Wi-Fi internet keeps dropping out despite strong signal

[Sakuya]

Hello, I have some Wi-Fi troubles and hope somebody can give me some pointers. I live in a condo and I have an Asus RT-N66U router in a bedroom and a Netgear N750 WNDR4300 as an access point in the living room. They're connected by hard wire ethernet through a wiring compartment through the wall. I set them both to have the same SSID and WPA2 password. Wi-Fi signal is strong 4-5 bars everywhere in the house.

Problem: When using Wi-Fi (2.4ghz band) on all my mobile devices, I notice the internet cuts out a lot. But I still see 5 bars of Wi-Fi. The webpages/social apps just refuses to load. I fix this by turning off Wi-Fi on the mobile device and turning it back on. It usually works for a few minutes until it happens again. This happens regardless of how many devices are using the Wi-Fi. Computers with hard wire ethernet work totally fine.

I have drawn a map of what my setup is like:

- Gray walls = concrete
- Pink area = My Wi-Fi problems are concentrated in this area, it happens less in the rooms with the routers.


[Attachment 43947 - Click to enlarge]

I have little to no router setting advanced customizations.
I have been upgrading the Asus stock firmware but it hasn't fixed the problems.
I also tried power cycling everything and switching to the least-used channel.
I'm wondering if it's because the pink area is between the 2 routers, and it's trying to decide which router's network to connect to?

Any help would be greatly appreciated! Thank you so much!

[pandy]

Current version will be shorter and not explaining why too strong signal may be not good (counter-intuitive to many people thoughts about WiFi) and many other things.

You have three options:
- first search for better channels and if possible switch to different spectrum (not 2.4GHz but for example 5GHz)
- second lowering output power in your AP's - this may help reduce polluting of EM spectrum and reduce observed misbehaviour
- third most recommended by me, remove one of AP's and try to find sweet spot for only one AP. This is most difficult task as it may require lot of work and tests.

To search best place you can use very good (IMHO one of best) open source software from Ruckus, Zap is available as command line but also as GUI (ZPerf - but it is less stable than Zap - for sure WLAN should be not tested with jperf and similar software), you will need two computers, one must work as server, second as client (they role can be changed so you can test both ways).
If there is interest i can provide software and scripts (used by me to test WLAN performance)

And my last comment - don't expect miracles, be reasonable in your expectations, if you searching for stable connectivity go for cable (copper, fibreglass).

Previous version was way longer and explain some common misconceptions related to RF signal but for now i lost patience to wrote this once again.

I'm electronic engineer with professional experience on this area.

[Case]

Make sure only one of them is doing DHCP and NAT (the other in “bridge mode”). Perhaps set different wi-fi channels for each and different from your neighbors (you can scan for that), so that interference is minimal.

[Sakuya]

Thank you for your in-depth replies. I downloaded inSSIDer and checked Wi-Fi channels. My apartment has plenty of neighbor networks. I changed my routers to 2 different channels (as seen in my screenshot below) that are the least occupied. As you can see, my signal isn't the best due to concrete walls. I seem to still get occasional drop outs but will continue to monitor it.

On both routers, Tx power is set to 100%. What would you recommend to lower it to? I am only able to set it in 25% increments.

Also, do you have a recommendation of the sweet spot to place one general router? I've tried before but due to power outlet and furniture arrangements, it was easier to setup an AP on opposite end of the house.


[Attachment 43959 - Click to enlarge]

The blue star network is my Netgear AP which my computer with inSSIDer is closer to. The other network listed is the main Asus router.
The Netgear AP is set to use Wireless Access Point mode (I'm guessing that is the same as bridge mode?) with fixed IP address.
I have tried the 5ghz band networks but their range is not that good and I'm usually disconnected when I leave the room the routers are in.

Thank you again for your help! Much appreciated.

[pandy]

For signal propagation methods similar to ray tracing can be used ( for example https://www.remcom.com/wireless-insite-em-propagation-software/ ; http://www.cwins.wpi.edu/projects/ray-tracing.html etc) however i'm not aware of any open source tool to perform such function (and ordinary graphical ray tracing will not work due significant difference in wavelengths).

Afraid that you have no option other than empirical and some intuitive approach. In longer version i've described some problems and why too strong signal can be as bad as too weak signal - briefly i can say this is about amplifier linearity, overall noise level and competition between different transmitters that must share same spectrum - imagine two screaming kids in one room - at some point there is nothing except high noise level.
From my experience for single AP best location looks like corridor nearby bathroom. You can also use single AP and with help of RF cable move antenna to required rooms (i know how this sound but at a cos of loosing MIMO you may gain stable connection with sufficient bitrate) - https://www.amazon.com/Antenna-RP-SMA-Extension-Cable-Router/dp/B005C8AYG0 - it doesn't look expensive, perhaps you can find it in proper colour that will fit in your expectations - cable attenuation may help in your case - IMHO definitely worth to try.
You can also try trick with different antenna polarization (place AP's with antenna orthogonally) - this should separate both source by approx 20dB however in environment where reflection dominating it may be not visible (but it may improve overall link as probably your mobile devices working in non-optimal positions anyway) - RF is not trivial things - lot of people thinks about RF like technology that just work out of the box where reality is very far from such expectation - RF works well in very narrow scenarios - in most of real life scenarios plenty of effort must be invested to get satisfactory results.

You just need to try various settings (reducing by 50% should cut range by 4 times), keeping power as low as possible (your network performance) is always recommended (even to protect pollution of the EM spectrum).

[Nelson37]

OK, a problem on a system with many components. First step is to simplify the system. AP's are often problematic, disconnect and power off the AP and test with router only. One wifi transmitter and signal to diagnose. For testing period, stay in the same room initially, second round go to alternate room, where AP is currently located. If feasible disable all other internet usage while testing. Not refrain from, DISABLE, as in power off the device. Not it's not a likely cause but if other devices are turned off, then they damn sure are NOT the problem.

Testing - what is the shortest period you have observed with no issues, and what is the longest? Day / Night a factor? Highly variable? Use standard time measurements, a rough guess is OK. Take the longest observed interval, double or triple it if feasible. Meaning use wifi with router ONLY, AP disconnected and powered off. This is to insure a true solution has been discovered.

The thick grey walls, are they by any chance stucco covered? Concrete not usually a problem, but stucco is. This is usually installed with a sort of chicken-wire backing and that crap is murder on wifi signal.

If possible, move router closer to, facing, or actually in the hallway connecting the two main areas. This is ONLY for AFTER initial test to determine if AP is the problem or not.

Signal strength being too strong is not a problem I have ever had on any installation. Signal interference, blockage from stucco, metal, or really old wood, faulty equipment, AP issues sure, but signal too strong? I would spend zero time on this, and I get paid to fix this stuff.

[pandy]

Signal strength being too strong is not a problem I have ever had on any installation. Signal interference, blockage from stucco, metal, or really old wood, faulty equipment, AP issues sure, but signal too strong? I would spend zero time on this, and I get paid to fix this stuff.

So i would highly recommend to you understand what is LNA and how it is affected by strong interferer present in same spectrum side to weak signal, what is IP3 and P1dB, understand linearity in power amplifiers and how linearize power amplifier (common problem in WLAN chips). This is basic stuff for every WLAN chip (usually hidden from user perspective). You can easily check all this empirically. Use Zap to test and describe link quality over time.

[october262]

do you have neighbors living around you in that condo ?? it could be interference from other wireless equipment and microwaves

[KarMa]

do you have neighbors living around you in that condo ?? it could be interference from other wireless equipment and microwaves

If the images in post #4 is anything to go by, yes there are a lot of other competing Access Points using channels 1 and 6. Channel 1 seems to have 8 other APs in range using that frequency, and Channel 6 has 5 other APs using that frequency along with 5 other APs using nearby frequencies that interfere with channel 6.

My advice to OP is that if moving to 5Ghz seems to work better, apart from the standard range issues, then 5Ghz might be the only option. As 5Ghz is much less used spectrum and is a less invasive frequency from external Access Points.

[Sakuya]

Testing - what is the shortest period you have observed with no issues, and what is the longest? Day / Night a factor? Highly variable? Use standard time measurements, a rough guess is OK. Take the longest observed interval, double or triple it if feasible. Meaning use wifi with router ONLY, AP disconnected and powered off. This is to insure a true solution has been discovered.

The thick grey walls, are they by any chance stucco covered? Concrete not usually a problem, but stucco is. This is usually installed with a sort of chicken-wire backing and that crap is murder on wifi signal.

The walls are not stucco covered. The shortest period when I have no issues while using the 2.4ghz band is 5 minutes. So far I haven't had any issues while using the 5ghz band and I have been using it for 10-15 minutes.

Thanks for everyone's help. I'll try to test the best location for the router. For now I'll be using the 5ghz network.