jks

I'm simply using SAM code from Warren Pratt's excellent wdsp package as pulled from the Teensy-ConvolutionSDR project. So I don't know too much about how it works. The code can be found on your Kiwi in the file /root/Beagle_SDR_GPS/rx/wdsp/SAM_demod.cpp

It has its own internal PLL code and does not use Christoph's C++ PLL code that is shared by the IQ and timecode extensions. Comments in the SAM code mention that the PLL zeta and bandwidth should probably be adjustable for varying conditions (e.g. dx, "medium", "fast"). Right now they default to a zeta of 0.65 and b/w of 200 Hz.

With the recent reorganization of the noise blanker/filter controls there is now some empty space on the audio tab of the main control panel. Perhaps some SAM PLL adjustment controls could be added there.

Antonio's excellent blog has information that WINB may indeed carry data on the LSB portion of the signal: http://i56578-swl.blogspot.com/search?q=winb

The "-s" in "curl -s" (silent mode) is preventing us from seeing any error message if the connection is failing.
What happens if you do:

ping 50.116.2.70

versus
ping kiwisdr.com

I'll bet this is a problem with DNS not working with WiFi for some reason (router configuration problem?)

The build count has stabilized at 204. So Seeed seems to have built the promised 200 units plus a few spares.

I'm only guessing here, and I don't mean to show favoritism towards particular distributors, but my bet would be ML&S (UK) and WiMo (DE). I wouldn't order from Mouser (USA) until they show units in stock (they had an issue earlier in the year with not filling a large backorder in the correct queue order).

As of this morning the total count of Kiwis that have been through factory final test since Aug 19 is 186. So depending on the delay shipping to the distributors we shouldn't have to wait much longer now.

So your Kiwi is not currently configured for the proxy service it seems. But after some checking I found I could connect to it at its current public ip address (on port 8073). On the admin page connect tab you have the menu entry set to "public ip" instead of "reverse proxy". Change it back, restart, and I'll bet you'll be fine.

Currently, when you connect with a browser, the default AGC threshold is -130 dB. In the past this default was -100 dB and there was disagreement about this choice and discussion about whether it needed to be automatically adjusted depending on passband etc. Remember that certain values in the UI (including the AGC values) are saved across connections via browser cookies. So what you see in the AGC tab of the control panel may not be the true default.

Kiwiclient (kiwirecorder et al) allows you to turn AGC on and use a fixed set of parameter values (set in the Python code) or turn AGC off and set a manual gain, all via the --agc-gain (or -g) [gain] argument. The fixed values as defined in the file kiwiclient/kiwi/client.py are:

    def set_agc(self, on=False, hang=False, thresh=-100, slope=6, decay=1000, gain=50):
        self._send_message('SET agc=%d hang=%d thresh=%d slope=%d decay=%d manGain=%d' % (on, hang, thresh, slope, decay, gain))

So Kiwiclient is using -100 dB as the threshold which is at odds with the current browser default. Although not with what some people may have it set to. There is for example no Kiwiclient --agc-threshold argument, but it would be trivial to add. If you peruse the code you will see other references to "thresholds" but those are for NBFM squelch and RSSI thresholds.

Please don't do this. But there is a much better way using that 6.3 VAC transformer. In fact, it's perfect.

Use a modern low-dropout linear regulator chip (LDO) to get exactly 5V. E.g. a modern LTC / Analog Devices LT1085-5 in a friendly 3-pin TO-220 package can handle 3A output. It's $2-3 USD. It will need to attach to a small heatsink. There are complete schematics in the data sheet and on the web.
https://www.analog.com/media/en/technical-documentation/data-sheets/1083ffe.pdf

Just because you have a 6.3VAC transformer doesn't mean you'd get 6.3 VDC. If you use a full-bridge rectifier you'd get 6.3*sqrt(2) = 6.3*1.4 = 8.9 VDC. But this is fine for the LDO input which always needs to be higher than the output anyway. Yes, the 8.9 to 5V difference will be burned away as heat by the regulation (and you'll need a good heatsink for the LDO to handle that). You don't need an output inductor with an LDO. Just the recommended input and output capacitors to the chip (see data sheet).

You would essentially be building one of those ~$60 Ebay power supplies we recommend on our webpage. They use the same design.

List updated to include new CNR frequencies: 15440, 17660, 17725, 17745, 17845, 17790

50 kHz wide? Good reception of a Russian or Chinese OTHR no doubt. You can better confirm this by zooming all the way in (z14) and looking for a "picket fence" structure to the signal on the spectrum display.