Date   

Re: CPU load on single board Linux computers with 12 bit SDR? #hardware #spyserver

Marcus D. Leech <mleech@...>
 

On 11/21/2017 05:21 PM, prog wrote:
On Tue, Nov 21, 2017 at 02:06 pm, Jeff Kelly wrote:
It does burp a little with the 8 mhz bandwidth of the airspy. so wouldn’t recommend that.
The Odroid XU4 has a slight edge on this use case.


Re: CPU load on single board Linux computers with 12 bit SDR? #hardware #spyserver

prog
 

On Tue, Nov 21, 2017 at 02:06 pm, Jeff Kelly wrote:
It does burp a little with the 8 mhz bandwidth of the airspy. so wouldn’t recommend that.
The Odroid XU4 has a slight edge on this use case.


moderated Re: EXTIO File for HF+ ?

prog
 

On Tue, Nov 21, 2017 at 02:12 pm, Patrick wrote:
I simply wish the combo SDR# HF+ is so great that I won't need other soft for these purposes.
 
Best regards
Patrick
Well. Make it great then! Learn how to use the plethora of plugins available to you - starting from the Aux VFO.
But of course, as long as you are unable to see DSP flaws in software, your perception of "great" will be flawed as well.
Unfortunately, we cannot spend time on producing quality DSP and educate everybody at the same time.


moderated Re: EXTIO File for HF+ ?

Patrick
 

I simply wish the combo SDR# HF+ is so great that I won't need other soft for these purposes.

Best regards
Patrick

2017-11-21 23:06 GMT+01:00 prog <info@...>:

On Tue, Nov 21, 2017 at 02:05 pm, Patrick wrote:
As a hardcore NDB & MW DXer, I also hope to use my HF+ with Studio 1 because of its very sharp CW filters and superior waterfall readability.
I made some tests using SDR# 1626 on the NDB band (where several beacons can be found within a few Hz) and found it quite poor. For MW carrier monitoring, I also found resolution disappointing (recorded files were used)
 
Will HF+ and SDR# allow to clearly see several carriers within 10 Hz and separate a couple of beacons within  a few Hz too ? I hope so !
 
Patrick
If you don't mind aliases (ie. signals that aren't there) you can use whatever software you wish.



moderated Re: EXTIO File for HF+ ?

prog
 

On Tue, Nov 21, 2017 at 02:05 pm, Patrick wrote:
As a hardcore NDB & MW DXer, I also hope to use my HF+ with Studio 1 because of its very sharp CW filters and superior waterfall readability.
I made some tests using SDR# 1626 on the NDB band (where several beacons can be found within a few Hz) and found it quite poor. For MW carrier monitoring, I also found resolution disappointing (recorded files were used)
 
Will HF+ and SDR# allow to clearly see several carriers within 10 Hz and separate a couple of beacons within  a few Hz too ? I hope so !
 
Patrick
If you don't mind aliases (ie. signals that aren't there) you can use whatever software you wish.


Re: CPU load on single board Linux computers with 12 bit SDR? #hardware #spyserver

Jeff Kelly
 

I am running the Asus Tinker board…one with an Airspy and one with a RTLSDR V3
It handles the 2.5 MHZ Airspy bandwidth and the 2.4 RTLSDR fine.  And I think 10 users would be fine.

It does burp a little with the 8 mhz bandwidth of the airspy. so wouldn’t recommend that.

I have 4 tinker boards stacked for projects and have pretty much put the RPI’s in the drawer! Love them.

Tinkers...




Jeff
K2SDR





On Nov 21, 2017, at 4:33 PM, weird_joe@... wrote:

Hi Everybody,

I have a few questions on the real life situation with mini computers and the Spyserver.
- I'd like to know how the CPU load of e.g. a Raspberry Pi 3 is when running the Spyserver with a device like the Airspy or Airspy Mini (more than 8 bits)?
- Do I have to reduce the FFT framerate in order to not overload the CPU? 
- What bandwidth can I practically use? Full spectrum? (full 10MHz on an Airspy?) I am really impressed how low the CPU impact on a second connected client is when running it under Windows. 
- Would you recommend a more powerful mini computer, like the Asus Tinker or Odroid octa core with twice as much RAM and much more performance?
- I believe the Asus Tinker has the exact same physical outlines of the Pi3, and there are plenty of inexpensive metal cases that should minimize the RF noise - so I guess that would be an advantage against the Odroid boards which are available with the faster memory modules than the fastest SD cards as far as I have been reading. Any recommendations on that?
- Do you have any recommendations on where to get a low RF-noise 5V 4A power supply for these more powerful boards?
- in case performance on these boards is no big issue, can I also run several Spyserver in separate directories addressing different receivers in parallel in Linux like I currently do on my Windows server?
- I have just been trying to run the Spyserver on my little Raspberry-Pi Zero W with the latest OS (slim version w/o GUI) but the only thing I get a a segmentation fault. The relatively dumb RTL-TCP server runs fine, so the dongle and driver setup works. So far I could not find out what the actual problem is. I have already been reformatting the SD card and put the latest image on.

In advance many thanks for any constructive feedback!


moderated Re: EXTIO File for HF+ ?

Patrick
 

As a hardcore NDB & MW DXer, I also hope to use my HF+ with Studio 1 because of its very sharp CW filters and superior waterfall readability.
I made some tests using SDR# 1626 on the NDB band (where several beacons can be found within a few Hz) and found it quite poor. For MW carrier monitoring, I also found resolution disappointing (recorded files were used)

Will HF+ and SDR# allow to clearly see several carriers within 10 Hz and separate a couple of beacons within  a few Hz too ? I hope so !

Patrick


Re: Direct Sampling frequency limit in r 1628 onwards #sdrsharp #spyserver

prog
 

On Tue, Nov 21, 2017 at 02:02 pm, Martin wrote:
OK, it's your call, and I understand why you may wish to do it for both technical and commercial reasons.
Nope. Correctness first. We don't have an s-meter for the same reason. We have an SNR-Meter.


Re: Direct Sampling frequency limit in r 1628 onwards #sdrsharp #spyserver

Martin - G8JNJ
 

Hi,

OK, it's your call, and I understand why you may wish to do it for both technical and commercial reasons.

As I said previously, there is a Nyquist Alias wrapped around 14.4MHz, but adding this restriction alone still won't completely stop alias signals from above this frequency appearing at lower frequencies. RTL dongles used in this way will always be a compromise, so I've never considered this particular issue to ever be a major problem. Especially as there are other sampling aliases present depending upon the sampling rate chosen in SDR Sharp (and others).

RTL dongles are never going to compete with hardware like Airspy's or HF+'s. But they do provide a good introduction to SDR's and facilitate a lot of experimentation and other applications that may otherwise be too costly to implement by casual or 'hobby' users.

I often use a diplexing filter to minimise this problem when running RTL dongles in direct sampling mode with SDR Sharp (and others), so my plan was to use three dongles connected by a triplexer to cover 0-14.4 / 14.4-28.8 / 28.8 and above.

The suggestion was intended to try and maintain the same functionality of SDR Sharp in both normal and Spyserver mode  wherever possible.

Personally I'd appreciate it if you didn't put this restriction in, but as I said before it's your call.

Regards,

Martin - G8JNJ


Re: Shipping dates - any updates/confirmations? #airspyhfplus

Alberto I2PHD
 

On 11/21/2017 4:18 PM, Joerg Bauerfeld wrote:

 

Anyway I guess none of you has received any confirmation from itead.cc on the order as well, apart from the money transfer via paypal?

You guess wrong...  I received the following order confirmation from Itead :



--
73 Alberto I2PHD
Credo Ut Intelligam



Re: CPU load on single board Linux computers with 12 bit SDR? #hardware #spyserver

prog
 

On Tue, Nov 21, 2017 at 01:35 pm, <weird_joe@...> wrote:
Hi Everybody,

I have a few questions on the real life situation with mini computers and the Spyserver.
Just ask!

- I'd like to know how the CPU load of e.g. a Raspberry Pi 3 is when running the Spyserver with a device like the Airspy or Airspy Mini (more than 8 bits)?
A Raspberry Pi 3 won't be able to handle more than 3MSPS regardless of the bit width. It's not the best platform for SDR for two reasons:
1) Shared bus between USB and the network controller
2) Modest processing power - not suitable for hosting a public server with a lot of clients

- Do I have to reduce the FFT framerate in order to not overload the CPU? 
That's one possible option.

- What bandwidth can I practically use? Full spectrum? (full 10MHz on an Airspy?) I am really impressed how low the CPU impact on a second connected client is when running it under Windows. 
The server is scalable. The FFT is calculated only once for all the clients, then it is sliced and streamed to the clients depending on their current "view". The DDC that tunes within the device's bandwidth uses newly developed algorithms that guarantee very high alias rejection while minimizing the CPU usage per session.

Getting many users on 10MSPS is easy on a TinkerBoard, Odroid XU4 or UpBoard (Atom or Celeron). Just don't expect the same level of performance from a RPi3. Very, very big difference.


- Would you recommend a more powerful mini computer, like the Asus Tinker or Odroid octa core with twice as much RAM and much more performance?
RAM is not a problem. But the CPU's of these beasts are really powerful and more than adequate for server use cases.

- I believe the Asus Tinker has the exact same physical outlines of the Pi3, and there are plenty of inexpensive metal cases that should minimize the RF
If it fits with a fan on the CPU, you should be fine.

noise - so I guess that would be an advantage against the Odroid boards which are available with the faster memory modules than the fastest SD cards as far as I have been reading. Any recommendations on that?
Most of the noise is generated by the PSU, not the board itself. Get a clean PSU, use good cables and move your antenna away from sources of interference. If you still have noise, get a metal enclosure.

- Do you have any recommendations on where to get a low RF-noise 5V 4A power supply for these more powerful boards?
No idea. If you find one, please let me know.

- in case performance on these boards is no big issue, can I also run several Spyserver in separate directories addressing different receivers in parallel in Linux like I currently do on my Windows server?
You can use the same directory with different config files. Just run the executable with the config file name as parameter. Make sure the ports are different.

- I have just been trying to run the Spyserver on my little Raspberry-Pi Zero W with the latest OS (slim version w/o GUI) but the only thing I get a a segmentation fault. The relatively dumb RTL-TCP server runs fine, so the dongle and driver setup works. So far I could not find out what the actual problem is. I have already been reformatting the SD card and put the latest image on.
The code was compiled so it can take advantage of more modern and capable hardware. Pi Zero should not be used as a DSP server anyway.

In advance many thanks for any constructive feedback!

You are welcome. Thank you for asking these questions. This will help many other people.


Re: Shipping dates - any updates/confirmations? #airspyhfplus

Joerg Bauerfeld
 

Thanks for the answer!

Hope it makes it within two weeks shipping, else I will be gone for a week.

 

From: main@airspy.groups.io [mailto:main@airspy.groups.io] On Behalf Of prog
Sent: Dienstag, 21. November 2017 21:20
To: main@airspy.groups.io
Subject: Re: [airspy] Shipping dates #airspyhfplus - any updates/confirmations?

 

On Tue, Nov 21, 2017 at 12:04 pm, Joerg Bauerfeld wrote:

I hope you guys will have happy birthdays including this brilliant gift J  but is there any official shipping date out yet?

 

When I ordered the unit, I hurried because it said units are in stock and then somewhere else pre-order?!

I would also have ordered it when it would have said no stock but available soon – here is your possibility for a pre-order …

I find this a bit misleading but maybe the shopping software is not able to allow orders without showing valid stock?

 

Anyway I guess none of you has received any confirmation from itead.cc on the order as well, apart from the money transfer via paypal?

 

Now I will spend the time before this new RX gets here to have my antennas and preamps in shape to be able to use it to its full potential.

 

Greetings

Joerg DL8EBM / N1VCA

Thank you for your purchase. The shipping starts in the first week of December. I asked the distributors to put a note on their expected delivery dates.
And yes, you will need a good antenna. Not sure you will need the preamps tho.


CPU load on single board Linux computers with 12 bit SDR? #hardware #spyserver

dl8ebm
 

Hi Everybody,

I have a few questions on the real life situation with mini computers and the Spyserver.
- I'd like to know how the CPU load of e.g. a Raspberry Pi 3 is when running the Spyserver with a device like the Airspy or Airspy Mini (more than 8 bits)?
- Do I have to reduce the FFT framerate in order to not overload the CPU? 
- What bandwidth can I practically use? Full spectrum? (full 10MHz on an Airspy?) I am really impressed how low the CPU impact on a second connected client is when running it under Windows. 
- Would you recommend a more powerful mini computer, like the Asus Tinker or Odroid octa core with twice as much RAM and much more performance?
- I believe the Asus Tinker has the exact same physical outlines of the Pi3, and there are plenty of inexpensive metal cases that should minimize the RF noise - so I guess that would be an advantage against the Odroid boards which are available with the faster memory modules than the fastest SD cards as far as I have been reading. Any recommendations on that?
- Do you have any recommendations on where to get a low RF-noise 5V 4A power supply for these more powerful boards?
- in case performance on these boards is no big issue, can I also run several Spyserver in separate directories addressing different receivers in parallel in Linux like I currently do on my Windows server?
- I have just been trying to run the Spyserver on my little Raspberry-Pi Zero W with the latest OS (slim version w/o GUI) but the only thing I get a a segmentation fault. The relatively dumb RTL-TCP server runs fine, so the dongle and driver setup works. So far I could not find out what the actual problem is. I have already been reformatting the SD card and put the latest image on.

In advance many thanks for any constructive feedback!


moderated Re: EXTIO File for HF+ ?

prog
 
Edited

On Tue, Nov 21, 2017 at 01:18 pm, <dx@...> wrote:
Hello, I've seen some discussion of a EXTIO DLL file for HF+ which one enthusiast has written, but evidently a working HF+ in-hand is still needed to finalize the file.

I know that all the early purchasers like myself are looking forward to using their receiver with their favorite programs. Mine are Studio 1, HDSDR, and occasionally the EXTIO version of SDRuno. Does AirSpy plan to release an "official" DLL and perhaps other supporting files specific to the HF+? I know that these things take time and manpower :^)

73,
   Guy
We tested Adrea's Extio and we confirm it just works fine with HDSDR.

Now that said, we recommend SDR# for best performance. You will get much better results by using our stack (hardware + software) than any other combination. It's not that we are locking features from other software, we just have a much more refined DSP. You will notice the difference once you have a good radio in hand :-)

I hope you will enjoy it.


moderated EXTIO File for HF+ ?

Guy Atkins
 

Hello, I've seen some discussion of a EXTIO DLL file for HF+ which one enthusiast has written, but evidently a working HF+ in-hand is still needed to finalize the file.

I know that all the early purchasers like myself are looking forward to using their receiver with their favorite programs. Mine are Studio 1, HDSDR, and occasionally the EXTIO version of SDRuno. Does AirSpy plan to release an "official" DLL and perhaps other supporting files specific to the HF+? I know that these things take time and manpower :^)

73,
   Guy


Re: Shipping dates - any updates/confirmations? #airspyhfplus

Mike Jackson
 

Yes indeed, I was just being optimistic!


Re: Shipping dates - any updates/confirmations? #airspyhfplus

prog
 

On Tue, Nov 21, 2017 at 12:41 pm, Mike Jackson wrote:
Thanks Youssef noting that the first week of December starts and finishes (business week that is) on the 1st Dec!!
If we were working only on business days, you would be probably still waiting for the announcement next year ;-)


Re: Shipping dates - any updates/confirmations? #airspyhfplus

Mike Jackson
 

Thanks Youssef noting that the first week of December starts and finishes (business week that is) on the 1st Dec!!


Re: Airspy Mini calibrate #hardware

prog
 
Edited

On Tue, Nov 21, 2017 at 05:47 am, Dave C aka Dave3825us wrote:
It was 462.something. I used a cobra walkie talkie . Tried from 5 feet away to 1 foot away. Get like 3.xx something ppm, then -1.04 ppm and after 4 or 5 tries its now at 0.06. 
Do you really believe a cobra walkie talkie is a good source for stable carrier? The signal is *frequency* modulated in function of your ambient audio. ie. completely random. Plus even without modulation, these talkies have worse frequency accuracy and stability than the Airspies.
I suggest you send it to Joe at airspy.us for proper calibration.


Re: Direct Sampling frequency limit in r 1628 onwards #sdrsharp #spyserver

prog
 

On Tue, Nov 21, 2017 at 10:05 am, Martin wrote:
Hi,


I see that in the release notes for r 11628 onward that "Set RTL max frequency to 14.4 MHz in direct sampling mode." has been added.

Is there any particular reason for this ?

I understand the desire use the tuner chip as much as possible in order to avoid Nyquist Alias signals and to have RF gain control. But I have an application where I wish to continue using direct sampling above 14.4MHz and this change has caused problems for me.

Could I suggest that it may be better if the user / admin could set this as required, as some configurations may need to use direct sampling above 14.4MHz and others may not.

It would be even better if it was possible to define frequency limits for direct / quadrature sampling in SDR Sharp and Spyserver, so that this could change automatically when the defined frequency boundary is crossed. This would save having to reconfigure things manually for example when you wish to move from HF to VHF and allow a mix of HF / VHF / UHF frequencies to be scanned.

It may also be possible to extend this proposed configuration option to automatically turn spyserver on-off (along with the frequency off-set) or turn the bias tee on / off (if folks are using this to switch antennas or pre-amps in / out).

Regards,

Martin - G8JNJ 

Know your SDR. RTL's sample at 28.8MSPS, thus offering a first nyquist zone between 0 and 14.4MHz. Nope. This won't change.

14481 - 14500 of 39948