Perimeter sender: weird voltage/current state

[romualdas]

After assembling perimeter sender I have performed tests in room (5m wire + 2x5Ohm resistors), then have setup trial (250m wire (4Ohm) on the ground + 1x6Ohm resistor) outside.
Tests were successful in the fact that perimeter was registered by receiver and robot behaved as expected.
During both tests, if I would increase the voltage ~ above total resistance of wire and resistor(s), perimeter sender would stop working (1A protection of MC33926 would kick-in?) and yellow LED would blink.
Then I installed final (ok, I have changed couple of times since the first installation) wire setup:
- ~1cm in the ground
- ~300+ m with wire resistance ~6Ohm
- Initially I have added 40W 6Ohm resistor
- and setup output voltage on LM2596S to 11V

After that signal was very-very weak on the robot. To my big amazement, when I increased voltage to 16V and 24V not only signal did not disappear (yellow LED would not start blinking), but actually got registered by robot.
I tried to remove resistor and perimeter started working as it did before.
For a moment I thought that voltage on the perimeter is different, but I got osci from a friend and checked the voltage and signal on the output.
At that moment: there was 12V for 6Ohm wire resistance (attached image, measurement cable had 10x multiplier).

Settings for the perimeter sender are the following:
USE_PER_FAULT 1
USE_PERI_CURRENT 1
USE_POT 0
USE_CHG_CURRENT 1

So, even the perimeter is working, the question is:
What is going on?
What should I check?

 

[roland]

Hi romualdas,

the maximum voltage for the Nano is 7-12V.
Mabe you destroyed it now.

"1A protection of MC33926"
I don't think so. It can handle 3A.
Maybe the LM2596S stops to work or the charger who can handle only 1.5A and recognise a shortcut?


How strong is your perimeter magnetude 15cm away form the perimeter wire inside the loop? I read here in the forum, that 1900 could be reached.

"What is going on? What should I check?"
I don't really understand the problem now. Is the signal working but too weak?

 

[romualdas]

When I was testing with 250m wire (on the ground, covering ~ 80% territory) with resistor and voltage set to 10V (current less than 1A) - I was getting decent signal throughout the field.
After changing the wire, if I'd use resistor and appropriate voltage (e.g. 12V) - signal is very weak even close to the wire.
Without the resistor and with voltages ~12-16V, perimeter signal is registered and received throughout the field and it reaches even 2500 near the wire.

In other words, I have a perimeter that is functioning, but using weird setup:
- Only wire installed (no resistors in line) - ~5Ohm
- LM2596S is set to 14V
I'm referring to http://wiki.ardumower.de/index.php?title=Perimeter_sender_(English), where it says: "Adjust the potentiometer so that not more than 1 Ampere current flows." In my case it it would be ~3A.

The main question: why the difference between test and final wire install?

Regarding Arduino Nano: it works and according to https://store.arduino.cc/arduino-nano - it should be able to handle up to 20V, but I guess the recommended voltage is up to 12V.

 

[romualdas]

It might be that everything is Ok, it is just different from the setup during initial tests and from the info that is available in wiki.

I have just attached to the perimeter sender, updated software (set USE_POT to 1 and ROBOT_OUT_OF_STATION_TIMEOUT_MINS to 720) and took the measures:
- DC-DC voltage: 13.6V
- Perimeter resistance: 4.5Ohm (shortened wire a bit recently)

Some screenshots attached.
One shows the pfod app, where the mower is furthest from wire, another - near it.

Any thoughts, comments, feedback.
Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/log-output.PNG/

 

[roland]

I think it is normal, that when you bury the wire, that the signal is lower than normal. And it will get more worse when it is raining.
I just test my software because of changes in the firmware. Therfore I have time and just soldered a new sender today, at which I want to increase the voltage of the perimeter signal for testing if i get a better receive at the reciever..
I will test in the future the following:
I adjust the voltage of the sender to 9V. Then I will disconnect the Vin and GND of the MC33926. The MC33926 should not have any contact of the motor output to the sender PCB. Then I use another external powersupply which supports 5A for example and up to 30V. I connect this to Vin and GND of the MC33926. I don't go over 20V becaue 50AC is dangerous.

 

[romualdas]

Burned 2 DC-DC converters
Figured out that they don't handle above 2A current well for longer periods of time.

Anyway, I've updated the perimeter. Before, I had kinder playground exclusion zone, because the mower could get stuck there.
After fixing the environment/playground I removed the exclusion zone (had a difficulty exiting it while tracking home anyway) and noted that perimeter signal improved even with lower voltages and resistor in series with wire.
And the perimeter is now basically a U shaped area ~1000sqm of cutting area.
So far, it looks good now

Also, a little side note (unrelated, but might be useful for others):
While experimenting with the perimeter I mixed up polarity on sender and mower was reporting outside, while it was inside the perimeter.
I thought I will use pfodApp to set `Swap coil polarity` to *Yes*.
After changing the setting mower was reporting inside, but for some reason perimeter signal was very weak.
After changing polarity on sender and setting `Swap coil polarity` to *No* everything went back to normal.