MattSR
Member
Is that usrp_p25_rx tuned to 455khz through a LFRX?
Complex Digital Demodulation Hardware Discussion
- Thread starter Mike_G_D
- Start date
- Status
KA1RBI
Member
Correct - the USRP/LFRx is tuned to 455 KHz. The usrp_p25_rx is heavily modified. I'll get around to booking in the code RSN... It uses a Gardner loop instead of the Muller/Mueller...
There's also a planned update (someday) for my 455 KHz I/Q soundcard downconverter-
A 455 KHz IF Downconverter for Digital Radio Reception
The new version will use the NE602/SA602 instead of toroids. Getting that setup to work would be the holy grail since that would finally enable legacy scanners to receive LSM reasonably well, using a hardware mod that's no harder to do than a disc tap...
73
Max
I thought the NE602 was discontinued long ago.
KA1RBI
Member
NE602 is now a colloquialism, isn't it... SA612AN should be a current replacement (AFAIK / YMMV / etc)
Max
Hi Max,
How hard would it be to modify the 455KHz IQ soundcard downconverter (or design a new one) to use the 10.7 MHz IF on receivers such as ICOM and AOR? These receivers have the port right on the back of the receiver and no surgery is required. The ICOM does have DC on the 10.7 IF so a DC block is required.
The I/Q output can be then input to the soundcard for processing by DSD, Unitrunker, etc.
thanks,
systron
Member
- Joined
- May 8, 2010
- Messages
- 28
Another idea. I came across a IQ mod/demod chip by RFMD that interfaces with IF between 100KHz to 250MHz. Price from Digikey is in the $8 range. Package is SOIC-14. There's also an Evaluation board for $150. Datasheet: http://www.rfmd.com/CS/Documents/2713DS.pdf
Anyone know if we can save parts by using this chip?
thanks,
Anyone know if we can save parts by using this chip?
thanks,
Great idea if starting from scratch with a new radio but I wanted to help design something that can interface with radios with 10.7 IF and ither IF freqs.
thanks,
It's been a while since the last post here.
Reading through these threads again, it seems that the GNU Radio with USRP (via IF out from a scanner/communications receiver) would be the best alternative for us to use to correctly decode all of the higher order modulation schemes that we need. Since there are already several apps running under GNU radio we have a good starting point. In addition, the DSD code is written in C under Linux so it should port over well - if DSDAuthor gives permission.
I would like to start a development project with key developers such as Unitrunker, KA1RBI, DSDAuthor, Slicerwizard, and any other developers with GNU radio and trunking/digital voice procotol knowledge, etc.
I would like to hear from the developers above if they are interested and if they have any comments. My idea would be a "Super" Unitrunker/DSD type application that runs under Linux and uses GNUradio with USRP connected to the IF out of a receiver.
thanks!
Reading through these threads again, it seems that the GNU Radio with USRP (via IF out from a scanner/communications receiver) would be the best alternative for us to use to correctly decode all of the higher order modulation schemes that we need. Since there are already several apps running under GNU radio we have a good starting point. In addition, the DSD code is written in C under Linux so it should port over well - if DSDAuthor gives permission.
I would like to start a development project with key developers such as Unitrunker, KA1RBI, DSDAuthor, Slicerwizard, and any other developers with GNU radio and trunking/digital voice procotol knowledge, etc.
I would like to hear from the developers above if they are interested and if they have any comments. My idea would be a "Super" Unitrunker/DSD type application that runs under Linux and uses GNUradio with USRP connected to the IF out of a receiver.
thanks!
Joe,
You pretty much took the words out of my mouth or rather fingers; those are my thoughts as well. I haven't posted on here in awhile for two reasons - 1) Been busy with other things (mostly computer oriented) and 2) because I wanted to do more research on the latest hardware and software (including your suggestions, BTW, thank you!).
I initially started this thread to provide a place to discuss real "from the ground up" hardware designs which could improve on and/or augment the current common approaches (i.e. basic discriminator tap and soundcard approach).
I really just wanted to get a discussion started and hoped some individuals smarter and better versed in current state of the art hardware design than myself would join in and carry it through. That seemed to have happened and I have learned a lot in the process. You were one of the first (the first, I think) to join in and I appreciate that, BTW.
I have concluded that I suffer from the following serious flaws that prevent me from adding much more to this discussion (at least until I can educate myself more fully):
1) Lack of knowledge about the current state-of-the art in terms of the hardware.
2) Lack of (pretty much any) software knowledge - nowadays, even serious hardware folks seem to code at least at the basic hack level; all the software discussion completely goes over my head, unfortunately.
3) Lack of (pretty much any, again) knowledge of Linux despite a few abortive attempts on my part to fix this.
When I was "growing up" in the profession I was raised by hardware folks who generally subscribed to the basic paradigm that essentially says that if it can be done in either hardware or software it is usually better done in hardware for stability and robustness reasons. I don't think that holds anymore. The current computer power available at low cost is so powerful that DSP processes can be run efficiently on (current level) home PC's that used to require dedicated hot-running large scale IC's when I was hardware-design active. I am actually surprised that apparently so many software folks seem to be DSP literate. I don't know if that is because it's common now for general software dabblers to have basic DSP knowledge or because there are now easily available "canned" or "cookbook" coded routines that any software person can grab off the internet and simply insert in their code with little modification. I suspect the later mostly but, again, my ignorance balloons out ridiculously here - obviously I can research this myself and need to do so (yes, I have heard of "codecs" but...?...like I said, I'm stupid here).
That thread about the LTR filter you and I both participated in also really slapped me across the face and beat me severely about the head and shoulders when the software gurus chimed in at the end claiming the designs were overly complex, unnecessary, and far easier to implement in software! From my point of view that was a serious "Holy Cow!" moment since, for me, even with six op-amps in the design it seemed WAAAY simpler than doing it in software; but, since software (especially DSP software) seems complex and "black magicy" to me, well, it's obvious that my limited "world view" is to blame here!
So I can't see me adding much here, at least for now. A few posts back I arrived at the same conclusion as you outlined in your last post that the USRP approach comes closest to my "dream box", not ideal but the closest I've come across yet when folding in cost and general approach. I still would like to see somebody like RFSpace or Winradio develop a reasonably cost effective narrowband LMR digital modulation decoding dedicated box which uses multiple demodulation schemes (i.e. FM discriminator and I/Q demodulator, etc.) and is easily controllable from software running on a PC and using a common interface (serial, USB, ethernet, etc.) and is fairly high performance compared to the current consumer scanner model. It still seems to me this would be an ideal platform to allow software folks to really "go nuts"! But I'm not holding my breath. Again, the USRP approach seems the best way to go for now. But I'm open to anything, really. As long as the discussion continues it's all still very interesting to me!
-Mike
D
DaveNF2G
Guest
Mike, you are not "stupid" despite your protestations to the contrary. "Ignorant" would be a more accurate description.
Ignorance means lack of knowledge.
Stupidity means inability to use knowledge that one has.
No insult intended. You are insulting yourself unnecessarily.
Ignorance means lack of knowledge.
Stupidity means inability to use knowledge that one has.
No insult intended. You are insulting yourself unnecessarily.
KA1RBI
Member
Lsm
Starting now to gain more respect for the 455 KHz IF signal in a scanner.
As an experiment, I tapped a R.S. PRO-2096 at "TP5" and fed the result through a 455 KHz JFET buffer to the SMA input jack of a (insert name of well-known all-open-source complex-I/Q-receiver here; no wish to sound like a commercial).
The results are attached. The first pic is a standard symbol display; the second shows the two interleaved QPSK constellations which are offset at 45 degrees relative to each other. 45 Degrees is equal to PI/4 radians, and LSM is known to be a form of PI/4 DQPSK.
The software receiver (capturing the tapped 455 KHz IF signal) copies LSM voice signals virtually perfectly; the scanner does its usual poor job. This is WITH THE SAME SIGNAL AT THE SAME TIME IN THE SAME RADIO. That is, the scanner is tuned to a LSM P25 voice channel and experiences the usual
-clicks
-buzzes
-chirps
-sounds like talking with mouthfuls-o-marbles
-occasional momentary losses-of-sync right during a juicy moment
-periodically dropping into FM-voice mode when it loses all P25 sync
-in short, anyone who has experienced the LSM crap will know *precisely* what I mean
73
Max
Starting now to gain more respect for the 455 KHz IF signal in a scanner.
As an experiment, I tapped a R.S. PRO-2096 at "TP5" and fed the result through a 455 KHz JFET buffer to the SMA input jack of a (insert name of well-known all-open-source complex-I/Q-receiver here; no wish to sound like a commercial).
The results are attached. The first pic is a standard symbol display; the second shows the two interleaved QPSK constellations which are offset at 45 degrees relative to each other. 45 Degrees is equal to PI/4 radians, and LSM is known to be a form of PI/4 DQPSK.
The software receiver (capturing the tapped 455 KHz IF signal) copies LSM voice signals virtually perfectly; the scanner does its usual poor job. This is WITH THE SAME SIGNAL AT THE SAME TIME IN THE SAME RADIO. That is, the scanner is tuned to a LSM P25 voice channel and experiences the usual
-clicks
-buzzes
-chirps
-sounds like talking with mouthfuls-o-marbles
-occasional momentary losses-of-sync right during a juicy moment
-periodically dropping into FM-voice mode when it loses all P25 sync
-in short, anyone who has experienced the LSM crap will know *precisely* what I mean
73
Max
Attachments
Hi Max,
Great job!
Did you use the softrock for this or the USRP (or something else)?
Would the same hold true if I used the 10.7 IF output instead of the 455KHz? The 10.7 IF is available on the rear of my ICOM radios.
Joe
Great job!
Did you use the softrock for this or the USRP (or something else)?
Would the same hold true if I used the 10.7 IF output instead of the 455KHz? The 10.7 IF is available on the rear of my ICOM radios.
Joe
KA1RBI
Member
USRP, using homebrewed software.
Probably, No theoretical reason why not - although it appears that a pre-amp would be required; and the one I have doesn't seem to offer sufficient gain...
Max
Excellent stuff! Thanks for the info Max!!
I guess the poor phase noise of the scanner's LO didn't mess things up too much for you, at least in this case.
Joe, the only thing I would worry about with the 10.7MHz tap is that you may experience sever adjacent and alternate channel interference - I think the 10.7MHz output is wide so as to allow the old NTSC analog TV decode and uses such as panoramic spectrum displays, etc. Ideally, you would need to add another filter following the signal to accommodate a 12.5KHz or 25KHz BW or mix it down to a lower freq. (like 455KHz - common) and use those common filters though you should, ideally, use one with good group delay characteristics to properly handle digital modulations having complex phase and amplitude variations.
Speaking of which, Max, was that tap following the 455KHz IF filter or before it (I don't have a schematic of a 2096 and don't know where TP5 is in the signal path)? I'm assuming it was after the filter. It might be interesting, if possible, to see those same plots at both filtered and non-filtered points to see what effect, if any, the common (and I assume, cheap; probably not having great group delay characteristics) scanner IF filter has on the constellation and symbol plots for LSM signals.
-Mike
I guess the poor phase noise of the scanner's LO didn't mess things up too much for you, at least in this case.
Joe, the only thing I would worry about with the 10.7MHz tap is that you may experience sever adjacent and alternate channel interference - I think the 10.7MHz output is wide so as to allow the old NTSC analog TV decode and uses such as panoramic spectrum displays, etc. Ideally, you would need to add another filter following the signal to accommodate a 12.5KHz or 25KHz BW or mix it down to a lower freq. (like 455KHz - common) and use those common filters though you should, ideally, use one with good group delay characteristics to properly handle digital modulations having complex phase and amplitude variations.
Speaking of which, Max, was that tap following the 455KHz IF filter or before it (I don't have a schematic of a 2096 and don't know where TP5 is in the signal path)? I'm assuming it was after the filter. It might be interesting, if possible, to see those same plots at both filtered and non-filtered points to see what effect, if any, the common (and I assume, cheap; probably not having great group delay characteristics) scanner IF filter has on the constellation and symbol plots for LSM signals.
-Mike
Last edited:
Hi,
I have been watching your discussions on SDR hardware for the new digital modes, has anyone had any experience with the new alinco DJ-x11? seems it has a built-in I/Q signal output and 10.7MHz IF.
http://www.alinco.com/Products/DJ-X11/DJ-X11brochure.pdf
looks like it will retail around $500 USD.
Scanhead.
I have been watching your discussions on SDR hardware for the new digital modes, has anyone had any experience with the new alinco DJ-x11? seems it has a built-in I/Q signal output and 10.7MHz IF.
http://www.alinco.com/Products/DJ-X11/DJ-X11brochure.pdf
looks like it will retail around $500 USD.
Scanhead.
In that particular case, the IF is less important than the I/Q output capability. I, too, would be interested in any results obtained from the x11; however, it's pretty new (as I understand it) and so it might be awhile before we hear anything.
-Mike
Hi Max,
Have gotten around to updating this schematic to work with LSM? Your page shows the old schematic with notes on the changes needed. Have you made any progress since 2010 (last update).
thanks!
Have gotten around to updating this schematic to work with LSM? Your page shows the old schematic with notes on the changes needed. Have you made any progress since 2010 (last update).
thanks!
KA1RBI
Member
No schematic update necessary- it's a standard frequency conversion stage (mixer) with an input "rf" freq of 455 Kc and an output at the sound card rate (typ. 96K). Any ARRL Radio Handbook will have data on how to build.
Max
thanks Max!
- Status
Similar threads
