2502 West Colorado Ave., Suite 203 · Colorado Springs, CO 80904 ·
Voice 719.636.2040 · Fax 719.528.5869 ·Wireless Web http://wireless.oldcolo.com
Progress Report 8
December 23d, 1996
Much has transpired since our last 'narrative' progress report in August. (we have posted here a number of Special Reports since then - among them reports on School District 20, Mongolian Wireless Project, and the Evans Elementary School project.
This report will summarize where we are in each of the original (Colorado Springs and San Luis Valley) installation and operations projects, point to the significant features of the studies we have made of already-installed systems which are seperately reported on here, and point to where we are going in early 1997.
Of course, overall, Wireless in the US is in a very volitile state. Technology itself - from academic research on the potential of process-gain digital wireless, new, or improved, radios - major Regulatory issues and debates, some of them issuing from the 1996 Telecommunications Act, foreign interest and products, to rising public and press awareness of the potential of no-licence data radios, all have kept us jumping. We have been diverted to a number of important conferences, technical, educational, and public policy oriented. The funded extension of our Project to August 1997, with its concommitant requirement that we analyze wireless networks already in place, has expanded our duties. And we have even had to divert our attention and efforts for almost two months to setting up a wireless network in Mongolia!
So we are here not only reporting on our own NSF project progress, but the progress of many others! Whose experience sheds even more light on the role wireless can play in education.
First, progress on our original NSF sites, three in the San Luis Valley and two in Colorado Springs.
MONTE VISTA MIDDLE SCHOOL
This 115Kbps spread spectrum, no-licence (FCC Part 15 rules), link using Free Wave radios, 17 miles between the Monte Vista Middle School - its NT server and 25 486 workstation classroom - and Rocky Mountain Internet's POP in Alamosa, has operated reliably from the time it was first set up in May 1996.
That is, the radio link has been reliable throughout. When it has gone down, it is because of either weather related problems with the 30 foot roof omni antenna, or improper mounting of the modified radio in its metal weatherproof case, or human errors at the router.
During the summertime, when school was out, the radio, which had been mounted in a non-standard way up on the antenna mast, in a weatherproof box in order to shorten the length of the RF cable between the radio and the antenna, was further modified by running led status-light wires down the RS232 cable to permit 'seeing' the status of the radio from its LED, as well as resetting the radio for reconfiguration, remotely.
All this, of course, added new connections, new wires, and new ways for things to go wrong. Which they did, as movement of the radio-board inside the metal weather-proof box by wind caused several shorts and gaps. Each time requiring us to bring down the 30 foot telescoping antenna, a two man job, removing the radio, and testing and reworking it.
So there is a labor 'trade-off' between mounting a radio high on the antenna mast to gain gain by a shorter antenna cable, and keeping it inside the building, requiring much longer RF cable line lengths, and suffering from line loss. Which, as the very low power levels .6 to 1 watt, can be the difference between a solid connection and a marginal one, at long ranges between radios.
But by the end of the summer, the 17 mile 115Kbps connection had been up for nearly a year, and performed well.
Center School is 22 miles (direct) to Alamosa. A number of attempts to connect it up, both by direct connection to the Alamosa, Rocky Mountain Internet POP and radio, and by relaying through the FreeWave Radio at Monte Vista failed.
Even though the San Luis Valley floor is pretty flat, it became clear that the rise of the ground between Center, and both Alamosa and Monte Vista, coupled with mature trees and stretched 'distance' (22 and 11 miles) caused just too much attenuation of the signal. Both at .6 and 1 watt of power.
Another factor bearing on the specific design of radios became evident in the tests. Because the Free Wave radio - although they are point to multipoint and can be individually set up as a relay radio - do have 'IP' intelligence (routing) in their firmware. This meant that, if the Monte Vista radio, as a 'link' between Center and Alamosa radios, was set up to 'relay', then all the packets destined for either Monte Vista or Center would all go to Center. For there was no way for the Monte Vista radio to 'know' that IP packets in the 188.8.131.52 to .254 should go to that router, as differentiated from 184.108.40.206 to .254.
By making the Monte Vista 'middle' radio, Multipoint Master, and both Alamosa and Center as Multipoint Slaves, it could handle the problem - but then reduce the Alamosa radio to connecting to a single radio - Monte Vista.
This illustrated clearly why future radios - or modified firmware in existing radios, if they are to be used flexibly as Internet links, need routing capabilities built into each radio.
By the time these series of tests, radio configurations, and high antenna placements had been made, the Center School District, which had recieved a large windfall from changed Colorado State school financing, was fully engaged in setting up, not only an extensive internal network, with workstations, it also could afford two 56Kbs Internet wired links to Colorado Supernet at its Adams State POP in Alamosa. One of these is for a specialized NovaNet programmed instruction link, the other more general.
As a consequence of these findings - which is what we needed to learn - I decided to stop pursuing a redundant (with Monte Vista) radio-internet link by wireless to Center, and instead make a test which we had not originally planned to do, a set of home-to-school wireless links in the small town of Center. To test what it takes to link either students or teachers to the servers at their school by a 'loaned' wireless device, connected to the teacher or student's home computer.
A report on the rationale and progress of this new test is reported on seperately. It will be done between January and May of 1997.
CENTENNIAL SCHOOL DISTRICT - SAN LUIS
The town of San Luis is 30 direct and 41 road miles from Alamosa. The school and the town also are hidden in a valley several hundred feet below a ridge that lies between it and the valley to the west, including Alamosa. There is no possibility of a single line-of-sight radio arrangement.
It is the most difficult of all the sites to reach. But in common with many a rural small town and school district, its size - which translates to need for bandwidth - is less also. The entire school has only 300+ students, and one old Novell network linking 286 computers in one room. Thus, even with the upgrade of PCs and network (NT server) planned by the school district in early 1997, we calculate that a 56Kbs Internet connection should be sufficient for quite a while.
Since a one-radio relay site usually halves the original bandwidth, if we used a Free Wave 115Kbs radio at the Base site, then a one-radio relay at some intermediate site, we can deliver 56Kbs to the radio at the school itself.
The task was to find that intermediate site that could directly see Alamosa and the School in San Luis. In the original proposal I considered putting a solar-rechargable battery-powered radio on the lower slopes of Mt Blanca - knowing its towering 14,000 foot peaks were visible from every point in the valley. But that would have posed some tricky problems - finding a site with permission from the owner - either the Forbes Baca land people if low on the south side of Blanca, or the Forest Service if higher up. And with vehicle access - at least within a reasonable hike. Then the construction, then the servicing - and finally taking into account the deep snow possible on the mountain - with ice on the yagi antennas, which can attenuate the signal. The legs of the relay would be 25 miles and about 20 miles - which is at the limits of the radio sensistivity.
Challenging, but not trivial, and hardly the thing that would be done every day by a school district.
Then there was the possibility of shooting from Alamosa to the Sangre de Cristo range east (behind) the school from an altitude that could see both. But that would make one leg 35 miles - a real stretch for 1 watt of power.
Examination of the ridge to the east of the town revealed several things.
First, there is a shelf about three quarters the way up the ridge, upon which there were several antennas, and a new one under construction. The new one was being built by Blanca Telephone company, to serve as a cellular phone system for San Luis and the valley surrounding it. It was 180 feet high.
Then besides a television relay antenna there was a tower about 120 feet high that belongs to Costilla County, is used by the Sheriffs Office for voice radio, and by the Centennial School District for its bus-route radios. So accessibility should not be a problem. But it was not clear at all from preliminary driving around the top of the ridge, that the top of the antenna could 'see' Alamosa.
A binocular's survey of the ridge found a stretch from which Alamosa and its water towers above the mature city trees, was clearly - though distantly - visible - and line of sight. The distance was a full 30 miles, however, so the question was - could the radio signal penetrate the trees in Alamosa at that distance?
Dewayne Hendricks, rigged an automobile cigarette lighter 12 volt supply, to power one of the 12 volt Free Wave radios with a hand-held Yagi antenna and configured the radio in Alamosa, with its 30 foot mast, to connect with the mobile radio.
He was able, from the top of the ridge to get a strong-enough signal to assure us we could reach one or the other of the tallest antenna towers on the lower 'shelf' - which were themselves short line of sight to the mast on Centennial School in San Luis. Less than a mile away.
After unsuccesfully negotiating with Blanca Telephone, to put, for the purposes of our NSF test, a 3lb antenna somewhere at least 150 feet high on their cellular tower (they began to think that we were going to 'compete' with their long range plans to be Internet providers themselves, using their phone circuits), we abandoned that approach, and concentrated on the County, lower tower.
As of this writing we are getting a Survey company to give us a precise height that lower tower has to reach in order to see over the ridge to Alamosa. A local tower company, SpectraCom has assured us they can extend the tower another 30 or more feet. We are engaged in contracting for that to be done, as well as place the yagi on the top of the tower, and adjust it while we test the link from the 'radio shack' below, linked by RF cable up the mast.
Meanwhile, at the school, and in San Luis itself, we are engaged in a parallel, but eventually to be connected, 'community' wireless network, that is a seperate funded (Colorado Advanced Technology Institute - CATI) project. And we await the installation of the NT server, LAN network, and new 486, Window 95 workstations at the school, now 8 months late, before we attempt to compelte the links. For there will be nothing for the radios to talk to at the school, until that network is installed, this spring.