Detecting H1 and pulsars with minimal equipment.

[Michiel Klaassen]

Hi All,

I read trials of other amateurs to detect pulsar signals. 

With a drift scan 60cm satellite dish and a sdr dongle is was possible for me to detect the H1 line, however I did not detect a pulsar with the 60cm dish beyond any reasonable doubt. 

Crucial is (for folding /stacking), I think, to know the precise period time of the pulsar, and the time correction for all the variations.

The best way to do that, it seems, is with Tempo or Tempo2. 

However I tried to install that, but with no success so far; Temo(1) cannot be found anymore.

All the instructions say it is "easy" and "simple". 

There are C/C++/fortram to be compiled versions, python start-up versions, youtube instructions and more, a windows version is not to be found (anymore)

Has anyone succeeded in installing and using this package. 

If you have, can you describe the step by step install procedure.

Upto now I tried to detect B0531+21 and B0329+54. 

I used (01-03-2013) the respective period times 0,03366848881872s and 0,714518664s for them. 

My observation location is 52.0435961N, 5.6666457E.

Michiel

[Steve Olney - VK2XV]

G'day Michiel,

Here is a link to the Windows/DOS version provided by John H Taylor (K1JT)....

http://www.k5so.com/Radio_astronomy_pulsars.html

Look for the link at the bottom of the buttons.

This is what I use.  I installed it as shown on here...

http://pulsar.vk2xv.net/n4_tempo.htm

These are not step-by-steps instructions, but if you read the Tempo manual as linked on the page (or the truncated version also linked on the page) it should be possible to install and setup for your own situation.

Tempo is not a program with a Windows GUI so a degree of determination is require to get the information you require.

My requirements are only for a static period value during a 4 hour observation period (as I use FFT analysis detection).   For folding/stacking you will need dynamic period values for any significant observation time.   For how to do that you will have to ask other people.   For example here...

http://www.moetronix.com/pulsar/index.htm

Cheers

Steve

[Marcus D. Leech]

With a tiny 60cm dish, the probability of your detecting *any* pulsar is
exceedingly low. You would need to integrate for *days* at a time,
which means
tracking with great accuracy, and your timebase would need to be
exceedingly stable, in order for things not to get smeared into the noise.




--
Marcus Leech
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org

[ARNO...@aol.com]

Hello Michiel ,

I have tried for years to detect the strongest Pulsar with a 2 m dish and last year with a 4 m dish ,using

the software tempo and all othe known methodes to anhance the signal with up to 6 hrs integration and tracking of the pulsar .I tried at frequencies from 8 GHz , 1296 GHz and 410 MHz .

All efforts did not produce a signal.What I know from others it look like a dish of at least 9 m is needed to get a signal at 410 MHz and than just a liitel bit over the noise.

 

regards Wolfgang , Dj3qd

 

 

 

In einer eMail vom 04.03.2013 00:19:57 Westeuropäische Normalzeit schreibt patchv...@gmail.com:

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

 

[Tom Crowley]

The 40 Ft (12.2M) telescope has been used to detect Pulsars at 1420 MHz.  The lower the frequency the brighter the Pulsar, to a point.  NRAO uses around 350 MHz to discover Pulsars with the GBT. We have one member who has discovered Pulsars with a 3 Meter dish at 408 MHz.

 

Tom Crowley

[Marcus D. Leech]

The 40 Ft (12.2M) telescope has been used to detect Pulsars at 1420 MHz.  The lower the frequency the brighter the Pulsar, to a point.  NRAO uses around 350 MHz to discover Pulsars with the GBT. We have one member who has discovered Pulsars with a 3 Meter dish at 408 MHz.

 

Tom Crowley

I'm pretty sure, Tom, that you mean "detected" rather than "discovered".  Discovering *new* pulsars with a 3 Meter dish at 408 Mhz seems exceedingly
  unlikely.

[Tom Crowley]

Right you are Marcus, Bad choose of words without a cup of Java this AM.

 

Tom

 

From: Marcus D. Leech

Sent: Monday, March 04, 2013 10:47 AM

To: sara...@googlegroups.com

Subject: Re: [SARA] Detecting H1 and pulsars with minimal equipment.

 

The 40 Ft (12.2M) telescope has been used to detect Pulsars at 1420 MHz.  The lower the frequency the brighter the Pulsar, to a point.  NRAO uses around 350 MHz to discover Pulsars with the GBT. We have one member who has discovered Pulsars with a 3 Meter dish at 408 MHz.

 

Tom Crowley

 

[Michiel Klaassen]

Hi All,

Thank you for sharing your experience.

A 60cm dish is very small indeed, but the H1 line was detected ok with it; see attachment.

I based this pulsar test on the information that with the 25m Dwingeloo telescope the individual pulses of the B0329 pulsar could be heard.

So, thats with 500m2. A 0.6m dish has 0.3m2 surface. For thesame result you have to stack 500/03=1666 pulses. 

This can be received in 1666*0.7= 1166s=19 minutes (P0329+54).

For a 1.2m dish; 500 pulses in 6 minutes.

Of course tracking is important, but I wanted to solve the period correction first(smearing), so the info from Steve is indeed very welcome. 

I am rewriting my python script, so it can send the dos commands and read the results back, so the period time can be adjusted.

It is not clear to me if the PC time/summertime/wintertime/UTC/location is taken into account by tempo.

Steve, do you know how it is done or can it be corrected?

As a pulsar simulator I use a signal generator with 1420 MHz inband frequency, and modulated with 1,4Hz. 

By lessening the outputlevel, I can optimize the receiving system.

Further, yes, also in the planning is the construction of a larger dish, but I think all amateur radio astronomers have such plans.

Michiel

The attachment show H1 drift results taken at DEC=30 over 24 hours.

Every line is the mean value of 300000 FFT transformations over a period of 15 minutes.

The receiver system consists of a 0.6m satellite dish, can type feedhorn, RAS LNA, 30dB amp, DVB dongle, PC.

     

2013/3/4 Tom Crowley <crow...@hotmail.com>

You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.

[Marcus Leech]

Nice Hydrogen plots.

In order to compare the Dwingleoo results with another setup, you also have to factor in the observing bandwidth.  Sensitivity scales with the square-root of the bandwidth, so it's important to factor *that* in as well.

[Steve - VK2XV]

G'day Michiel,

>
> I based this pulsar test on the information that with the 25m Dwingeloo
> telescope the individual pulses of the B0329 pulsar could be heard.
> So, thats with 500m2. A 0.6m dish has 0.3m2 surface. For thesame result
> you
> have to stack 500/03=1666 pulses.
> This can be received in 1666*0.7= 1166s=19 minutes (P0329+54).
>

Unfortunately detection sensitivity is increased in direct proportion to
aperature - so the 25m telescope will be, say, ~1700 times more sensitive
than a 0.3m telescope.

To regain that sensitivity by stacking pulses requires approx. (1700 *1700)
pulses, because sensitivity increases by the square root of the number of
pulses stacked so you need 3,000,000 seconds = ~840 hours for P0329+54.

That relationship of aperture and integration time is the killer for small
antennas.

Also, as Marcus has pointed, there is the question of pre-detection
bandwidth. Here the news is a little better because if you are using a
bandwidth 100 times narrower, you only lose sensitivity by a factor of the
square root of 100 = 10.

>
> I am rewriting my python script, so it can send the dos commands and read
> the results back, so the period time can be adjusted.
> It is not clear to me if the PC time/summertime/wintertime/UTC/location is
> taken into account by tempo.
> Steve, do you know how it is done or can it be corrected?
>

I do not need to correct in real time using my FFT 'sweetspot' method -
needing only the correction for orbital motion, not diurnal (earth rotation)
correction. However, if you read the information on my website you can
see how your local location is fed into Tempo. The output is in UTC which
you can relate to local PC time if needed. It took me about a week or so
of study to understand how Tempo works - there are no shortcuts to tinkering
with it and doing some hand calculations and analysing the results. What I
did to make sure I was getting the correct results was to work out the
rising and setting times for a pulsar was here and check that those times
corresponded to the shortest period and longest period swings over the
diurnal cycle as output by Tempo.

Cheers

Steve

[Dave Typinski]

On 03/06/2013 15:42, Michiel Klaassen wrote:
> Hi All,
> Thank you for sharing your experience.
> A 60cm dish is very small indeed, but the H1 line was detected ok with it; see
> attachment.
>
> I based this pulsar test on the information that with the 25m Dwingeloo
> telescope the individual pulses of the B0329 pulsar could be heard.
> So, thats with 500m2. A 0.6m dish has 0.3m2 surface. For thesame result you have
> to stack 500/03=1666 pulses.
> This can be received in 1666*0.7= 1166s=19 minutes (P0329+54).

S/N ratio goes as the square root of integration time -- and this goes for
folded or stacked data as well as real time data. Point being, if you halve the
aperture, you must quadruple the integration time to make up for it.

So, if:
-- pulsar B0329+54 has a period of 715 ms,
-- Dwingeloo can see single pulses,
-- your aperture is a factor of 1670 smaller,
-- your bandwidth is the same as Dwingeloo,
-- your receiver's noise figure is the same as Dwingeloo,
-- your de-dispersion is the same as Dwingeloo,
then:
to achieve the same S/N ratio as Dwingeloo, you'd need to fold 0.715 * 1670^2
seconds of observation, or about 554 hours.
--
Dave

[Michiel Klaassen]

Hi Steve and Dave,

I do not agree with your calculation, but perhaps you did not see the square meter notation I used. 

I am only interested in the amount of power received.

You can also use the diameter of the two dishes; Power ratio=(25/0.6)^2=1700.

So when the radio rain comes down at ..W/m2/Hz, the only thing that matters is the amount of square meters of your (umbrella) antenna and bandwith of course.

Michiel

2013/3/6 Dave Typinski <dav...@typnet.net>

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to

sara-list-unsubscribe@googlegroups.com

For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
--- You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.

To unsubscribe from this group and all its topics, send an email to sara-list+unsubscribe@googlegroups.com.

[Steve Olney - VK2XV]

G'day Michiel,

Of course you are perfectly entitled to your opinion - but if we are talking about science then I feel I must point out that the calculations are correct.

The calculations are not ours, it is the basic radiometer equation to be found, in various forms, in radio astronomy textbooks.    I was going to scan several for you, but being not sure of copyright, I have inserted a graphic I created myself for pulsars which you will find on my website.  I am not sure how it will come out as this is the first time I have inserted a graphic on this forum.   If it fails I will send you a link to my website in context.

The factors relevant to our discussion are G,  delta V and tint - all on the bottom line.   Smin is the minimum flux density detectable.     G is proportional to aperture (or the square of the dish diameter).    Delta V is the pre-detection bandwidth and tint is the integration time.  Please take special note that these two factors are inside a square root sign.    So - holding all other factors equal, to regain the same sensitivity lost by a reduction of 1700 in G, you must increase the product of the bandwidth and integration time by a factor of 1700*1700.

As I said these are not our calculations.   Any good radio astronomy book will have them in one form or another.

Cheers

Steve

2013/3/6 Dave Typinski <dav...@typnet.net>

sara-list-...@googlegroups.com

For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
--- You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.

To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.

[Dave Typinski]

Hi Michiel,

Here is a video about the radiometer equation that does an absolutely
outstanding job of explaining why it works as Steve and I described in previous
emails.

http://www.youtube.com/watch?v=dwT-tMoscsY

So, to achieve the same S:N ratio as Dwingeloo, you will have to stack 1670^2 =
2.8 million times the data that Dwingeloo uses.

--
Dave

On 03/08/2013 13:51, Michiel Klaassen wrote:
> Hi Steve and Dave,
> I do not agree with your calculation, but perhaps you did not see the square
> meter notation I used.
> I am only interested in the amount of power received.
> You can also use the diameter of the two dishes; Power ratio=(25/0.6)^2=1700.
> So when the radio rain comes down at ..W/m2/Hz, the only thing that matters is
> the amount of square meters of your (umbrella) antenna and bandwith of course.
> Michiel
>
>
>

> 2013/3/6 Dave Typinski <dav...@typnet.net <mailto:dav...@typnet.net>>

> <mailto:sara...@googlegroups.com>

> To unsubscribe from this group, send email to

> sara-list-unsubscribe@__googlegroups.com
> <mailto:sara-list-...@googlegroups.com>

> For more options, visit this group at

> http://groups.google.com/__group/sara-list?hl=en

> <http://groups.google.com/group/sara-list?hl=en>
> --- You received this message because you are subscribed to a topic in the
> Google Groups "Society of Amateur Radio Astronomers" group.
> To unsubscribe from this topic, visit

> https://groups.google.com/d/__topic/sara-list/sMqCmg-5O_Q/__unsubscribe?hl=en <https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en>.

> To unsubscribe from this group and all its topics, send an email to

> sara-list+unsubscribe@__googlegroups.com
> <mailto:sara-list%2Bunsu...@googlegroups.com>.
> For more options, visit https://groups.google.com/__groups/opt_out
> <https://groups.google.com/groups/opt_out>.

>
>
>
> --
> --
> You received this message because you are subscribed to the Google
> Groups "Society of Amateur Radio Astronomers" group.
> To post to this group, send email to sara...@googlegroups.com
> To unsubscribe from this group, send email to

> sara-list-...@googlegroups.com

> For more options, visit this group at
> http://groups.google.com/group/sara-list?hl=en
> ---

> You received this message because you are subscribed to the Google Groups

> "Society of Amateur Radio Astronomers" group.

> To unsubscribe from this group and stop receiving emails from it, send an email
> to sara-list+...@googlegroups.com.

[Paul Oxley]

Dave   I watched the video. It looks correct for the assumptions made. One of the first assumptions is that the source is a random noise source. For pulsars as well as some other objects, this assumption is only partially correct. The component of the signal related to the pulse of a pulsar is non random. It repeats on a regular basis. It also is the item of interest for pulsar detection.   Without getting in over my head, this means that to be able to detect a pulse, you can use less integration time. The signal levels from the pulse add coherently with the integration. The random noise does not add coherently. In the telephone business where I retired, we often used dB as a measure of levels. For a purely random source, the signal adds on a 10 Log basis. For a purely coherent signal, the addition across the integration time is a 20 Log function. For something inbetween (ie pulsar), the addition would be somewhere between 10 and 20 Log. The 20 Log function applies when the voltage of the signal is in phase. The 10 Log function applies when only the RMS power is present.   Thus I think you both are correct poissibly with different assumptions on the nature of the desired signal.   For pulsars, the integration is in the folding of the signal over time. When the period of the folding matches the period of the pulse, the signal is enhanced.   Paul --- On Sat, 3/9/13, Dave Typinski <dav...@typnet.net> wrote:

From: Dave Typinski <dav...@typnet.net> Subject: Re: [SARA] Detecting H1 and pulsars with minimal equipment.

> to sara-list+unsub...@googlegroups.com.

> For more options, visit https://groups.google.com/groups/opt_out. > >

-- -- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group. To post to this group, send email to sara...@googlegroups.com To unsubscribe from this group, send email to sara-list-...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sara-list?hl=en --- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+unsub...@googlegroups.com.

[Dave Typinski]

Hi Paul,

Thanks for the input! I don't understand how to apply your 10 log and 20 log
functions to the present scenario, but my hunch is that we're saying the same
thing in two different ways.

I'm sure you already know all of the following, Paul, but for those who might not:

We start with the fact that we're not comparing the pulse's peak power to the
background noise power, but rather to the /variation/ in background noise power.
That is, the N part of the S:N ratio is not the noise power of the background
noise, but rather N is the /variation/ in the background noise power.

For pulsar pulses, the pulse amplitude remains at the mean value as integration
time increases, precisely because it is near enough a coherent signal. Sample
your AC mains voltage for as long as you like and you still get 160 volts peak
to peak.

The /fluctuation/ in the background noise in which the pulsar pules are
embedded, however, tends toward zero as integration time increases because it is
random noise with a nominally Gaussian distribution. Integrate the galactic
background longer and you don't reduce it's amplitude, but you do make the strip
chart trace flatter.

This works because the standard deviation of a Gaussian distribution (e.g.,
random RF noise) is reduced as the number of samples is increased. This
reduction goes as the square root of the number of samples. In
statistics-speak, the standard deviation (sigma) of a Gaussian distribution is
inversely proportional to the root of the number of samples.

Quadruple the number of samples, and you cut the standard deviation in half.

Thus, quadruple the integration time and double the S:N ratio.

So, increasing the integration time doesn't do anything to the S part of the S:N
ratio. A longer integration time improves the S:N ratio by reducing N, not by
increasing S.
--
Dave

On 03/09/2013 10:54, Paul Oxley wrote:
> Dave
> I watched the video. It looks correct for the assumptions made. One of the first
> assumptions is that the source is a random noise source. For pulsars as well as
> some other objects, this assumption is only partially correct. The component of
> the signal related to the pulse of a pulsar is non random. It repeats on a
> regular basis. It also is the item of interest for pulsar detection.
> Without getting in over my head, this means that to be able to detect a pulse,
> you can use less integration time. The signal levels from the pulse add
> coherently with the integration. The random noise does not add coherently.
> In the telephone business where I retired, we often used dB as a measure of
> levels. For a purely random source, the signal adds on a 10 Log basis. For a
> purely coherent signal, the addition across the integration time is a 20 Log
> function. For something inbetween (ie pulsar), the addition would be somewhere
> between 10 and 20 Log. The 20 Log function applies when the voltage of the
> signal is in phase. The 10 Log function applies when only the RMS power is present.
> Thus I think you both are correct poissibly with different assumptions on the
> nature of the desired signal.
> For pulsars, the integration is in the folding of the signal over time. When the
> period of the folding matches the period of the pulse, the signal is enhanced.
> Paul
>

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=dav...@typnet.net>
> <mailto:dav...@typnet.net
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=dav...@typnet.net>>>

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara...@googlegroups.com>
> > <mailto:sara...@googlegroups.com
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara...@googlegroups.com>>

> > To unsubscribe from this group, send email to
> > sara-list-unsubscribe@__googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara-list-unsubscribe@__googlegroups.com>
> > <mailto:sara-list-...@googlegroups.com
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara-list-...@googlegroups.com>>

> > For more options, visit this group at
> > http://groups.google.com/__group/sara-list?hl=en
> > <http://groups.google.com/group/sara-list?hl=en>
> > --- You received this message because you are subscribed to a topic
> in the
> > Google Groups "Society of Amateur Radio Astronomers" group.
> > To unsubscribe from this topic, visit
> >
> https://groups.google.com/d/__topic/sara-list/sMqCmg-5O_Q/__unsubscribe?hl=en <https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en>.
> > To unsubscribe from this group and all its topics, send an email to
> > sara-list+unsubscribe@__googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=unsubscribe@__googlegroups.com>
> > <mailto:sara-list%2Bunsu...@googlegroups.com
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=2Bunsu...@googlegroups.com>>.

> > For more options, visit https://groups.google.com/__groups/opt_out
> > <https://groups.google.com/groups/opt_out>.
> >
> >
> >
> > --
> > --
> > You received this message because you are subscribed to the Google
> > Groups "Society of Amateur Radio Astronomers" group.
> > To post to this group, send email to sara...@googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara...@googlegroups.com>

> > To unsubscribe from this group, send email to
> > sara-list-...@googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara-list-...@googlegroups.com>

> > For more options, visit this group at
> > http://groups.google.com/group/sara-list?hl=en
> > ---
> > You received this message because you are subscribed to the Google Groups
> > "Society of Amateur Radio Astronomers" group.
> > To unsubscribe from this group and stop receiving emails from it, send an
> email

> > to sara-list+...@googlegroups.com
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=unsub...@googlegroups.com>.

> > For more options, visit https://groups.google.com/groups/opt_out.
> >
> >
>
> --
> --
> You received this message because you are subscribed to the Google
> Groups "Society of Amateur Radio Astronomers" group.
> To post to this group, send email to sara...@googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara...@googlegroups.com>

> To unsubscribe from this group, send email to
> sara-list-...@googlegroups.com

> <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara-list-...@googlegroups.com>

> For more options, visit this group at
> http://groups.google.com/group/sara-list?hl=en
> ---
> You received this message because you are subscribed to the Google Groups
> "Society of Amateur Radio Astronomers" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to sara-list+...@googlegroups.com
> <http://us.mc1815.mail.yahoo.com/mc/compose?to=unsub...@googlegroups.com>.

> For more options, visit https://groups.google.com/groups/opt_out.
>
>
> --
> --
> You received this message because you are subscribed to the Google
> Groups "Society of Amateur Radio Astronomers" group.
> To post to this group, send email to sara...@googlegroups.com
> To unsubscribe from this group, send email to
> sara-list-...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/sara-list?hl=en
> ---
> You received this message because you are subscribed to the Google Groups
> "Society of Amateur Radio Astronomers" group.
> To unsubscribe from this group and stop receiving emails from it, send an email

> to sara-list+...@googlegroups.com.

[Paul Oxley]

Dave I agree we are saying the same thing from a different perspective.   On additional thought - The calculations assume that the intermodulation in the system is insignificant. Since intermod is also coherent (coming from the same products), it would also add in the integration. The longer the integration time, the more the intermodulation becomes significant. Likewise for tones or spurs in the system.   Paul

>      > to sara-list+unsub...@googlegroups.com

>     <http://us.mc1815.mail.yahoo.com/mc/compose?to=unsub...@googlegroups.com>. >      > For more options, visit https://groups.google.com/groups/opt_out. >      > >      > > >     -- >     -- >     You received this message because you are subscribed to the Google >     Groups "Society of Amateur Radio Astronomers" group. >     To post to this group, send email to sara...@googlegroups.com >     <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara...@googlegroups.com> >     To unsubscribe from this group, send email to >     sara-list-...@googlegroups.com >     <http://us.mc1815.mail.yahoo.com/mc/compose?to=sara-list-...@googlegroups.com> >     For more options, visit this group at >     http://groups.google.com/group/sara-list?hl=en >     --- >     You received this message because you are subscribed to the Google Groups >     "Society of Amateur Radio Astronomers" group. >     To unsubscribe from this group and stop receiving emails from it, send an

>     email to sara-list+unsub...@googlegroups.com

>     <http://us.mc1815.mail.yahoo.com/mc/compose?to=unsub...@googlegroups.com>. >     For more options, visit https://groups.google.com/groups/opt_out. > > > -- > -- > You received this message because you are subscribed to the Google > Groups "Society of Amateur Radio Astronomers" group. > To post to this group, send email to sara...@googlegroups.com > To unsubscribe from this group, send email to > sara-list-...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/sara-list?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Society of Amateur Radio Astronomers" group. > To unsubscribe from this group and stop receiving emails from it, send an email

> to sara-list+unsub...@googlegroups.com.

> For more options, visit https://groups.google.com/groups/opt_out. > > -- -- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group. To post to this group, send email to sara...@googlegroups.com To unsubscribe from this group, send email to sara-list-...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sara-list?hl=en --- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+unsub...@googlegroups.com.

[Steve Olney - VK2XV]

Paul,

In fact the pulsar signal is a random noise source - it is in no way coherent as a tone/s used in communications (except for spread spectrum).    As far as the receiver is concerned a pulsar signal is a periodic rise in wideband random noise over the bandwidth.    As such it is not different from the sensitivity problem of continuum measurements - except for the pulse peak to average nature of the pulsar signal (which you will see is included in the pulsar sensitivity equation I sent previously (where W is pulse width) and also in the equation in the linked document below (where W is the duty-cycle).

I have researched the pulsar literature (done by professional astronomers who actually do pulsar work as their job) and they are all agreed on the relationship as David and I have pointed out (or rather - they are the source of our information on this).

As I said before - we all entitled to our opinions - but if we are serious about understanding and perhaps, in our wildest dreams, of succeeding, then I for one am going to go with the professional expertise.

In addition I have done a number of simulations which agree with the relationship as given in the professional arena.

I would be interested in your research results, especially the simulations you might have done.

Working out the required characteristics for pulsar detection with minimal equipment is heavy work.   If you can find the time to read through the process gone through here...

http://pulsar.vk2xv.net/

you will see I am still not at the point, after months of study, at the point of doing a formal search.   If it was easy it wouldn't be so attractive a project.

Here is my final submission on the subject...

http://www.atnf.csiro.au/research/graduate/theses/2010/cameron_thesis.pdf

See section 3.3.1 on page 28.

I found this reference by two minutes of googling.

I have added it to my references collection as it is a new one for me.

Best of luck.

Steve

[Paul Oxley]

Steve   I dont care to debate your simulations. The point I was making is that the pulse is coherent at a very low frequency rate (pulses per second or seconds per pulse).   In your research, dont forget to also investigate dispersion which is also coherent since it is the same signal delayed.   Paul --- On Sat, 3/9/13, Steve Olney - VK2XV <pul...@vk2xv.net> wrote:

From: Steve Olney - VK2XV <pul...@vk2xv.net> Subject: Re: [SARA] Detecting H1 and pulsars with minimal equipment. To: sara...@googlegroups.com

--

-- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group. To post to this group, send email to sara...@googlegroups.com To unsubscribe from this group, send email to sara-list-...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sara-list?hl=en --- You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

[Steve Olney - VK2XV]

G'day Paul,

I understand what you mean by coherent - the point is the original post by Michiel about integration times needed to compensate for a difference in aperture was incorrect.    The correction offered by David and myself is correct whichever colour you want to paint the barn.   It is derived from what the professionals say (I presume you don't care to debate them either... ;-)

In any case, as they say, "you can lead a horse to water..."

As this has turned into an argument rather than a discussion of facts I see no use in continuing.

I am happy, though, to be educated, via this forum, by facts.   As a case in point, I had an lack of understanding that, in addition to dispersion, you must take into account scattering effects at low frequencies when attempting pulsar detecttion.

I was ignorant on this point, but fortunately I had the ability to see my deficiency and so I benefited from the knowledge of others.   The embarrassment of quickly acknowledging my ignorance has the compensation, for me, of a faster rate of learning.

Best of luck.

Steve

[Randall family]

An odd thought experiment:

When the beacon of the pulsar turn towards you, it is a higher temperature.  Still random noise.

Imaginary pulsar is 10 second period.

Imaginary pulsar is 1 second at higher temperature

We record 100 pulses with 1 second bins in exact sync with pulsar. 

Total measure time = 1000 seconds.

100 bins of 1 second averaged together is 100 seconds of integration time for pulse.

Does temperature at the pulse peak count as level for sensitivity calculation?

Is it best to use largest bin size that can properly resolve pulse?

I suspect coherents is of little help in detection.  Supect is NOT the same as know.

 

We all are searching for the answer.  Don't take it personal.

Professionals have been wrong before.

Thinking is required.

Relax, drink a beer & look at it again another day.

 

Bruce Randall

----- Original Message -----

From: Steve Olney - VK2XV

To: sara...@googlegroups.com

Cc: oxl...@att.net

Sent: Saturday, March 09, 2013 6:05 PM

Subject: Re: [SARA] Detecting H1 and pulsars with minimal equipment.

--

[Steve Olney - VK2XV]

G'day Bruce,

>
>We all are searching for the answer.

>

Agreed.

>
> Don't take it personal.
>

I don't.  I think it is more likely my curt and to the point style in my replies is more likely to cause others to take it personally... ;-)

>

>Professionals have been wrong before.

>

True - but not all of them on one subject.   The equations presented as the basis for David's and my statement are backed by practical measurements by the professionals who actually do do this stuff.    In fact, it is used to design their systems.

Please note - the following is not related directly to Michiel's original question, but is a side comment only.

The question of coherence is a tricky - a similar effect is found when using FFT analysis on signal tones in noise (i.e., coherent signals).  

In practical systems there is a limit to how long a data record you can use for FFT analysis.    Too long a record in one FFT can result in diminishing returns due to issues of receiver and transmitter drift and propagation effects where the signal energy gets smeared over adjacent frequency bins.   One way around this is decimate in time such that the individual frequency bins are wider and then stack individual decimations.   The problem is then, once again, you are stacking incoherently and so the S/N improvement goes as the square root of the number of stacks.   Deciding which is the best decimation in time requires careful examination of the channel characteristics of the signal.

I sympathise with those struggling to accept that folding or stacking pulsar pulses is an incoherent process.   I was reluctant to let it go (because it is such a killer for small antenna systems) myself.  That's why I did the simulations (hoping the professional were wrong), but found they were correct - arrogant for me to think otherwise, or could I excuse myself by characterising it as a massive attack of wishful thinking on my part.

As for me I cannot think about it further as I now understand it and have verified it for myself.   Others will follow their own path.

Cheers

Steve

[Michiel Klaassen]

Hi All.

I have placed my reaction in the wrong section, sorry for this. So here is a copy.

I read a lot of reactions on my last message. Thank you all for your comments and advice again.

There was a lot of discussion about the S/N ratio, however it was never mentioned in my message to be the goal.

The only thing is, that I want to try to detect a pulsar; yes or no, with minimal equipment. In thesame way as was proven that detecting H1 can be done with minimal equipment. (btw this is a very cheap way for schools to replicate)

Thanks to Steve and Tempo now a precise period can be programmed into the sw, so this problem was solved.

Next week or so, the dish will be mounted more firmly, because now it is still laying on the ground for the H1 detection test, and it was blown away by the wind al the time, and filled with snow because of the upside down position (see attachment)

The dish will be mounted on the wall and pointed to el=50 degrees (50 is mech. max. possible) and az=60 degree.

In this way the B0329+54 pulsar drifts by every day for some time.

Further, the python sw is modified from the H1 detection to the pulsar version. In a loop, 240 buckets/registers are filled with 4096 samples every period. The trigger time is set by a loop wich is waiting for the beginning of each second period (the non trigger/waiting loop is done about 100000 times). There are some further dependencies because of the processor speed, so the sampling time of 3.2MS/s is not effectively used. The bandwidth is about 1.6MHz, but there are time gaps because of the usb communication and the initial data manipulation. Yes there are also other dongle comm methods, but I wanted to try this method first.

Now, in one minute time, the buckets are filled 41 times(2*0.7*41=58s). After that all the buckets are saved as a group of buckets, and the next group of buckets are in place to be filled. 

The data of the group of buckets is not written to disk yet, because that will take time and then we lose synchronisation.

For an hour or so, capturing will be going on. After that he pulsar will be out of the beam. 

Now the data is written to disk and analysing can begin.

 

Of each goup of buckets the FFT is calculated, resulting in bins with half the number of buckets. The first bin gives the .5 frequency of the pulsar frequency etc.

Now, the second bin value is added to the 4th and the 6th bin to give a value (Y), which can be compared to sum of the 1the, 3th and 5th bin, the N value.

If there is no pulsar signal, then all bins will have thesame value. If there is more energy in the 2, 4, and 6 bin than in the 1, 3, 5 bin than we have detected a pulsar (with the given period time) (asymmetric pulse).

So this will be the first thing we will try to find. After a Yes/No result we will see what can be done to improve the result.

I will will report back to you soon

Michiel

2013/3/10 Steve Olney - VK2XV <pul...@vk2xv.net>

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---

You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.

[Michiel Klaassen]

Hi Dave,

I will try to get some more answers, however it was a measurement done by amateurs in 2009.

The only thing I know extra is that they said that the b0329 pulsar was strong and present every cycle.

The crab pulsar was weak and heard only a short time every 15 minutes or so.

Michiel

2013/3/13 Michiel Klaassen <vmin...@gmail.com>

[Michiel Klaassen]

Hi Steve,

I am finalising my python sw script; the communication between the dos files is going fine.

Just to check with you my interpretation of the final list of frequencies generated in polychk.out.

I see from top to bottom first an increase in frequency then a max, then a decrease and at the bottom a minimum.

I think that the top max is because when the pulsar is just above the horizon, the approach velocity of that point of the earth is maximum to the pulsar.

When the pulsar is going just below the horizon the frequency is minimum. Just the doppler effect again. 

So when observing at an elevation of 60 degrees west we have to select the frequency 4H past the maximum.

So, I have to use 1.3994278614776.

Michiel

  

mes0: 

C:\tempo>set TEMPO=c:\tempo\ 

C:\tempo>tempo -z 

TZ source list for site = j

    PSR        Nspan  Ncoeffs  Maxha    Freq

----------------------------------------------

 0329+54         1080       5    9.00    432.00000

 Current date is 13-MAR-13, or MJD 56364.806

 Enter first and last MJD, or hit return to run for today: 

259 5

260 6

261 3

262 6

263 4

264 .

265 8

266 0

267 6

mes2: 

56364.806

mes5: 

C:\tempo>polychk 0329+54 56364.806 

 0329+54 0329+54    56364.8060000000        56364.6875000000               19

          19

Read   1 set of coefficients from polyco.dat

Wrote    18 TOAs to polychk.out

mes8: 

19   1 0329+54  430.0000 56364.3131885291659     1.3994277879689

19   2 0329+54  430.0000 56364.3548556884271     1.3994279971364

19   3 0329+54  430.0000 56364.3965228427332     1.3994281226280

19   4 0329+54  430.0000 56364.4381899944201     1.3994281749160

19   5 0329+54  430.0000 56364.4798571455103     1.3994281644726

19   6 0329+54  430.0000 56364.5215242977138     1.3994281017703

19   7 0329+54  430.0000 56364.5631914524347     1.3994279972812

19   8 0329+54  430.0000 56364.6048586107572     1.3994278614776

19   9 0329+54  430.0000 56364.6465257734599     1.3994277048320

19  10 0329+54  430.0000 56364.6881929410083     1.3994275378165

19  11 0329+54  430.0000 56364.7298601135481     1.3994273709034

19  12 0329+54  430.0000 56364.7715272909336     1.3994272145651

19  13 0329+54  430.0000 56364.8131862021037     1.3994270792981

19  14 0329+54  430.0000 56364.8548533874346     1.3994269755191

19  15 0329+54  430.0000 56364.8965205752611     1.3994269137297

19  16 0329+54  430.0000 56364.9381877641717     1.3994269044025

19  17 0329+54  430.0000 56364.9798549524494     1.3994269580095

19  18 0329+54  430.0000 56365.0215221380640     1.3994270850233

mes10: 

19   1 0329+54  430.0000 56364.3131885291659     1.3994277879689

mes10c: 

1.3994277879689

0.714577778573

mes11: 

19   2 0329+54  430.0000 56364.3548556884271     1.3994279971364

[Steve Olney - VK2XV]

G'day Michiel,

To use the Fo given in the output you need to find the relevant MJD when the pulsar in question is passing through your antenna beam rather than deduce from the Doppler shift alone.   This dependent on your location, your beam heading and the RA of the pulsar.

Cheers

Steve

[Steve Berl]

I'm new to this. Is that 1.3994278614776 Hz? Or Seconds? Either way, what is the clock source that drives the sampling and how stable is it? Seems like all those decimal places will get messed up by even a tiny clock jitter. 

Steve

You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
-steve

[Michiel Klaassen]

Hi All,

For a project I am stationed now in portugal for 3 months , so now I have  very very minimal equipment.

In the last message I mentioned a sampling of 3.2MS/s but that is intended for the detection of pulsars. For the H1 a rate of 2.024MS/s was used; resulting in a BW of 1 MHz.

Attached is a presentation of thesame data in a contour map.

It can be seen that the beam of the 60cm satellite dish passes the galaxy rim twice.

More clearly than in the line presentation it can be seen that the top and the bottom of the graph lines match.

Some extra info of the way this map is made can be given.

1. Every spectrum line is treated so that the spikes in the FFT outcome are filterd. The reason is that these spikes are not needed in the end presentation. It is of no use to avarage the data, because the line then will be elevated always. Best is to  find the lowest point in the noise line and connect this to the lowest next value trough linear interpolation. The bridging distance of the gap can be adjusted to get a smooth line. Result; no vertical lines in our spectrum.

2. The band pass curve of the total system is not flat, so some 25% of the data is lost. To compensate that, a correction factor over the BW is applied. This is actually a noise measurement over the total BW without any external signal present. Now this bandwidth spectrum is multiplied as 1/BW curve with every spectrum line. Here we use two "silent" spectra for this BW correction factor. Result; (almost) no BW amplitude variation in our spectrum.

3. We want to display in the contour map just the frequency dependent variations; so we do not want to display broadband sources because it will corrupt the H1 curve too much. To correct that we measure the enerby content in the first 10 bins of the 1024 bins and compare it to a fixed value. Next we sum or subtract that value from/to every bin. Result; (almost) no horizontal lines in our spectrum.

Finally some general advise for people who want to get these or better results.

1 Trim your can type horn probe with a directional coupler (+sign. gen/receiver) and take at least 2 days time to get the best results. See page 85 to 90 of the book "The Radio Sky" by Jeff Lashley.

2 Position your dish at least 5 meters from your dongle and PC; do not face the dish or the horn to your PC (or TV set)

3. Position your horn front at thesame distance centre and at thesame front angle as the removed LNB.

4 Add amplifiers to your system so for the digital output value you get the full range (1...254).

5. Do not expect results on a rainy/snowy day.

Michiel

---------- Forwarded message ----------
From: Jonathan Rawlinson <jonatha...@gmail.com>
Date: 2013/3/24
Subject: Re: [SARA] Detecting H1 and pulsars with minimal equipment.

To: Michiel Klaassen <vmin...@gmail.com>

Hello Michiel

Sorry about the slow reply, schoolwork seems to get in the way of my hobbies

Many thanks for the comprehensive reply, I enjoyed reading it!

I have had some experience in computer programming but not much in this area(I.e. sdr) and I am still learning the 'basics' behind how sdrs work. I am thinking about trying your approach but in gnu radio (with a rtlsdr), dumping the data out as fast as possible and post processing. I have a little spare time coming up and will try to implement it then. 

I am still trying to understand the sdr data format, are there any good sources of info for this?

I am also modifying a feed design for my antenna. Hopefully it will work!

I will keep you posted

Many thanks for your time 

Best regards 

Jonathan - M0ZJO

Sent from my iPad

On 21 Mar 2013, at 14:44, Michiel Klaassen <vmin...@gmail.com> wrote:

Hi Jonathan,

Of course you can ask; thats the reason why we all are here.

I started with an even more simple sampler; the Velleman digital oscilloscope PCS500. 

With that system I learned a lot about sampling, and why excel is too slow to use. See attachment.

A portugese friend of mine; Ricardo Gama advised me to use PythonXY with the included Spyder editor. 

This was really fast but can be simply written like Basic.

So now all my sw is written in Python.

At the moment I use the DVB-T+DAB+FM usb 2.0 dongle. To initiate it I use SDT# (sharp) v1001066.

In the recording mode you can get data streams out. Important is that you use the entire range of the a/d converters.

So in the output the maximum values have to be from 1 to 254 (idealy). 

To do that, you have to increase your gain from the antenna with amplifiers suitable for 1420MHz.(total about 50dB)

Also set in the configure screen the RTL AGC to "on"(checked), and set the gain slide to max.

Further I use the quadrature sampling setting and 3.2MS/s. In my dongle I had to set the freq correction to 58.

I do not use the recording mode to dump the data, but perhaps you can try it. 

It gives you a lot of data for a short time, and you have to start it manually every time.

Also the files are getting a different name which you have to alter manually.

That is why I decided to do the capturing in python. 

With the usb support files from python the usb port is "sniffed". 

Because the received H1 line BW is about 1MHz, and when sampling 3.2MS/s gives you 1,6MHz, it is some waste of BW.

Still we want to use all the samples, so I split the quadrature array into two parts; I call them the even and odd part.

Now on both of them the FFT function is done and after that the results are summed.

Here is the most important part:

data_odd= [data_all[i] for i in range(4096) if i % 2 == 0]#odd    

data_even= [data_all[i] for i in range(4096) if i % 2 == 1]#even

Pfft = 1*abs(np.fft.rfft(data_even)) 

Mfft = 1*abs(np.fft.rfft(data_odd))

Callsets[sets-1,:]=Pfft+Mfft

The script is repeating that as fast as possible. There are even faster FFT methods (FFTW), but I have not tried that yet.

So in 15 minutes some 300000 FFT outputs are avaraged and the resulting spectrum written to disk. 

The system is now performing thesame measurements over 24H.

After that you can begin with the plotting of the results; also in python.

I did not mention the antenna because you have already detected the sun, but remember to optimize/trim the antenna for 1420MHz.

Michiel

2013/3/20 Jonathan Rawlinson <jonatha...@gmail.com>

Hi Michiel, sorry to bother you

I (like you) am attempting to do 'proper' radio astronomy with minimalistic equipment

I am a budding new radio astronomer and have been trying to get a hydrogen line profile for a long time but with little success so far. I have acquired a 1.2 M dish and have done some successful solar observations but have been unable to acquire any signals from outside the solar system . 

I am very interested with the images that you posted on the 6th with a 60cm dish. Would you mind elaborating on what equipment you are using and how? Also how are you scanning the range of frequencies, I assume that there is some SDR software that will do it for you? 

I have a few RTLSDR dongles and a Funcube dongle pro+ and have not had any luck with any of them so far.

Many thanks for your time

Jonathan - M0ZJO (17)

<Measuring 21cm03.pdf>

[Marcus D. Leech]

I'd like more details on your feed arrangement, since I'm going down the
same road as you, but with a larger dish (1M), but still an offset-fed
type satellite dish.

Also, if you're using an RTLSDR, it produces *complex* samples, which
means the bandwidth == sample_rate. So if it's delivering 2Msps, you're
seeing 2Msps (complex) bandwidth centered about DC. The I/Q samples
*together* give you information, so you should compute a
complex FFT on the complete set of I/Q samples.

My simple_ra application does this.

But I'm having poor results, likely due to local noise and too much
ground noise getting into my feed, so I'd like more information about your
instrumental setup.

[Michiel Klaassen]

Marcus

Can you send me some 50 (bin) spectra you have recorded, with also a galaxy plane transition.

Also give me the format of the file. 

I will analyse it and give you an answer.

Michiel

2013/4/2 Marcus D. Leech <patchv...@gmail.com>

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to

sara-list-unsubscribe@googlegroups.com

For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
--- You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.

To unsubscribe from this group and all its topics, send an email to sara-list+unsubscribe@googlegroups.com.

[Marcus D. Leech]

> Marcus
> Can you send me some 50 (bin) spectra you have recorded, with also a
> galaxy plane transition.
> Also give me the format of the file.
> I will analyse it and give you an answer.
> Michiel
>

The attached file should contain some galactic plane spectral data,
around 20:00 or so LMST, at a declination of roughly 30deg.

The PARAMS line gives data for the observation: Fc, pre-detector
bandwidth (that's a bug, it should just be the 1.5Mhz of spectral
bandwidth), and the
declination of the observation.

Lines with times on them have UTC first, then LMST, then the spectral
data between [ and ] in dB, from -bw/2 to +bw/2. My FFT runs 4096 points.

[Michiel Klaassen]

Marcus,

The number of bins/line in your file is 8192, not 4096.

The plot shows no evidence of H1.

The plot shows drift of your bandwidth.

The power graph (power of the energy added of all the bins in a "line") shows no evidence of a passing GP.

There was no reason for me to make a contour map.

My advice to you is to;

Decrease the number of bins; we use 1024.

Decrease the captured (total bin) bandwidth; we use 1Mhz (see our graph; we sample at 2MS/s)

Increase the number of captured and added spectra/sec as a "line"; we use 333spectra/s and in 15 minutes we add 300000 spectra for one "15 minute line".

Use bandwidth correction.

Michiel

2013/4/5 Marcus D. Leech <patchv...@gmail.com>

[Marcus D. Leech]

> Marcus,
> The number of bins/line in your file is 8192, not 4096.
> The plot shows no evidence of H1.
> The plot shows drift of your bandwidth.
> The power graph (power of the energy added of all the bins in a
> "line") shows no evidence of a passing GP.
> There was no reason for me to make a contour map.
>
> My advice to you is to;
> Decrease the number of bins; we use 1024.
> Decrease the captured (total bin) bandwidth; we use 1Mhz (see our
> graph; we sample at 2MS/s)
> Increase the number of captured and added spectra/sec as a "line"; we
> use 333spectra/s and in 15 minutes we add 300000 spectra for one "15
> minute line".
> Use bandwidth correction.
>
> Michiel
>

I'm going to put in spectral flattening sometime this weeknd.

I use spectral integration already. I've used all of this stuff on a
previous radio telescope I had a few years back. It all worked just fine.
The software isn't the issue.

[Michiel Klaassen]

Marcus

What kind if backend do you use hw/sw and how many spectra do you average

 per second or per 30 seconds ("line") in your .dat file.

Attached you find another sesion; however now with a line time of 1 minute, so 330*60; about 20000 spectra averaged.

There is more detail, but consequently more noise.

Our system is only 30% effective in comparation with analog or equivalent systems.

Michiel

2013/4/5 Marcus D. Leech <patchv...@gmail.com>

Marcus,

[Marcus D. Leech]

> Marcus
> What kind if backend do you use hw/sw and how many spectra do you average
> per second or per 30 seconds ("line") in your .dat file.
>

I average spectra at 100% efficiency -- that is, my FFT rate is exactly
sample-rate/FFT-size. The magnitude outputs are then averaged using
a single-pole IIR filter.

This is the same approach I have used since 2005 on my various radio
telescopes, and it has worked well, giving me results that are close to
what theory would predict for such a set-up.

My problems, I'm fairly sure, are hardware related, or perhaps
environment related.

> Attached you find another sesion; however now with a line time of 1
> minute, so 330*60; about 20000 spectra averaged.
> There is more detail, but consequently more noise.
> Our system is only 30% effective in comparation with analog or
> equivalent systems.
> Michiel
>
>

There's no reason a digital system can't be competitive with a
comparable analog system. My previous SDR-based endeavours, when I
lived out in
the middle of nowhere used the same approach and were quite
successful. I had a 1.3m prime-focus dish, and an SDR back-end, and it
worked just
fine, and then I upgraded to a 3.8m dish, and it worked even better.

[Marcus D. Leech]

Marcus

What kind if backend do you use hw/sw and how many spectra do you average

 per second or per 30 seconds ("line") in your .dat file.

Attached you find another sesion; however now with a line time of 1 minute, so 330*60; about 20000 spectra averaged.

There is more detail, but consequently more noise.

Our system is only 30% effective in comparation with analog or equivalent systems.

Michiel

Here's a graph showing a hydrogen line response in the spectrum.




This is, obviously, without any spectral correction.  I'm just playing with that now, I can correct the spectrum in real-time by insert a FFT filter
  into the signal-processing chain, based on a computed correction from "nothing there" spectrum taken at a time when there's nothing--interesting
  in the spectrum.

[Michiel Klaassen]

Marcus,

So your minimal system is working now. What was the problem/solution. Still you did not tell me what hw/sw you used to get your result.

About the bandwidth flattening; you are "playing" with now; see my remarks earlier as an advice for other experimenters.

.......

2. The band pass curve of the total system is not flat, so some 25% of the data is lost. To compensate that, a correction factor over the BW is applied. This is actually a noise measurement over the total BW without any external signal present. Now this bandwidth spectrum is multiplied as 1/BW curve with every spectrum line. Here we use two "silent" spectra for this BW correction factor. Result; (almost) no BW amplitude variation in our spectrum.

.......

I corrected your previous offered data also with thesama idea.

So I do not see a reason for you to repeat it/present it as your invention.

Well, if you have more data, I can make a contour plot for it.

However, I hope that we still are talking about this subject; "minimal systems" (intended for school programs)

Michiel

2013/4/7 Marcus D. Leech <patchv...@gmail.com>

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to

sara-list-...@googlegroups.com

For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/sMqCmg-5O_Q/unsubscribe?hl=en.

To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.

[Marcus D. Leech]

> Marcus,
> So your minimal system is working now. What was the problem/solution.
> Still you did not tell me what hw/sw you used to get your result.

I'm using RTLSDR, with Gnu Radio based software, called "simple_ra" that
I'm the author of.

If you do a search on my name in the context of Radio Astronomy and
SETI, you'll see that I have been doing SDR-based radio astronomy
observing for many years. Previously, in a much-quieter environment,
with a much-larger dish. I also am the P.I. for SBRAC observatory,
which is a 17m dish located near Ottawa, Ontario. I'm no stranger to
SDR and Radio Astronomy, I just haven't had to do it with a tiny
dish in a suburban back-yard before.

>
> About the bandwidth flattening; you are "playing" with now; see my
> remarks earlier as an advice for other experimenters.
>
> .......
> 2. The band pass curve of the total system is not flat, so some 25% of
> the data is lost. To compensate that, a correction factor over the BW
> is applied. This is actually a noise measurement over the total BW
> without any external signal present. Now this bandwidth spectrum is
> multiplied as 1/BW curve with every spectrum line. Here we use two
> "silent" spectra for this BW correction factor. Result; (almost) no BW
> amplitude variation in our spectrum.
> .......
>
> I corrected your previous offered data also with thesama idea.
> So I do not see a reason for you to repeat it/present it as your
> invention.

Um, actually, I've been doing spectral flattening in other software I
wrote that I presented back in 2006 at the SARA meeting, and the idea
for that
came from much-earlier papers on the subject. I'm not trying to
"impinge on anybodies territory". I thought this was a cooperative
organization
where we shared ideas, and expanded on them.

>
> Well, if you have more data, I can make a contour plot for it.
> However, I hope that we still are talking about this subject; "minimal
> systems" (intended for school programs)
> Michiel
>
>

[Jonathan Rawlinson]

Hi Marcus,

I am very interested in using your simple_ra software although not sure how to get it running on my end
Would you mind giving a quick run through on it, I.e. where to download it, where to install it, which permissions to change and how to use it??

My laptop is setting up Gnuradio (which I assume that I need) at this very moment using the script on this page:
http://superkuh.com/rtlsdr.html

Is this correct thing to do?

Many thanks
Jon

[Marcus D. Leech]

Yes, use the build-gnuradio script, once that runs to completion, you
can install the simple_ra software:

svn co https://www.cgran.org/svn/projects/simple_ra

Then:

cd simple_ra/trunk
make
make install

There's a README file that you should look over. It's slightly
out-of-date with respect to the post-processing tools.
Must fix that.



--

[Michiel Klaassen]

Hi All
Because I noticed that some people had difficulty in using python or
linux, I have transferred my python script; "capture and fast fourier
radio astronomy data" to an exe file.
The link to download the zip file is here
https://mail.google.com/mail/u/0/h/1fkfqlmc720du/?&v=c&th=13e4af555ea492bc

The user instructions are also in the zip file.
The file is posted on the newly started CAMRAS forum; the meeting
place of radio amateurs and amateur radio astronomers of the
netherlands. Our main instrument is the famous 25m Dwingeloo radio
telescope, which is undergoing a restoration now.
Just use the translate button to understand the comments.
In schort; the program uses a dongle and is initiated with sharp#.
When all things are set to your content, then kill the sharp# with the
taskmanager.
The dongle is still sending its data, which now can be read by the
cfrad program.
Every 5 minutes 92550 spectra are added and the result spectrum is
wriiten to you HD.


I hope this is of some help.
Michiel

2013/4/9, Marcus D. Leech <patchv...@gmail.com>: