HDHomeRun FLEX 4K and Negative SNRs

[WNYOTA]

I've been using the HDHomeRun FLEX 4K DEV for quite some time and I noticed what appears to be a glitch. Whenever an ATSC 3.0 signal drops below the noise floor (0.0dB) in HDHomeRun Config GUI, the L1 Signaling information from the preamble disappears. The preamble is supposed to be available until -7.8 dB AWGN. I should visibly see the HDHomeRun report -1 dB, -2 dB, and so on until almost -8 dB, while still showing the characteristics about the PLPs. Technically, the HDHomeRun should still be giving signal information until almost -10 dB, since that's where the bootstrap ends.

The HDHomeRun acts like ATSC 3.0 signals don't work below the noise floor. My question is, what if a signal is QPSK 2/15? Will the HDHomeRun not work if it's below 0dB?

Is this something that can be fixed? And has anyone else encountered this issue?

[nickk]

Initial thoughts...

The SNR value has a lower reporting bound of 0dB. This is not related to maintaining or reporting lock, just the reported signal quality value.

The SNR value will report 0dB if there is no lock.

Looking at the code - L1Detail must be received for lock to be reported. It also needs to be able to pull the FecModCod information from the tuner which I think should only require L1.

You should still get L1 lock, you just won't know how far below 0dB you are.

Deeper investigation will require setting up a lab test and seeing exactly what the demod is reporting.

[WNYOTA]

Interesting. Thanks for your response Nick. This is my experience with WTVJ in Miami, FL: At around -99 dBm it was reporting a ~1 dB SNR and at -101 dBm everything was gone. That "PLPs" box vanished. Since the lower reporting bound is 0dB, once it hits 0dB and the dBm value continues to weaken--it should stay at 0 dB for at least another few dBm, but it doesn't.


The fact that it is not related to maintaining or reporting a lock is a good thing. However, if the channel is coming in below the noise floor and let's say PLP0 is QPSK 2/15, would the HDHomeRun include that station in the channel lineup?

Lastly, is there a way to change the reporting bound for development users?

[drmpeg]

This is not super scientific, but it's not too horrible. With my open source ATSC 3.0 transmitter, I can adjust the transmit power in 0.25 dB steps with an attenuator that's built into the Analog Devices AD9361 transceiver chip on my Ettus B210 SDR. I believe that attenuator is fairly accurate.

In the the first screenshot, the attenuator is set to 30.0 and the SNR is 3.3 dB.



In the the second screenshot, the attenuator is set to 27.0 (3 dB less) and the SNR is 0.1 dB.



In the the third screenshot, the attenuator is set to 22.0 (5 dB less) and the SNR is 0.0 dB.



If we believe the SNR reading is fairly accurate, then the last screenshot is running at -5 dB SNR.

Note: use "Open Image in New Tab" to see the full size images.

[WNYOTA]

If we believe the SNR reading is fairly accurate, then the last screenshot is running at -5 dB SNR.

I can't get my HDHomeRun FLEX 4K to go a single dBm lower with my over the air stations once it hits 0dB. I also tried mine with a (physical) attenuator.

Once that QPSK 2/15 drops out, is the PLP info still listed with even more attenuation?

[WNYOTA]

I did an experiment again just to confirm. Riddle me this:

You should still get L1 lock, you just won't know how far below 0dB you are.

How can WTVJ be 0.2dB, and be past the point where any preamble data would be displayed while both at -102 dBm. When I increase attenuation, it doesn't just sit at 0dB for a while. It instantly drops out completely. There is obviously something wrong; it shouldn't be doing this.

[drmpeg]

Once that QPSK 2/15 drops out, is the PLP info still listed with even more attenuation?

The Linux version of hdhomerun_config_gui doesn't have the PLP checkbox (and I don't have any Windows machines here).

It would be nice if SiliconDust could update the Linux version.

The signal strength indication doesn't seem too accurate. In my example above, it shows -92, -93 and -95 dBm even though the attenuation was 3 and 5 dB. However, the SDR is transmitting a big out of channel carrier that may be upsetting the readings. It's an artifact of how the AD9361 chip works (it's the DC offset from the direct conversion architecture). I've shifted it as far out of channel as I could (3.4 MHz from the center of the channel).

[WNYOTA]

If SiliconDust is able to do this:

- Fix L1 Lock so it locks below the noise floor
- Change the lower reporting bound to display negative SNRs

It would be much appreciated!

[drmpeg]

If SiliconDust is able to do this:

- Fix L1 Lock so it locks below the noise floor
- Change the lower reporting bound to display negative SNRs

It would be much appreciated!

The Sony demod chip may not support indicating negative SNR.