v1.696
From the CHANGE_LOG file:
v1.696 August 5, 2024
Antenna switch:
Progress trying to get the frequency offset settings of the admin config (global)
and antenna switch (per-antenna) to play nice with each other.
In the antenna description area of the admin > extensions > antenna switch page there is
a new menu called "Mode" for setting what kind of frequency offset and mixer injection
should be applied when each antenna is selected.
"no offset" means nothing about the offset will be changed and the
"frequency scale offset" field to the right will be ignored.
"offset" means the frequency scale offset will be applied and user connections will be
reloaded (and autoruns restarted) to account for the change. Low-side mixer injection
for any downconverter is assumed for the "offset" mode.
"hi side inj" is also an offset mode but sets up tuning of the Kiwi mixers to accomodate
downconverters using high-side mixer injection.
The mode menu will be preset based on your pre-v1.696 settings. But you should review it
to make sure it is set to do what you want.
The "cURL command arguments" field allows an arbitrary cURL command to be run when an
antenna is selected. This could be a command causing a network-enabled downconverter to
change setup such as frequency/band etc. More info: kiwisdr.com/info#id-antsw
(thanks N6GN)
SNR measurements are now skipped if the antenna is grounded because no users are connected
(if that option has been enabled) or if in thunderstorm mode. However, after a restart the
initial SNR measurement (to set the values on the user page and optional public listing)
is performed before any antenna grounding takes place. Also, an error is shown if forced
SNR measurement is attempted via the "/snr?meas" URL or "Measure SNR now" button on the
admin control tab if the antenna is grounded. (thanks Stefan)
Admin console: Option on admin security tab restricting admin console to local network
connections only has been replaced with a new mechanism: See kiwisdr.com/info/#id-opt-dot
for details. (thanks LA6LU)
Admin DX tab: Fixed problem where changes to non-DX label info (e.g. band bar info) was not
being seen by user connections consistently. (thanks Nhosker)
Added new LED pattern if communication with FPGA fails. See kiwisdr.com/info for details.
Primarily a Kiwi manufacturing aid.
Comments
Seems to work well, Thanks John.
About the only frill I can think of is to have a new frequency request search the available 20 MHz blocks for a match and, if found, do the selection for you.
With that doesn't the Kiwi become pretty much a 0-2 GHz all-mode receiver?
Maybe there's a reason for not going to that degree of agility that I don't see yet.
Maybe time for me to revisit and update the FE frequency extender kit for the Kiwi and Kiwi2 and not just wait for Kiwi3? Maybe put it in a Kiwi2-size enclosure and add an antenna switch so that a 'favorite antenna' for any of the 30MHz-2GHz frequencies might be selected too?
I'm not sure how much interest there might be but maybe someone would want to take on fabricating and distributing at a slight profit?
Hi Glenn,
Yes, no reason not to have some UI that is more specific for the FE. I'd be happy to do that.
I've been struggling with the idea of which to do: An "accessory box" that supports Kiwi-2/-1. Or Kiwi-3.
I don't think it's possible to do both because with each introduced product comes a tremendous amount of overhead on top of just the design and prototyping (e.g. manufacturing, sales, support).
For the same reason I think it's important to include everything you'd want in a single product. For an accessory box that means not only FE functionality and an V/UHF antenna switch as you mention. But also a few goodies for HF. Such as the 10 MHz shelving filter you guys developed and some series traps for the very common problem of strong AM BCB signals needing some attenuation. And an HF antenna switch too.
The counter-argument is that doing an accessory box would help "get those pieces right" for eventual inclusion as modules for the Kiwi-3 concept.
I've wondered the same things.
Suppose we were to follow that last path; update the 'accessory box' originally made with the FE for Kiwi-1 but with the design updates from the REference2 & TRansverter. There's plenty of room in the original Kiwi-1 enclosure but it could be put into a new one the same size as Kiwi-2 if it matters. Perhaps Peter could help with that and still keep it as a kit that could be built by end users locally?
If/when that kit had some time on it to discover refinements and corrections, perhaps someone might come along and produce, take orders and distribute and support it in larger quantities? This would take a lot of ongoing hours, wouldn't make anyone rich but it might make for a very small cottage industry. Maybe it should never become too tightly coupled with any Kiwi version.
Part of a new spin probably needs to be finding a non-Chinese fabricator where all the existing parts, including the presently sanctioned SI5351 v C, can be obtained and assembled. That remains a problem with JLCPCB, last turn I had to reflow the part myself for the Reference2. Parts availability and obsolescence is also a continuing headache.
Glenn n6gn
We should probably take this to email (again).
A few things. An accessory box should definitely be a formal KiwiSDR NZ product. This is because of a counter-counter-argument. Yes a second product is more overhead for us. But it can also be implemented with great efficiency using the current structure we have in place for Kiwi-2:
I haven't looked at RE/TR BOM component prices, but if it's like the Kiwi just a few components can account for a large share of the BOM cost. So there is tremendous advantage is buying those in bulk, in advance, at a large discount. We are now able to do this for Kiwi-2.
We can do some polling of the customer base in advance to see how much interest there is to get a sense of production quantities.
v1.696 Antenna switch with arduino netshield.
The indication of choosen antenns seems to work properly now.
BUT the grounding of antennas when no user connected do not work. (thunder season here now)
Switch to default was to set to YES after autoupdate, I changed to NO and restarted server
Ground antennas was correct set to YES after autoupdate (previous setting)
73:s de Kenneth / SM0OHC
Works fine here testing with the Beagle-GPIO backend.
Are you sure this isn't just a case of the code in your Netshield not working correctly for the "ground antennas" command? Or it could be a bug in the Netshield backend script I suppose.
Once again I would need access to a Kiwi with a Netshield to test. But I can't do anything about Arduino code..
@sm0ohc, @jks;
I am not seeing that issue on my version of the Netshield Antenna Switch. (v1.696) When I close the user page the antennas go to ground in about 20 seconds. When I connect to the Kiwi User page, I see about 10 seconds of empty waterfall / no signal before the Arduino energized the default antenna. I am ok with those delays.
JKS, I don't claim this to be a problem, Just an unexpected side effect. When I call up the Kiwi Admin page the Kiwi server triggers the Arduino to connect in the the default antenna. That's with no user connections prior to starting the admin page.
This does put the kiwi at some risk if admin activity is being done during a thunderstorm.
Mike N8OOU 73
Yeah, that's impossible. It doesn't happen here using the GPIO backend and there is no code that would do that. It's got to be more Netshield weirdness.
@jks
OK, I am 99 and 44/100 percent sure I saw that happen in the past. Antennas in the grounded state, switching in, when the admin page was activated. I was using a computer at the kiwi and heard relays click and saw LED indicators change state when I started the page.
I just tested that again, and you are correct. I was watching debug information created by my Netshield and the default antenna was NOT activated when I started the admin page.
So... My apologies for making that statement without testing the current version again to insure the issue is real.
However; in my quick little test just now, I did find this situation with this release. The Netshield can be controlled externally by using web browsers, or any other tool that creates an HTTP Get Request.
After calling up the Admin page I went to the "Control" tab and tried the Measure SNR Now Button. The Antennas were in the grounded state due to the "No User" state of the Kiwi. The admin page reported back that the antennas were grounded. -- That's Good.
I issued a Curl command to the Netshield from this computer to activate antenna 1. I verified the switch state with a "Q" Query command. I then clicked the Measure SNR Button again and was told again that the antennas were grounded. I waited several minutes and tried again. I still got the grounded error. The Query result from my computer again says Antenna 1 is connected.
Is the admin page measure button only testing if a user is connected? Should it query the switch to find the current antenna state and go from there?
Mike N8OOU
Not sure if that's a bug or a feature, but I don't see the frequencies anymore on the /?camp site
Yeah, just a bug. Already fixed for the next release.
Thanks
Revisiting the arduino sheild antenna switch.
Now v1.698, still the same.
I can not see that there is a bug in the arduino-code, since it is the same switch i used for years since oh1kk made the netshield-code and it worked/works flawlessly through the years and when I send the commands from browser direct eg "192.168.xxx.xx/?cmd=3" it reacts and gives med number 3 and so forth...
I tried to set "switch to default antenna when no user" and use cURL option with-:
"192.168.xxx.xx/?cmd=g" to make "ground when no users" and the netshield responds as intended.
You are doing an amazing job John !
Best 73:s / Kenneth SM0OHC