Odd behaviour?? (not Drop and Mag sensor related)

[Brett]

[EDIT: Thread title changed for clarity]

I have an VX Signature Pro here that has strange behaviour. It appears at first to be working normally, but occasionally will be travelling and stop so suddenly that the rear rollers lift up and 'clack' down on the floor again. Often it will then say it is "determining it's position" again before moving off. It soon becomes apparent that all is not well and it starts getting lost and confused trying valiantly to find a way through a wall and then pops up with a "vision is blocked" error. If prompted to return to dock it heads off in the opposite direction and after losing an argument with another wall, the error pops up again.

Even though this is a new unit (5 months old) and I doubted the LIDAR motor is worn but I swapped it anyway. No change.

While fiddling with NeatoControl (what a great tool!!!) I noticed that the Drop sensor readings stay on 0 and the mag sensor reading does not change even if a bring a magnet close to the sensor.

While this seems wrong to me, it does not sound consistent with the above odd behaviour?

Any ideas?

Cheers.

[Lewiy]

You should first turn on the test mode (TestMode on) in the Commands tab. Then you can test the sensors. More info -- http://www.robotreviews.com/chat/viewto ... 64#p131022. In addition the sensors can be checked in the Neatos log (sometimes it is more informative), but it's a little more complicated.

[glnc222]

I think Test Mode is needed only for motor activation commands. Neato Control polls sensors with GetAnalogSensors etc., which do not need Test Mode.
The Cliff Sensors and the Wall Sensor can be tested by observing the change in reported distance moving a hand in and out within a couple inches of the sensor.
The problem reported, however, is intermittent. Testing the sensor may not detect such a failure. Such problems are often the result of a weak power supply to the digital IC's. The power depends on more than the batteries, including voltage regulator components etc. between the battery and the IC's. The common approach would be to replace the system board to see if that made any difference in the behavior. Troubleshooting digital systems is expensive. A battery known to be good is essential.

I am not sure the cliff sensor should show zero instead of some small number. It will go high when a cliff is detected.

The Mag Sensors need to be tested with a piece of Neato Magnetic Strip to get the correct magnetic pole against the sensor, I think a South pole -- specification in the Hall Effect sensor data sheets.

Any unit so new with malfunctions would normally be replaced under warranty, though this is difficult in some places outside the U.S.

[Lewiy]

I think Test Mode is needed only for motor activation commands. Neato Control polls sensors with GetAnalogSensors etc., which do not need Test Mode.

Without TestMode 'GetAnalogSensors ' gets static sensors values, the dimension of which is unknown (as Brett shows). In TestMode the same command gets the sensors values in the proper dimension (e.g. for WallSensor -- mm) and with proper reaction to obstacle. Finally, 'GetAnalogSensors raw' in TestMode gets the sensors values in mV.

The problem reported, however, is intermittent. Testing the sensor may not detect such a failure.

So the Neatos log may prove more useful for troubleshooting of sensors and critical voltage dip under real running conditions.

[glnc222]

Thanks for the important detail on Test Mode. I will add it to other USB instructions.

[glnc222]

Neato Control program author Hex advises an on screen switch for Test Mode will be added in a future version.

[Medtech]

It seems to me that the author simply can not put the connector to the board or the board has no + 5V all the red wire connector P5. Check it is necessary.
If there is 5V then you can get them from the C19 or L3 which are adjacent m connector P5, just solder + 5 V to any red wire P5 on Russian it sounds like: brosit peremichku - throw jumper :+))
apparently there is a problem on the board in one of the tracks under the connector P5

[Brett]

I have an VX Signature Pro here that has strange behaviour. It appears at first to be working normally, but occasionally will be travelling and stop so suddenly that the rear rollers lift up and 'clack' down on the floor again. Often it will then say it is "determining it's position" again before moving off. It soon becomes apparent that all is not well and it starts getting lost and confused trying valiantly to find a way through a wall and then pops up with a "vision is blocked" error. If prompted to return to dock it heads off in the opposite direction and after losing an argument with another wall, the error pops up again ...

I never did sort that one out, and I now have another doing just the same. (Also a newish Sig Pro)

Also I have an XV21 that is playing up. (I'm busy uploading a clip - I'll add it when it's up). The unit does a bit of wondering around the walls, starts to follow the wall (as it should to draw an area to clean) but then it breaks away from the wall, and wanders around a bit more, then will do the same bit of wall following again and eventually bumble it's way through cleaning the whole room and then move on to the next room - all the time moving in a hap-hazard Roomba like way! It is clear from the movement that the LIDAR is working, as are the IRs. It is further clear that all the nav bits are working because this one goes straight back to the dock when commanded to do so, just as it should?!

I have not run NeatoControl on it yet* will do that once the video is uploaded etc.

* I'm running Linux (Ubuntu) and as far as I can tell I need to swap back to "Windows mode" to run NeatoControl - is this true?

Video:


(-8

[JEfromCanada]

It looks to me like your Neato is having trouble with your carpet. Check out the threads about sliders.

[Brett]

Nope - does exactly the same on the hardwood floors...

???

[andrew881]

Mine has exactely same problem as yours. I have a neato xv essential so i think same SW. It travels with weird patterns roomba-style but yet follows walls avoids obstacles and tends to move parallel to walls. In spot cleaning it is ok, well, sometimes does a bit of zig zag and does not do totally straight lines, but generally is ok.
And mine too almost always gets stuck in a corner or simply tries to fight with a wall then gets error "free my path". Exactely same behaviour. Mine too enters a room does some cleaning going over and over the samw spot then goes out...this is not intelligent cleaning at all.
I am thinking to replace the unit, i am out of warranty (bought used item) so i should buy a new one for 300 selling mine for about 150. No idea why but here in Italy they cost one half more.