Only Mac OS Users can connect on AC5300 but not Android

[Phoenix]

So in the cafe I go to daily, I convinced them to buy the ASUS RT-AC5300 as their old router was a POS.

anyway, these days they are complaining that customers with an Android phone cannot connect but ones with iOS such as iPhones or iPads have no problems connecting. Laptops connect fine.

They are using Smart Connect on Firmware 3.0.0.4.380.3341

Any suggestions to solve this please?

 

[sfx2000]

Any suggestions to solve this please?

Put the old device back in service and return the AC5300...

Next time you'll know better

 

[garyd9]

I also had problems with the AC5300 (and AC3100) with android devices. Both of those devices were returned and I ended up going with the AC3200 which works fine.

 

[Phoenix]

I also had problems with the AC5300 (and AC3100) with android devices. Both of those devices were returned and I ended up going with the AC3200 which works fine.

any technical explanation to this? I can't wait for the next firmware update from ASUS hoping it would fix this BS for such an expensive router.

 

[Shanley]

I have no problem connecting my Android devices to the AC5300, could be the settings.

 

[garyd9]

any technical explanation to this? I can't wait for the next firmware update from ASUS hoping it would fix this BS for such an expensive router.

I can't give a technical explanation - only a Best Guess: The firmware related to the new radio has some issue that's only hit in SOME (but not all) circumstances. Based on reviews from this site, Asus usually takes a year or so to iron out the issues in their firmware...

 

[sfx2000]

I can't give a technical explanation - only a Best Guess: The firmware related to the new radio has some issue that's only hit in SOME (but not all) circumstances. Based on reviews from this site, Asus usually takes a year or so to iron out the issues in their firmware.

Perhaps a lack of Proper Quality Assurance Testing?

No excuse for something like this... devs mean well, but if QA is missing things... then customers suffer.

 

[garyd9]

Perhaps a lack of Proper Quality Assurance Testing?

No excuse for something like this... devs mean well, but if QA is missing things... then customers suffer.

As a software engineer who just happens to work with wireless technologies for a living, this is a sore spot with me.

I'm extremely sick of my "peers" blaming QE/QA departments for flaws in software. Too often, devs are forming an "us vs them" mentality when dealing with Q/A, and that is WRONG. Developers need to think of QA and testers as people who are HELPING the developers to find issues before they get released.

Ultimately, any bug is the responsibility of development or product/project management. Nine times out of ten, a tester / QA person actually has noted the issue, but a developer blew it off as "unable to repeat", "unwilling to repeat", "I'm too lazy to fix that", or some other BS. Or, some product/project manager decides that it's a corner case and unlikely to be seen in the wild.

Coming out of my own shop, QA might completely miss 2-3% of the bugs, and might be unable to reliably repeat another 10%. Management might blow off 20% of the bugs/issues. However, 100% of them are the responsibility of the developer/development team.

(No, I'm not a tester/QA person myself. I don't have the patience for the repetitiveness of it. I'm a developer who sorely misses the days of doing console game development two generations ago when no patches could be made - the first pressing of the game disc/cartridge was it. We strived for 100% bug free (and usually came damn close.))

 

[sfx2000]

Ultimately, any bug is the responsibility of development or product/project management. Nine times out of ten, a tester / QA person actually has noted the issue, but a developer blew it off as "unable to repeat", "unwilling to repeat", "I'm too lazy to fix that", or some other BS. Or, some product/project manager decides that it's a corner case and unlikely to be seen in the wild.

Can not duplicate means someone isn't trying hard enough...

Dev's usually will unit test their contributions - and integration should be walking thru the builds (and what breaks them).

But a lack of clear understanding of the requirements going in - this means that QA is kind of on their own with their test case development and workflow....

Wave 1/Wave 2 802.11ac devices need a lot more work - from the systems engineers up front, to the developers in the middle, to QA on the back end - and this is a problem when a chipset company drops an SDK on to an OEM for integration/customization into something that ships...

Asus is good at building stuff - and perhaps with their WebUI tricks, but not end-to-end, as QA can't pull the requirements from the chipset folks...

Some vendors do a better job at this...

 

[RMerlin]

Can not duplicate means someone isn't trying hard enough...

When it comes to wifi issues, full QA testing would require the QA department to have access to every single possible clients available, which is simply not realistic. And to be given sufficient time to test every single one of them over an extended period of time.

No amount of internal testing can reach the same level of thoroughness as releasing in the wild. That's why some manufactuers will go with "soak tests" to complement the series of tests done by their own QA departments.

 

[garyd9]

Steering back on topic.. (sorry about my soapbox up there...)

For Asus (or any other consumer router that incorporates "bleeding edge" technology) the best advice I can offer is never buy anything right after it comes out. Wait until a product is at least a year old (if not older.)

Not only does this give time for (most of) the firmware bugs to be fixed (and they all have 'em), but it also lets you, the consumer, decide if anything that the router offers is really worth paying for.

Sure, having 4x4 streams and other "new" technologies might sound nice, but NO CLIENTS USE IT. Nothing. The BEST client wifi card available is 3x3 (and even that is fairly rare to see in actual use. Mainstream notebooks come with single and double stream wifi cards.) So, that 3 year old AC-1900 router is still more than 95% of the clients could use.

In REAL use, for a single router/AP, there's very little difference between the Asus AC3200 and AC5300 other than a faster processor in the latter. (Both of these routers differ from others in that they both have two 5GHz radios... and the firmware has band/radio steering that works for some and not for others.) Anyway, that faster processor isn't likely to get you much unless you happen to run something processor intensive on the router. On the other hand, the older processor and wireless chipsets on the 3200 have most of the issues ironed out by now. (I strongly disagree with anyone who suggests that the faster processors on the newer routers have any noticeable impact on routing/switching/wireless performance.)

There are some that suggest that the newer radios/wifi chips in the 5300 are somehow better than the older radios/chipsets in the 3200 for even 2 and 3 stream devices, but unless that newer stuff works properly for you, how can they be better? Is "broken" better than "old but works"?

With all that being said, I have to also tuck my tail between my legs and admit that I, also, went all out and paid for the overpriced AC5300 just because it was the newest thing out there. Thankfully, Amazon has a wonderful return policy. I'm using a 3200 now. It works.

 

[sfx2000]

When it comes to wifi issues, full QA testing would require the QA department to have access to every single possible clients available, which is simply not realistic. And to be given sufficient time to test every single one of them over an extended period of time.

Actually no - while you haven't worked at a chipset vendor - I have...

And there's not that many chipset vendors out there - and the ones that are, they're easily characterized by looking at logs and packet traces and how they behave.

Depends on the systems folks on the front end, the QA folks on the back end, and the investment in the labs in the middle, along with testplan/testcase development.

It's a people thing perhaps - and some vendors are better at this than others...

Some OEM's would rather focus on check sheet features and dead spider style rather than security, stability, and performance...

And they leverage communities to this end - and depend on third party devs to find/fix their deficiencies...

@Eric - I'd rather you spend time contributing to something useful rather than fixing someone else's bugs - consider your time spent developing those patches and supporting someone else's code that you have little influence over...

 

[L&LD]

Does everything work with smart connect disabled? If yes, then disable it.

 

[Phoenix]

I can't give a technical explanation - only a Best Guess: The firmware related to the new radio has some issue that's only hit in SOME (but not all) circumstances. Based on reviews from this site, Asus usually takes a year or so to iron out the issues in their firmware...

well this will be last time I buy the latest router especially from ASUS then.

 

[Phoenix]

Does the latest Nighthawk which is basically the same router but rebranded have the same issues you think or is their firmware better?

 

[SoCalReviews]

So in the cafe I go to daily, I convinced them to buy the ASUS RT-AC5300 as their old router was a POS.

anyway, these days they are complaining that customers with an Android phone cannot connect but ones with iOS such as iPhones or iPads have no problems connecting. Laptops connect fine.

They are using Smart Connect on Firmware 3.0.0.4.380.3341

Any suggestions to solve this please?

Either advise them to return the RT-ac5300... and go with another option (go with another Asus router, go with another brand, etc.) or try adjusting the settings (turn off smart connect, etc.)... or if they really wanted to keep the 5300 and wait for a firmware upgrade that fixes the problem... then as a temporary work around solution you could add the old router to the LAN port on the 5300 (double NAT) and set the old router to a different 2.4Ghz channel for the customer's Android devices (turn off 5Ghz completely)... rename the connection NAME-2 or NAME-Android.

... I realize this is a barbaric solution considering the 5300 is expensive and new but it would be easy and get things working until the Android connection issue could be resolved ... which it likely will be in a future firmware update (but fix is not a guarantee). If you haven't already done so or if you want to try to elevate a support ticket... I would definitely persist to contact Asus support and let them know what's going on.

If you've tried to adjust the router settings (turn off smart connect, etc.) with no solution... and you don't want to hassle with still using the second router... and there is limited time Window to return the RT-ac5300 I'd probably do the easiest thing and recommend returning it and going with a different model router. Look at other brands but the Asus ac3200 and ac1900p are good options. If tri-band with smart connect isn't needed I'd lean towards the simple and reliable design of the dual band ac1900p model with the 1.4Ghz processor.

 

[garyd9]

Does the latest Nighthawk which is basically the same router but rebranded have the same issues you think or is their firmware better?

While my own experience has been that netgear firmware tends to be slightly more reliable, it also has it's problems (and is less flexible in my opinion.) On the other hand, I've found netgear (in general) is less responsive to fixing issues in their consumer and prosumer products than Asus. Even with netgear, you'll be better off getting something that's been out in the wild for at least a year.

Changing brands isn't going to help with bleeding edge routers.

 

[RMerlin]

And there's not that many chipset vendors out there - and the ones that are, they're easily characterized by looking at logs and packet traces and how they behave.

With manufacturers like Apple doing all kind of weird things, just validating the chipset isn't enough - you have to validate the software around it if you want to be really torough.

 

[sfx2000]

With manufacturers like Apple doing all kind of weird things, just validating the chipset isn't enough - you have to validate the software around it if you want to be really torough.

Sure - and that would not fix the issue that someone has deployed the RT-AC5300 into a service environment that it is ill-suited to support. A cafe hotspot should probably be running something a bit more robust - to support more connections (and variety of clients), along with long term stability and security, along with offering captive portal support and the like.

The RT-AC5300 is perhaps a nice consumer router for the home, but probably isn't for hospitality or office use...

Platform support for 3rd parties - A good QA team is aware of various client software issues, and should be working with the Systems Engineering folks to define the requirements for the various edge cases so that code can be written, unit tested, integrated, and verified by QA prior to release to the public.

Apparently this hasn't occurred inside the Asus development team, as these issues just keep creeping up...