Second build sold out. Message will appear here when store is ready for third build ordering.

DRM forces BW [fixed in v1.385]

edited April 2020 in Problems Now Fixed
R. Marti runs 9 KHz BW on their DRM and I'd like to edit the DX Label to use 9500 Hz but specifying DRM forces 10 KHz BW

Comments

  • jksjks
    edited February 2020
    Already on the DRM wish-list is setting the passband based on the detected DRM mode (4.5, 5, 9, 10, [18, 20] kHz).
    A generalization of how the passband is customized for WINB. For example CNR 6030 always seems to use 9 kHz mode (i.e. +/- 4.5 kHz) to avoid the AM carrier 5 kHz lower in frequency.
  • adaptive would be great but just allowing the DX label PB value to force the BW rather than DRM forcing it would suffice
    Powernumpty
  • I think the 9 kHz mode has more to do with the 9 kHz broadcast channel spacing in certain countries.

    In DRM robustness mode B (the most common on HF, as far as I can tell) the OFDM carrier spacing is 46.875 Hz. If the center (nominal) frequency is 0, then the highest carrier numbers used in the 10 kHz bandwidth mode are +/-103. That's +/-4828.125 Hz. Allowing an extra carrier on each side for modulation sidebands, the signal still extends only from -4875 to +4875 Hz. That's consistent with my waterfall observations. That means a 10 KHz DRM signal should not be bothered by any interfering AM carriers at +/-5 KHz. Mode B also leaves the 0 Hz carrier unused to avoid any QRM by the carrier of a co-channel AM station.

    Now interference from the sidebands of an AM signal at +/- 5 kHz is a different story. There would still be sideband overlap in the 9 kHz mode, but the amount of QRM depends on the spectrum of the modulating signal, which is hard to predict.
  • The day I ran the BW test, there was a strong modulated AM station right at the + 5 KHz position
  • >>I think the 9 kHz mode has more to do with the 9 kHz broadcast channel spacing in certain countries.

    Not necessarily. The VOA is transmitting DRM on 7345 kHz from their Greenville, NC site using an old transmitter. They chose a 9 Khz bandwidth so as to allow more power with a little less distortion than with the normal but slightly wider 10 kHz bandwidth.

    WINB has been transmitting a combo DRM/data signal using 5 kHz BW for DRM plus 5 kHz for (non DRM) Data - not exactly as specified in the DRM engineering guidelines.

    73,

    -Zyg- AF4MP
  • today shown below, the previous day the signal at 7350 was much stronger
  • The VOACAP hourly signal coverage charts for VOA 7345 kHz are shown at drmnainfo.blogspot.com/

    On an unrelated topic your WSPR station receives daily my 200 mW 40M WSPR TX from my mobile antenna in the car!

    73,

    -Zyg- AF4MP
  • edited February 2020
    I have yet to decode Marti on 7345 here in San Diego but that's a combination of propagation, their schedule, and my location.

    Remember Shannon: channel capacity increases with power and bandwidth

    Narrower bandwidths put you at a definite disadvantage, power wise. To maintain the same data rate in 5 KHz as you had in 10, you have to go to higher order modulation, e.g., 64QAM instead of 16QAM, and even with FEC that incurs a substantial SNR penalty that exceeds the 3 dB noise saving from going narrower. I can get you the minimum SNR figures from the DRM30 specs.

    WINB is a good example. As far as I can tell, they're also wasting half their transmitter power on those lower sideband tones (what are they anyway?) which is another 3 dB hit on their SNR. I can only occasionally decode them here, though that's probably for the better considering their program material. Looking at the SNR tables, if they got rid of those tones, went to 10 kHz, and dropped from 64QAM to 16QAM, I'd decode them most of the time.

    Caveat: this assumes (thermal) noise as the background. Things are totally different in the presence of QRM. If half your 10 kHz channel is QRM, then going to 5 kHz to avoid it is definitely the right thing to do.
  • Only slightly related to the BW issue, though Shannon says things stop improving very much after a while, once inter-symbol interference pretty much goes away, I guess.
    I'm really intrigued by the spectral plots of DRM. Having a bandwidth filled fairly evenly by a 'noise-like' signal, as it should be for efficiency, lets one see those interesting cancellations which sweep through and in a way I haven't previously observed. I suppose what is happening is that we're seeing multi-path cancellation. Previously, just listening to AM SW broadcasts, it was obvious when these things swept through and took out a carrier, leaving largely DSB. Every SWL has probably heard that. But to see it happening even across only 10 kHz is fascinating. Most previous modulation types had a lot of coherent lines and I missed this view of what was happening.

    I'm not sure what combination of O and X wave or N, N+1 hops this commonly may be but it sure is interesting.
    Today I looked at Radio Marti from here in Colorado state and also from a kiwi in the Florida keys. Not only was the 'local' seeing less cancellation but also of a different sort. I'm not sure if that kiwi is truly close enough to be 24 hour 'groundwave' and if there will be a relatively clean signal once the ionosphere isn't involved, but I hope to find out.

    Very cool the things we can see with broader band representations.

    Glenn n6gn
  • > WINB is a good example. As far as I can tell, they're also wasting half their transmitter power on those lower sideband tones (what are they anyway?)

    It is data of some kind. Speculation is that WINB is using the "Latency Arbitrage" phenomenon where vast amounts of money, it is said, can be made by reducing the time it takes to send financial trading information, over a normal overseas fiber optic cable, by using HF radio. The rationale being that electromagnetic waves travel slower through glass fiber than through the ionosphere.

    More discussion at http://i56578-swl.blogspot.com/2018/08/winb-red-lion-to-test-drm-single.html

    -Zyg- AF4MP
Sign In or Register to comment.