Noise Blanker and panadapter

  • 3
  • Question
  • Updated 5 years ago
  • Answered
Looking for input from other users who may experience pulse type interference such as that caused by an electric fence. When my fence is or, or my neighbor's, the panadapter display is almost useless. I know the NB has not been optimized yet along with ANF and NR (NR causes more noise than it removes) but the current implementation of the NB is within the slice bandwidth. I sent a screenshot to Steve Hicks of how bad the panadapter is effected. I really think it is important, required, for the NB to work on the entire display for any panadapter open, not just the slice bandwidth. This is how it works on the other Flex radios and any other SDR I have used.

Please chime in if you have similar thoughts or concerns.

73
Greg
AB7R
Photo of Greg

Greg

  • 656 Posts
  • 63 Reply Likes

Posted 6 years ago

  • 3
Photo of Steve - N5AC

Steve - N5AC, VP Engineering / CTO

  • 1053 Posts
  • 1076 Reply Likes
Official Response
Today the NB is in the slice receivers only -- not in the panadapter. We are looking into putting it in the panadapter.
Photo of Michael - N5TGL

Michael - N5TGL

  • 308 Posts
  • 101 Reply Likes
Steve,

If I understand correctly, with the NB in the slice RX, it is not affected by adjacent strong signals like every other radio on the planet? If so, this is a huge advantage for flex. Seems to me if we go to a panadapter NB, then we'd lose that advantage? I guess it would be switch selectable, so we could chose to have it on or off.

On the subject of the NB, it still doesn't seem quite as effective as the one on the PSDR. That one was solid, effective and worked every time. With this one, I have to click click click click it before it seems to take a good "bite". Even then, after about 5 seconds it seems to lose its effectiveness. A few more clicks and it works again, only to repeat the previous performance. I know it's a learning NB, but it appears to be "unlearning" as it goes, rather than learning. It is also wholly ineffective unless the slider is 90+.

ANF works great, maybe almost too well as it takes a pretty good whack out of the audio, even when there are no carriers to remove. PSDR didn't seem this heavy handed, and the slider doesn't fix it either, it just removes the carrier less.

NR, well, I never really used NR, but it wipes out the highs so much, its utility is debatable. I haven't played around with AGC-T and NR balancing, but then the UI makes that cumbersome due to the controls being on separate pages. In the interest of presenting a "cleaner UI", Flex has made the radio more cumbersome to use. I'll probably start a separate thread on this one.

73,
Michael
Photo of Tim K8XS

Tim K8XS

  • 48 Posts
  • 8 Reply Likes
Saying that the Noise Blanker (NB) is in the slice receiver only is conjuring up all kinds of interpretations. Here is my shot at it. Saying that the NB is in the slice receiver only to me means that it is applied to the data being processed that will eventually be converted into the audio we hear. It doesn't necessarily mean that the data used is limited to the 2.1 or 2.8 or 3.2 kHz bandwidth we pass through the digital filters. I have always been told that you want to perform noise detection and blanking before passing the signal through any narrow bandwidth filters (NBF). Before the NBF, the pulse noise has nice sharp rise and fall times. Once it is passed through a NBF, all the sharpness is smeared which makes it difficult to determine what is noise and what is the desired signal.

Now, having said all that, a simpler interpretation is that the NB is not applied to the data used to create the pan display. That is why, when you enable the NB, you see no change in the display. Not even in the small part of the display that is within the band pass filter. The data used for the pan display is not noise processed or even averaged (the devil made me say it).

I think of it as two separate streams of data. One for the display (frequency domain) and one to create audio (time domain). The audio path has the NB. I don't know at which point the original data stream splits.

I agree with the OP that the NB should be applied to all the data used to generate the pan display when the NB is enabled. I don't even think that this concept has really been an issue, just one of the many (mis)-interpretations. Just applying it to a small piece doesn't make sense. I want to see the noise level drop and reveal the signals underneath. Applying the NB to the entire display does not have to affect the audio or generate any undesired side effects. If they are treated as two separate streams, what you do to one doesn't have to affect the other.

Here is where we have a real opportunity to be creative. We have to do two things. We have to identify the noise and we have to blank it out. We could possibly use the wideband time domain data (some - all - or even more ) of the data that is used to create the pan displayed data to identify the noise and then possibly blank it somewhere else. I think the only issue is that you have to maintain time continuity between the data used to ID the noise and the data used to blank the noise. If we pass the data through a NBF, we will smear the noise and the start and stop times of the pulse will no longer be meaningful. However, we might be able to do something like automatically create and apply the PSDR "like" Tracking Notch Filters to reduce or eliminate strong signals that cause the NB to generate awful distortion. There are lots of possibilities.
Photo of Michael N6BRP

Michael N6BRP

  • 48 Posts
  • 5 Reply Likes
I too would like to see the noise blanker applied to the panadapter display. When the local noise increases in the evening, I cannot see many signals and have to tune through the band (with the NB on), as I would using my old IC706. I know there are technical problems with this, but it's number 1 on my wish list!
Photo of Larry Davis

Larry Davis

  • 82 Posts
  • 12 Reply Likes
i have a 5000 the nb works great this 6700 i am having problems with it i works a little or not at all not getting rid of the 5000 till they fix that
Photo of Ned K1NJ

Ned K1NJ

  • 314 Posts
  • 81 Reply Likes
    Still needs work.  Signals in panadapter below blanked noise can still not be seen.
Would be especially helpful during marginal E-skip conditions.

Ned,  K1NJ

This conversation is no longer open for comments or replies.