[FADCA] 9.6KB Throughput vs 1.2KB

[bud thompson]

----- Original Message ----- 
From: "Chuck Hast" <wchast at gmail.com>
To: "Russell Oder" <oderr at bellsouth.net>
Cc: "Florida Amateur Digital Communication Association" 
<fadca at mailman.qth.net>
Sent: Sunday, November 21, 2004 22:56
Subject: Re: [FADCA] 9.6KB Throughput vs 1.2KB


<snip>

> The MSR2K and MSF5K should do data with few issues, Motorola designed
> them to do so. I am not sure where the data points are on them, but we 
> have
> a lot of people who know them and should be able to point the way.
>
> Mitreks do pretty well, if you can replace the filters they do even 
> better. With
> some additional work you can make them transisiton from TX to RX quite 
> fast,
> it has been a while since I have looked at that issue but I believe it
> is published
> one the web somewhere.
>
>

Chuck - Thanks for the details  in another note... our subjective testing 
(timing the transfer of a 10K file with no noise - just signal - bench 
tests) between two Kenwood Rice radios resulted in 9.6K being about 1.8X as 
fast as 1200b with the same radios. Putting a single vhf Mitrek converted 
for 9.6kb in the mix with each Kenwood  did not improve things a lot.. Our 
next test will be Mitrek-to-Mitrek on 1200b and 9.6kb which we expect to do 
better, but no way will that be 5x as fast.. I don't think we have 
experienced that on our Mitrek to Mitrek 9.6kb UHF backbones we've had in 
service for almost two years.  We'll sure be anxious to report anything like 
that!  8X is probably not attainable w/o a lot of effort.

We do replace four of the monolithic filters in the Mitreks with Digikey 
filters.  If the fifth filter is in the stock Mitrek (some are some are 
not), we remove it and jumper across.  We have detailed destrucitons written 
up for making  9.6kb mods, including two or three different things to do 
such as cleaning up the low frequency response coming out of the detector 
and biasing the TX data feed point, etc.

By the way, these last tests reported above were made w/o doing any fine 
tuning of the IFs after replacing the filters! We'll do that next.  Our 
experience is that fine tuning has, perhaps, not been needed - and we often 
wonder how much time we should spend fiddling with that.

We are now working on dual speed VHF Mitrek for LANS - users would be able 
to access on 1200 or 9.6kb.  We have successfully used a KPC9612 for initial 
testing, but have not yet subjected the single radio to use by both 1200 and 
9.6kb transfers at the same time (sharing the channel of course, not TX /RX 
both at the same transmission.).  Unless/until we can get FPAC to support 
the KPC9612 we'll not be very interested in pursuing this. Making this work 
with two different TNCs is a lot more effort for impedance/level matching - 
I've not found any documentation on that.

I've not run across any special mods for speeding up the Mitrek TR switching 
times... but faintly remember a mod to keep the RX oscillator running 100% 
duty cycle so that RX startup time would be reduced.... All that effort to 
optimize TR switching times is very important on dedicated Mitrek to Mitrek 
backbones - but won't help much where there are other several different 
breed of user radios on a LAN.  (If we learn how to speed up the Mitreks 
will do that no matter, but the gain would be mostly on the backbones.)

bud