new SDR 2 channel maybe coherent?

[Hamish Barker]

The antsdr200 sdr transceiver is about to be released. Supposedly 2 channel MIMO, 20 MHz bandwidth, same chip as ettus 205mini. I don't know if the 2 channels are coherent or not. if they are, the minimum $369 sounds competitive and might be good for interferometry and/or pulsar observing.

https://www.crowdsupply.com/microphase-technology/antsdr-e200#products

Cheers,

Hamish

[Marcus D. Leech]

The AD936x family are inherently two-channel coherent, because there's only a single LO in each direction.   So the two
  RX channels share an LO, as do the two TX channels.

Now, having said that, if the FPGA code that processes the two channels isn't careful, there can be digitally-induced
  phase offsets or even loss of coherence, depending on how bad the implementation is.  But if they claim MIMO as
  a goal, then they will have maintained coherence.

The AD9363 offers a smaller frequency range than the AD9361.  The 9361 can directly tune from 70Mhz to 6GHz, whereas
  the AD9363 is from around 300Mhz to around 3.8GHz.

--
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/26e3aafa-6e17-42af-9bad-afae22030795n%40googlegroups.com.

[Dave Typinski]

Looks very nice -- but has a 70 MHz or 325 MHz lower cutoff depending on the
model. Unfortunately not useful for HF band stuff until one adds an upconverter
ahead of it.
--
Dave

> --
> --
> 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

> <mailto:sara-list+...@googlegroups.com>.

> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sara-list/26e3aafa-6e17-42af-9bad-afae22030795n%40googlegroups.com

> <https://groups.google.com/d/msgid/sara-list/26e3aafa-6e17-42af-9bad-afae22030795n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Marcus D. Leech]

On 19/09/2023 16:55, Dave Typinski wrote:
> Looks very nice -- but has a 70 MHz or 325 MHz lower cutoff depending
> on the model.  Unfortunately not useful for HF band stuff until one
> adds an upconverter ahead of it.
> --
> Dave

Most SDRs in this general category don't do HF.   Because the RFFE chips
they're based on are optimized for telecom
  (cellular, etc) and public-service (VHF, UHF) radio applications.  
Mixers that work OK at DC-30MHz generally don't do
  well as you move up through VHF and UHF.   Modern PLL synthesizers
don't have as much of a problem here--they're
  usually based on a VFO that operates over a small range (2:1
typically) but at eye-watering frequencies (several GHz),
  and then various stages of divider are used to bring the LO frequency
down to the target frequency.  It's easy to see
  that you could just add more divider stages, but then you end up with
mixer issues.   That's not insurmountable--you could
  have different mixers switched-in for different frequency bands. But
that adds cost to the chip, and since there's much much
  less of a market in the commercial world for HF stuff, the ADIs and
LIMEs of the world just don't bother.

[Dave Typinski]

On 9/19/23 17:04, Marcus D. Leech wrote:
>
> Most SDRs in this general category don't do HF.

Indeed. There's a large product availability gap between inexpensive sound card
digitizers (up to 192 kHz) and inexpensive SDRs (30+ MHz). Apparently not
enough people want to play in the LF/MF/HF bands.
--
Dave

[Marcus D. Leech]

On 19/09/2023 21:42, Dave Typinski wrote:
>
> Indeed.  There's a large product availability gap between inexpensive
> sound card digitizers (up to 192 kHz) and inexpensive SDRs (30+ MHz). 
> Apparently not enough people want to play in the LF/MF/HF bands.
> --
> Dave
>

I'll point out that the available science bandwidths in HF are small.   
Although being able to bring in the entirety of
  the HF band into your system, so that it can "do stuff" is appealing...

[Dave Typinski]

On 9/19/23 21:46, Marcus D. Leech wrote:
>
> I'll point out that the available science bandwidths in HF are small. Although
> being able to bring in the entirety of
> the HF band into your system, so that it can "do stuff" is appealing...

Depends on the science. By law there are only a few narrow bandwidths here and
there supposedly reserved for RA. Nevertheless, if one is doing Jovian DAM
work, or solar work, or terrestrial propagation work, the whole HF band is
useful -- or at least the upper half of it in the case of planetary and solar
emission.
--
Dave

[Marcus D. Leech]

15MHz (or thereabouts) can be brought into a PC and get "churned on"
without too much trouble these days.    A LimeSDR mini
  prefixed with a decent upconverter (like the SpyVerter -- I've used
that for the WWV phase distortion observations) would work
  just fine I think.

[Dave Typinski]

On 9/19/23 22:34, Marcus D. Leech wrote:
>
> 15MHz (or thereabouts) can be brought into a PC and get "churned on" without too
> much trouble these days. A LimeSDR mini
> prefixed with a decent upconverter (like the SpyVerter -- I've used that for
> the WWV phase distortion observations) would work
> just fine I think.

It might work, but "just fine" is a little iffy depending on the OS, the USB
hardware, and the DSP software.

We've been dreaming of a software defined radiometer for a while. Apparently
it's much easier said than done. Block diagram PDF attached.

Some semi-thought-out prelim desired specs are here:

https://radiojove.net/SUG/Pubs/Radio%20JOVE%20SUG%20SDR%20Design%20Specifications%202019%2010%2024,%20Dodd,%20et%20al.pdf

--
Dave

[Marcus D. Leech]

I'm guessing that the other "unspoken" requirement is low cost.

Because something like a USRP X310 with a Basic_RX card would, I think,
satisfy most of your requirements.  I don't think
  you'd need hardware phase shifters to form the circular polarizations.

NO USB involved at all, and the X310 supports 10Gbit fiber connection to
its host where you can do all the DSP, or you
  can learn the RFNoC framework, and do some of the DSP on the X310 itself.

But, again, I suspect a large unspoken requirement is cost...

[Lester Veenstra]

Actually there are a lot that do HF

Lester B Veenstra K1YCM MØYCM W8YCM 6Y6Y W8YCM/6Y 6Y8LV (Reformed USNSG CTM1)
les...@veenstras.com

452 Stable Ln
Keyser WV 26726 USA

GPS: 39.336826 N 78.982287 W (Google)
GPS: 39.33682 N 78.9823741 W (GPSDO)


Telephones:
Home: +1-304-289-6057
US cell +1-304-790-9192
Jamaica cell: +1-876-456-8898

--
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/650A4E23.9020207%40typnet.net.

[Jim Abshier]

30 MHz? That's base band... But seriously, a few years ago there was an
article in QST on how to build an up-converter. I built one of these and
used it with an RTL SDR unit. It seemed to work ok. Then I discovered
that my ICOM R71 A HF receiver has a 70 MHz IF output that can be
digitized directly with the RTL SDR. So there are some options available
for digitizing HF with an inexpensive SDR.

Jim Abshier

[Captain Anne Flint]

Thanks for the block diagram - Wende

> --
> --
> 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.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/650A65F9.7080306%40typnet.net.
> <Software Defined Radiometer (SDRm) - Block Diagram 2019 07 14.pdf>

[Dave Typinski]

On 9/19/23 23:42, Marcus D. Leech wrote:
>
> I'm guessing that the other "unspoken" requirement is low cost.
>
> Because something like a USRP X310 with a Basic_RX card would, I think, satisfy
> most of your requirements. I don't think
> you'd need hardware phase shifters to form the circular polarizations.
>
> NO USB involved at all, and the X310 supports 10Gbit fiber connection to its
> host where you can do all the DSP, or you
> can learn the RFNoC framework, and do some of the DSP on the X310 itself.
>
> But, again, I suspect a large unspoken requirement is cost...

Lower than the $15k they want for a dual-receiver setup, yes. Under $5k would
be nice and $2500 would be better. Not exactly "low" -- but certainly lower
than what seems to be on offer.
--
Dave

[Dave Typinski]

Can you get a 15 to 20 MHz wide slice of spectrum from the R71's IF output?
Say, 15 to 30 MHz, or 14 to 34 MHz? That seems awfully wide for an IF output in
the HF band and a lot wider than an RTL-SDR can handle.
--
Dave

[Marcus D. Leech]

On 20/09/2023 12:31, Jim Abshier wrote:
> 30 MHz? That's base band... But seriously, a few years ago there was
> an article in QST on how to build an up-converter. I built one of
> these and used it with an RTL SDR unit. It seemed to work ok. Then I
> discovered that my ICOM R71 A HF receiver has a 70 MHz IF output that
> can be digitized directly with the RTL SDR. So there are some options
> available for digitizing HF with an inexpensive SDR.
>
> Jim Abshier

One of the listed requirements that Dave has is "no up/downconversion"  
antenna-->RF gain--->ADC is preferrable.

[Nathan]

Sorry Marcus, I did not get this message sent to the list the first time.

I want to mention the RX888mk2 receiver.  It can do direct sampling at
120 MHz (60 MHz analog bandwidth), and costs about $200 US.  I have one
running with good results.  A souped up computer is not necessary.

The receiver also has an internal connector for an external clock that
can be brought out to an SMA connector.  Two such units driven by one
clock can synchronously sample two linear polarizations.  From their
complex Fourier transforms, one can compute the four Stokes parameters. 
Such a configuration would really be something for Jupiter and solar
observing.

So I think this receiver has a lot of potential.

Nathan

[Jim Abshier]

Dave - Attached is a diagram showing the R71 A RF unit. As you can see,
the mixer is preceded by a bank of filters. The IF output is taken
directly from the output of the mixer so the bandwidth available is
limited only by whichever input filter is used. There are filters
covering 15 to 22 MHz and 22 to 30 MHz but not continuously from 15 to
30 MHz. Also note that the output spectrum is inverted in frequency.

Jim Abshier

[Dave Typinski]

Hi Nathan,

Thanks for that! Didn't know a Mk II unit was available. That's a pretty
capable unit for the price of $160. The spurs aren't great, but for $160, what
does one expect.

https://elekitsorparts.com/some-truths-about-rx-888-mk2-sdr/

The idea is to be able to do various amounts of DSP (all the way through
integrated spectra if desired) in an FPGA before the data gets sent to whatever
recording software is being used. That way one could use a Ras Pi or other
small single board computer to do nothing but record the data to an SD card for
long term operation in remote areas where AC power -- and hopefully RFI -- are
unavailable.

The original open source design by IK1XPV (the BBRF103) might be a good starting
point for the front end and ADCs for such a software defined radiometer.
--
Dave

[Dave Typinski]

Hi Jim,

High side injection vs low side injection -- the freq inversion is just math
and/or plot labeling. Easy to handle. :)

I'm surprised the IF output BW is as large as it is. Neat info, thanks!
--
Dave

[djl]

Are we talking about the same SDR?
https://www.rtl-sdr.com/highlighting-major-software-issues-with-the-rx888-sdr/

> --

------------
"It's always something."
Roseanne Rosannadanna
----------------------"
Dr. Don Latham AJ7LL
PO Box 404, Frenchtown, MT, 59834
VOX: 406-626-4304

[Nathan]

Hi Dave,

Good luck with the FPGA approach.  As you know I was a proponent as well in past years with my 2015 receiver in logic and that it requires a modest computer and USB 2 interface.  But in the last 15 years or so the development tools and intellectual property have gotten much more complicated and I have since given up on logic.  I am not trying to discourage logic.  It is the way to go when the pieces can be put together. 

I do think that your comment about spurs in the rx888 overstates the problem.  I've attached a spectrum from my receiver taken with the a load on the input.  There are a few intense points out of a 2048 point spectrum, while others are small and can be removed in background subtraction.  Your eventual receiver in logic may or may not be as good - performance depends on many things.

You can see some dynamic spectra from the new receiver at this link.  Scroll to the bottom and open files with the pattern *towne26.png.  I am still working on controlling electrical noise from the equipment, so don't expect perfection.

It is true that the wider bandwidths require more electrical power and the PV panels and batteries get bigger.  I don't know what will be required so that my remote system will be able to run for longer periods with the new hardware and software.  I am currently running less than a day at a time requiring frequent trips to the site with recharged batteries.

For reference, the computer requires 37 W to run at 120 MHz, compared with 5 W for an RPI 4B and an SDRPlay at 8 MHz.  These figures are consistent at about 0.65 W/MHz analog bandwidth.  Compare with logic when you have something running.

Nathan

On 9/21/23 9:14 AM, Dave Typinski wrote:

[Marcus D. Leech]

On 21/09/2023 15:50, Nathan wrote:

Hi Dave,

Good luck with the FPGA approach.  As you know I was a proponent as well in past years with my 2015 receiver in logic and that it requires a modest computer and USB 2 interface.  But in the last 15 years or so the development tools and intellectual property have gotten much more complicated and I have since given up on logic.  I am not trying to discourage logic.  It is the way to go when the pieces can be put together. 

I do think that your comment about spurs in the rx888 overstates the problem.  I've attached a spectrum from my receiver taken with the a load on the input.  There are a few intense points out of a 2048 point spectrum, while others are small and can be removed in background subtraction.  Your eventual receiver in logic may or may not be as good - performance depends on many things.

You can see some dynamic spectra from the new receiver at this link.  Scroll to the bottom and open files with the pattern *towne26.png.  I am still working on controlling electrical noise from the equipment, so don't expect perfection.

It is true that the wider bandwidths require more electrical power and the PV panels and batteries get bigger.  I don't know what will be required so that my remote system will be able to run for longer periods with the new hardware and software.  I am currently running less than a day at a time requiring frequent trips to the site with recharged batteries.

For reference, the computer requires 37 W to run at 120 MHz, compared with 5 W for an RPI 4B and an SDRPlay at 8 MHz.  These figures are consistent at about 0.65 W/MHz analog bandwidth.  Compare with logic when you have something running.

Nathan

The question to ask oneself about spurs is the degree to which "science information" is lost if you just zero them out in the
  FFT?

I have a riometer being deployed to antarctica, and it uses a number of techniques for excising both narrow-band RFI spurs,
  and sudden changes in apparent total power.    Spurs in this case would just be considered persistent RFI bins, and
  excised.  The quality of the "output science" remains quite high in this case.

If the spurs are "loud" enough that they're noticeably eating into dynamic range, then, yeah, you have a problem.

On 9/21/23 9:14 AM, Dave Typinski wrote:

Hi Nathan,

Thanks for that!  Didn't know a Mk II unit was available.  That's a pretty capable unit for the price of $160.  The spurs aren't great, but for $160, what does one expect.

https://elekitsorparts.com/some-truths-about-rx-888-mk2-sdr/

The idea is to be able to do various amounts of DSP (all the way through integrated spectra if desired) in an FPGA before the data gets sent to whatever recording software is being used.  That way one could use a Ras Pi or other small single board computer to do nothing but record the data to an SD card for long term operation in remote areas where AC power -- and hopefully RFI -- are unavailable.

The original open source design by IK1XPV (the BBRF103) might be a good starting point for the front end and ADCs for such a software defined radiometer.
--
Dave

--
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/c0c7110b-b216-f371-a28a-9b00b3954f34%40ownmail.net.

[Nathan]

Marcus,

I did not want to be long winded in that post.  I am rephrasing some of
what you said: little information is lost if there are few spectral bins
lost out of many.  On one hand, intense lines destroy essentially all
information in bins they cover.  But on the other hand, weak lines of
constant (in time) intensity can have their intensities subtracted from
the bins they cover with partial loss of their information related to
the new signal-to-noise ratios of those bins (through the noisy Tsys and
the radiometer equation).

That is why I think that the noise of the rx888 I measured is not
particularly severe.

In the same spirit, there can be noise of constant (in time) and modest
intensity that covers a band.  In such a case the noise does not
completely extinguish the information in the band, but similarly reduces
the information through the reduced the S/N ratio.  In such cases
subtraction of an estimate of that noise intensity can provide useful
spectra across the band.

This is often the problem with the electronic noise I am contending
with, which illustrates more clearly the utility of background subtraction.

Nathan

On 9/21/23 2:08 PM, Marcus D. Leech wrote:
> On 21/09/2023 15:50, Nathan wrote:
>>
>> Hi Dave,
>>

>> ...

>>
>> I do think that your comment about spurs in the rx888 overstates the
>> problem.  I've attached a spectrum from my receiver taken with the a
>> load on the input.  There are a few intense points out of a 2048
>> point spectrum, while others are small and can be removed in
>> background subtraction.  Your eventual receiver in logic may or may
>> not be as good - performance depends on many things.
>>

>> ...

[Dave Typinski]

Not having spurs is nice. However, for a lot of science, they're not a
dealbreaker. As you said, it's difficult unto impossible to get *no* spurs.
--
Dave

[Dave Typinski]

Hi Nathan,

Those dynamic spectra look great! If there are spurs there, I can't tell them
from radio stations.

Beats me how hard it would be to fab a software defined radiometer using an
FPGA. At present I have neither the time to build one, wishful thinking
notwithstanding, nor the desire to drop $15k on a receiver when the RFI here
gets worse every year.

It occurs to me that building a software defined radiometer from scratch would
surely cost more than $15k in material, IP licenses, effort, and especially
/time/. Which is why I keep on waiting for someone else to do it.

Using four RPi's and four SDRPlay receivers to cover RCP and LCP over the upper
half of the HF band has long been an idea... and may in fact be the most cost
effective way to do it, even if the lower & upper 8 MHz bands don't marry
together perfectly in amplitude. The desirable lower power requirements would
offset that somewhat.
--
Dave

On 9/21/23 15:50, Nathan wrote:
> Hi Dave,
>
> Good luck with the FPGA approach. As you know I was a proponent as well in past
> years with my 2015 receiver in logic and that it requires a modest computer and
> USB 2 interface. But in the last 15 years or so the development tools and
> intellectual property have gotten much more complicated and I have since given
> up on logic. I am not trying to discourage logic. It is the way to go when the
> pieces can be put together.
>
> I do think that your comment about spurs in the rx888 overstates the problem.
> I've attached a spectrum from my receiver taken with the a load on the input.
> There are a few intense points out of a 2048 point spectrum, while others are
> small and can be removed in background subtraction. Your eventual receiver in
> logic may or may not be as good - performance depends on many things.
>
> You can see some dynamic spectra from the new receiver at this link

> <http://towne56.ownmail.net/dynamicspectra/Sun/>. Scroll to the bottom and open

> --
> --
> 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
> <mailto:sara-list+...@googlegroups.com>.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/sara-list/c0c7110b-b216-f371-a28a-9b00b3954f34%40ownmail.net
> <https://groups.google.com/d/msgid/sara-list/c0c7110b-b216-f371-a28a-9b00b3954f34%40ownmail.net?utm_medium=email&utm_source=footer>.

[Nathan]

Dave,

There are definitely spurs in the dynamic spectra you viewed. But in my
opinion there are more in the SDRPlay receivers.  A few are intense, but
there are many more (I think) at much lower intensities that become
visible when the contrast is stretched.

Given the bandwidth you are aiming for, the power consumption issue is
not that different between four SDRPlay/RPis and the Dell Optiplex 3050
I am using.  Assuming a 70 MHz sample rate for 35 MHz maximum frequency,
the Dell's power consumption is about 26 W (see the graph).  Compare
this figure with four RPis at 5 W each = 20 W.

One other possibility is to run the rx888 on a single RPi 4B at 60 MHz
(second attachment).  This would not meet your spec at only 30 MHz
maximum analog frequency, but it might be an interesting option for some
users and remote sites on a budget. I could measure the power
consumption if you are thinking about it.

Both of these options would avoid the spectra-splicing clumsiness.

Nathan

On 9/21/23 8:11 PM, Dave Typinski wrote:
> Hi Nathan,
>

[Dave Typinski]

Hi Nathan,

Have you need able to save integrated spectra for 12 or 24 hours? Or just short
bursts? I'm wondering how USB3 holds up over a day's worth of observation --
especially at the wider bandwidths.

By the way, how are you getting 50 MHz of BW out of the unit?

http://towne56.ownmail.net.user.fm/dynamicspectra/Sun/20230811T162600_towne26.png

This web page says it can only do 32 MHz of BW -- I assume because it doesn't
use complex sampling.

https://elekitsorparts.com/some-truths-about-rx-888-mk2-sdr/
--
Dave

[Marcus D. Leech]

On 22/09/2023 07:57, Dave Typinski wrote:
> Hi Nathan,
>

> Have you need able to save integrated spectra for 12 or 24 hours? Or
> just short bursts?  I'm wondering how USB3 holds up over a day's worth
> of observation -- especially at the wider bandwidths.

I have friends who are using USRP B210 and LImeSDR-USB for long-term
radio astronomy observations, and they run for
  extended periods of a couple of weeks.   The USB on ARM-based SBCs
seems to be flakier than on X86, though.   My riometer
  was originally based on an ARM-based SBC, but when we migrated to
X86, it would just run and run and run forever.

[Nathan]

Dave,

I've been running for several hours at a time limited by battery
charge.  The longest was 15 hours at 120 MHz.  It didn't show any sign
of lost samples.

The rx888mk2 samples at 120 MHz, real samples.  It is the older model
that ran more slowly.

Nathan

On 9/22/23 5:57 AM, Dave Typinski wrote:
> Hi Nathan,
>

[Dave Typinski]

Thanks Nathan. This is looking more attractive by the minute.
--
Dave