Samsung Galaxy Nexus doesn’t support 20/40 MHz Wi-Fi channel bandwidth

Stick to 20, kids
Stick to 20, kids

Learned this the hard way over the past few days after I switched my ASUS RT-N66U‘s channel bandwidth setting from 20 MHz to 20/40 MHz. For what it’s worth, here’s what ASUS’ official documentation says on the setting (emphasis mine):

802.11n can combine two 20 MHz channels to form an effective bandwidth of 40 MHz. 40 MHz enables higher data transmission rates to be achieved as compared to 20 MHz. When you select 20/40 MHz mode, the router decide to use 20 or 40 MHz based on the interference/contention the router detected. Care should be taken when using 40MHz mode, the legacy client may not be connected to the router.
However, when using a wider channel bandwidth, there are fewer channels available for other devices, making more interference/contention with neighboring WLANs due to increasing overlap. In order to avoid excessive interference, the Wi-Fi Alliance develops an advice: setting 20 MHz in 2.4GHz as default. 40MHz is still appropriate for some situations, e.g. in a warehouse, but we do not recommend that using 40MHz in the 2.4GHz band for dense residential areas.

I enabled it anyway, and noticed my Galaxy Nexus frequently being unable to access the internet despite being connected to the router. Switching back to 20 MHz solved the problem. Given the Galaxy Nexus’ release date, apparently “legacy” for ASUS means anything from late 2011 or preceding. For the record, the phone’s Wi-Fi radio is the Broadcom BCM4330XB2KFFBG. Broadcom radios have been suspect recently, so this isn’t surprising.

And yes, I know ASUS recommends 20/40 MHz not be used for residential settings, but my Dell Precision M4600, Amazon Kindle Paperwhite, and Sony Vaio Fit 15 E all handled it without complaint.

UPDATE: Also, disable “b/g protection” for 2.4 GHz frequency. Then, forget the network on the Galaxy Nexus and then reconnect to it.

The ASUS PA248Q doesn’t play well with active adapters (Updated)

UPDATE: After experiencing a similar problem with my Dell U2713HM, I’ve concluded that the real issue isn’t with the monitors, but with the cable length. Basically, try to avoid anything over 6 feet long for DVI-D.

I don’t normally write entire blog posts on the shortcoming(s) of a particular product, but when implementing a simple setup takes 3+ weeks* it’s probably a good idea to write about it so others don’t make the same mistake(s). The short story here is that the ASUS PA248Q doesn’t play well with (mDP -> DVI) active adapters, so don’t bother trying to use them with the it.

The particular setup I tried was AMD Radeon 7870 -> StarTech mDP to DVI active adapter -> DVI cable -> PA248Q. This worked for about 5 minutes, after which the monitor would repeatedly go in and out of sleep, claiming it detected no signal. No amount of input cycling or power cycling of both the monitor and PC helped. I wound up having to connect the PA248Q directly to my PC via DVI and then use the active adapter for a bizarre AMD Radeon HD 7870 -> mDP to DVI active adapter -> DVI to HDMI cable -> Dell S2240L connection for 1 of the remaining 2 monitors. I’m honestly shocked the latter worked with a consumer monitor while the supposedly professional grade PA248Q choked on an ostensibly simpler connection. The incompatibility is also disappointing considering AMD actively recommends the StarTech adapter.

Everything works now, got the PA248Q and 2 Dell S2240Ls up and running, but yeah … don’t use active adapters with the former.

This monitor has been exonerated of all charges (see Update)
This monitor has been exonerated of all charges (see Update)

I should point out that this is the only issue I’ve had with the ASUS, it’s been absolutely stellar otherwise.

*The simple setup I’m referring to is using 3 monitors with my Radeon HD 7870-equipped XPS 8500. The 1st week I bought the wrong cable. The 2nd week I bought the right cable but ran into the Radeon’s inability to support more than 2 non-DP monitors without using an active adapter. The 3rd week I tried the active adapter solution written above.