went through and had two IF frequencies: for example, a standard 30 for when you're transmitting and listening for a collision coming back, and say maybe 50. You'd move your IF down another 20 meg so that you're not transmitting in the reception bandwidth of the other gunnplexor, of which ever one is the central one--again, picturing something more like a repeater, or higher-located one, a central point for any particular node getting into--and when you go to transmit you jump your gunnplexor in to the 30 meg off so you can hear it and go ahead and do your transmitting. When you're listening to that node, if that were the terminus of a link over to another side of the mountain, when your listening to it you're off by 50 meg, when you're transmitting to it you're off by 30 meg and go ahead and make your connection to get in there. That way you could hear when it was busy and do your hold-off, and when you want to go ahead and insert your transmission, you're ready to go through and try to capture it.
WA6BFH: Okay, while you've been describing that, I've been looking back at the CA3189E layout and what-all is available on that as a receiver. That's the integrated circuit that I think would be the optimum choice. That's the thing that's part'n'parcel of the main working part, if you will, the main active component for the IF system that we're talking about. Otherwise ONZ, did that answer the question there? WA6BFH.
KE6ONZ: Yeah, that helped out quite a bit, cuz that was my main question; how we were going to route it around. So if you did use it as a router, I take it that you would have to have more than one Ethernet card, is that correct?
W6QS: Yeah, at the sites where you wanted to do routing where you had an "IN" and an "OUT" you'd need to have more than one card. If you're doing just the point-to-point communication then you wouldn't need that, but I think that the longer-term one is gonna be more of a router. If you went through and you had two dishes, then you'll probably need two cards to go through and support 'em, but the cards are the much cheaper portion of that.
KE6ONZ: Okay, well that was my main question. Back to you John. KE6ONZ.
WA6BFH: Okay. I am thinking that probably for the time that we have used so far on the air here--an hour and twenty minutes now--we've probably covered at least the bulk of I figured this first discussion would want to cover, as far as what the basic hardware is, where it's coming from, what it would be, what it would cost, sorta how you'd do this, in a sense of whether you're going to be soldering this together, or--if Fred will excuse my analogy--you're going to be hack-sawing it together, anyway, any more questions about that, before we perhaps move towards wrapping this up. WA6BFH listening for call-signs.
WA6BFH: Yeah, AC6RM--Go ahead, Martin.
AC6RM: Okay. I just want to clarify a couple of things with Don. Don, you're just a wealth of information; answering questions that I've had left-right-and-center. Just so I can understand a little bit more of the practicality of setting up such a wide-band data link; 12 to 15 megahertz. If we want a reasonable data link, you mentioned the distance of approximately five miles to deal with some of the timing issues on the size of the packets that the Ethernet card is generating. I looked at the bandwidth as possibly another limiting factor. The beam coming out of those dishes is ... we'll it's not two degrees or three degrees; it's a little wider than that so it's easier to point, but if you're having to send and receive that kind of band width, I'm thinking you're going to have to have a pretty rock-solid type of affair, and whether that's realistic for what we're looking at . ...
WA6BFH: Okay, yeah and I don't mean to be Discussion Control, I just figure that's probably a better way to work things under the current constraints. I have another frequency in mind for the next discussion, and I think I'll refrain from mentioning that at this particular time; many of you may realize why I am deciding to do that. But we'll get the word out to ya!
WQ6S: You scared me, John; I thought you were going to ask me to do some FFTs! Umm, the front end is two megahertz wide. The IF, as Chuck designed it, was 220 kHz but I ended up with a filter that is 180 kHz. So somewhere in that ball-park anyway.
WA6BFH: Okay, close enough for government work. Yeah, I wanted some Fast Furrier Transforms from ya, and you better get out that calculator and your fingers better be moving at lightning speed! (laugh). So, Don, let's say for argument here, and I think this is a very practical number for us to work with, in terms of the ten milliwatt signal that we're gonna be getting out of these gunnplexors, probably wouldn't want to go to any wider bandwidth than this right now; I'd like to make this work. But you're talking in terms of a 10 MHz wide band pass for the sine wave for the detection signal, and we're talking about .2 MHz, but I'm thinking about this on the basis of voltage detect transform. If we can switch fast enough, if we can capture fast enough, that should work for any data that we can jam down our, basically, .2 MHz wide IF, should it not? W6QS, WA6BFH.
WA6BFH, W6QS. I believe you what you're looking at is what comes up as a bezel function, which is a ratio of your modulation bandwidth to your modulating bandwidth, and your minimum is 1:1 which is you having a very small deviation, but deviating very quickly then you come up close to your modulating bandwidth. If you go from there, then it takes off from there. If you're, obviously, modulating with 20 MHz wide with 1 Hz data then you still have 20 MHz bandwidth. But if you're doing a very narrow modulation but have a wide modulation signal, the modulating signal is a fairly high frequency, you still contain the bandwidth of the modulating signal in there as a minimum. So you'll not be able to go through and put it down a 100 or 200 kHz bandwidth with the high rate signal through it. The one that I'm looking at as my model for trying to go through and trying look at the power levels and some of the other stuff is the 10 gig inputs on some of the ATVs, where, my understanding is that with 50 milliwatts and a reasonable dish you can go through and pick up Santiago Peak which is a reasonable number of miles away and it's running a video signal which is on the order of 6 MHz or 6 and a half MHz. So in the ballpark of what we're talking about; I'm not sure what their deviation is on it so I am not sure how much wider it is in the microwave spectrum, but we're like a factor of 2:1 or somewhere in around there as an approximation. Trying to get some ideas there; I know that at 1.7 gig, they're coming out with a ten-watt transmitter; I was looking around at the various ones coming out; 360 degree beam, but fairly narrow to the horizon, coming out of that one that ends up turning a very nice video signal. Not sure of the exact signal/noise, like I say I have not worked that much.
WA6BFH: Okay, I wuz gonna toss in something, but I'll yield. Let me go to AC6RM first, and why don't you, Martin, pass it over to Fred, KF6HQC.
<KF6HQC: Martin, can I have it real quick please?>
<AC6RM: Age before beauty>
KF6HQC: (laugh) I have to check out, guys; I'm just in here to tell you that. I appreciate all comments. Enjoyed all the good information, and wanna thank each and every one of you guys for checking in and showing an interest. Get in touch with me if you guys out there do have an interest and desire to get involved with this and we'll work out a gunnplexor for you. seven-three to the group. AC6RM, KF6HQC.
AC6RM: Okay, 73 Fred, and I'll talk to you tomorrow, probably. Okay, John I just wanted to interject a clarifying question on what Don just said: so I take it this IF board that we're looking at, the one that Steve has built, is not the project that we want to do, if we wanna move high speed data. If we want to build something we can play with that's wide-band FM, it IS the product. If we wanna do data, we're looking at something more akin to ATV stuff, like the stuff I brought down the other night, using that stuff, and having a board, or whatever, with THAT kind of band-width on it. That's my question.
WA6BFH: I don't think so. What Don and I were both talking about is something relative to his comments about a bezel'd null function or a bezel function. And with the kind of band-pass that Steve mentioned on the way HIS board is built. This design could be broadened if we need to. The problem with making it too wide, is you loose signal-to-noise ratio. We've only got ten milliwatts of RF power to deal with. Optimally we don't want to throw signal away by spreading it over band width. Likewise, with the other hardware that we've been talking about there may be a point of criticality in needing more deviation, needing more voltage swing. But to put it in more realistic terms, these boards are gonna be a half-megacycle wide! That's what happens with that bezel function. With the 200 kHz bandwidth, they're realistically gonna be looking at a modulation scheme that is a half-megacycle wide, about. So Don, is that doable? What kind of data or, how do these boards relate to dumping them down a half-megacycle wide band-pass?
W6QS: As they're set up, at the ten-megabit rate, they wouldn't work too well. They are all keyed off a 20 MHz crystal in there, so you go through and drop that down to 2 MHz and you'd be all set up to do 1 megabit per second instead of ten and you might be able to drop something like that down the width of channel that you're talking about. A little lower data rate, but that's what you're getting versus your bandwidth, there. Again, using basically the same card with a crystal change. So it is possible with out too ugly of a set up, there. The transceiver on them, you probably have to go through what's called an AUI connector on there before they go through and do some of the buffering on it which may be just as good to separate it out. Most of the older cards have those, because in side the transceiver that gets you out to the Ethernet; it's usually transformer-coupled, and if you dropped it down by a factor of ten, I'm sure you'd be running in to some of the transformer saturations in there, the microsecond volt type limitations in the transformer. So yeah, if we drop it down, you're doing a little more surgery on the Ethernet card rather than using standard ones, but, hey, all things be possible. WA6BFH, W6QS.
WA6BFH: Okay, very good. Yeah, that's what I was thinking about; how much adaptation of the cards might be necessary because, here's the way I think it boils down: we should probably build the IFs for this band width, like Steve was talking about before. This will give us a useable bench mark for signal-to-noise ratio with them built this way; without Q spoiling the band width on the card down to some considerably wider band width, and throwing away some of the ten milliwatts we've got at the transmit end. Also it allows us to do some of the things that Fred was talking about, like plugging a microphone in to the bloody thing and talking to each other. We would probably want to, later, set up a couple of the Ethernet cards, re-crystal them, so that they would work more narrowly, and find out how much data we can put down there, and find out if that'd be useable in the practical terms of what we wanna do, and what we wanna send. If we find out that it's not, we widen it back out, we widen the bandwidth on the IF and probably we also come up with more RF, or shorter paths. Either shorten the square of the distance, or we come up with more gain. There's gonna be a balance point in there someplace, but I think probably building the IF boards as we're talking about doing NOW, and just making THAT work, and then playing with it at the data system end of it, is probably the way to go. Who posed that question anyway? AC6RM, I think that was yours. WA6BFH.
AC6RM: ... ... 115.2 kilobaud, which is what you're looking at in terms of the highest data rate on a serial port, at least as shown in MY Windows95(tm) dialog box; I'm sure they've got it higher ... So [we would be operating at] essentially ten times that data rate, if the math I do in my head is right, and that'd be -- with four or five nodes, that's okay. Say you set up your architecture so you've got five nodes on a bus and if you go beyond five, you start routing between buses, or whatever, when the network starts to get big -- you're still okay; sure beats the heck out of 9600 baud packet, that's for sure! Certainly, if at some point we do come up with a link in to an actual T1 site that's providing services, just for giggles and grins just to see what would happen, your screen updates on your browser would be nice'n'quick, from what I can see, unless everybody was downloading or FTPing large files all at the same time. Looks good to me. My questions are now answered. What we need to do is fairly clear. Minus the exact specifics, but I have been steered correctly, John. WA6BFH, AC6RM.
WA6BFH: Yeah, my basic thought is that the signal-to-noise ratio situation that we need to deal with because we don't know how that's going to be, and--believe me, in a lot of practical situations, just around our houses--we're gonna have RF path problems. I mean you've got ... we don't have a lot of RF power anyway, relative to what is possible with gunnplexors. Gunnplexors that are out there available on the market do 100 milliwatts or better; 10 dBs to START with, and actually probably closer to 16 dBs better than what we're running, and these may not be 10 milliwatts, because they probably didn't need to be for their original purpose. Many of them are probably six or eight milliwatts.
KE6VHM: Yes I do.
WA6BFH: Okay. I wonder if AC6S... well I know he does and I know how to get a hold of him. AC6UV is not on the Internet yet. KE6ONZ, how about you?
KE6ONZ: Negative on the Internet, John, and I'm gonna go ahead and check out at this time. seventy-threes. Thank you for all the information. KE6ONZ's gonna be clear.
WA6BFH: Before you run away, are you still interested in doing this?
KE6ONZ: Definitely yes.
WA6BFH: Okay, some how we need to work out a way to get a hold of each other via telephone or something like that. We don't have to do that real soon, but you know what I mean.
<AC6RM BREAK: I can get 'im.>
WA6BFH: Cool -- that works. All right, let's see. W6QS, I have, HQC can get a hold of me on the twisted pair, RM can get a hold of me on the twisted pair, and WQ6S can get a hold of me on the twisted pair. So, anyway, all right, we have that in hand because I think we do a chat session like this over radio, we'll probably wanna do it on another frequency. And beyond that, probably my guess is, after that next chat session, we'll probably wanna physically get together, maybe over a Denver omelet or something like that, across a restaurant table from each other and scream at each other and have "real time full duplex!"
WA6BFH: AC6UV, Go ahead!
AC6UV: John, if there is a prototype of the IF circuit already built and you're figuring on getting a board reasonably priced, there is a fella I can turn you on to that has done quite a bit of kitting; actually two of them: Doug Hendrix
WA6BFH: Okay that's fine, Gody, because-- what I'm gonna be doing after we get down to some numbers, after we get down to some realistic numbers, when we know we're looking for 10 boards, 20 boards, 30 boards, whatever -- I'm gonna be shopping amongst the different contacts that I have for some commercial boards, some people that I've dealt with in the past, and also some sources that I know of now that build prototypes and stuff for hams that are not necessarily of commercial quality but decent boards none-the-less, and getting some price quotes all the way around, so I can find out thresholds of pain as far as price and what-not. If we can get a decent number of boards made, I think I can get some real good prices on real good quality boards, on commercial quality boards. But I'll be looking at everything to make sure that we can get this done. I just don't want to do it the way Steve had to. The thought of having to drill holes in boards, you know, to drill a hundred-some-odd holes, probably a few hundred holes, leaves me cold. I wanna just have the board in hand and be able to put parts in it and solder them in place. So anyway, that's where my thinking is. But yeah, get that information to me. It's not critical that it come in the next week, but in the next few days or something like that; what ever's handy for you. You've got my land-line there, so it shouldn't be a problem.
WA6BFH: Okay, W6QS, hasta lombego. WQ6S go ahead.
WQ6S: Just real briefly, before we submit a final thing for somebody to make some boards, we might wanna consider a few changes to Chuck's original design to better suit ours. That's all I had. Something we can hash over in the future.
WA6BFH: Ahh yeah! Hash that over with me; let me know what you're thinking of. Not that I have the artwork from him in hand yet, but shoot, I didn't wanna hear about things like that! But anyway, okay, well, talk to me about that via e-mail or something?
WQ6S: I wish, but don't have it yet John. But basically it boils down to -- he has the modulator built on to that IF board using a different chip. I would suggest using the modulator that YOU built. Incorporate that design in lieu of what's on the board. Just that kind of stuff, nothing real drastic.
WA6BFH: Oh GOOD! You mean using a 1458 instead of a CA3130?
WQ6S: Umm, well, either way. That or the 3130; but on the board he has ... oh what was it ... I'm lost; a 741? I don't think we wanna go that way.
WA6BFH: Okay, I'll look at that on the schematic. That's not a big problem, because a way I thought about doing this ANYWAY, it'd probably help people out, my next modus operandi, is getting people building a power supply to power up their gunnplexors--the power supply modulator, and building that as a little vector board. That way if they screw that up, learning to solder or whatever, they can go "Oh, feces!" and throw it away! Let's build another one! Then, when the manufactured board comes out, which in many respects will be easier to work with--plug in a part and solder it--it'll be a lot easier than vector board kinda stuff. They'll have a real leg up on that. I was thinking about coaching that along the way. Literally getting people together and saying "this is how you do this. You buy this board RS number whatever, and you take this solvent stuff and you wipe it off, you clean it off first, and then here's how you do the next!" Anybody who's built that is gonna have no trouble at all building this finished circuit board, because they'll apply the same general technique to building the finished circuit board; it'll just go together in a snap.
WQ6S: Okay, one more quick comment, then I'll get outa here. Umm In addition to changing out the modulator, the 741 that Chuck originally had, John, we might wanna change out the 7805 he has in there and put in the 317, since we don't we have varactors on these Solfan units we wanna change that line a little bit, too, but keep it available for the time when everyone does have a varactor available. So a couple subtle little things, but we should think about it before we invest money in a board. That's about all I had. WA6BFH, WQ6S.
WA6BFH: Okay, probably what I need to do is get together with you and take a look at yours. And tacking on parts, you or I could probably do these changes with the existing board easily enough, but it might be difficult to lead other people in to easily doing this. Tacking parts on to the board, cutting traces, doing stuff like that. That's the beauty of the circuit board, not having to do stuff like that, but I hear what you're saying, so we aughta get together some time and take a look at your board.
KD6RHD: I got the gist of the time you're talking about building gunnplexors and the IF board and all that. I turned it on for a while, and I didn't catch their overall price. Once you get a manufactured board, is that gonna be in the $30 range or are you talking less than 30 or what? And components, I've got other boards around here I can probably pull components off. I've had experience soldering and desoldering components on PC boards. That's my electronics experience and working with computers. But I just wondered if there was gonna be a special configuration board or did I have to go down to RS and get the PC board, get it to make my own, etching my own board for this project? WA6BFH, KD6RHD.
WA6BFH: The parts we're thinking about putting together in this kit, which would include a commercial-grade etched circuit board, much like you would find in any commercial device that you would buy, along with all the parts to put in that board, the gunnplexor, be around $150 or less. Beyond that, any wizz-bang stuff, any prettiness you applied to that, as far as how fancy a box you decided to buy; literally a metal box, LNB type metal box or cabinet, however fancy that was, and whatever wizz-bang knobs you put on the front panel and stuff like that might add to that cost. But the internal guts of it will probably be in the $150 level or probably less.
KD6RHD: QSL John. Okay, I'll think about it. Money's kinda tight right now with being out of work, but that's a good project. I'm kind of interested in the IF project and the IF board and all, and building the gunnplexor and getting in to the data mode that you're talking about. I really hadn't had much feedback on it, or gotten too much questions on it, other than what you guys were telling about it and maybe I can get back to you later on doing some more thinking on my part, and maybe get in to it a little bit later. I'll say seven-three for tonight and good net, what I caught of it. WA6BFH, KD6RHD clear on your final.
WA6BFH: Somebody else wanna get in there?
WA6BFH: Okay, all right, every good John. Anybody else with any other questions before we wrap things up here? WA6BFH.
WA6BFH: Yeah, go ahead Gody.
AC6UV: Go for it, RM.
AC6RM: Thanks Gody, and thank you John. Learned a lot, good stuff! My preferred method of contact is e-mail. I think you know how to reach me, and I can get to Craig. Anyway, 73. Great evening. I'll talk to you guys later. AC6RM is clear and QRT at this time.
WA6BFH: Okay, very good Frank, seventy-three. And AC6UV, go ahead, WA6BFH.
AC6UV: John, from what I have gleaned here, since I came in late, and I have no access to whatever illustrations and paraphenalias that were distributed through the Internet, what I have gleaned here is that there is a gunnplexor, and/or a Solfan unit. Question regarding the availability of the same item, and how easy is it to modify? I was given the impression according to the article that I have been reading it's more of a mechanical type of job to create all the modification required and, number two, there is a power supply; do you have a schematic of how to build one, and this vector board that you've been talking about, and number three, regarding the IF circuit board, I'd like to see it, and I think after building some of the stuff I have been building lately, I should not have any difficulties putting one of those together. WA6BFH, AC6UV.
WA6BFH: Okay, yeah, all of the questions: Tuning the gunnplexor is a very mechanical type of thing, and believe me, for the ones that we have done so far, it's almost a "monkey-wrench" simple kind of thing. It's not difficult at all, and there'll be lots of hands-on help with that, because it requires the use of a tuning device otherwise known as an absorption wave meter, that just requires a little bit of attention as far as looking it over, to dial these all in on frequency, so that we know where they are. And believe me, they can only be dialed in so close. Believe me, this is not hundred-cycle read-out, this is not KC read-out. But that's okay, that doesn't matter, because what we're gonna be doing is cataloging where everybody's gunnplexor is tuned as closely as we can read it on the wave meter we're using. So, anyway, be lots of hands-on help to do that; you won't be on your own, and kinda like "Here's a gunnplexor! Tune it up!" It's not going to be like that. That would yield bad results for the overall end of our project here. So there'll be lots of help on that. And the gunnplexors are available, basically for FREE! But--you don't get a gunnplexor unless you're committed to the project. And a good indication of commitment to the project is the desire to build the IF board, which will NOT be for free. So anyway, that's the way that works! You don't have to worry about getting one or tuning it up because there'll be lots of help on both those regards.
AC6UV: Some time next week, I'd like to hook up with you. I'm ready to build the power supply modulator, if you have the stuff you can give me. If it's something I can pick up at RS, I can do that. The vector board, I should say, and I certainly would like to look at the IF circuit board and maybe I could do ugly construction on a prototype before we get a real nice finished board, commercial grade type.
WA6BFH: Okay, I don't think you'd wanna bread-board this IF system; not unless you build things a lot better than I do! Not unless you throw together a prototype on vector board a lot better than I do! Cuz, I looked at this thing and said "Yeah, I guess I could do this, but it would take me a month!" to build it right. I'd much rather build something of that level having a pre-punched nicely laid out circuit board, and when I get together with Steve here and look over his, and we've talked about whatever modifications he's recommending there, and see how much modification might be necessary to this, we'll go from there. But ultimately, before too much longer here, we're gonna get everybody together at the local Denny's restaurant or something like that and talk about the finer points of doing this and at that time I will have circuit boards that I can bring along, un-stuffed circuit boards of designs that I have built before that show you what a nice clean FR4 board looks like when ya get it back from the board fabrication house, as well as Steve will have this board that he got from Chuck, which is similar except that he had to do things like drill the holes in it and his isn't solder-masked and does not have the component layout on it. Mine may not have the component layout on it either.
AC6UV: No John, that's about it, and--hey, thanks! I appreciate everybody's participation tonight, and this was very elucidating. So I'll chat with you later! WA6BFH, AC6UV, 73.
WA6BFH: Okay, very good. I think I'm about ready to wrap it up and turn in to a pumpkin or something' myself! Steve, do you have anything else? WQ6S, WA6BFH.
WQ6S: Sure do! I'll second that motion. WA6BFH--Been fun!--WQ6S.
WA6BFH: Okay guys, thanks for participating in the first discussion of the high-speed, wide-band data link ten gigahertz data link project, and for those that have been participating in the net, in this discussion, your presence is duly noted and we'll be getting a hold of you there and letting you know about future frequencies and future get-togethers as far as developing the project further.
Basically, what we've done so far is managed to lay our hands on a few Solfan gunnplexors. Solfan is the company involved, and a gunnplexor is basically a microwave device that is built around a Gunn diode. That's G-U-N-N. I probably should've looked this up for the net, but it might be better explanatory. But a Gunn diode is a diode, much in the sense of any other diode, built around a gallium device, like gallium arcenide, like any of the other high-speed switching devices that are real typically used now-a-days. The only difference is that a Gunn diode has been around since World War II or so, and gunnplexors--a device built to simultaneously (or in a duplex fashion) be able to both transmit and receive a wide-band microwave signal--has been around since about that time. Basically, a gunnplexor is a microwave cavity device built in a piece of microwave <sic> -- a very short piece of microwave
So that's basically what goes on. The hardware that you need is the gunnplexor up at the antenna end that would be--perhaps not on a tower (though a tower wouldn't be a bad idea, if you could arrange that) probably mounted at about roof-top height level; perhaps bolted to the side of a chimney, or something like that, with a dish antenna; maybe 18 or 24 inches in diameter. Figure if you can manage that it would be nice, but probably not too much bigger, because the larger dish antenna makes for pointing problems: the larger the dish, the higher the gain, but that means also the narrower the beam width, so the more accurately you have to point it.
For our purposes, probably an 18 to 24-inch dish is probably gonna be quite desirable, because that's going to get us up in to the 30db gain figure, and provide us a signal that can be pointed at other nodes or other stations in this network without too much criticality of real accurate pointing. We might even get lucky and be able to point a dish at one guy and be able to talk to another; kinda split the difference in between the two. Those are all things that are going to have to be worked out. In fact none of us I think--even in terms of a lot of people that have been using these devices and know a lot more about it than probably anyone here on frequency, in the San Bernardino Microwave Society and what they've been able to tell us--this is a new kind of application; this'll be a different kind of application.
Gunnplexor communications have been used over hundreds-of-mile paths. Gunnplexors have been used for DX communication over, basically, 200-mile paths. Ten-milliwatt gunnplexors in to about a two-foot diameter dish. So, this is not exotic communication in THAT regard. It's just putting the dynamics of the system together, and making it a real usable system. That's the kind of thing that's a little bit different than has been done, pretty typically, before. Other groups, I understand--and I don't have the details on this--other groups, like TAPR (the Tucson Amateur Packet Radio people) are doing similar things. Don't know the details on the whys and wherefores of what they're doing, but we're certainly going to be talking to them and finding out what they're doing so as not to reinvent the wheel, but these kinds of things are in the works and this is something that is pretty new to ham radio; it's a pretty "cutting-edge" kinda thing to do.
So, that's what we're up to. Another part of the project is to build the IF--or intermediate frequency--portion of the receiver section, which we have circuit board artwork virtually in hand and available to us from Chuck Swedbaum<spell>? WA6EXE, a very talented individual who has been involved in microwave for something between 40 and 50 years. A really nice guy and good person to know for such sources of pursuit. So that's coming our way and we'll also be talking about building that IF system as a kit. Basically that IF would live in your ham shack along side your other radios; whatever else--the gunnplexor and perhaps a small dish antenna--would live up on the side of the chimney, on a tower or on a mast or something like that.
Anyway, I think I've painted a general picture, Fred. Let me turn it over to you to talk about the potential for providing, down the line, gunnplexors for those people who would commit to building this IF system. Basically, my thoughts on the way this would probably work--and you can correct me on this--my idea would be that anyone who commits to building the IF system, may well be able to see down the line a gunnplexor, if that would work out. KF6HQC, and the group here, WA6BFH.
Other than that, I'm not really interested in the digital aspect, but I am willing to participate in this experiment with the digital link. It's intriguing. The speeds that you can get over this link are pretty high, and such that you can transmit video and voice at the same time, it's pretty crazy. Other than that I don't really know what else there is to say, other than getting these gunnplexors tuned up. That's another thing that needs to be done. Once this gunnplexor is in your hands it's going to need to be tuned up and there's a few people that have the wave guide meters, and hopefully next weekend, I'll be the owner of one too. So if that does come about, once I get my hands on four or five gunnplexors I can tune them up and they'd probably be ready to go.
John, I'm not sure if there's anything else you wanted me to comment on, so I'll hand it back to you and we'll go from there. KF6HQC over to WA6BFH.
Fred did you want to make a comment? KF6HQC, WA6BFH.
Let me pass this next to WQ6S. Steve, what I'd like you to do is--within 50,000 words or less, kind of--paint a thumb-nail sketch of the IF system that you're working on; the board that Chuck Swedbaum
Ah, yeah -- lemme back up a little bit, John. An 807--you're an old-timer--IS a bottle of beer! For the new guys that don't know, the old timers used to call a bottle of beer an 807. That goes back a few comments.
Anyway, I think it would be premature to say that I am really gung-ho about a high speed digital link. I want to learn about it, I want to learn how to do it, but I think back to my ham career; I started out as a Novice, got a General, got an Extra, and the small steps, the small successes were kinda good for me, so what I plan to do is learn the audio mode first--wide band, go to narrow band, and then get hot'n'heavy in to the digital stuff, which I know nothing about. If I can do that concurrently, it'll be fun. Lookin' forward to it!
The other thing I was going to mention was that G. B. Gunn came up with his famous diode in about 1963, I think, not WWII time.
But to the IF board: physically it's about three and one half inches by six inches, it probably has, I dunno: 125, 150 parts on it. Not REAL complicated. What I did was I got the board from Chuck, bought the parts from various sources and, trying to thumb-nail it, I probably have about $120 in to it right now as it sits, and I'm not done yet. But that's buying all new parts, with the exception of a few minor things that I had sitting around. Some of the major things I still need to acquire are an enclosure and the "bell'n'wistle" part of this board, which make it so nice, the meters, etc.
Putting it together is basically a matter of soldering the pieces on, and winding a few--eight--torroid coils, which wasn't my favorite thing to do, but was a good experience. Time-wise: I'm slow, so I've got HOURS invested, but they've been enjoyable hours. I dunno what else to say . . .
ARR Company offers a board very similar board this ...
<timer time-out; it appears there's malicious interference>
But anyway, my thought is that a number of circuit boards could be made that would be of commercial grade, commercial quality, all drilled, all trimmed, all ready to go, through-hole plated, fiberglass material boards; what is generally called in the industry: FR4, flame-retardent-4 circuit boards, probably with solder masks, possibly tinned over bare copper--either way, and I'll do what ever I have to do to keep the price in range--but I'm thinking that the circuit boards could be provided for thirty dollars a piece or less. Trying to get the ultimate price on the whole thing down to the area where $150 would be top-dollar money, and maybe less than that. That might mean that you'd invest a little bit more than that to make it a real pretty box if you wanted to have real custom-shiny knobs on it and a real pretty looking cabinet enclosure for the ham-shack end of the arrangement, but between the IF system and the gunnplexor I think that a ball-park of not more than $150 is probably a very doable kind of thing, and quite likely it's less than that. So that may yield some information for those who are wondering what that end of the project might get in to.
Lemme see if some other people have come on frequency and then I wanna turn it over towards Don here, and talk a little bit about the digital scheme that he has proposed. He has some information up on the Tech Bench Elmer's web page that you may want to look in to. That's available on a hypertext link; if you pull up the web page, you will find a page there for a ten-gigahertz data link; something in the front page that you can click on. When you get to that page it'll say "Ethernet Interface for RF Data Link." I'll ask Don to talk about that in a little bit.
So lemme see if some of the other people might be on frequency that we called for before:
WN6EKZ, Alvin are you around?
<no response>
How about KE6OLH?
<no response>
AC6UV?
<no response>
Okay, lemme go to Don. Don, give us an overview of your idea on this, and the adaptability, and also a little bit broader thumb-nail sketch of perhaps using Ethernet hardware; what Ethernet is and adapting that from its coaxial cable origins--as far as I understand Ethernet--to something that would be useable, or adaptable for this RF point-to-point kinda data system. W6QS, WA6BFH.
The way that they tell that there's more than one [card] trying to talk at once is they have what is called collision detection. In addition to a square wave--I'll get in to the encoding a little bit further down--they put a DC current in to the coax. If it sees more than one of those DC levels together you end up getting past a magic threshold of about 1.75 volts negative; if it gets past that then they [the cards] both determine that they've had a collision and back off for a pseudo-random time that each has in its algorithm. Then they go through and try again, unless someone else [another card] has beaten them to the draw! Very very similar to packet radio--except it's all going through a wire. What were looking for here is for it to NOT go through a wire, but to bounce it through the air. Once you go ahead and have something going at this sort of rate they have a little fancier modulation schemes that'll push it all the way up to one hundred megabit per second. But the basic simple cards that are available at swap-meets in the $20-$25 dollar range, you end up picking up a very nice modulator-demodulator setup--a modem--capable of ten megabit per second with all its threshholding, all the buffering and--more importantly--a whole lot of your software already available on your standard platforms. Under Windows(tm) you can get some TCP/IP software, or Windows95(tm) it already comes in it. You can go through and run Linux; same type of thing: you've got a whole lot of software already there. So that's not going to be as much of an obstacle towards getting what we're looking at doing!
A starting type of thing would be just doing some point-to-point, and then--if it were possible to get it going--possibly getting a high site and set up something akin to a repeat node or that type of operation; [that] type of operation will allow a little bit more of a wide-coverage, wider star-type of connection instead of just straight point-to-point.
The way it basically sends signals down the line is called "Manchester encoding." What this basically is is: at ten megabit per second, you've got a hundred nano-second period set aside for each bit that's coming out, and it's set that the transition in the middle of the bit period, at the fifty nano-second point of the bit period, if it goes in the "one" direction--from a zero to a one-- you've got a "one" bit, [and conversely] if it goes from a "one" to a "zero" you've encoded a "zero" bit. What it basically looks at is what ever it was at the end of the last period; if it just got done sending a "one" it will be high, if you're trying to send another one, at the zero-nano-second point it goes low and fifty nano-seconds later it goes high, so it has that rising edge that says that you are sending a "one" bit. And vise-versa; if you're sending a "zero" after sending a "one," there's no transition until it gets to the fifty-nano-second point, and then it drops.
So you have your basic signaling that's sitting in there for a modem pretty much in a pretty cheap card that already has a whole lot of the AC threshholding built in to go through and pick up the threshold point, the zero-cross point, on the AC wave form; and all that already built for you at a very modest [price].
The only other big piece of it is how to interface the coax out to the gunnplexor, and I'll come back to that in a moment. John, did that go through and fill it, or did you need some further elaboration on that? WA6BFH, W6QS.
I wonder if there's anybody in our group now that has any questions about other things about how to do this or what we have in mind, or things that have come up in the discussion that I haven't thought of to elaborate on, that might want to ask a question at this point in time. So listening for call-signs for questions in that area. WA6BFH standing by.
<AC6RM, KE6VHM, HQC>
My question is on the setting up of a star-type system, as opposed to setting up a link, and my vision of this is basically; if somebody is going to be a mid-point then they're essentially gonna have two setups: one at ten-two-fifty and another at ten-two-eighty. Basically, you'd end up joining a bunch of people that way. That means, from what I can see, two IFs, maybe--I don't know. Maybe possibly somehow the information that's received, is not only received by the IF board and then, thus, the Ethernet decoding hardware, but it's also automatically passed on to the transverter
Anyway, I'm gonna refer the question on the whys and wherefores of getting that net linking system, the data part of it, to some other people, because I'm kind of curious about the best way to do that myself.
Anyway, KE6VHM also had a question there. Frank, what was your question? Go ahead.
But anything like that. It's gotta be a solid-surface dish; not anything made out of welded rods or anything like that.
Lemme see ... Fred, is there anything I may not have touched upon, or may not have touched upon clearly enough, and I think you also had a question there. KF6HQC, WA6BFH.
It's also possible to build horns like that if you're in to the idea of sweat-soldering brass shim stock together. If you made a horn antenna, which is kind of a rectangular, funnel-shaped device that would be about, maybe, five inches by eight inches at it's open aperture end, and tapered down to the same standard flange size; something on the order of 30 or 40 decibels of gain would be available that way too.
Also, for the people who might be thinking about different antennas, I've come by a ten-foot Andrew dish, very nice communication dish, that cost me one third or one fourth of the rental of a Ryder rental stake truck, so I think I got that dish for about thirty bucks, or maybe less. I've gotten a four-foot dish, a five-foot dish, and a two-foot dish for free, for just going to pick them up. Just saying it's here, you wannit, come get it. Definitely keep your eyes open for those sorts of things as they come available as well. As I say, it MAY not be needed, but wouldn't be a bad idea; something on the order of an eighteen-inch dish would be a nice thing to have.
All right, let's see if there are any other general questions on the subject as we've talked about it so far. WA6BFH listening for call-signs.
<WQ6S>
On the error correction, I don't believe they've got ... well, I know they've got error detection; they've got CRCs on them ... but I don't believe they've got any forward corrections on them (the simple Ethernet cards, 'cuz they are considering they are running down a reasonable channel). But, if you're running something like TCP/IP over it, there's lots of retry logic in there: they'll go through and send a frame, send a couple frames ahead, and if they don't hear acknowledgment of it, they'll go ahead and resend it, 'cuz they're used to lots of frames being tossed or lost here and there. For something like this, we might set the retry rate fairly high, just to get through, according to how much traffic there is.
I think those were the two questions you had there. WQ6S, W6QS.
Lemme see if there are any other general questions about anything we've talked about so far here. Listening for call-signs at this time. WA6BFH.
<KE6VHM, "one more, John">
All right, anybody else with any questions on anything we've talked about relative to this project. WA6BFH listening for call-signs.
<ONZ>
The other one, as far as the duplex'n'all, you could do some of that a couple of different ways: you could go ahead and put a couple of cards in it and configure your system for multiple cards, and have one transmitting and one receiving, if you want to do that. You only basically get about two times the data rate, which is going to take a while to use up, in exchange for a whole lot of extra hardware. I don't know if that would be that advisable. A minimum packet length, like 512 bits in a packet, short little packet, is on the order of 25 micro-seconds to get in and out with. So you're fairly fast in the process of doing it. In fact that's one of the main modifiers to straight Ethernet we're going to be doing, is normally any collision is picked up by all transmitters on it, so that they know that they are fine. That's one of the things having to do with the length of it, so that they know they've had a collision. That's going to limit you to like a five-mile distance, or something like that--speed of light type of thing. For our set-up here, you'll have to live with a few more lost packets in there; just go ahead and do the retry. If you know how many packets get retried when you're going across Internet trying to pull something out of England, it's a whole lot more loss than I think we're going to get on a link like this.
The other one that you have in there, as far as whether or not you have to transmit simultaneously, if you do go through point-to-point-to-point, you can either retransmit the signal as you received it directly, or you can do more like a digipeater, where you actually pull it in and retransmit it. That's basically the function of a router that you could run on the PC so that it would know where to send it back out. WA6BFH, W6QS.
<AC6RM>
Let's see, what was the other one: I forgot! Oh, you were going to mention, just briefly if you could, the practical ... how we would go about connecting these Ethernet cards to the IF and Gunn. Not in so complicated terms, just briefly, so I have some sort of comfort that it's a reasonable project to undertake in my garage, and not something that has to be done in a white sterilized room with laser-precision tools, or whatever.
Anyway, those are my two questions. WA6BFH, Discussion Control, this is AC6RM.
Don, lemme field that question back to you, but also let me toss in the idea that this is all Furrier stuff, right? Lemme ask a question of Steve first, and then I'll frame a question that will add to Martin's. Steve, WQ6S, what's the basic band width that your IF board is set up for now, from the design you got from Chuck?
The other one, as far as interfacing the coax with the gunnplexor, to separate the two direction on the coax, it's a pair of op-amps and six or eight resistors; that's what John was saying was up on the Tech Bench Elmer's [web page]. And then, basically, another high speed op amp to run your modulation signal in, and whatever you need to pull the IF down in the other direction. WA6BFH, W6QS ... did that go through and explain it reasonably?
<HQC, AC6RM>
So anyway, I don't think we want to get in to a situation where we are throwing RF away by saying "well, let's widen the band pass on the receiver." Let's leave the band-pass at the kind of level that Steve has built his system for right now, and find out how far we can get away from each other and just TALK. Just carry simple analog voice communication. And then start doing the data stuff, and find out ... if we've got good reasonable signals, if we can present some reasonable paths over that kind of bandwidth, over that relatively narrow bandwidth, which is already broader than some, from there we can go "Well, okay ...", scratch our heads a little bit and maybe do something else. But I think that would be the way to attack the problem in due course.
All right! Lemme go down the list here and find out how many people that are taking part in this discussion right now have Internet capabilities, so that we can exchange some information on a little bit more private basis.
KE6VHM, how about you? WA6BFH.
Anyway, anybody have anything to toss in to the pot at this point, before we start thinking about wrapping this up. WA6BFH listening for call-signs.
<AC6UV>
Okay, anybody with any questions, comments, etc. on the project. WA6BFH listening for call-signs.
<WQ6S> <W6QS just checking out, 73's>
Anyway, but I'm glad you're only talking about changing the modulator, cuz, yeah ... I figured we'd wanna start ahead of the game anyway! WQ6S, WA6BFH.
<BREAK for a call ... not transcribed ... John talks about the gunnplexors used on the 91 freeway in/out tracking gates>
<KD6RHD have a comment, but I'll wait for WQ6S>
All right, RHD you had a question, why don't you go ahead. RHD, WA6BFH.
<No response>
<AC6UV, with a question, John.>
<AC6RM ... I gotta check out>
<WA6BFH Seventy Three!>
<KE6VHM: I think I'm gonna hit the hay too, John! It was pretty neat! We'll talk atcha tomorrow. Seven-threes to all that's left. KE6VHM clear.>
What else did you ask about?
So the catch here, the little carrot here, is the gunnplexor! I was worried about that! There is one for free if anyone is willing to build--I certainly am! Very very good sounding project, and I also could get my hands on one of those dishes; I have a friend who is willing to part company with one. So, yep, good deal. WA6BFH, AC6UV.
But anyway, we'll talk about all that stuff. It'll be easy to do. It'll not be a cumbersome project, or I would not get in to it. If there were an excess amount of hand-holding that would be necessary, I wouldn't get involved in doing this. As far as stuffing parts in this board and building it, with a little bit of care and attention to it, and the help of friends who have done these kinds of things before that won't be a problem. Same with the gunnplexor, etc. etc. As far as the dish antennas and stuff like that, you're kind of on your own -- find yer own dish! Everybody's gonna be out hunting for dishes!
Anyway, any more questions on that? AC6UV, WA6BFH.
Seventy-three, I'm gonna be clear of the frequency at about 11:09 or 07:09 Greenwich; we'll catch you later on! WA6BFH.
Return to Microwave Data Project page
Tech Bench Elmers![[beta.gif]](beta.gif)
© 1996-98