home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Cream of the Crop 1
/
Cream of the Crop 1.iso
/
EDUCATE
/
TN210.ARJ
/
NETWORK.EXE
/
NET_14.TXT
< prev
next >
Wrap
Text File
|
1992-06-28
|
8KB
|
131 lines
NET_14.TXT
HF NETWORKING AND PRACTICES
---------------------------
Casual observations of HF networking reveals it's even more vulnerable to poor
operating practices than are the VHF systems. Somewhere along the line it
appears packet has been vastly oversold as to what can and what cannot be
accomplished. Due to the longer amount of time it takes to pass a given amount
of data at 300 baud, as compared to 1200 baud, transgressions become painfully
obvious.
We have often succesfully passed packets on HF for an hour or more at signal
levels approaching the noise with scarcely a retry. In every case this occured
when there was only ONE SET of stations on frequency. The instant another
packet station came on, the retry rate went up and throughput dropped. If the
new station's signals were very much stronger, we typically would retry out and
become disconnected.
From this it's concluded marginal signals on a MULTI-USER ACCESSED HF CHANNEL
have little chance of successfully passing much data. Even though packet
is a multiple user access mode, there are very definate limitations as to the
number of users a channel can accomadate.
Many packet operators believe it's okay to zero in on an active HF packet
channel. They connect to other stations and go about their business quite
oblivious to the damage created to the original on-channel stations. While
communications can be achieved on a lightly loaded HF channel, in general,
throughput between one or more sets of stations will suffer. This is true even
if all stations have good signal strengths.
When conducting direct on-channel QSO's on a node\gateway frequency it's proper
operating practice once contact is made to move off the channel. This practice
expedites not only YOUR communications, but those through the gateway as well.
By the same token, operating MULTIPLE STREAMS on an active channel can cause
serious disruptions to other conversations.
Maximum throughput occurs between two HF forwarding BBSes when THEY ALONE
occupy the channel and have signal strengths somewhat above the noise level.
In this regard, packet is little different from having multiple voice, CW, or
RTTY stations attempting to pass on-channel traffic at the same time. In many
instances, channel discipline has broken down. Multiple stations attempt to
forward at the same time. Some BBS forwarding stations use incredibly long
TXDs on the order of two to three seconds. This helps them to "capture" the
channel, but at the cost of vastly reduced throughput for everyone concerned.
A TXD of 100 milliseconds is typical for quiet channel operation with modern
radios.
Where possible, network managers should assign forwarding stations time blocks
tailored to the best propagation periods between partners. The practice of
allowing BBS/nodes to send beacons and NODES broadcasts on forwarding channels
should be discouraged as it hinders throughput.
Besides using a QRM-free frequency, successful forwarding operations requires
thought be given to propagation paths and station equipment. Forwarding
partners should be geographically established to take advantage of optimum
skip conditions for the band(s) involved. A helpful aid is the MINIPROP
program. Inputting HF station locations and current sunspot data into
MINIPROP will easily reveal optimum distances and frequency bands for this
purpose.
HF transceivers should have sufficient frequency stability. Excessive retries
WILL occur if both partners aren't netted to each other within 30 Hz. Crystal
control will solve many stability problems. A 500 Hz IF filter is mandatory
for optimum performance. TNCs will give superior results if equipped with the
HF DCD modification. Nearly any TNC will work equally well on HF if it has the
DCD mods and is preceeded by a 500 Hz IF filter. It's desirable to use a gain
antenna, preferably a beam, for the higher frequencies. Antenna height
should be matched to the forwarding partner's optimum take-off angle (MINIPROP
displays take-off angles).
Even modern high priced digital radios may have inaccurate dial calibration. A
simple way to check calibration accuracy uses a frequency counter connected to
receiver audio. In one of the SSB modes, tune to a dial setting 1 KHz off
an UNMODULATED carrier known to be accurate (WWV, service monitor, etc.). The
counter will read 1.0000 KHz if the radio's reference oscillator is correct.
While the lower HF bands have borne the brunt of forwarding, the ten meter band
has been under utilized. 1200 baud operation is authorized above 28 MHz. Yet
very little forwarding activity has taken place. Around 28.105 MHz, 300 baud
operation is popular. At 28.190 MHz, 1200 baud forwarding occurs. Just as on
lower frequencies, the "herd instinct" prevails, thus overloading the channel
at times.
A BBS-free node/gateway ragchew 1200 baud channel exists at 28.195. From time
to time, BBS SysOps, TCP/IPers, or DXCluster Ops insist on uninvitedly setting
up camp on the channel. While everyone has a "legal right" to operate on any
frequency within the amateur bands, the standards of courtesy are no different
with packet then they are for any other mode. Even on the "dog-eat-dog" 20
meter band, one often hears a voice politely asking: "Is this frequency in
use?" Packet operators, especially those with servers capable of causing
considerable QRM, should enquire whether their activity is wanted prior to
establishing operation on ANY frequency.
Compared to the lower frequencies, 10 meters has a vast amount of room for
packet operation. With proper antennas, and attention paid to propagation
paths, inexpensive low power converted CB radios at 1200 baud will move an
impressive amount of data. Considerable presssure on the lower bands could
have been eased had more effort been made to coordinate individual forwarding
partners on 10 meters. This band would make an ideal 1200 baud wide area
network (WAN) for TCP/IP experimentation.
A drawback with 10 meters is that the band is open only during a few years
corresponding to the peak of the 7 year sunspot cycle, and then, mostly during
daylight hours. However, during this period great propagation exists that
could be used to good advantage.
A significant chunk of the digital band is taken up with unattended CW beacons.
When the band is open, these tend to cause harmfull interference to the other
authorized modes. An interesting propagation experiment would be to replace
the existing CW beacons with packet nodes. Initially the nodes could operate
on one frequency and be widely spaced over the globe. TheNet's dynamic routing
will automatically list other nodes when propagation allows their broadcasts
to be heard. Node routes listings would be available for inspection by local
packet stations at any time.
An advantage of using packet nodes vice CW beacons here is that propagation
info will remain in the routes list for several hours after the condition goes
away. Those interested in recording propagation conditions need only to check
the routes list on their local node from time to time. Otherwise, with CW
beacons, one has to monitor or tape the channel constantly. In addition to
propagation info, this concept would allow networking channels when conditions
allowed. If and when automatic unattended packet operation is allowed, these
nodes could be gatewayed into existing VHF/UHF packet networks.
To summarize, should this proposal be adopted, 200 KHz or so of CW beacons
would be phased out in favor of packet nodes on perhaps, only 1 - 5 descrete
channels. Such implimentation would conserve frequency resources, and possibly
prove (or disprove) the theorem that 10 meters is open to someplace in the
world at all times.