Adding one more photon to an area causes other photon to go offine


#1

I am a very interesting problem with four photons.

In a residence I have four photons on the first floor. The router/wifi is one floor down.

The first floor has three garage doors each with a photon. Each of these three photons have an external Antenna and the code initializes the Antenna as:

WiFi.selectAntenna(ANT_EXTERNAL);

Now the fourth photon is about 20 feet away in the home office. The fourth photon has only the internal Antenna.

Here is the issue.

When the fourth photon in the office is powered up. One of the three garage door photons will immediately go off line and stay offline. The office photon will stay on line.

Unfortunately, at this home I cannot get access in the router since it’s a router modem using a 5GHz link. Outside the home an antenna is mounted and the ISP connects to this home. Moving the router to upstairs would require the ISP to come on site and relocate. Since it’s winter this work has to be done with considerable snow on the ground and roof.

Now the three photons in the garage are very stable and stay on line until the office photon is powered up. I am not at the location so I can’t run a NetSpot program to see how crowed the WiFi b/g channel is. I don’t think other WiFi networks are the issue since the closest hour is several hundred feet away.

The RSSI for the three Photons in the garage are

D58 -60 RSSI / Firmware 0.8.rc2 ( closest to the router and the home office)
D57 -67 RSSI / Firmware 0.8.rc1 ( between D58 & D56 D57 goes off line)
D56 -65 RSSI / Firmware 0.8.rc1 ( longest distance to the router)

For the office particle I don’t have an RSSi read. Office Particle Firmware is 0.7.0.

So what is the cause and how can I get Door57 to come online and stay on line?

ChipMonk


#2

Could this be an IP collision issue?
Either one (or both) devices are set to use the same static IP - or were previously an since it’s sticky setting still are.
Or your DHCP server has run out of IPs to serve out to new devices an hence a more powerful device will eventually claim and hog the last free IP address.


#3

Can you ask one of the home owners to connect their mobile to the same network to see what response they get to the connection attempt?


#4

Hi ScruffR,
Good Questions.

I don’t know how to set a static IP on a particle so I assume the particles must be using DCHP.

The other theory that the DCHP server has run out of IP not the case. Presenlty about 14 people are in the residence and all have at least one smart phone using WiFi.

When I go to this location I need to see if someway I can get the ISP to allow me inside the router and see if the number of IP addresses are limited for some reason.

Thanks for the ideas.

Chipmonk


#5

Good idea. I will give this a try when the home owners are awake.

ChipMonk


#6

I checked and all the cell phones and tablets are able to connect to the network. No users are reporting any issues with the WiFi network.

One additional photon is too many.

ChipMonk


#7

I understand now why you said this was an interesting problem!

There are at least two things different between the garage and office photons, internal antenna and Device OS release on the office unit.

I know that it does not make any sense but probably a good idea to bring all four devices up to the same, latest, release candidate so as to reduce the number of variables, ie 0.8.0.rc.12.

Q1. Have you set WiFi.selectAntenna(ANT_EXTERNAL); on the office photon by mistake? Perhaps doing so causes some weird interference???

Q2. Is there any pattern to which of the three photons goes offline?

Another way forward is to use the logging facility from 0.8.0 to view what error messages Device OS is outputting but I understand that getting access to the device is problematic…

Q3. Another trail that you can go down is to pursue this statement “When the fourth photon in the office is powered up”. How is this done exactly? Am thinking if it is being powered by an RF noisy plug pack this could be supplying the interference. This is easily tested. Just turn on the plug pack, without having the office photon plugged into it.


#8

Here are my answers.

Q1. Have you set WiFi.selectAntenna(ANT_EXTERNAL); on the office photon by mistake? Perhaps doing so causes some weird interference???

CM: I rechecked the code and the office photon does not have the external Ant Selected

Q2. Is there any pattern to which of the three photons goes offline?

CM: The only pattern is that the D57 will go off line. D56, D58 do not go off line
I have not used the logging function of 0.8.0. Have to look up how to use this. When I am on site it may be possible to get physical access and hook up the serial port for logging ( I need to read up on how to use the logging)
Q3. Another trail that you can go down is to pursue this statement “When the fourth photon in the office is powered up”.

The office photon is powered by an AC wall wart 115V plugged into the wall. To get the office photon to come up the power had to be cycled. Currently I can’t get the occupants to disconnect the AC wall adapter from the photon and leave the power adapter connected. Getting the power connection re-connected to the photon correctly may not happen.

Thanks for the questions. I hope to make it to this location early next year.

What i would like to do is run wireshark and see what the photons are doing.

Can someone point me to a good particle reference on using Wireshark to monitor photon traffic?
I did this several years ago when I was in Asia and had a problem with a core.

ChipMonk


#9

No, just when you haven’t attached an external antenna reception would be very bad.

Antenna selection is a sticky setting, set once it will remain that way till the setting is altered, even if it’s not part of your current firmware.


#10

Yes you are correct!

I will try to re-flash the office photon today and select the Internal Antenna.

This will eliminate this issue. The other photon are all flashed with code selecting an external Antenna.

CM


#11

Has the issue been sorted?


#12

I still have not been able to access these photons. Hopefully soon I will be at this site.

Do you have a wireshark setup to monitor traffic on the WiFi? I had a setup several years ago from a particle engineer when we were debugging an issue.

If you or ScruffR can point me to a post that discusses wireshark monitoring that would be helpful when I go to this site.

ChipMonk


#13

The mystery continues.

i recently had all four devices power cycled and only one was able to connect to the cloud.

Next I had the router power cycled and not a single photon was able to connect.

What port(s) are used by the cloud?

I have a feeling that the ISP may be blocking and I can’t get inside the router modem. The iSP used here has a 5GHz microwave link to the house with it’s own equipment.

i have an upcoming trip to this site soon and will be able to get more data.

Chipmonk


#14

Made the trip to the residence and found all of the photons with a solid cyan LEDs. Power cycled each one photon. Each photon was okay and working as expected.

All units were updated to 1.01 firmware so I have more information. I also installed better external antennas that improved the RSSI. The lowest is -69 which is the longest distance to the router.

So no conclusions. Thanks to all who replied.

ChipMonk