Pages: 1 2 3 4 5 [6] 7 8   Go Down

Author Topic: Sherwood Engineering Receiver Test Data  (Read 2264 times)

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #75 on: March 27, 2021, 05:00:55 PM »

The airspy discovery being so small (its really small) its interesting to think about it as the RX part of a modern rig.
The airspy uses about 5% cpu on my old I5 windoze computer.
If you wanted to do a box with knobs like the 7300 or the M series Flex radios, how much cpu power would it call for?
Could you do it with a tablet and some knobs?

It is very interesting that there seems to be a lot of space and not much parts inside of many modern radios.


Logged

W6RZ

  • Member
  • Posts: 449
Re: Sherwood Engineering Receiver Test Data
« Reply #76 on: March 27, 2021, 07:15:41 PM »

If you wanted to do a box with knobs like the 7300 or the M series Flex radios, how much cpu power would it call for?

You might be interested in the RX888 (latest version is MK II). It's a direct sampling receiver for under $200. It doesn't have an FPGA, so it can stream the full 32 MHz of bandwidth to the PC (and give it a workout).

Here's DC4KU's review (unfortunately in German) with the 3rd order DR at 102 dB.

http://www.dc4ku.darc.de/RX888.pdf

Manufacturer.

https://twitter.com/fei666888
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #77 on: March 27, 2021, 08:09:48 PM »

No, I don't think I would want that.
I am just thinking how small and inexpensive a good RX system could be.
Ideally, The airspy with an ethernet interface (or something with less/no buffering then usb 2) and a tablet in a box with some knobs like flex or the 7300.
I wonder if the airspy is as good or better then the 7300.

Airspy says: -141 mds,
110 blocking dynamic range,
660 KHz bandwidth,
.5 ppb clock with low phase noise.

 All in the size of a packet of sugar...plus the computer of course.
Logged

W6RZ

  • Member
  • Posts: 449
Re: Sherwood Engineering Receiver Test Data
« Reply #78 on: March 27, 2021, 09:46:21 PM »

No, I don't think I would want that.
I am just thinking how small and inexpensive a good RX system could be.
Ideally, The airspy with an ethernet interface (or something with less/no buffering then usb 2) and a tablet in a box with some knobs like flex or the 7300.
I wonder if the airspy is as good or better then the 7300.

Airspy says: -141 mds,
110 blocking dynamic range,
660 KHz bandwidth,
.5 ppb clock with low phase noise.

 All in the size of a packet of sugar...plus the computer of course.

Got it. I thought you were interested in what it would take to process a direct sampling architecture like the IC-7300 or Flex 6-series. The RX888 would probably require a pretty beefy computer, and it only works with USB 3 (you can't get 64 Megasamples/sec through USB 2).

For my high rate SDR experiments, I use a big Dell 5810 workstation. Not exactly portable.
Logged

N6YWU

  • Posts: 362
    • HomeURL
Re: Sherwood Engineering Receiver Test Data
« Reply #79 on: March 27, 2021, 11:12:31 PM »

The airspy discovery being so small (its really small) its interesting to think about it as the RX part of a modern rig.
The airspy uses about 5% cpu on my old I5 windoze computer.
If you wanted to do a box with knobs like the 7300 or the M series Flex radios, how much cpu power would it call for?
Could you do it with a tablet and some knobs?

I’ve run my HF+ using a gumstick-sized Raspberry Pi Zero W to connect to an iPhone for the SDR, UI, and panadapter display.  There’s a tiny add-on board for the Pi, usually used for WSPR, but could also be used for QRPp CW & some digital modes as well (20 dBm Tx). It looks possible to assemble a full SDR & digital mode transceiver that would fit in a coat pocket (including the powerful SDR computer = iPhone) for SOTA ops using these currently available items.
« Last Edit: March 27, 2021, 11:19:09 PM by N6YWU »
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #80 on: March 28, 2021, 09:52:33 AM »

I did some actual operating on 40 meters using the Airspy as the receiver today.
I did A/B/C comparisons on weak signals and strong signals between my homebrew RX (my benchmark) the Flex 3000 and the Airspy.
If I turned the preamp on the airspy off, it was as good as the others, the 3000 might have a bit more background noise.
With the preamp on, the noise level comes up slightly to equal the Flex 3000.
Strong signals just out of the filter passband did not impact the signal listened to, even just 50 Hz out of the filter passband.

The noise reduction in the airspy or is it SDR sharp program? rocks!
Impulse noise can be eliminated and its easy to use.
You have AF, IF and Broad Band blanker's and reduction, the IF noise blanker works well on my noise.

On AM you have an interesting and very well working sync AM or ECSS type stuff, its tagged lock carrier and anti-fading.
It does not seem to squeal on lockup like most do.
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #81 on: March 28, 2021, 05:04:14 PM »

I actually had KE9NS power sdr and sdr sharp running on my old I5 computer at the same time, about 20% cpu usage.

Airspy audio comes out of the computer, the Flex 3000 has its own audio output...

I guess its about 200 ms latency through the airspy, likely because of the buffered usb 2 port.
Logged

N6YWU

  • Posts: 362
    • HomeURL
Re: Sherwood Engineering Receiver Test Data
« Reply #82 on: March 28, 2021, 09:41:36 PM »

I guess its about 200 ms latency through the airspy, likely because of the buffered usb 2 port.

That’s likely mostly software latency.

I measured the size of the data packets used to send IQ samples from my Airspy HF+ Discovery to my Raspberry Pi 4 over the USB port, and they were only 5.3 mS long, and fairly evenly spaced.

My guess is that software specifically designed for transceiver QSK operation is usually much better optimized for latency than generic SWL SDR code.
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #83 on: April 02, 2021, 06:16:12 PM »

I got a brand new fast computer and the latency is the same as the old one.
Its not the cpu slowing things down.
The flex 3000 runs at 5% cpu max on it.
Logged

N6YWU

  • Posts: 362
    • HomeURL
Re: Sherwood Engineering Receiver Test Data
« Reply #84 on: April 02, 2021, 09:45:38 PM »

I got a brand new fast computer and the latency is the same as the old one.
Its not the cpu slowing things down.

Yup.  A faster computer won't help if your chosen SDR software is waiting to capture or process long buffers of data (at the same sample rate) with long FIR filters.  Developing software for real-time audio is non-trivial.
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #85 on: April 03, 2021, 06:59:55 AM »

It runs my Flex 3000 with NO latency that I can detect.
The difference between firewire and usb 2....
Logged

N6YWU

  • Posts: 362
    • HomeURL
Re: Sherwood Engineering Receiver Test Data
« Reply #86 on: April 03, 2021, 01:25:56 PM »

It runs my Flex 3000 with NO latency that I can detect.
The difference between firewire and usb 2....

There's not that much difference in the hardware.  Older Pro-audio recording gear had less than 10 mS of latency over USB 2 (using short audio sample buffers).

Are you using the same sample rates on both?  The HF+ has a lot more latency at lower sample rate settings, as it has to wait a lot longer to filter, resample, and buffer up a fixed 2k IQ samples before sending them over USB.

There may also be some sort of Control Panel / Device Manager / Ports / USB latency timer setting on your Windows PC which came with too long a default.

Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #87 on: April 03, 2021, 03:36:43 PM »

The airspy is running as high a sample rate as it can.
I have never been able to find any settings for USB ports on any computer.
I know its a buffered interface...
Logged

KF5LJW

  • Posts: 577
    • HomeURL
Re: Sherwood Engineering Receiver Test Data
« Reply #88 on: April 04, 2021, 08:46:26 AM »

I did some actual operating on 40 meters using the Airspy as the receiver today.
I did A/B/C comparisons on weak signals and strong signals between my homebrew RX (my benchmark) the Flex 3000 and the Airspy.
If I turned the preamp on the airspy off, it was as good as the others, the 3000 might have a bit more background noise.
With the preamp on, the noise level comes up slightly to equal the Flex 3000.
Strong signals just out of the filter passband did not impact the signal listened to, even just 50 Hz out of the filter passband.

The noise reduction in the airspy or is it SDR sharp program? rocks!
Impulse noise can be eliminated and its easy to use.
You have AF, IF and Broad Band blanker's and reduction, the IF noise blanker works well on my noise.

You discovered what I and many other younger hams know. You do not have to spend several thousand dollars for a good RX. For less than $200 will do things a ham radio can never do. I have two friends with a Flex radio's including myself. We all use the Airspy as our main RX.
Logged

N2DTS

  • Member
  • Posts: 1368
Re: Sherwood Engineering Receiver Test Data
« Reply #89 on: April 04, 2021, 05:37:17 PM »

I must say, I am pleased with it.
I thought it would be another crappy dongle or sdr play type radio that overloaded easy and did odd things in the strong signal environment but it seems great to me on HF.

Quiet, right on frequency, great dynamic range.
I have no idea where it would stand on the Sherwood list, but it seems as good as anything else I have had.

I wish it was not USB but it could not be anything else being the size it is.
That is another amazing thing, just how small it is....
Logged
Pages: 1 2 3 4 5 [6] 7 8   Go Up