Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1]   Go Down
  Print  
Author Topic: Spurious signals in SDR  (Read 7581 times)
WX2S
Member

Posts: 735




Ignore
« on: April 05, 2013, 04:06:14 AM »

Does anyone know of any technical references to the types and causes of spurious signals in SDRs? I looked on Amazon.com for a good book, but all the good ones seem to cost $150 or more.  Sad

73,
Wx2s.


Logged

73, - Steve WX2S.
I subscribe to the DX Code of Conduct. http://dx-code.org/
WX2S
Member

Posts: 735




Ignore
« Reply #1 on: April 06, 2013, 07:39:45 AM »

Wow, 409 reads and no responses. That means there's a lot of pent-up demand.  Grin

73, WX2S.
Logged

73, - Steve WX2S.
I subscribe to the DX Code of Conduct. http://dx-code.org/
NI0Z
Member

Posts: 569


WWW

Ignore
« Reply #2 on: April 06, 2013, 08:04:44 AM »

Do you me RF interference?  Or the little chirping birdies that sometimes fly across the scream, or other?
Logged

WD5GWY
Member

Posts: 403




Ignore
« Reply #3 on: April 06, 2013, 02:23:28 PM »

Do you me RF interference?  Or the little chirping birdies that sometimes fly across the scream, or other?
I agree with the above. You need to clarify what you mean. With an SDR receiver with a display that shows a section of the spectrum that you are tuned to, you can see a lot of things you might not notice with analog radio. Although you might hear it as you tune around or it goes screaming by!
 A lot of the things you see and hear, might well be coming from inside your home or very close
by. (neighbors)
james
WD5GWY
   
Logged
N0YXB
Member

Posts: 321




Ignore
« Reply #4 on: April 07, 2013, 10:39:54 AM »

I believe I understand the question and was hoping one of the very knowledgeable folks would respond, because it's an interesting question.

There is some information contained in the Bores Signal Processing online course, "Introduction to DSP".  The frequency leakage and windowing sections touch on this.  Although this may not directly answer your question, this course is a must for those who really want to understand SDR, and it's free.  (http://www.bores.com/courses/intro/index.htm)

The SM5BSZ homepage contains an article "IMD In Digital Receivers" that appears to be a copy of an article from QEX, Nov/Dec 2006. 

Other than this, I have not been able find much information outside of those expensive books you mention.  There are a couple of folks who can probably explain this phenomenon in their sleep, but I have not seen them posting lately.  Hopefully the wannabes have not run them off...
Logged
W6RMK
Member

Posts: 656




Ignore
« Reply #5 on: April 07, 2013, 11:16:35 AM »

Does anyone know of any technical references to the types and causes of spurious signals in SDRs? I looked on Amazon.com for a good book, but all the good ones seem to cost $150 or more.  Sad

An interesting question.
There's not a lot of good books on DSP/etc tailored to SDRs in a general sense. There are textbooks like those by Proakis or Oppenheim & Schaefer (and used versions should be available less expensively than new), but they tend not to talk as much about non-ideal behavior (e.g. what does it look like when you clip into the ADC).  Partly this is because classical signal processing makes the assumption of linear behavior (or, at least, idealized non-linear: multipliers are perfect). 

When it gets to non-ideal behavior, you're getting into specific implementations.  Journal articles on various aspects of this are where the information is: for instance, spurs in DDS output is the subject of a variety of articles, and even a couple specialized books. 

I don't know that there's a SDR equivalent of, say, Rhode's "Communications Receivers" which is a great overall review of analog receiver design. (with some DSP, too).  There are a fair number of newer books out on SDR, but they tend to focus on data radios, given the fact that commercial radios are virtually all carrying data, not analog signals, and so the analysis tends to be in terms of things like effect on bit error rate (BER), or perhaps on error vector magnitude(EVM).

There are also an enormous number of useful applications notes out there from TI and Analog Devices on various pieces of the signal processing chain: ADCs and DACs are where the rubber hits the road in terms of performance, and there's lots of ap notes on them and their various non-idealness.

But.. if you're looking for a discussion of "typical spurious signals" you might find it a bit hard to find.  ARRL handbook has some.   In general, you're going to see the following kinds of spurious sources (over and above the usual ones from the analog circuitry):

1) clock leakage - intermods and sampling artifacts with the clock frequency and multiples thereof.  I have a receiver which shows an interesting spur that turns out to result from the 66 MHz CPU processor clock leaking into the 49 MHz sampling clock (or the power supply), so I see intermod products at combinations of 66 and 49 and multiples thereof, some of which appear in band.
2) aliases - there's no such thing as an ideal anti-aliasing filter, so you wind up with what would normally be out of band spurs showing up at in band frequencies as the result of aliasing.
3) DDS spurs  from phase truncation or from finite DAC resolution - there's a whole literature on this.
4) Spurs from overloading the ADC.  This can result in what are essentially intermods with the sample clock rate or multiples, that alias back into the sample bandwidth.
5) I/Q modulator leakage. Either the carrier getting through, or the image showing up.  40 dB is doing quite well for I/Q modulators, but 40dB down can be pretty big, if the desired signal is large, and it's often in an unexpected place, compared to what people expect from conventional radios.
Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #6 on: April 07, 2013, 01:40:41 PM »

Add to those the various word length issues where folk have truncated without appropriate dithering (Tends to be more of an issue in DDC where correctly handling word length reduction down the CIC decimation chain can be a headache).

In general any word length reduction should be accompanied by appropriate dither if you wish to avoid weird spurs in places you would not necessarily expect, but doing this in the high rate stages is **expensive** in terms of FPGA area.

This is discussed in the literature relating to the use of fixed point math for DSP, sometimes it appears in the chapter dealing with word length in IIR filters. 

73 Dan.
Logged
WX2S
Member

Posts: 735




Ignore
« Reply #7 on: April 12, 2013, 06:37:13 AM »

The reason I asked is that I've been looking over some data on duplicate spots (same time, same station, different frequencies) in RBN and trying to deduce the cause of the duplicates from the spot frequencies.

73,
- WX2S
Logged

73, - Steve WX2S.
I subscribe to the DX Code of Conduct. http://dx-code.org/
K9IUQ
Member

Posts: 1957




Ignore
« Reply #8 on: April 12, 2013, 09:42:47 AM »

The reason I asked is that I've been looking over some data on duplicate spots (same time, same station, different frequencies) in RBN and trying to deduce the cause of the duplicates from the spot frequencies.

You may want to make sure that the duplicates are not coming from the xmitting station. A few years ago I owned a Flexradio 5K. At the time this radio was known to produce xmitting spurs on 160/80 mtrs. When I got it I first tested at low power it by listening in another RX, Sure enough my CW signal was on several freqs within the same band I was xmitting on. I then fired up the amp and running 1500 watts checked the results on RBN. Sure enough my signal had multiple freq reports, in fact the spur strength was almost as good as the fundamental signal reports.

As is well known in this SDR forum I complained repeatedly about the spur problem here on eham and on the Flex reflectors. I also had discussions with Flexradio. For my efforts I got banned from the Flex reflectors because of my constant complaining. Finally Flexradio offered to fix my radio with a hardware Fix. This was the first customer radio they did the spur fix on. It worked. They kept this fix quiet for 6 months until I decided to sell my Flex 5K. I advertised it as having the "coveted 160/80 mtr spur fix.) That day Flexradio announced that the spur fix was available to current owners and would be implemented in subsequent production radios.

I suspect not all radios were not sent in for the Fix. Not everyone works 160/80 mtr CW. Anyone buying a used Flexradio 5K should make sure the Radio has the spur fix before purchase. I would guess any Flexradio 5K from when it was first released (2008) thru about early 2011 may not have the spur Fix. My Flex 5K had the spur mod done Sept-Oct 2010.

I sold my Flex 5K on 3-8-2011. I just looked up the eham Ad and this is what I put in the ad "Also has the coveted 160/80 mtr spur elimination Factory Modification. Do Not Buy ANY 5K without this modification."

Since I am a 160/80 cw enthusiast I did note when I had the FLexradio 5k that I would sometimes RX multiple signals on different freqs from one call. On the Flexradio Pan-adapter this is very easy to spot. I just assumed that these multiple sigs were from other Flexers that did not have the spur fix.
 
Stan K9IUQ
« Last Edit: April 12, 2013, 10:03:54 AM by K9IUQ » Logged
ZENKI
Member

Posts: 960




Ignore
« Reply #9 on: April 13, 2013, 08:50:59 PM »

Aliasing  is known to cause spurious signals, however you cant generalize because there are so many variables hardware and software.
Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #10 on: April 14, 2013, 04:44:31 AM »

SDR is not magic, if you want good spurious performance you still have to ensure linearity, including linearity in the digital domain, just like you do for a conventional rig.

To give one simple example, multiplying two N bit numbers (Say implementing a mixer) gives a result having a length of 2N, if we then want to shorten that to say just the upper N bits, there are two ways we can do it.

1: Just take the top N bits to the next stage - Unfortunately this results in non linearity as anything smaller then 1 LSB of the output will be truncated to zero, additionally weak components that happen to force the quantization between levels will be effectively mixed with other signals present and will cause discreet mixer spurs.

2: Add a pseudo random noise having triangular probability to the 2N long word before truncating, this adds a constant noise floor but makes the truncation perfectly linear, an enhancement is to shape this noise in the frequency domain to reduce the energy in whatever our eventual passband will be, but in either case it takes much FPGA area to pull off especially in the high rate bits of a DUC/DDC design.

It sounds however like the problem flex had was probably something mundane and hardware related rather then being a problem in the vhdl/verilog.

Regards, Dan.
Logged
W6RMK
Member

Posts: 656




Ignore
« Reply #11 on: April 14, 2013, 11:31:19 AM »

It sounds however like the problem flex had was probably something mundane and hardware related rather then being a problem in the vhdl/verilog.

No verilog or vhdl in a Flex. It's all C running on the PC.

The problem with low bands on Flex comes from the basic architecture.  It uses a form of vector modulator on a DDS produced carrier.  The I/Q signal is setup as a few kHz offset from zero (to get around problems with DC response of sound cards in the first flex).  If the I/Q balance isn't perfect, you get an image below the carrier, as well as some carrier feedthrough.

The way the (analog) modulator works uses a I/Q version of the carrier generated by a DDS that puts out quadrature outputs, but there are discrete component 7th order filters on the output of the DDS. They match pretty well, but not perfectly.  Ditto with the audio band I/Q from the sound card.

All potentially calibrate-out-able, but the calibration changes with temperature, carrier frequency, and isn't constant across the I/Q passband.   It's easy to get 20-30 dB carrier and image suppression with a single point cal (ignoring the above), and that works for typical HF usage, where the person receiving is seeing the signal at 20 dB above the noise floor (they don't see the low level carrier and image). When folks doing top band CW started running high power amps, all of a sudden, that 30dB image is pretty noticeable.

The original SDR-1K with the original sound card interface is doomed in this situation, because it has no DC response to take out the offset term in the I/Q.  I think the F5K has DC coupled DACs to the modulator, but I haven't seen a schematic, so I don't know for sure.

There were several schemes for a loopback "self cal" in the F5k design, but I'm not sure they ever worked well, or were really thought through fully. Essentially, you need to have a good receiver with good instantaneous dynamic range and good image control with which you can look at the transmitter output and adjust the image and carrier down to zero. You can theoretically do this with the receiver that's already in the radio, but then you have to calibrate the receiver first.  And you really need to be able to set the center frequency of the receiver to something different than the center frequency of the transmitter (so you can use the unmodulated Tx as a tone source to calibrate the receiver).  The SDR1K can't do this (it's a half duplex radio).  The published F5K block diagram is unclear whether a single receiver F5K can have different frequencies for Tx and Rx chains (the dual receiver certainly can), and I'm not motivated enough to go look at the controller source code to see if it can.
Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #12 on: April 14, 2013, 01:46:47 PM »

Ah, at frequency phasing method, doable in DUC/DDC but always going to be a little interesting in an analogue mixer based SDR, better to use a fixed IF if doing that.

73 Dan.
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!