Independent tests prove lack of frequency hopping with XPS

[ricoalonso]

Yes, the 'receiver reboot' can be another explanation and with the fast reboot of XPS he may just haven't noticed it.

[sweetpea]

Or it could have hopped and he didn't notice it. The 400ms hop is as fast or faster than the reboot/reconnect time

[Wojcigitty]

If it takes 400ms to hop, and it reboots so quickly, why didn't Jim just make it reboot instead of hop when it finds interference?

It would reboot, find a clean channel, tell the transmitter to use that channel, and go on with its business. That is, if it reboots that quickly.

[Julez]

This procedure would have to rely on a firm backchannel. If the backchannel is unreliable, you might end up with a RX lintening on a different channel than the TX is transmitting on.

[wrightme]

	Quote: Originally Posted by sweetpea
	Or it could have hopped and he didn't notice it. The 400ms hop is as fast or faster than the reboot/reconnect time

I just did a quick unscientific test of reboot/reconnect time, and it is virtually instantaneous, from my GUESS, it is less than 1/2 sec.

XDP on for Analyzer. TX on. RX on. Links up. (Ch 50 on the spec anal).
Unplug the RX. Upon battery connection, the light goes directly green, and servo movement matches stick position.
With a stopwatch, I got 340ms, but that was mostly the time it took to push the on/off button on the stopwatch twice. The connection and servo movement were arguably LESS than that timeframe. It did stay on the same channel that the units were on before disconnecting the RX battery.

[sweetpea]

That would explain why it was so fast.

So in this case it doesn't appear to be a reboot.......since the system appears to restart on the same channel (like the new software upgrade for Spektrum)

[Wojcigitty]

	Quote: Originally Posted by Julez
	This procedure would have to rely on a firm backchannel. If the backchannel is unreliable, you might end up with a RX lintening on a different channel than the TX is transmitting on.

I thought that the RX was the "brains" of it...On start up, the RX finds a clean channel, starts listening/broadcasting on it, and the TX finds it and they link up.

Maybe this would require a TX reboot too, though.

[Julez]

	Quote:
	and the TX finds it and they link up.

That is the critical thing.
If the TX does not find it, you might end up in the dreaded "Zombie State".

[Daemon]

	Quote: Originally Posted by Wojcigitty
	I thought that the RX was the "brains" of it...On start up, the RX finds a clean channel, starts listening/broadcasting on it, and the TX finds it and they link up. Maybe this would require a TX reboot too, though.

The Tx doesn't know that it should start looking for the Rx until it's power cycled itself.
There's no guarantee that the Rx has moved, even if the Tx isn't getting acks back.

Read my thread on RCG that goes through all of this, with videos and such.

The first thing the Rx does whenever it's powered up, is check to see if the Tx is transmitting
on the last channel it was using. If it finds it there, it links instantly.
I call that hot link.

If it doesn't hear the Tx transmitting on the last known channel for about 1 second,
it then goes off and does a channel scan, finds a low noise channel, and starts sending
out "I am here" chirps, waiting for the Tx to come find it.

Assuming the Tx is on and actively scanning for the Rx, it will eventually find it, and link up.
I call that cold link. The whole process of finding a channel, and the Tx finding the Rx on
that channel usually takes about 3 seconds. BTW, when you first turn on the Tx it
also tries sending packets on the last used channel. If it gets an ack back from the Rx
it hot links instantly. If it gets nothing back, it goes quiet and starts scanning
all the channels for the Rx's "I am here" signal.

Zombie state occurs when you start already linked (originally through cold link)
and then something happens to cause the Rx to reboot or is power cycled outright,
while there is some reason that the Rx can't hear the Tx (noise, or range, or shadowing.. etc).
The Rx reboots/powers up tries to hear the Tx on the last known channel.
If it fails for more than a second, it then goes into cold link sequence.
Scans for a low noise channel finds it, and starts saying "Hi, I'm here, come find me".
Problem is, throughout this whole time, the Tx is still assuming that the Rx is on
the original channel. It can't know that the Rx rebooted, couldn't hear it,
and went off looking for another channel. For all it knows, the Rx *might* still be listening
on the original channel, may even be still receiving packets, but the Tx just isn't
able to hear the ACKs.
So the Tx just keeps retransmitting packets forever on the old channel. The Rx just sits on its
new channel, forever.
Zombie state.

In case it isn't clear why this is relevant.
If the Rx reboots for any reason, and does not hot-link, but goes and finds a new channel,
then the plane falls out of the sky. Period. If XPS has real hopping, this is not the
mechanism it uses.


ian

[ricoalonso]

So then it is 'agreed' that the Rx reboot can never be accounted to cause for what to have been observed change of channel and therefore could really had been a valid 'hop'. So the question remains: What caused it to hop. The poster stated that there were other 2.4GHz (ie. SPEKTRUM & FASST) flyers only that he did not know who/what was on the air or powered ON. Either it hopped because: (a) it perceived a 'potential interference'; or, (b) because of the presence of FASST (as another poster had shown that FASST caused it to hop).

Anyway, another XPS user promised to do another experiment in the field with his SA and with some friends with other 2.4GHz systems. It will be interesting to see his post in RCG later.

[Franz2002_01]

about a week ago someone on this forum posted that he would do some more testing on the hopping. Were there any results posted and did I miss it or were no results posted? BTW - I did 15 more flights this weekend w/o any problems with XPS (of course I dont know if it hopped or not...)

[sweetpea]

Kiwi and Xjet have been busy and have not posted about any other tests yet.

I saw this "hop" thread and thought it fit here perfect. As it appears the system did hop. As Xjet states; depending on why it hopped it may not be the most robust method. I'll grant him that until we have more proof. In the same breath though until we can prove it didn't actually hop this thread has been busted (on the premise the system didn't hop) The system hopped (at least onece) in flight with no ill effects to the user.

I'm eagerly awaiting more results from Kiwi. I'm also very interested on Kiwi doing his intial tests again with the powerlevel of an R/C radio and not what appears to me to be a higher level than an R/C radio.


Last bit of news......I'm trying to download a video and photos off a friends camera. At the North Las Vegas Field a 50% Hempel Edge 540 was flown yesterday in some viscous winds. It was powered by the XPS module with the new case and 10ch rx (not sure what version).

I got there after it had already flown once. After the 3rd flight they were packing up and I got pics of the install. RX was very low in the fuse and placed ontop of SM power exapander. The wires on the sides of the SM unit went straight up. Basically surrounding the RX with wires on the sides. No chance of them touching the RX but they appear to be inside the 2inch rule.

I suggested to the owner to move the RX higher and move the wires. I hadn't seen a GS plane fly XPS at my field since I crashed a long time ago. I have seen numerous smaller planes fly (including my own) with no issues though.