[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

linrad =running

Hi Jurgen,

I do not yet have Linrad running. I am building the
front end now and planning to use an RX2500 as well.
We will both be converting from 144 to 2.5 Mhz.

The local oscillator is a comprimise that allows
frequency agility and reasonable phase performance.
A CB radio PLL operating at 25 Mhz is mixed with
a crystal oscillator to produce 141.55 Mhz.
This gives a 2.5 Mhz IF. The mixer is built and
uses a TUF-1h. This is a comprimise as well. Ideally
it should be about 6db stronger. The LO will drive each
mixer input with +20dbm power.
I plan to start with an ATF33143 running at 23db gain
and .2db noise figure. This will drive a low loss (.2db)
dual LC network to provide some protection against FM BCB
and commercial two-way. I am currently building this filter and
getting unloaded Q's of over 500 with #10 wire (painted with
krylon clear spray paint) and common rectangular ceramic silver plated
6 plate screw driver adjusted 20 pf trimmers. Coil diameter
is 1" and 2 turns. With only .2db loss, I am getting -30db
at 100 Mhz and -13db at 158 Mhz. The 3 db bandwidth is 12 Mhz.

The second stage is to be 3 x J310's
in parrallel, much as Leif is using in his circuits. The cost of these
devices are very low. Each J310 stage will have a single LC network
on the input and provide some small additional selectivity.  The
gain will be minimal, perhaps 8db using noiseless feedback
and a 2db noise figure. All of this will be located at the
array. Stage two will drive a 2db feedline loss. In the shack,
Stage three will be another 3 x J310, roughly 8db gain, 2db
nf pre-amp. This will drive a 6 pole helical filter robbed
from a Motorla Micor. Its loss is about 4db, but provides
-50db at the image frequency (139 Mhz). Stage 3 will be similar
to stage 2 and drive a second 6 pole helical filter similar
to the first. At this point, image rejection is over 100db.
Stage 4 will be similar to the other two and only running
6db gain and limited to +26dbm power output. A 6db attenuator
will follow to drive the mixer with a maximum +20db power level.
This is the safe maximum for the mixer.

Overall front end gain is around +34db, so there is no trouble
driving the 14db noise figure produced by the TUF-1h, RX2500,
Delta 44 (running maximum gain) combination. If a simple
pre-amp were used, only 32 db gain is needed ahead of the mixer.
The comprimise in gain is made due to the need for more selectivity.
There is a slight nf degradation by cascading more stages,
each sporting a 2db nf, so the gain is set a few db higher
to accomdate. Multiple stages are used to keep any one
stage from saturating on a strong signal. The 3 x J310's
are expected to produce around +27dbm. One or more stages
may use more then 3 x J310's to accomodate possible higher
power levels linearly.

This front end is mostly overkill. The weaknesses are only the
local oscillator and mixer. For signals inside the displayed
window of Linrad, the signals in the front end, will still be well below the
level needed to overload the Delta44. For signals outside of
the window, the extra dynamic range and filtering may prove useful.
Rf tools is used to explore optimal gain and noise figure trade-offs.
Each stage will be in its own box so I can easily change my mind
as the need arises.

Currently my ATF33143 pre-amp is displaying a .5db noise figure
and a +16dbm, 1 db compression point. Work is currently being
done to improve noise figure. Of course everything but the LO
is duplicated for Xpol.

73, Jim Shaffer, WB9UWA.

On 30 May 2003 at 14:08, Jurgen van de Glind wrote:

> I am glad to announce that I have Linrad running.
> Thank Leif for the excellent software.
> Next goal: Downconverter 144Mhz to 2,5Mhz  some ideas on this
> reflector? Would be realy fun to see the different between my present
> setup (which is working pretty well) and the Linrad setup. Jurgen
> PE1LWT 2x 3WL xpol H/V 144Mhz
> Linrad setup:
> pentium 1,5Mhz
> Delta44
> WSE2500