|
I appreciate the work put into the tests but without actually testing the responsiveness of the system
during the test, I don't see that it proves much of anything. Yes it could not be made to hop,
but what we don't know is if it really *needed* to hop. I assume XPS Tx antenna was on
and it was running at full power during all the tests shown in the video.
If the Rx was still maintaining full control (which we don't know because nobody checked), then there's no reason it would need to hop. Why would we expect it to?
If it was totally blocked by the noise (which we also don't know), then it
would not hop, by design, per admission of JD. And frankly, while JD and apparently Kiwi too
thinks it's a silly idea to hop *after* onset of failsafe, that is frankly the only
case I care about. That's the only time it makes sense to me. "Potential noise" is just
too nebulous and you could easily get yourself in more trouble hopping when it's
not absolutely necessary (still maintaining control but there's more noise than
there was before) than waiting until the link is actually lost long enough
to cause failsafe when you *know* that it can't get any worse.
And as for "first tier" versus "second tier" systems, with the implication that
Spektrum having two frequencies to XPS's one (100% better) makes it first tier is still
a bit short sighted. I watched a Spektrum DX7/AR6100 combo lock out and fall out of
the sky twice the other day in a non-motorized Easy Glider. After our ground testing
with very surprising results, and my subsequent thread in the Radio forum on RCG, I come to learn
that if a Spektrum AR6100 Rx (maybe all models) loses the signal on both channels for any
reason for more than about a second, it goes off into la la land and scans for the Tx
on all 80 channels (assuming that the Tx has been turned off and back on) and may not find it
for another 5, 10, 15 (and in my tests 25 and 40) seconds before it finds it. People are
telling me this is "normal" behavior. And no, we're not talking about the known
low voltage reboot issue, for which there is a fix. We reproduced this behavior on a fully
charged 6V battery pack, on video. All it needs is anything blocking the Rf signal
(range/orientation/shadowing of Rx or Tx) for more than 1 second. At *least* XPS always
hot-links instantly as soon as the Rf signal can get through again. The other model Spektrum
Rxs basically try to avoid this issue by having better diversity (more Rxs and/or more widely
spaced antennas), and that's better, but if the link is lost anyway, then you're still screwed for
that 5-15 seconds while it scans and I don't know of many planes that can survive that long
without control. That's not Tier 1 behavior as far as I'm concerned.
BTW, with regard to 2.4Ghz video transmitters. *Most* people still use 100-200mW.
More people these days are buying 500mW units, and only a small handful of people use
1-2W systems. Anything above 100mW requires a HAM license to use legally. A 500mW video Tx
placed between XPS Tx and Rx will easily swamp the signal if it's on an overlapping "channel". And
contrary to what Kiwi said, the video transmission is fully analog and consumes pretty much 100%
of the bandwidth on its channel 100% of the time.. It may look similar to the rectangular
shaped spectrum of say a Wifi router but it is *not* spread spectrum, it is much higher power
density, it does not play nicely with *anything* close up. If I turn my video Tx on on certain
channels inside the house, it will just kick everything off my Wifi router instantly. In terms of
making XPS upset, a single 2.4Ghz video Tx of sufficient power is a pretty good noise generator
(I've got a bunch of videos showing how I used mine). You can vary the distance between them
to increase and decrease the noise floor gradually.
Anyway, back to the topic at hand.
I still think the "How do these systems actually react to noise" (as in what they actually
do, rather than what the manufacturer claims they do or don't) would be more useful.
Test like this.
1. Increase noise floor on all channels used by the system until onset of failsafe.
How much does it take?
2. Increase noise on all channels used by system gradually while monitoring actual
responsiveness. Do they slow down? Do they get glitchy? Intermittent.. etc
3. Repeat test 1, and then immediately shut off all sources of interference and see how long it
takes before they become responsive again (based on my testing, we'd see XPS come back
instantly, and Spektrum take however long it takes.. 5-15 seconds). After the change that JD is talking about in the software to make it hop after onset of failsafe, that might make XPS take
longer to relink because it has to scan (like Spektrum does now).
4. Introduce noise that is narrower than the SS channel, overlapping that channel, and see how
it reacts. Basically testing the spread spectrum ability of each system. XPS has 12 wide
SS channels so may be more resistant to narrow band interference, than Spektrum for instance.
FASST has very narrow invidividual channels, but it sort of simulates one big fat channel
that covers the whole spectrum. I agree with XJet's assessement of how it'd probably react
to interference. Saturate half the spectrum, it'll run at 50% responsiveness.
We're talking about ability to work through noise to gauge whether the designed in capabilies
are any good. If we discovered that XPS could function normally through twice the level of
noise on its single channel, than Spektrum can on its two channels, would that
level the playing field?
ian |
|