[karla]

@Cliff,

... For a test, do you get disconnections at a distance where you have about -40dBm?

It seems in one of my earlier posts here, using the first of three bad OPLMs, it did drop out at -40dBm even -35.
Thats when I was flying it.
Look again here https://forum.librepilot.org/index.php?topic=3999.msg27694#msg27694

[TheOtherCliff]

I will take that with just a grain of salt considering that (I recall) you have no dropouts when bench testing close, but do have dropouts when bench testing a bit farther or with moving around.

I imagine that antenna orientations etc when flying can make it vary a lot. 

[f5soh]

Splitting 10mhz of band (from 430.000mhz to 440.000mhz) into separate 40khz channels yields exactly 250 channels, not 251.

First seems you do a mistake between channel spacing and "40khz channels" so related to the frequency bandwidth while transmitting or receiving.

Dividing a 10Mhz band by 40KHz (the channel spacing we use, see bellow why) give channels numbers from channel 0 to channel 250 (as displayed in GUI) and the total channel count is 251.
Those 251 channels are possible hopping channels available for link, and the spacing between this hopping channels is not related to tx/rx bandwidth.
Simply we have a 10Mhz band (430-440), hopping channel value from 0 to 255 and 10Khz step size: answer is divide by 40Khz and get the maximum of hopping channels.
Next a set of channels is generated, at this point the tx/rx bandwidth is taken in account assuming the bandwidth is bigger at higher baudrates. This means two channels effectively used cannot be #10 and #11 @100Kbps but more likely #10 and #13 @100Kbps.

[TheOtherCliff]

But you are not allowed to go outside the band and using 251 channels forces it outside the band

Imagine that the band is 10mhz wide and we use 1mhz channels.  You can fit 10 such channels "inside" the band, but we fit 11 channels (in this example it would be channels 0 through 10).  Both the bottom channel and top channel overflow outside of the band.  This is not legal band usage.

I have no idea whether or not this causes any RF problem other than band compliance.  I would guess not, but it's possible that these bad clone RFM22B would work correctly if this were changed.

https://forum.librepilot.org/index.php?topic=3999.msg28276;topicseen#msg28276

[f5soh]

But you are not allowed to go outside the band and using 251 channels forces it outside the band
This is not legal band usage.

Already answered:

We simply talk about the nominal frequency, like every transmitter/receiver in the world and user need to define the min/max channel according to the local regulations.

Last time you buy a legal FM Tx/Rx you may ask the seller to disable first and latest channel into the band used.
For reference, the UHF bandplan from ARRL:

[TheOtherCliff]

This is wrong.  Ask Brian.

Default should be what is right.  And if the GCS describes channel centers as it seems, then the channel centers are displaced by 20khz.  There is a very minor chance that is what causes the issue with these particular clone RFM22B.

[f5soh]

This is wrong.

What's wrong ? The fact the first channel (0) is 430Mhz ?

[TheOtherCliff]

Yes.  First (40hkz) channel should be centered at 430.020 so that the first channel extends from 430.000 to 430.040

In addition, there should be only 250 channels and the last channel should be centered at 439.980

Users (especially non-technical users) whose country allows 430.000 to 440.000 will think the 430.000 to 440.000 default is set up correctly for them when it is actually illegal in their country because GCS saying 430.000 to 440.000 actually means 429.980 to 440.020

Yes, you can say that people living in such countries should just override the default and skip channel 0 and channel 250 and change it to start at channel 1 and end at channel 249.

[f5soh]

40Khz spacing is not related to frequency deviation (frequency deviation can be 30KHz, up to 150Khz), read message above.
So your logic about the channel centered at 430.020 or 439.980 still not "legal" from your point of view.

Just a remark about the so precise 430-440Mhz band where you can't have any frequency deviation outside: Why the 430 or 440 limits are not even mentioned in ARRL bandplan ?
According to the bandplan, the 430 and 440Mhz frequencies are in ranges dedicated to ATV. Especially the upper limit should be avoided to transmit because the 439.250Mhz used for input in ATV repeaters.
Again, your 40Khz limitation at upper limit changes nothing here assuming the ATV uses more bandwidth and you are transmitting in frequency range where a repeater can receive your signal.

User should select carefully the channel range, not only thinking about a 40Khz (unrelated freq. deviation) but according to the bandplan and local regulations.

[TheOtherCliff]

No matter how you do channel width vs. separation, putting channel center on the exact edge of the band is wrong.

If someone were trying to get this certified, say in a country where the band was 430 to 440, the certification would fail because emissions extend outside the band.

It looks like UK allows exactly 430 to 440.  The average UK user would think the default is perfect for them.  But it isn't if it actually uses 430.000 as a channel center.  (at any baud rate)

The best way to handle this would be to input the min and max frequency and have the code determine min and max channel centers and channel separation based on air baud rate.

I do understand that the deviation goes up as air data rate goes up.

[f5soh]

Glad you finally understand your "40Khz rule" was unrelated.

Since you talk about the "average UK user", here is what the Ofcom (the communications regulator in the UK) says:

Can I use an Amateur Radio licence for model control including airborne use?

No. Amateur Radio is a service intended to allow hobbyists and enthusiasts to experiment with radio. As licensed Radio Amateurs may operate transmitters at relatively high power levels in various frequency bands, the licence is available only to those who have demonstrated the necessary competence by studying for and passing a special examination.
We receive enquiries from those (principally non-amateurs) who wish to fly radio-controlled aircraft, such as quad-copters, fitted with video cameras (used to view the ground from the aircraft or to provide First Person View (FPV) to aid the control of the aircraft.).
There is a belief that the use of higher power equipment can be authorised by applying for an Amateur Radio licence. This is wrong. Amateur Radio licence expressly prohibits use in any aircraft or airborne vehicle. This restriction is not relaxed for radio-controlled models, airplanes and balloons.

This is not only UK, but most countries do not allow transmitting from a aircraft.
There is many others rules violated implying "the certification would fail", and anyway you transmit in ranges where the digital mode / hopping / bandwidth we use do not match the bandplan rules.

[TheOtherCliff]

The point I have been making all along is that using a center frequency that is actually at the exact edge of a band is illegal and that the way it is currently written it looks to the less technical user that he is set up according to his country's rules, but it is not.  I also would guess that no one else (other than OP derivatives) breaks up a 10mhz band into 251 channels.

In 16.09 where the user does not even know the air data rate (and thus can't determine the bandwidth), he can't set it up without knowing the baud rate relationship to the air data rate.  He would have to read the code or ask someone who knows what the air data rate is to cross reference the bandwidth in the RFM22B doc to be able to determine what channels to use.

In next where the user can see the air data rate, he guesses at air data rate that is necessary to make his baud rate work.

[f5soh]

Where you see the band is "430-440Mhz only" in the place you live ?
Assuming you are is ITU region 2, the UHF band limits are 420-450MHz as detailed in previous ARRL bandplan posted where a 430 or 440 number do not even appears.

[f5soh]

I also would guess that no one else (other than OP derivatives) breaks up a 10mhz band into 251 channels.

At least for OpenLRSng, used in common UHF rc modules, the min/max limits you can use are 413-463MHz...

[TheOtherCliff]

Where you see the band is "430-440Mhz only" in the place you live ?
Assuming you are is ITU region 2, the UHF band limits are 420-450MHz as detailed in previous ARRL bandplan posted where a 430 or 440 number do not even appears.

I never said this was the range where I live, but someone may live where there is either a 430 min or a 440 max.  And for those, refer to my previous statement about looking like it is configured for them but it is not.