Synology RT1900AC Router

[thiggins]

Review posted.

 

[Rascally Jack]

I read all of that Pete I am considering #8 Minor bug fixes... that covers so many things they may not want to mention. It is a commonly utilized neutral phrase saying we fixed something but we do not want to say what it is by many businesses.

 

[pete y testing]

why do you think that?

the wifi chipsets are different to other d1900ac routers and in your own testing with a linksys access point Linksys LAPAC1750PRO with the same wifi chipset it also had poor wifi coverage and throughput , its those BCM43460 chipset thats the issue

see https://wikidevi.com/wiki/List_of_802.11ac_Hardware





i have also had lots of contact with synology support who really dont have any answers for the wifi issue , so unless for some reason you guys find a cause i cant see the rt1900ac improving

i will check out your review shortly

-----------------------

so for reference here are my wifi test results

5 gig tested on ch 153

10 feet same room location A

read 40.1MB/s write 22.5MB/s

10 meters 1 room away location C

read 26.3MB/s write 22.0MB/s

25 meters 3 rooms away location E

read 1.77MB/s write 4.27MB/s

-------------------------------------

2.4 gig tested on ch 1

10 feet same room location A

read 7.35MB/s write 9.55MB/s

10 meters 1 room away location C

read 10.7MB/s write 12.4MB/s

25 meters 3 rooms away location E

read 1.96MB/s write 1.88MB/s

just a note with 2.4 gig there is no current way to disable obss coexistence and thus 2.4 gig will only run at 20mhz

--------------------------------

as a comparison with other routers including the r7000 see my results in the link below

https://forums.whirlpool.net.au/forum-replies.cfm?t=2384995

 

[sfx2000]

the rt1900ac uses the Broadcom BCM43460 for both its 2.4 and 5 gig transmissions and may explain why both are showing the same issue

the others use the__ Broadcom BCM4360

they dont seem like they would be that much different but its proven to be the case that there are some issues with the BCM43460

The BCM4360 and BCM43460 were released at the same time - and essentially the same chip under the skin, the difference is the targeted market - 4360 is consumer, 43460 is for enterprise - which brings to mind that the performance issues are likely due to config on the WLAN driver within Synology's board support package in general - my best guess is either chip used would have the same problem.

 

[pete y testing]

ether way atm both tim's and my data show the wifi is sub standard even for early wireless AC standard and its usb write performance is quite sub par as well , since they have had 6 months with this in the wild and havnt been able to improve of "fix" the wifi issues i cant see it getting fixed any time soon

 

[sfx2000]

ether way atm both tim's and my data show the wifi is sub standard even for early wireless AC standard and its usb write performance is quite sub par as well , since they have had 6 months with this in the wild and havnt been able to improve of "fix" the wifi issues i cant see it getting fixed any time soon

Totally get it - something's odd - but it ain't the chips - could be calibration inside the drivers, could be self-jamming on the RF - the numbers don't make sense, which is a good reason for Synology to put some resources on the issue...

They're new to the Wireless side... and Wireless ain't easy...

 

[pete y testing]

could be calibration inside the drivers, could be self-jamming on the RF - the numbers don't make sense, which is a good reason for Synology to put some resources on the issue...

they have had 6 months to figure it out yet have not

im pretty sure they will move onto something newer with link aggregation etc

 

[thiggins]

Could be self-jamming on the RF

The "Fixes" in the R1, shielding tape, antenna cable re-routing, would point to attempts to control RF leakage and/or spurious. Their thermal design also needs work, as pointed out in the review.

 

[avtella]

The BCM4360 and BCM43460 were released at the same time - and essentially the same chip under the skin, the difference is the targeted market - 4360 is consumer, 43460 is for enterprise - which brings to mind that the performance issues are likely due to config on the WLAN driver within Synology's board support package in general - my best guess is either chip used would have the same problem.

Out of curiosity what actually are the differences between the consumer and enterprise versions if they are the same chip?

 

[RMerlin]

The "Fixes" in the R1, shielding tape, antenna cable re-routing, would point to attempts to control RF leakage and/or spurious. Their thermal design also needs work, as pointed out in the review.

Do you have SSH access to one of these? I could send you a few SSH commands to read the temperature values, so we could compare them to an R7000 or RT-AC68U. The RT-AC68U was already generally considered as running quite hot, yet still within BCM's specifications.

 

[RMerlin]

Totally get it - something's odd - but it ain't the chips - could be calibration inside the drivers, could be self-jamming on the RF - the numbers don't make sense, which is a good reason for Synology to put some resources on the issue...

They're new to the Wireless side... and Wireless ain't easy...

And they also use the same PAs according to Tim's review - that was one of the areas where I thought they might have gone with something cheaper.

 

[thiggins]

Do you have SSH access to one of these? I could send you a few SSH commands to read the temperature values, so we could compare them to an R7000 or RT-AC68U. The RT-AC68U was already generally considered as running quite hot, yet still within BCM's specifications.

Not unless SSH access is standard. Besides, I've moved on to beating on eero.

 

[thiggins]

And they also use the same PAs according to Tim's review - that was one of the areas where I thought they might have gone with something cheaper.

Actually, I think it's my first sighting of the RFMD RFFM4552 5 GHz front ends

 

[pete y testing]

Do you have SSH access to one of these? I could send you a few SSH commands to read the temperature values, so we could compare them to an R7000 or RT-AC68U. The RT-AC68U was already generally considered as running quite hot, yet still within BCM's specifications.

i can test that for you if you like , am away till the weekend but can test once i get home

 

[sfx2000]

Out of curiosity what actually are the differences between the consumer and enterprise versions if they are the same chip?

From what I've been told - it's the same chip, just different part number to differentiate between enterprise and consumer - so it's not even bin-sorted... there's some firmware inside the chip, low level, that activate functions that are relevant to enterprise/carrier grade, but that's all abstracted and not visible from the main system firmware - not much different from what AMD (Radeon vs. FireGL) or nVidia (gForce vs. Quadro) on the GPU side - same chip, different firmware/drivers...

 

[sfx2000]

Do you have SSH access to one of these? I could send you a few SSH commands to read the temperature values, so we could compare them to an R7000 or RT-AC68U. The RT-AC68U was already generally considered as running quite hot, yet still within BCM's specifications.

Actually, if SSH is available, it would be also interesting to get a read on other items inside the /proc directory as well...

 

[sfx2000]

Besides, I've moved on to beating on eero.

Amen - there are a few that are very interested in what those guys are doing...

 

[RMerlin]

Actually, I think it's my first sighting of the RFMD RFFM4552 5 GHz front ends

I thought it was two Skyworks, unless I've misread it.

 

[RMerlin]

i can test that for you if you like , am away till the weekend but can test once i get home

Assuming they used the standard BCM SDK/kernel:

http://www.snbforums.com/threads/rt-ac68u-temperatures.17418/#post-122705

I'm not 100% sure on the radio temperature tho. Asus applies this weird formula to the results returned from querying the SoC. I'm not 100% sure the same formula also applies to other manufacturers. Might still give us some general idea of how it compares with an RT-AC68U.

 

[pete y testing]

Actually, I think it's my first sighting of the RFMD RFFM4552 5 GHz front ends

hi tim

did you by any chance test with a different set of antennas just to eliminate poor antenna design or mis match