[linrad] Re: Linrad new user

2007-10-08 Thread Leif Asbrink
Hello Dave,

   I don't see anything on the screen...no waterfall display or anything 
 that indicates the program is running. At one point I did get a SDR-14 
 detected message.
OK.

 I run a logging program and an email program most of the time while I am 
 at the radio. I don't think these two applications are too taxing on the 
 CPU.
No, total load on the CPU is no problem at all. The problem is latency.
The routine that reads the USB has to be serviced in time EVERY SINGLE
TIME when you run under Windows. Linux is forgiving, it will just
loose some data. This means that any software that you run in parallel
with Linrad has to be started before Linrad and that you must not resize
any window or do other things that lock up the CPU for a long time.
(If you run near the speed limit where the time window within which
Linrad must respond is VERY small.)

With two CPUs in the system it might be much better, but I really
do not know.

Just set a low frequency by setting big decimation numbers like CIC2=16, 
CIC5=10 and RCF=3 for a total decimation of 480 and a sampling frequency 
of 138.9 kHz. This should work on any computer with any operating system.
Once you see spectra and know Linrad is running you can step the sampling
speed upwards. (It will be easier to edit par_sdr14 than to go through
the Linrad setup when you try different speeds.)

  Once you have decided what bandwidth to go for, you should find a pulse
  generator to do calibration. Read here:
  http://www.sm5bsz.com/linuxdsp/flat/ampcal.htm
  Actually this should not be required with an SDR-14 because the
  filter response is exactly known since it is determined by the
  digital filters (except for the small slope on gain vs frequency
  that your converters may give) but because of my laziness, Linrad
  does not know what filter response the digital filters inside
  the AD6620 will supply so you will have to calibrate to get
  an absolutely flat frequency response.
  
 I have a test bench full of equipment good through 20 ghz, but I'm not 
 sure any of it pulses :-)
A low frequency square wave will be fine:-) Anything that emits a square
wave with fast rise times. You will need an attenuator if the output level
is TTL. Do the calibration somewhere between 7 and 14 MHz on the 0.1 to 30
MHz input.

 OK--Spectravue is the same...the larger the FFT sample size, the slower 
 the waterfall and the more sensitive it is.
I think you will find that Linrad differs. I will be interested in
your comments on that later on:-)

 I am leaving on a one week trip tomorrow and when I 
 get back am leaving one day later to go the MUD (Microwave Update) in 
 Pennsylvania. I will not have much time to experiment until after MUD. 
 If all else fails, I have a CD with Ubuntu 7.04 and a friend who is 
 really, really good with computers who will help me install it on my 
 machine.
There is no reason anything should fail. You can run under any
Linux distribution under X11 or with svgalib and you can run under
Windows 2000, XP and probably 98 and Vista as well. Finding the optimum
operating system could allow a slightly larger bandwidth, but if you
are happy with something like 160kHz the OS should not make any difference
at all to Linrad.

At the MUD you will have a chance to hear Roger, W3SZ. After that
I guess you will have no problem to get Linrad running. 

It is another thing to optimize parameters for a specific task.

As I understand it, first priority is finding every single station
that could perhaps be copied. Once this problem has an optimum 
solution for your environment, the next step is to find out how 
to filter out a weak CW signal as well as possible.

None of the two problems has been discussed in detail anywhere
as far as I know. It is easy in a specific situation, but a
general analyzis would be extremely complicated due to the
almost infinite number of combinations hardwares, interferences 
bandwidths of the desired signal(s) and timing issues on 
particular bands. 

73

Leif

#
This message is sent to you because you are subscribed to
  the mailing list linrad@antennspecialisten.se.
To unsubscribe, E-mail to: [EMAIL PROTECTED]
To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED]
To switch to the INDEX mode, E-mail to [EMAIL PROTECTED]
Send administrative queries to  [EMAIL PROTECTED]



[linrad] Re: Linrad new user

2007-10-07 Thread Leif Asbrink
Hi Dave,

 I have just subscribed to this list. I would like to use LINRAD with my 
 SDR-14 to look for and demodulate weak CW signals on the microwave bands.
OK:-)

 I have downloaded the software for LINRAD. I have scanned through the 
 help file. I have attempted to run LINRAD and have entered some initial 
 values in the setups. (They are probably wrong.)
If you can see and hear signals they are not wrong - but they could
be far from optimum for your purpose

 I would like to have someone help me through the initial learning curve, 
 if only to direct me to which articles I should read to begin with. The 
 learning curve appears to be very steep to me at this point.
This seems to be a reasonable assumption;-)
To start with, it would be a good idea to master the SDR-14,
in other words to get an idea about what the AD6620 is doing.
In Linrad you have full freedom to select the decimation factors 
and to get various compromises between spur rejection and visible
bandwidth. For microwaves I guess spurs would not be any problöem at
all, and then you should look for a small value for RCF decimation.
Just play around with CIC2, CIC5 and RCF decimation to see what your 
computer can handle. The A/(D clock is 66.6 MHz so you need
the product of all three decimation factors to be somewhere around
400 to 300 for a bandwidth between 166.66 kHz to 222.22 kHz.
Maybe you will find that CIC2=16, CIC5=10 and RCF=2 for a total
decimation of 320 and a sampling frequency of 208.33 kHz will be OK
for your computer. Most likely under Linux, but under Windows you
would probably have to use some more decimation.

On my computer, a 2.66 GHz PIV, the limit under Linux is about 230 
kHz, but under Windows it is about 180 kHz.

It will be a good idea to verify that Linrad does not fail when you
do other things that you intend to do on the same computer in case you
want to use it simultaneously for other purposes. This is more likely to
be a problem under Windows than under X11.

Once you have decided what bandwidth to go for, you should find a pulse
generator to do calibration. Read here:
http://www.sm5bsz.com/linuxdsp/flat/ampcal.htm
Actually this should not be required with an SDR-14 because the
filter response is exactly known since it is determined by the
digital filters (except for the small slope on gain vs frequency
that your converters may give) but because of my laziness, Linrad
does not know what filter response the digital filters inside
the AD6620 will supply so you will have to calibrate to get
an absolutely flat frequency response.

Once you have a flat frequency response (after calibration) you
will find that a slower waterfall will supply better sensitivity.
The bin bandwidth as well as the averaging is important. I suggest
you start by getting a calibrated system at the bandwidth of your 
choice. The next step would be to decide whether to use the second 
FFT or not and that will depend on what you can observe with a 
calibrated system.

This mailing list is far from overloaded and your problem is the
same as many other operators face so I think it will be a good 
idea to discuss it on the list.

73

Leif / SM5BSZ







 
 I have instant messenger capability if you want to try a real time 
 session. I suspect we should email directly so as to keep the clutter 
 down on the reflector.
 
 I appreciate any help.
 
 Thanks,
 
 Dave, K4TO
 
 
 #
 This message is sent to you because you are subscribed to
   the mailing list linrad@antennspecialisten.se.
 To unsubscribe, E-mail to: [EMAIL PROTECTED]
 To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED]
 To switch to the INDEX mode, E-mail to [EMAIL PROTECTED]
 Send administrative queries to  [EMAIL PROTECTED]
 

#
This message is sent to you because you are subscribed to
  the mailing list linrad@antennspecialisten.se.
To unsubscribe, E-mail to: [EMAIL PROTECTED]
To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED]
To switch to the INDEX mode, E-mail to [EMAIL PROTECTED]
Send administrative queries to  [EMAIL PROTECTED]



[linrad] Re: Linrad new user

2007-10-07 Thread Dave Sublette

Hello Leif,

Well, Well! There is nothing like getting advice from the author of the 
software!! Thanks for answering.


Leif Asbrink wrote:



I have downloaded the software for LINRAD. I have scanned through the 
help file. I have attempted to run LINRAD and have entered some initial 
values in the setups. (They are probably wrong.)

If you can see and hear signals they are not wrong - but they could
be far from optimum for your purpose
 I don't see anything on the screen...no waterfall display or anything 
that indicates the program is running. At one point I did get a SDR-14 
detected message.





I would like to have someone help me through the initial learning curve, 
if only to direct me to which articles I should read to begin with. The 
learning curve appears to be very steep to me at this point.

This seems to be a reasonable assumption;-)
To start with, it would be a good idea to master the SDR-14,
in other words to get an idea about what the AD6620 is doing.
 I have used the SDR14 for over a year with the SpectraVue software and 
I am disappointed in the results. That is what led me to try Linrad. I 
am a retired electronics engineer and so I so have some understanding of 
FFT processes. However, I have never worked with them professionally and 
would not claim to know much about it.


In Linrad you have full freedom to select the decimation factors 
and to get various compromises between spur rejection and visible

bandwidth. For microwaves I guess spurs would not be any problöem at
all, and then you should look for a small value for RCF decimation.
Just play around with CIC2, CIC5 and RCF decimation to see what your 
computer can handle. The A/(D clock is 66.6 MHz so you need

the product of all three decimation factors to be somewhere around
400 to 300 for a bandwidth between 166.66 kHz to 222.22 kHz.
Maybe you will find that CIC2=16, CIC5=10 and RCF=2 for a total
decimation of 320 and a sampling frequency of 208.33 kHz will be OK
for your computer. Most likely under Linux, but under Windows you
would probably have to use some more decimation.

This appears to be a good starting point for me. My computer is a dual 
CPU DELL XPS, about a year old.


On my computer, a 2.66 GHz PIV, the limit under Linux is about 230 
kHz, but under Windows it is about 180 kHz.



Ok on this. I will look for this.


It will be a good idea to verify that Linrad does not fail when you
do other things that you intend to do on the same computer in case you
want to use it simultaneously for other purposes. This is more likely to
be a problem under Windows than under X11.


I run a logging program and an email program most of the time while I am 
at the radio. I don't think these two applications are too taxing on the 
CPU.


Once you have decided what bandwidth to go for, you should find a pulse
generator to do calibration. Read here:
http://www.sm5bsz.com/linuxdsp/flat/ampcal.htm
Actually this should not be required with an SDR-14 because the
filter response is exactly known since it is determined by the
digital filters (except for the small slope on gain vs frequency
that your converters may give) but because of my laziness, Linrad
does not know what filter response the digital filters inside
the AD6620 will supply so you will have to calibrate to get
an absolutely flat frequency response.

I have a test bench full of equipment good through 20 ghz, but I'm not 
sure any of it pulses :-)




Once you have a flat frequency response (after calibration) you
will find that a slower waterfall will supply better sensitivity.
The bin bandwidth as well as the averaging is important. I suggest
you start by getting a calibrated system at the bandwidth of your 
choice. The next step would be to decide whether to use the second 
FFT or not and that will depend on what you can observe with a 
calibrated system.


OK--Spectravue is the same...the larger the FFT sample size, the slower 
the waterfall and the more sensitive it is.


This mailing list is far from overloaded and your problem is the
same as many other operators face so I think it will be a good 
idea to discuss it on the list.


Wonderful! Thanks. I am leaving on a one week trip tomorrow and when I 
get back am leaving one day later to go the MUD (Microwave Update) in 
Pennsylvania. I will not have much time to experiment until after MUD. 
If all else fails, I have a CD with Ubuntu 7.04 and a friend who is 
really, really good with computers who will help me install it on my 
machine.


Thanks again, Leif. I will be continuing to work on Linrad as time 
allows. Being retired is very hard. There is so much to do and so little 
time.


73,

Dave, K4TO

#
This message is sent to you because you are subscribed to
 the mailing list linrad@antennspecialisten.se.
To unsubscribe, E-mail to: [EMAIL PROTECTED]
To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED]
To switch to the INDEX mode, E-mail to