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

RE: compiling/linking Linrad's C-source code...etc

Hi, Roger, Peter and All,

> > I understand that You use Leif's WSE 2500.
> > Does this make a significant difference compared to direct
> > conversion ( if properly designed and not overloaded ) 
> > from 144 MHz to audio ?
> Leif is a better person to answer this than I, even under
> the best of circumstances.  And here it is 2300 and I am
> soon due back at work.  So my answer is even poorer.
If you compare the RX2500 (with some suitable 144 to 2.5 MHz 
conversion in front of it) to the simple 144MHz direct conversion
receiver http://antennspecialisten.se/~sm5bsz/linuxdsp/hware/sbl1.htm
you will find that there is no difference whatsoever if the
unit is not overloaded.

The big difference is how strong a signal can be before it overloads
the receiver:)

> But to answer your question.  I think it is very much more
> difficult to design a properly functioning direct conversion
> receiver at 144 MHz that to do a multiple conversion design.
>  Leif has some webpages that cover this.  Practically
> speaking, I played around with direct conversion a little
> bit without good results.  It was easy to do the multiple
> conversion models.
The words "properly functioning" should be replaced by 
"high performance";)

It is easy to make a well functioning I/Q mixer for 144MHz to
audio. It is not so easy to make a good local oscillator
to use with it. I have written a bit about it but since then
I learned a bit more about low noise local oscillators.
You may use the LO of RX70 and a crystal at 48 or 72 MHz,
then a tripler or doubler to feed a direct conversion rx 
at 144 MHz.

> > What is the net profit of WSE 2500 above appropiate direct
> > conversion 
There is no advantage whatsoever. The problem is that you
will not be able to make an appropriate direct conversion radio
for very demanding situations. 

> In which respects ?
> > Selectivity ? ( sure.. )
I do not understand what you mean here. There is selectivity
at many different levels, but the final selectivity is within
Linrad and the ultimate limit is set by the delay you can
accept. If you want a filter at 500 Hz audio with a Q of
500000 the time delay through it will be in the order of
1000 seconds. The analog hardware you use in front of the
computer is another thing.

> > IP3/dyn. range ? ( questionable ? )
> > Total system noise / noise floor  ? ( doesn't multiple
> > conversions add noise compared to DC ? ).
These two questions are the two sides of the same problem.
If the A/D converter has a dynamic range of 147dBc/Hz which
is what the Delta44 provides, a converter in front of it
that has the same dynamic range will degrade performance by
6dB. At the low end the two noise floors will add to raise
the noise floor by 3dB and at the high end, the third order
IM will add and make IM3 6dB stronger which will bring IP3
down by 3dB. The WSE converters have a much larger dynamic
range than the Delta44 and therefore each one of them degrades
the dynamic range by about one dB only. The price to pay
is a lot of power dissipation from all the RF power amplifiers. 

> > I studied Leif's papers on the WSE system's four modules .
> > I am fascinated by the stuff but I am still a bit confused
> > with regard to above paragraph.
> > Please, any help is welcome.
> I am not an RF Engineer, but I know that there is a reason
> there are many, many multiple conversion receivers out there
> and very few direct conversion designs. 
The WSE RX2500 and the converters to use with it is something very
special. The performance is about 20dB above "state of the art"
except for IP3 which is below "state of the art". It is designed
this way because of the great unbalance between BDR and IM3
limitations in a conventional top class radio.

There will soon be a lot of software receivers out there, 
there is already the SDR-14 http://www.rfspace.com and there 
will be many more in the future. These receivers will probably
require additional RF filters to provide useful performance
on HF bands because of spurious responses and limited dynamic
range. I fear that too many hams will go ahead and use SDRs
without adequate roofing filters and they would then generate
a rumour "the good old analog radio is much cleaner than that
modern digital stuff". I decided to make a digital radio that
will invalidate such a statement with a broad margin.

It is fun to connect a conventional 144MHz transmitter to the 
RX144 input (through an attenuator). I tried the IC 706 already
and I will carry the equipment to two meetings this summer
and collect spectra of various rigs. SSB and CW mode:)

Unfortunately the ARRL lab test data is not very reliable
(understatement). We do have a lot of qrm from each other 
that could be avoided. Every owner of a WSE system can
easily do accurate transmitter tests and tell his friends
what to buy and what to refuse.

> > It is not a bargain, though I do understand fully that it
> > must be costly for a very limited production .
This is not true. The price you have to pay is as if the
production was made in volume.

> > In Holland the WSE 2500 will amount to ~USD 730.
> > Is it worth the money already now ( as separate WSE 2500 )
> > or does it reward only in the chain 144/70/10.7/2.5 ?
This is not for me to answer.
Most probably something very much simpler would be adequate
but you will have to build it yourself.
If you have severe qrn problems, a very high dynamic range
is required to make the Linrad clever blanker operate

> > I guess You could have made comparisons between separate
> > WSE 2500  yes/no ?
compare what to what?

> Value is different to each of us.  I am hearing better with
> the homebrew front end I have coupled to Leif's WSE2500 and
> then the Delta44 and Linrad than I have ever heard with any
> other receiver / receive system on 144 MHz eme.  And to me,
> that is value to the nth degree.  It is beyond dollars, to
> me it is priceless ;)
I am happy to hear - but I do not quite understand. When
properly (gain levels) used, the time machine should provide
the same noise figure and bandwidth. Maybe the dynamic range
is not adequate for the blanker with the Time Machine?

> I have not yet had the advantage of the Leif Chain of
> modules, but I hope to use it when they are available.
The main difference will be BDR. If you can not notice any
loss of S/N when strong local signals enter your system,
then there will be no difference at all. The difficult unit 
is the RX2500. The way you use it now does not take advantage
of it's full potential and something much simpler would have
been equally good. (but maybe not quite as simple as 
the Time Machine)
When the series of converters is added you will get a system
in which the parts are matched to each other. It will be 
overqualified, the locals who cause qrm will have sideband 
noise that will be the limitation. The system will be an adequate
spectrum analyzer for evaluating your neighbour's transmitters
and could be used to clean up the RF environment. 
I do not want to use the term value. The price is low if you
count the cost of the parts that go into a system. Wheter it
is worth all the money to use an expensive system like this is 
the real question for which I do not have a general answer. 


Leif / SM5BSZ