Apache Labs ANAN-200D a competitor?

Bill -VA3WTB

Still no information on the 7000 radios, can you please send it to me, I'm interested.

np2g

Sorry to disappoint you Actually not garry but edited from 15 e mails .I wonder if one euro could hear garry ??? And again what difference does it make It doesn't change a thing . as long as there is a human with a choice the one they bought is better no matter what . For me this is why I look to Sherwood on empirical testing , The ARRL Labs for test's . Most others are far too personal to believe . The "no comparison Comment" could be taken many ways I would like to believe that you cannot compare function for function . So you shouldn't compare And the disclaimer to the reply does not reflect my opinion is the truth . Sorry to disappoint you on that too.

np2g

Sorry EA4gli I do not participate in the Yahoo ANAN forum . And I could not give you the link these comments came from E Mails . NOT conversations I disclosed this fact . Not comments from Alpha,Beta testers or company people just e mails directed to ME . I also told you and everyone that I edited them to take out any resentment . Both ways . Oh yes I do not chase windmills . But would use the structure for HF antennas .

Bill -VA3WTB

Walt that is interesting. I wonder how many angels can dance on a head of a pin. Maybe we can keep this going another week on that...lol

Dave - W6OVP

If everyone stopped posting, this silly thread would go away.  It's way past time to shut it off and make the children go outside and play.

Ed baker

As a update since 1.5 flex is out there. The nr. Or noise blanker on the 200d 100 d. 10 And 10e are vastly superior . The most recent beta out there is even better than the last. Saying this flex you can do it

David Decoons, wo2x

Actually for the noise I get here sporadically the Flex was 100% effective without any audio distortion. The 200D with 03.2.27 still has distortion. Not bad but it is there.

Ed baker

Yes the version.you used is not the newest but even that one is a more effective noise reduction event on a wide variety of interferences The latest mod corrected that audio
It also knows the difference between a cw dit. And a noise burst. Or. Better putting it state of the art noise abatement selectively defining communications from interference.

Flex has the ability to provide the noise abatement system we all should have. 1.5 has not. Simple. Don't settle for almost or better than before. How about as good as what you can presently compare it with. And for everyone out there anan is just a science project. Well almost.

David Decoons, wo2x

Actually I do have the latest beta from Warren. Have not had a reoccurrence of my line noise lately to test it.

W9OY

The Anan NB is ok, certainly not amazing. In my opinion as an owner of both radios the entire Anan interface experience while adequate is quite dated. The Flex blanker applies a different theorhetical basis to noise mitigation and thats what i find amazing. New ways of doing things lead to newer ways. I used to have radios with with hundreds of dollars of roofing filters Today I can't even remember what a roofing filter radio sounds like. My new technology has turned roofing filter radios into history. I expect a similiar experience with Anan's NB over time. The present Anan architecture is determent of that. I hardly ever turn on my Anan, but when I do I find it to be a good radio but the interface is without doubt clunky and bounded. 73. W9OY

David Decoons, wo2x

Warren is working on a new noise blanker for PSDR mRX. Supposed to be very good. I have it here but don't have the line noise present the last week to really do any testing. It is VERY CPU intensive. That coupled with PS and other things in PSDR mRX is one of the reasons the radio doesn't support gigabit Ethernet and only two rx slices.

The 200D isn't a bad radio and there are plenty of things it does well but for how I operate the 6700 environment is my go-to radio.

Dave wo2x

Bill -VA3WTB

Lee, I have a good friend who lives close. He has a 200D. I sometimes I go over to visit. We play radio a lot. I also find the software on the Annan to be very dated and clunky. He still has many off the problems that Dave mentioned he has with his 200D. Personally I would never be interested in a radio without SSDR. Slick, clean. His NB works ok there too. I find the NB in 1.5 to be wonderful, as it is wide band sampling. I have a big problem with noise here on 40M, but with the NB in 1.5 it goes away 100% and 0% distortion in the audio. I can now hear a pin drop out there on a noisy day. I'm more then happy....

Walt - KZ1F

Drax, I think the answer to your question and the reason this issue invokes so much, shall we call it, passion in folks is 'it depends'. On this forum, you're hear from people with a heavy bias towards Flex radios. The dozen or so most vocal frequenters of this board are a heavily homogenous group. If you go to an Elecraft or Anan or kenwood reflector you'd see that same heavy bias toward another brand. The reality of it is the Anan 200D is a competitor if you perceive it has features more important to you or you like better than some other brand. It's all personal choice.

Ed baker

Dave this one required c+. To be loaded as it has not been assembled

Ed baker

This comparison wasn't to make someone buy. It was a direct comparison to gauge the noise 1.5. Progress.
If you love 1.5. Well ok.

If you really compare
1.5. To this one flex Has more work to do Any short coming should not be used as a excuse. As we know in the USA you can spin just about anything into some reason. Use what excels to compare in this case if my kx3 noise system was "better" I would suggest that to be used as a benchmark.

Ed baker

I'm going to try a different way to say Idea. Use a benchmark. (Whatever you want) to compare a function on the software ) The results should be comments directly related to that radio. Not which is better Here. Simple. Well on xyz. Noise it seems to need more y. Or. On the wave file I sent you on mitigating noise it seems to work but it most likely needs 123 Or with your new version it is 100 percent. And is as good as whatever. Or better. This is the result you want. Not a is better than b. but flex 1.5. It right there Now this results in a improvement. And. Also can change the benchmark if it is a improvement. And as you might have seen the benchmark also listens and quests to be better. (For its own sake not for bragging rites)

Ernest - W4EG

Just as I said on my letter to the ARRL review of and the comparison of the ANAN to the Flex-6x00 series radio. 
"Is like comparing a tube radio to a crystal set radio." 

Walt - KZ1F

I just read this thread, start to finish. Ah, it was,in deed, a simpler time then (5months ago). :-)

W9OY

Or you can let the radio do it for you as opposed to spending half an hour screwing around with XYX, wave files and all that.  SSDR 1.5 does the noise mitigation tuning for you, and exactly for the noise you are experiencing at that minute, not for the noise on some standard wave file. If the noise signature changes, WNB turns off, recalculates the best noise mask, and reapplies the mask to the new levels.  It doesn't rely on some arbitrary "standard" it just eliminates the periodic noise in a wide bandwidth pretty much on the fly and just in time using just a couple seconds to set up correctly.  This is the result I WANT.  I don't really care if noise mitigation is perfect I just care that it is adequate and clearly WNB is more than adequate especially since there is NO AGC pumping which is the thing I hate most about noise blankers.  The other advantage is, because it's wideband, it clears the noise in the panadapters as well including the noise across several MHZ.  Amazing piece of software!

73  W9OY

Joe, KQ1Q

Re "use a benchmark", this is a great idea and there has been discussion about why ARRL does not devise and publish noise reduction benchmarks in QST. Manufacturers are spending more effort on this area, tapping ever-increasing DSP hardware and software.  Increasingly how well NB/NR works is a key purchasing decision. Yet to my knowledge there are no standardized NR benchmarks published by the ARRL.

In 2013 at the Huntsville Hamfest, K8KI gave a presentation about NR benchmarks he had devised. He showed charts and graphs of how several popular rigs performed. I don't remember the results. However this very thread shows the need for increased effort in this area.

W9OY

Oh great now we can start NOISE BLANKER WARS just as we get over roofing filter (dynamic range) wars.  Manufacturers will build their rigs to eliminate "standard noise" instead of my noise because winning the noise blanker wars (eliminating "stand noise") will become the primary ad feature in QST, and the state of the art will once again be frozen for 20 years while we fight over "my nb is 2db better on standard A than your nb".  Not a future I'm much interested in.

If polio had been left to the engineers we would have super miniature iron lung machines, (kind of like we have ever miniaturizing CPAP machines).  Instead the sciences of virology and medicine solved the issue and eliminated polio.  


73  W9OY

N7AIG

Guys! Take a look at what K8KI actually wrote in QST, Feb 2015. He shows how to measure what you have, in your very own noise environment. Some algorithms work better than others, depending on bandwidth, noise environment, etc.

Try making actual measurements for yourself, and we'll probably discover the sweet spots for various conditions if we share the results we get - electric fences, switcher motor controllers, etc.

73 de Dave, N7AIG

N7AIG

Actually... just thinking about the measurement process, it is very similar to measuring infrared signals of stars -- perfect application for automated measurements using synchronous detection.

Way back in the ancient 1970's we used a wobbling secondary mirror to modulate the starlight impinging on the detector, and an old PAR "Boxcar Integrator" phase locked to the wobble driver signal to perform the detection.

There we were trying to read signals at -70 dB SNR. This ought to be even easier. I have an old Stanford Research SR-510 Lock-In Amplifier sitting in the closet, that could be used here. But automating through the audio port would be even easier to do. Just use an FFT and look for the peak amplitude around the modulation frequency (a few Hz) applied to the RF tone atop the noise.

73 de Dave, N7AIG

Alan C

Dave, your infrared measurement process from the 70's reminds me of the 'turbo encabulator' https://www.youtube.com/watch?v=Ac7G7xOG2Ag

:-)

Alan, W4FBI

Ed baker

On bench marks. This is exactly why since no data it becomes personal. Rather than analytical. And for some more than they can fathom or stand. As band conditions further drop It. becomes just one more valuable tool. To allow hams play time

N7AIG

Just did this test using an available noise level metering system I had already written for another application. Some interesting insights into the Icom way... It is my contention that Icom has found some magic that makes their NR one of the very best out there.

For the test, I rigged up an air injected RF signal from an Agilent 33522A function generator. The Icom rig, an IC-7700, was tuned to an empty portion of 160 m, USB 1.9 MHz, initially with no noise reduction, no noise blanking, and RF gain was turned down till it read S9 to avoid AGC action. (You can't disable the AGC. But this does the same thing for weak signals.)

I automated the injected RF signal by modulating with a 0.05 Hz square wave, to produce an on/off signal with a 10 sec duration on, then 10 sec off. My noise level meter has a leaky RMS detector with about a 3 second 1/e time. The noise level is displayed on a wrapping chart recorder graph.

The first test used 3 kHz bandwidth while changing the NR level in 1 tick dial steps. That is shown in the first graph, and we see that around dial setting 5 (midway) the algorithm appears to change over to a more aggressive noise reduction. Below that level, it appear that you get 4 dB of signal improvement per dial tick. At mid setting it becomes very aggressive.

(the data points should probably be viewed as having error bars of +/- 1 dB)

The next test placed NR at mid range, and varied the audio bandwidth. In the following graph the light red line shows the expected reduction in the noise floor with bandwidth (-10 Log(BW)). Again, at around 500 Hz bandwidth, something odd occurs in the slope.

But interestingly, there is the expected reduction at narrow bandwidths. And to me, this shows that they must be applying the noise reduction prior to narrowband audio filtering. In other words, they understand that the Widrow algorithm works best at wide bandwidth, and they don't try to stupidly push narrowband audio through the algorithm... (Elecraft, are you listening?)

I have no ready explanation for the shallower slope at wider bandwidths...

73 de Dave, N7AIG

W9OY

Try it with the AGC turned off if you can

N7AIG

Heh! Turboencabulator.. well, we didn't need any dingalongs or whatever he said. That's a hoot!

Surely you EE's know all about synchronous detection. That's one of the first things we showed the kids in "Scopes for Dopes", aka "Astronomy for NonScience Majors". I know it is discussed pretty thoroughly in Horowitz and Hill.

N7AIG

No, cannot turn off the AGC in the IC-7700. But I verified linearity of response in the peak signal level at the injection levels I was using. My signal was down around S2-S3, and the AGC wouldn't kick in until up around S9.

The IC-7700 needs the AGC engaged at all times to avoid overrunning the ADC in the final pseudo-SDR chain.

(I shouldn't say pseudo-SDR. It is a full SDR running at 38 kHz IIRC. It is heterodyne up to that point.)

N7AIG

FWIW... I have written many implementations of the Widrow noise reduction algorithm. I have never seen that algorithm produce more than 20 dB of noise reduction. So whatever Icom is doing above dial setting 5 has to be something more elegant.

.. at the same time, I have never seen the need to increase the Icom NR much above 3-4 on the dial. And that looks to be the region of Widrow-Hough.