Intel Chip(!): Rivet Networks Launches the Killer Wireless-AC 1550

I just ordered an AC 1550 from Amazon which was sold by Killer. It should arrive on Wednesday to my US Freight Forwarder address and I probably would have it next week so I can test its speed and range.

 

Gladly waiting for ur review.

 

This is what I've learned so far with the Killer 1550ac:
15' line of sight R7800 to GS60
HT160 is enabled in the R7800
Channel 161 selected,has the least interference.
The Killer Control Center beta software/driver for Windows 10x64 Insider Preview doesn't work,reverted to the normal version.
Windows 10 Insider Preview doesn't have a Home group option and all future versions of W10 will not have it. .

Used Normal W10x64 for tests.

Using a near 4 year old MSI GS60 running Windows 10x64 Pro and a 750GB hard drive in a USB 3.0 case attached to the Netgear R7800 router:
From the Router attached storage using Tera Copy a 5gb W10 iso file reached a high of 101 MB/s from the router to the GS60 but averaged at about 85 MB/s.Going from the GS60 to the router attached storage it reached a high of 80 MB/s but averaged about <70 MB/s.
A file of 10 MKV videos reached a high of 82 MB/s from the attached storage and 62 MB/s from the GS60 to the storage.
A file of 10 AVI videos was just a bit slower than the MKV folder.

I was able to play/watch the AVI,MKV and watch movies off the attached storage on all my devices without issue.

At no time do I see a data rate higher than 866.7 Mbps in inSSIDer with a 100 link score and 30 dBm signal strength.I don't have a source to achieve a higher data rate.

Spectrum's(Time Warner) Internet service has a max of 400 Mbps in this area.

I don't get faster downloads from the Internet using the Killer 1550ac than I did using an Intel 8265ac

 

My 1550’s will be here tomorrow. I’ll run som analysis and test and report back I have a Asus rapture router, linksys wrt-32x, synology router and a sonicwall I’ll be testing with.

 

Unfortunately, we have run into some snags with the UK availability. It's still in the works, and still our next market, but we do not have an ETA at this time.

-- Anthony with Killer Networking

 

To those here who own a Killer 1550: Is it worth it to upgrade from the 1535? Would I notice a difference? Or should I save my $30?

Thanks

 

If you are using same 802.11 ac router then you won't notice any difference at all! If you're already having 802.11 ad router then Killer 1550 might help when your ISP line has more than 100Megs of Bandwidth up/down.

 

I have purchased one from HIDevolution on amazon.com as they ship to UK.
https://www.amazon.com/gp/product/B07933HLRF/ref=od_aui_detailpages00?ie=UTF8&psc=1

TOTAL
$45.49

That is if I am not hit by duty/import charges.
Will find out in a few weeks time.

 

If you are looking for faster downloads from The Internet I would say probably 'No" but that will be dependent on whether you can get >Gigabyte service from your provider and sites you are downloading from have the resources to provide faster speeds.

I don't know if services like Netflix throttles or limits speeds but I do know that apps like uTorrent does limit speeds.

Now that that's said I'll state if you have 2 or more devices that are capable of 1,733 Mbps on your home network and you transfer files between those devices plus to/from 160mhz router's network attached storage you should see increased speeds.

Only my i7 powered Surface Pro 4 has difficulty with seeing the USB drive attached to the R7800 router.

My android phones and tablet use ES File Explorer to connect to the USB storage attached to the R7800

 

To see a big difference between the 1535 and the 1550, you will need one or more of these things:

• Internet speeds of over 800 Mbps
• A 160 MHz MU-MIMO router
• A MU-MIMO router that is capable of theoretical speeds greater than 867 Mbps (this is basically mandatory to see any difference at all)
• A need to transfer files around your local network, wirelessly, at speeds greater than 800 Mbps
• $30 and the urge the future-proof your machine since Wi-Fi adapters aren't going to get faster for at least a few more years

-- Anthony with Killer Networking

 

If I were to recommend anyone other than us, it would be them, as they are present here on the forums and are only charging $5 more than we are. So if they are shipping to places that we aren't, hooray!

-- Anthony with Killer Networking

 

Just received it! The fun starts shortly...

 

So who was lying when they said that card only works on Z370 ?

 

lemme install it and we'll see, just tweaking someone's taptop now when I finish I will jump on this and I also have my Coconuat TM

 

Killer Wireless-AC 1550 Review-@Phoenix

 

So i got my 1550 and installed it and I am not impressed so far. Below are my speed test results at 80mhz and 160mhz. I am using the Linksys WRT32X which is also Killer branded.

80mhz:


160mhz:


As you can see my speed and signal quality went to **** when i turned on 160mhz on my router. I am running in 80mhz with Auto channel select right now and getting speeds comparable to the 1535. Not sure what is wrong with the 1550 but without working 160mhz I dont see it as worth the upgrade.

 

@kqmaverick What you have is essentialy an Intel 9260 and a Linksys WRT3200ACM with some (expensive) stickers on them, and there is a known incompatibility between the two in 160 MHz mode.
It is hard to tell whose fault is it but it is supposedly fixable via driver or firmware update. I'm not sure if it's been fixed already or is the fix on its way but make sure you have the latest firmware for your router and the latest drivers for your card.

 

I am running the latest firmware and have the latest non-beta drivers from killer.

 

This is what's wrong: Killer Wireless-AC 1550 Review

 

When it comes to HT160 some routers don’t handle it well, same goes for MU-MIMO, where Qualcomm based routers are the only ones that give a real benefit. On Broadcom based routers (ie AC88U/R8500) MU-MIMO can actually reduce performance of MU clients.

Now for the Intel card, I have 9260ac in my Inspiron 7577 I get at a (HT160) 1.3 Gbps Link Rate: 108 MB/s Down and 45 MB/s up in transfers to my NAS. This at a distance of 15ft with two walls. Even If I were at a 1.7 Gbps link it wouldn’t improve much as Ethernet bottleneck means you are limited to <125 MB/s including overhead. At HT80 I’m at 866 Link rate almost everywhere in my 2600 sqft home that is 3 levels including a basement where it can go down to a 650 Link rate. One floor below the router I get 72-75 MB/s Down and 45-50 MB/s Up in transfers to my NAS at HT80.

My router is a Netgear R7800 which ironically has better 5Ghz performance than the R9000 despite the same QCA9984 chip (I am a Netgear Beta tester so have pretty much all their high end routers).

Also the 60 GHz band is useless on the R9000 unless you actually have a card that supports it like the Intel 17265ac/18265ac and that too 60Ghz doesn’t have much range at all like it can barely get through paper as as an obstacle (bit of an exaggeration but you get the point).

The Killer 1535 is pretty much a rebranded Qualcomm 6174A and uses the same drivers but it had stability issues which made quite a few people I know replace them for 8265ac cards.

If you see a difference in range/performance between the 9260ac and 1550 that means you got a defective unit or a unit that didn’t meet standards and was sold illegally by a contracted factory and this actually happens believe it or not.

 

So I did a HT160 test directly one floor below the router and it connects to 1.7Ghz Link rate guess 1 floor is better than 2 walls (my room) makes sense since the router is on the ground, I get 112-114 MB/s Down and 69-70 MB/s Up

Please also take into account your router placement/model and home build materials will have a huge impact on range and link rates. My home is mostly wood interior so I get very high speeds in most places on my 3 floors. If you have more concrete walls and granite flooring expect massive hits in range and speeds.

 

New version 1.5.1852, of Killer Control Center 64-bit released today,Friday 03/09/2018

https://www.killernetworking.com/driver-downloads

 

It’s the 20.30.1.2 drivers available on Intel’s site as expected, in the suite according to the Killer PDF, I’ve been using that on my 9260ac not much in terms of change log from Intel.

I guess unless there is an ID check you could get newer drivers from Intel directly as they also occasional have beta drivers.

 

The Killer Control Center software was updated.

 

I opened a ticket with Killer and they got back to me stating the 160mhz speeds was a known issue with the Linksys WRT32X. They gave me a few SSH commands which had no affect but promised that a Linksys fix was forthcoming. Just wait and see i guess.

 

I thought the WRT32X was enhanced for Killer?

 

Yes - unlike Killer branded cards that prioritize traffic from your own PC over other traffic from your own PC, this actually makes sense.
Once the router assigns priority you at least get prioritized over whoever is watching cat videos in your household at the moment. Here is some info.

Mind you, this is literally a WRT3200ACM with new chassis color and new firmware feature - they even used simplified process to get if through FCC by simply admitting it's not new in any way (scroll all the way down here)

One can actually flash WRT3200ACM with WRT32X firmware although a serial cable is needed - considering that WRT32X retails for $246 while a refurbished WRT3200ACM costs $120, that might not be a bad idea for the adventurous ones.

 

Yeah that’s like the new “Gaming” Netgear XR500 which is literally a Netgear R7800 with 128 MB extra flash. Netgear made us beta test the XR500 firmware on a retail R7800 given to us as the casing and minor changes weren’t ready for test period.

For extra price for sticking on extra QoS and monitoring features you are better of buying a Qotom pFsense box with an Intel x86 CPU (which is significantly more powerful than any ARM chip in consumer routers) and using the “non gaming” version of a router as an AP. Better QoS and VPN performance with such a combo.

 

I bought a Netgear R7800 factory refurb from Amazon and couldn't be happier with over a year of ownership.

 

But-but-but, cat videos are top priorty - they deliver hope, joy and infinite cuteness in this grim and violent world!

 

Says a cat. Hmmm....

-- Anthony with Killer Networking
Check out this review of the Killer Wireless 1535!

 

How do you test your NAS speed? I use a program someone wrote: Nas performance tester 1.7. Would the write speed be the "up" and the read speed be the "down"? I get about 70MB/s read and write using HT80 866 link using my R9000 router and Killer 1550. I'm not sure what the best way is to test Max transfer speeds. When I do actual transfers with about 7gigs of large and small files I get about 40 MB/s.

 

You simply take a large file, I used the Office 2003 or 7 install exe which is around 700+ MB as the test file (you can use any large file honestly). I transferred it from the ReadyNAS 524X to my Dell Inspiron 7577 (Down) and Laptop to NAS (Up). I repeated those tests on HT80/160 from my room which is 15ft from the router (2 walls in between) and also directly one floor below. Use a 1-2 GB file, as the larger/longer the transfer the more you can look for consistency in transfer rates. Hope that clarifies things.

 

When testing speed I would recommend sticking to movie files as transferring ISOs might cause AV software to attempt to scan cabs inside them on the fly and slow things down. Another thing - I use ftp not samba shares. FTP is as simple as it gets and with almost no overhead and compatibility issues. As for samba different versions have different performance and the version you have on your NAS depends on firmware updates so you have no control over it.

I use FileZilla - at the and of the transfer it shows the exact size of the file in bytes and elapsed time in seconds so you can easily calculate average transfer speed.

 

Thanks for the input!
@ Aivxtla it seems you get nice speeds despite double walls and floors which is awesome. Using my handy dandy Killer Control Center WiFi Analyzer shows a lot of other signals around me using my channel (48). The good thing about HT80 was that I was on a DFS channel (148) all by myself. Have you checked for interfering signals around you? What channel do you use?

@downloads I like to use FastCopy because it shows the average MB/s after the transfer is done. I have no experience with using FTP but I'll have to look into setting it up and trying out FileZilla.

 

@nosauce I used TeraCopy for the same reason but I copied from a network share and that was done by samba, so there was a chance that samba might the slowing down the transfer. FTP is pretty much as fast as it gets and it's still actual transfer rather than artificial throughput.

 

View attachment 157300

I use the upper channels set at 153, in reality I believe the actual center channel is 155 at HT80 that is used as 153 is actually a 20 MHz center channel but for simplicity sake they keep it like that in the router UI so consumers don’t get confused. I only used lower channels to test 160Mhz as otherwise that doesn’t work on upper channels with the 9260ac. Honestly unless you do tons of transfers to a NAS/Computer that can handle/saturate Gigabit speeds stick to HT80 for better signal/performance which also due that results in lower latency/more stable latency if you are a gamer.

Look at the table in the link below for actual center channels used by routers at different widths:
http://www.revolutionwifi.net/revolutionwifi/2013/03/80211ac-channel-planning.html

 

After doing a lot of testing this week, I agree with you. I decided to stick with HT80 for the stable, solid performance. The performance at HT160 is very volatile. I'm able to achieve speeds that I never see with HT80 (up to low 100MB/s) but I don't see these speeds the majority of the time. The speed seemed more consistent in the morning and became really inconsistent in the evening perhaps due to people around me using their networks more maybe causing more interference? I got a wide range of speeds (as low as 20-30MB/s). I'm not sure why. There are too many variables. It may be the driver/software which I updated today. It maybe be that HT160 is more prone to being interfered by the signals around me. Ch48 is really congested, but HT80 seems to do okay at this same channel, slightly lower than normal and some variability but nowhere near what I see with HT160.

Looking at the table in the article, are we trying to hit the center of one of these 5 non-overlapping 80MHz channels for optimal performance? If we're trying to hit Ch155 then are we supposed to choose 153? or would 157 be equally as good? There's a ton of traffic in the upper and lower non DFS Channels around me. And there's absolutely no traffic in the middle DFS region. I'm trying to aim for the Ch106 middle. Are Ch104 or 108 equally good?

 

I took some time today to setup FTP and FileZilla for the first time. I did some transfer tests using HT160 and got similar speeds for FileZilla vs Windows Explorer vs FastCopy...

Test 1 - 28.2 GiB movie file
FileZilla: 69.5 MiB/s down, 60.1 MiB/s up
Windows: 66.6 MiB/s down, 62.8 MiB/s up
FastCopy: 73.9 MiB/s down, 61.2 MiB/s up

FastCopy using HT80, DFS channel: 78.1 MiB/s down

Test 2 - folder containing various file sizes: 1,138 Files, 222 Folders
FileZilla: 33.4 MiB/s down
*4 files didn't copy down and >250 files didn't copy up due to file length (>260 characters)
Windows: 31.5 MiB/s down
* >250 files didn't copy up due to file length
FastCopy: 36.2 MiB/s down, 48.0 MiB/s up

NAS to NAS copy of the same folder using Fastcopy: 23.65 MiB/s

Clearly, when copying a lot of smaller files, the bottleneck isn't the wifi speed. I'm assuming the slow processor on my NAS has something to do with it.

Anyway, it's pretty cool that I can access my NAS from anywhere now using FTP. Is it a bad idea to do this on public wifi (starbucks, library)? There's no encryption so does that mean people can easily see my username and password?

 

@nosauce I wouldn't use ftp as such from outside - either implement something similar but more secure like SSH or - which is what I would do if I needed that - you can set up a VPN server on your router which will allow you to access all devices on your LAN from outside of the LAN in a secure manner.

 

Regardless of if you pick 36,40,44,48 the router will use the necessary center channel for 80Mhz you don’t need to worry about that. Same with the other channels.

 

I informed Intel on their online forum about the HT160 on upper channels issue with the R7800 (possibly other routers) and I guess they are looking into it. I kinda hijacked someone’s thread lol.

https://communities.intel.com/thread/124298

 

Does the 1550 have the option in the advanced driver properties pane inside device manager to disable the Wi-Fi portion when Ethernet is connected? This isn't something I can do with the Killer 1535/E2500 combo that my 16L13 uses (it would be great if this could be added to the drivers). I'd rather not rely on software or a task/script to do this. @Killer_Networking

 

Just for the record, I have very little to no understanding of how these waves work, what theses channels mean, how the technology works, etc. So what I'm saying may make no sense whatsoever. A lot of the article is over my head. And I'm speculating based on a poor understanding of how all this works.

http://www.revolutionwifi.net/revolutionwifi/2013/03/80211ac-channel-planning.html

According to the article you linked me (above), "160 MHz wide channels allow for one (1) non-overlapping channel in the U.S. and two (2) in the UK/EU, with DFS being mandatory for their use in all circumstances." According to the picture in the article it seems the centers for the two channels are at 50 and 114. Do you know which 1 of the 2 is allowed in the US? The lower channels that you can switch to overlaps with the 50 range but the upper channels that you're allowed to switch to are too high to hit that 114 range if you have your router's region set to North America (if I recall correctly, I'm not at home right now). Is your router's region set to North America? or Europe? For my R9000 I had to use a hack suggested by Anthony from Killer to enable HT160. It required that I change my region to Europe. I don't think you have to use a hack for R7800. Did you just enable HT160 (checkbox) and it worked at the lower channels?

Based on the channels allowed for North America, I'm assuming that the one (1) allowed 160Mhz channel in the US is the one centered around 50. I can imagine Netgear using a few ways to follow the "regulations" for North America. e.g. making these routers either totally incapable of broadcasting(?) 160MHz at channel 114, or just not letting us switch to that range. The later scenario is more hopeful, because it'll mean that we'll just have to switch to the higher range somehow to use the second 160MHz channel centered around 114. I'd really like to try to use 160MHz at this higher channel to see if I get a more stable, consistent connection for HT160 because I can't use it at the lower channel because the connection is too unstable and only get HT160 speeds sometimes, usually not. It might be because there's a ton of traffic around that lower range, but around 114 there's a wide berth, absolutely no interference.

I did try a similar hack to try and use HT160 at the higher channel. I'm not 100% sure but I think you can switch to the channel in the 114 range if you're router's region is set to Europe. However, the HT160 option becomes unchecked and greyed out. I used the hack to un-grey the setting but when I checked the box and hit apply, the HT160 enable option did not take (it just goes back to being unchecked and greyed out. Maybe the router is just incapable of broadcasting 160MHz in that channel range.

Again, none of this might be making sense as I don't know how any of this works.

 

I'm not privy on using the advanced options but these are the options available in Device Manager --> advanced.

 

Yeah on the R7800 you just set it at the lower channels an it works without any convoluted methods unlike the R9000. The R7800 still is probably the best consumer unit for 5Ghz performance and range. The R9000 uses the same WIFi chip (QCA9984 by Qualcomm) but uses a different CPU, a stock Quad core ARM Cortex A15 made by an Amazon (yes that same Amazon you know) subsidiary called Annapurna labs. The R7800 uses a dual core Qualcomm Krait 300 CPU (IPQ8065); same Krait 300 cores found in the Snapdragon 600 Quad Core found in 2011-2012 mobile phones. It's closer to the ARM Cortex A9 architecture wise but performance is very similar to the ARM Cortex A15 as Qualcomm utilizes ARM instructions but create custom CPUs to boost performance. I assume since the R7800 is fully Qualcomm there are no CPU-WiFi chpset integration issues like with the R9000.

The channels are basically a spectrum in a frequency range for example lets say you have 5170-5835 Mhz of available spectrum on 5Ghz (lets ignore unusable parts) so for 80 Mhz channels; one 80 Mhz channel would encompass 5170-5250 Mhz, 5250-5330 Mhz would be another and so on. The channels for 5Ghz that you see in routers are 20Mhz channels so lets say you picked any channel from 36-48, it would use all those channels as combined they make up 80Mhz of spectrum when in AC mode, it would only use the individual selected channel if you were in 20Mhz mode. In 40 Mhz mode it would use an additional channel along with the one you selected. I'm terrible at explanations but I hope that was simple enough.

For DFS channels the router has to check for prioritized radios like military radars, civil aviation and weather radars. If any are detected it will prevent the use of those channels, so the router does intermittent checks. So even if you select 160Mhz the router may switch to 80Mhz or less due to that. Also since covering more spectrum reduces range and conversely more susceptible interference that is another reason why you may not always maintain a HT160 link at 1.7Gbps.

 

Thanks a lot for the explanation. I understand things better now.

Is the Quad core ARM Cortex A15 1.7GHz in my router slower than the processor in my NAS: Intel Atom C2538 Quad Core 2.4GHz? I have my VPN set up to work through my NAS but I think I can set it up through the router as well. I’m not sure which is better. I’m assuming a faster processor would matter in doing all encryption/decryption.

 

I’m back home now and I now see that my Router allows me to switch to the middling DFS channels when the region is set to North America. I have mine currently set to 104(DFS) right now and the signal is awesome.

1. Can you confirm that the US has one channel for 160MHz and that this channel is centered around channel 50? The US does not have a second channel for 160MHz?

2. In your post in the Intel thread you seem to imply that HT160 should work on channels 149-165. But wouldn’t it make sense that it does not work on these higher channels? This would be for 2 reasons: 1 - For North America (US) settings there shouldn’t be a second 160MHz channel centered at channel 114. 2 - Channels 149-165 are not in the range where HT160 operates. This would explain why HT160 only works on the lower channels (48) on our routers.

3. On your R7800, are you able to switch to channel 112(DFS) or 116(DFS)? If you switch to these channels does your HT160 setting become disabled? For me this setting is under “Advanced Setup” à “wireless settings”. On the bottom of the page there’s an “Enable HT160” checkbox that becomes disabled (greyed out) if I switch out of the lower channels. Is this the case for you as well?

I’m thinking if we can get HT160 working on that second channel, centered at 114, we may be able to get a more stable signal. For me at least there’s a ton of signals around me at those lower channels:



Wouldn’t all those signals interfere with HT160 and perhaps may be causing all the disconnections? Do you have a lot of interference in that range? If I recall correctly, when I switch my router’s region to Europe, the software doesn’t let me switch to a middling channel near 114. This may be because I have to be on a lower firmware to get HT160 to work. I think I read in the changelog for the latest firmware that it just started allowing DFS channels which is required by HT160 according to the article you linked me.

@Killer_Networking – Anthony do you mind chiming in on this? Are we only supposed to get one signal for HT160 at the lower range? Were you able to use HT160 on higher channels (near 114) for any of the routers you tested the Killer 1550 on? Also, what exactly does “short guard interval” mean for Killer control center’s WiFi app? Does the listing of 160 MHz (red arrow) simply mean that the Killer 1550 is capable of HT160? or is it saying something about the signal that I’m connected to – e.g. HT160 signal is being broadcasted by my router?

 

Yes the Atom is faster in general plus in addition it’s VPN has hardware AES acceleration which the A15 doesn’t have.

Even the newer low power ARM Cortex A53 which is slower than the A15 in almost everything is actually faster at VPN due to the AES acceleration also. That’s why the Asus AC86U, GT5300AC (Quad Core) and Linksys E9500v2 does so well in VPN though still not as good as the Intel CPUs. I believe the Netgear R7000P and R8000P use the same modified Cortex A53 with AES acceleration as well made by Broadcom. Overall though OpenVPN is single threaded so more cores won’t help with that other than spreading the load of other tasks.

 

Yes there is only one true contiguous 160 Mhz Channel with the rest being split bonding of non contiguous 80Mhz Channels. The R7800 can do split bonding but I don’t think the 9260ac likes that. I informed Intel and a rep on the forum forwarded the report about the upper channel issue with HT160.


Honestly though best performance will be at the two non DFS 80 Mhz bands one at 36-48 and the other at 149-165. In the remaining channels power limits are much lower. So unless you do a lot of NAS transfers of large files via WiFi stick to HT80 at those two HT80 bands. Unlike 2.4 GHz 5Ghz has much lower range so don’t worry too much about interference from neighbors.