ASUS RT-AC5300

[12darline]

I knew exactly what I was buying in that it had tech that either wasn't enabled on it yet or tech that my devices wouldn't take advantage of. So I'm good with the fact that yes, it has tech not needed by anyone out there. I bought things with this router that will do me absolutely zero good. At peace with that.

I tried it to see if it delivered better range, reliability and a working smart connect, all things I didn't get out of the ac3200. To that end its delivering lights out on all fronts. I wanted to see if a newer generation build could deliver a better performing router in real world every day moments and it's what I'm getting so far. I get a connection that holds at the far end of the house where it would not with a ac3200, a ac68 before that and so on down the line in generations of Routers that I have tried. At that far end, I'm also getting good enough performance that I have ditched the repeater I inevitably went to in other generations. Thirdly, the smart connect is working much better then on any fw with the ac3200, not getting the disconnects and I'm getting a more then workable speed wherever I work on a device from.

I know the hardware is all known beforehand, I know most of it has been seen in other devices so to that end I don't know why my performance is noticeably better. But I can't deny that a newer generation on that known and previously used hardware is giving me the router I've endlessly been pursuing for years. All I know is it's been two weeks or so, no major hiccups, stable performance and that damn repeater hasn't been plugged in once. I haven't rebooted because of faltering or disapearing performance once. For the first time, I'm not curious or looking forward to what their next router is going to be like.

The singular most profound realization I've had about this router? Not one person of my family(kids, wife, parents when they visit daily) have complained one bit that 'it's not working again! Not one singular peep. The obsessive techie in me will always find a fault here and there and true to form I have with this one too but in that crucial real world run, the family isn't seeing one interruption in their lives due to wireless being down, connection this or that and so on. We have one ssid that is covering the entire house, I can't tell you how happy that makes me.

I totally agree with 99.9% of everything that you've stated. You just said it better than I could. I had a RT-AC87R, and upgraded to this RT-AC5300. I have only had it running for 5 days, but I could not be more pleased. Can't wait for Merlin to work his magic on this model.

 

[Embattle]

I know that some people had a bad experience with the 87U but the driver like most new routers has taken a long time to come good but I rarely ever had to restart it because of instability or loss of signal issues, most the time it was to change a setting with the longest period I had between a setting tweak being 35 days uptime.

As I know have two 5Ghz networks due to running a router and an access point the only router I might consider which would also be the AP would be the 88 but that is just for the extra ports but even then I really can't justify it at the moment.

 

[L&LD]

I knew exactly what I was buying in that it had tech that either wasn't enabled on it yet or tech that my devices wouldn't take advantage of. So I'm good with the fact that yes, it has tech not needed by anyone out there. I bought things with this router that will do me absolutely zero good. At peace with that.

I tried it to see if it delivered better range, reliability and a working smart connect, all things I didn't get out of the ac3200. To that end its delivering lights out on all fronts. I wanted to see if a newer generation build could deliver a better performing router in real world every day moments and it's what I'm getting so far. I get a connection that holds at the far end of the house where it would not with a ac3200, a ac68 before that and so on down the line in generations of Routers that I have tried. At that far end, I'm also getting good enough performance that I have ditched the repeater I inevitably went to in other generations. Thirdly, the smart connect is working much better then on any fw with the ac3200, not getting the disconnects and I'm getting a more then workable speed wherever I work on a device from.

I know the hardware is all known beforehand, I know most of it has been seen in other devices so to that end I don't know why my performance is noticeably better. But I can't deny that a newer generation on that known and previously used hardware is giving me the router I've endlessly been pursuing for years. All I know is it's been two weeks or so, no major hiccups, stable performance and that damn repeater hasn't been plugged in once. I haven't rebooted because of faltering or disapearing performance once. For the first time, I'm not curious or looking forward to what their next router is going to be like.

The singular most profound realization I've had about this router? Not one person of my family(kids, wife, parents when they visit daily) have complained one bit that 'it's not working again! Not one singular peep. The obsessive techie in me will always find a fault here and there and true to form I have with this one too but in that crucial real world run, the family isn't seeing one interruption in their lives due to wireless being down, connection this or that and so on. We have one ssid that is covering the entire house, I can't tell you how happy that makes me.

This is the only kind of 'review' I personally like. Comparative in real world usage. I may or may not get the same experience in my own network setup, but at least I know which path to follow for myself and my customers.

Thank you for stating it so well.

 

[sfx2000]

MU-MIMO is not enabled yet in Broadcom-based 4x4 routers.

You do NOT benefit by buying into MU-MIMO routers early. It is a hellaciously complex technology and will need a good deal of time to move down the learning curve. Broadcom has thrown non-standard technology into its implementation (1024-QAM) that provides no practical benefit and could cause compatibility problems with QCA-based MU-MIMO devices.

MU is likely not enabled for some very good reasons... as Tim Higgins has mentioned above, it's hella complicated, and interoperability testing across different chipset vendors is still ongoing... Broadcom MU works great with Broadcom clients, but they're not the only vendor doing MU right now...

I know this reminds folks of Asus' promises with the whole Quantenna mess on the RT-AC87U, but it's better to delay a feature slightly, and ensure that it works as promised than to deliver something that breaks everything.

As a 4*4:4 Router/AP, the AC5300/88U/3100 products should have outstanding performance and range across all clients - I've seen this with other four radio AP's, and there's no doubt in my mind that these Asus products will perform similarly.

Even the 87U with the QTN radio, most complaints were about stability and client connectiivity, not range/speed when things worked..

Remember, MU isn't faster on a client basis, in some cases it will be slower perhaps, but it makes the air interface more efficient in certain conditions - those being where the clients are fairly stationary, if client moves about the house, and the transmit beamforming matrix has to be rebuilt each time for all clients in the MU set, and this takes time/adds latency..

Being that MU is downlink only (AP to clients), it's really just that last itty bitty little bit on networks that might be heavily loaded - most folks are actually probably better off without MU...

my $0.02 thoughts..

 

[sfx2000]

well the AC88U is more sensible even if it had half the ports

That's called an RT-AC3100

I'm a bit surprised that the 5300 didn't add the extra switch with 4 ports - for the price, it's not a big BOM addition..

 

[sfx2000]

I can't believe Asus can't be bothered to send Merlin a new router every time it's launched while Netgear gave away tons of X8s for free under their Friends and Family program.

For RMerlin's outstanding contributions to the Asus community, I'm also a bit surprised, as many of his bug fixes and detailed debug make it back into the mainline builds..

 

[System Error Message]

That's called an RT-AC3100

I'm a bit surprised that the 5300 didn't add the extra switch with 4 ports - for the price, it's not a big BOM addition..

The problem is when the AC5300 isnt the center and only device of your network. If you had 2 AC5300s you would need to bond all ports between them so thats a lot of ethernet ports that need combining so at the very least having more ports would help. wifi bridging just like that isnt very effective compared to ethernet and to properly use it you would need to start bonding ports or just see low performance and packet losses.

 

[RMerlin]

Broadcom has thrown non-standard technology into its implementation (1024-QAM) that provides no practical benefit and could cause compatibility problems with QCA-based MU-MIMO devices.

Fortunately, this can be disabled. You can select the maximum MCS level on both bands. Personally, I recommend people disable marketingQAM levels unless they know for a fact they have a compatible client (which is very uncommon right now).

 

[RMerlin]

That's called an RT-AC3100

I'm a bit surprised that the 5300 didn't add the extra switch with 4 ports - for the price, it's not a big BOM addition..

I suspect that someone who has enough wireless clients to need the three radios is less likely to have a lot of wired clients.

BOM and R&D cost difference seems minimal indeed, there's only 20$ difference between the AC88 and the AC3100. I was expecting Asus to better differentiate the two by perhaps dropping RAM to 256 MB and underclock the CPU a bit. This would have allowed to perhaps drop the price by a bit more, perhaps bring the RT-AC3100 down to the RT-AC87U pricing. Right now, for 20$ difference, there's little incentive to get the RT-AC3100 over the RT-AC88U IMHO. 20$ is what it would cost you to buy an external switch if your network were to grow in the future. And a 20$ difference out of a 300$ purchase is quite minimal.

 

[RMerlin]

For RMerlin's outstanding contributions to the Asus community, I'm also a bit surprised, as many of his bug fixes and detailed debug make it back into the mainline builds..

I did get two different routers over the past few weeks (one of them is a new hardware revision of a previous model, which will require firmware-level changes). I haven't received an RT-AC5300 most likely because the hardware is virtually identical to the RT-AC88U, so it might not be necessary. The main reason why I haven't blindly tried to support it yet (like I did for the AC3100) is the absence of GPL for that specific model.

Beside, the two shelves devoted to my mini lab setup are getting a tad crowded right now.

 

[karkor]

Hello i have some questions about this router is it worth to wait for EU one or should i just buy the US? and about the difference is it only will be having US with 1-11 channels and EU with 1-13 channels and the other thing is the ac adapter that's it? or there is more like maybe wireless range? wireless speed? or anything else maybe?

Am i asking my questions at the wrong place? can someone please answer me and help me out here please.

 

[RMerlin]

Am i asking my questions at the wrong place? can someone please answer me and help me out here please.

Never buy a router not intended for your region. Beside potentially being outside of the legal specs of your country, you might also end up having problems connecting some of your wireless clients to it if the region doesn't match.

 

[karkor]

well about that legal specs of my country... the funny thing there is no legal specs in my country i think or even there was i never heard about it or maybe unless someone from the neighbors complain i think and that never happened too

all i wanted to have is a strong wireless router and i heard alot about Asus routers and i never tried one before but i tried all the other brands the only one left is Asus

and i checked about Asus there is no option that i can change my region in the router like Netgear and some other routers.

 

[dlandiss]

This is the only kind of 'review' I personally like. Comparative in real world usage. I may or may not get the same experience in my own network setup, but at least I know which path to follow for myself and my customers.

Thank you for stating it so well.

As an engineering professor for 34 years I have learned that in the scientific world anecdotal evidence alone is not enough. It must be backed up with measurements or it becomes a "my horse is faster than your horse" spitting contest.

My test setup cannot match that of Tim Higgins but I did my best to remove external variables. The router under test sits on a windowsill in my living room, and transmits through 3 panes of glass to a workbench in my garage about 75 feet away. This is enough distance that the peak speed of the WiFi connection is not the limiting factor and line-of-site communication minimizes reflections and other external influences. This is real-world in the sense that one of my goals is to get decent bandwidth in the garage.

So this afternoon I made 100 measurements comparing an RT-N66U with an RT-AC68U. I also compared 2.4GHz against 5GHz, and 5dBi antennas against 9dBi (advertised) antennas. Ten download Mbps measurements were made for each test condition and the averages computed. Both routers used John's fork 15E1.

The comparative results are more informative than the absolute numbers. These may look like speed tests but they are actually range tests because at a distance of more than 30 or 40 feet speed depends on attenuation, which in turn depends on range and intervening obstacles. Download speed is just an easily measurable parameter like volts or amps and it ensures that the router and the client device are not moved during the testing.

The greatest improvement was beamforming in the AC68U -- 39% better speed on 5G.

Next biggest was replacing the stock 5dBi antenna with 9dBi (advertised) -- 28% improvement on 2.4G.

Because of the variability of the test measurements differences smaller than 8 or 10% are probably not statistically significant. I'm not going to bother running a full statistical panel on these but from years of experience analyzing lab measurements that's a pretty good guess.

My 7-word summary from all of this is "turn beamforming on and use good antennas."

 

[katinatez]

Well I'm exited to learn @RMerlin is supporting the AC3100! I am in dire need of a upgrade, still running rt-n66u here. And my Home Entertainment center has 6 wired clients and will leave me with 2 open ports available. Was considering the 5300 but need compatible clients, so my choice is the AC3100 because of the 8 G ports. These news made my day.

 

[tl2k2]

Well I'm exited to learn @RMerlin is supporting the AC3100! I am in dire need of a upgrade, still running rt-n66u here. And my Home Entertainment center has 6 wired clients and will leave me with 2 open ports available. Was considering the 5300 but need compatible clients, so my choice is the AC3100 because of the 8 G ports. These news made my day.

you meant the AC3100 class rt-ac88u? because the AC3100 model selling at BB only have 4 gigabit ports.

 

[L&LD]

As an engineering professor for 34 years I have learned that in the scientific world anecdotal evidence alone is not enough. It must be backed up with measurements or it becomes a "my horse is faster than your horse" spitting contest.

My test setup cannot match that of Tim Higgins but I did my best to remove external variables. The router under test sits on a windowsill in my living room, and transmits through 3 panes of glass to a workbench in my garage about 75 feet away. This is enough distance that the peak speed of the WiFi connection is not the limiting factor and line-of-site communication minimizes reflections and other external influences. This is real-world in the sense that one of my goals is to get decent bandwidth in the garage.

So this afternoon I made 100 measurements comparing an RT-N66U with an RT-AC68U. I also compared 2.4GHz against 5GHz, and 5dBi antennas against 9dBi (advertised) antennas. Ten download Mbps measurements were made for each test condition and the averages computed. Both routers used John's fork 15E1.

The comparative results are more informative than the absolute numbers. These may look like speed tests but they are actually range tests because at a distance of more than 30 or 40 feet speed depends on attenuation, which in turn depends on range and intervening obstacles. Download speed is just an easily measurable parameter like volts or amps and it ensures that the router and the client device are not moved during the testing.

View attachment 4921

The greatest improvement was beamforming in the AC68U -- 39% better speed on 5G.

Next biggest was replacing the stock 5dBi antenna with 9dBi (advertised) -- 28% improvement on 2.4G.

Because of the variability of the test measurements differences smaller than 8 or 10% are probably not statistically significant. I'm not going to bother running a full statistical panel on these but from years of experience analyzing lab measurements that's a pretty good guess.

My 7-word summary from all of this is "turn beamforming on and use good antennas."

Using just 'scientific' or just 'anecdotal' evidence is not enough. I agree.

But the problem with scientific 'evidence' is that it rarely translates to the real world environments I work in.

Not to mention the marketing and other lies that take advantage of specific 'scientific' evidence to further mangle the truth.

Your tests just prove the point too; all else being equal, the newer model usually gives the better (real world) results.

 

[sfx2000]

The problem is when the AC5300 isnt the center and only device of your network. If you had 2 AC5300s you would need to bond all ports between them so thats a lot of ethernet ports that need combining so at the very least having more ports would help

That was a topic of exploration some time back - bonding two layer 2 AC1300 connections across a WiFi bridge with two AC3200's, but the firmware wasn't capable back then, and it's likely not capable now...

But going back to my original comment - odd that the Ac5300 limits the lan ports to 4...

perhaps the BRCM SoC is just plain out bandwidth limited between either 3 Wireless Ports/4LAN or 2 Wireless Ports/5Lan (4 + 1 Realtek adding 4 more)?

 

[System Error Message]

Actually ever since the AC68U the broadcom platform has had 8 ports available, 4 of those are ethernet and 4 of those are RGMII. A typical consumer router has 4 ethernet ports, 1 RGMII link to the CPU and 1 port connected to a CPU to make 5 ports. The AC88U uses all the 8 ports, of the switch chip or it may have 2 switch chips (i havent seen the architecture yet). In some routers RGMII is used to connect wifi to the switch. Its not only to do with the number of ports but there are other hardware stuff you can add to a router that a lot of controllers support. SD card or sata is one. Sata/esata would reduce the load on CPU for hard drives but the problem is that esata didnt really take off compared to usb and usb3 not to mention that esata enclosures were a lot more expensive than cheap usb drive enclosures that are bound to fail within a year.

Another thing that i dont see in consumer routers that could really make a difference are VOIP ports and SFP. SFP can be used for fibre optics eliminating the need for a modem. Some mikrotik access points are inexpensive and have SFP, a gigabit ethernet port or 2, mini PCIe and usb. SFP on consumer routers would mean less power use in total for no modem and make it last longer for newer things. With the cost of the AC5300 its very disappointing that it lacks SFP or even the firmware to bond ethernet ports.

Essentially the switch in a router is either a 5 port switch (4 ports physical, 1 CPU), or an 8 port switch a bit like adding a seperate switch behind a 2 port router.

 

[dlandiss]

Your tests just prove the point too; all else being equal, the newer model usually gives the better (real world) results.

More interesting to me is the major gains of the newer model came from the beamforming feature. Without it the results were only about 5% apart, well within the probable margin of error. And there lies the advantage of systematic testing over anecdotal comments. I had hoped that the newer routers might benefit from better circuit design or higher quality fabrication, but without that particular feature there is little difference.