AX6000 1 to 2 meters proximity ping from clients

[tashcz]

Thanks. That's a bit more than the distance I'm using it from and same barriers, but it performs much worse for me.

Are you on Merlin or stock firmware, and could you tell me with what firmware you had issues and which one solved your issues?

 

[elp]

Thanks. That's a bit more than the distance I'm using it from and same barriers, but it performs much worse for me.

Are you on Merlin or stock firmware, and could you tell me with what firmware you had issues and which one solved your issues?

I am on latest Merlin. I had issues with the factory firmware the router had when purchased, any other firmware i tried was fine.

Sorry if it was allready mentioned but is Enable WMM No-Acknowledgement disabled on your router?

 

[RMinNJ]

I have old i3 laptop..connects with wirelessN (AC) to my GTAx6000 on the 5GHz..pings are 1 to 3 ms...this 30 feet and one floor away. What is the test case here..an AX connection at 2 meters?

 

[tashcz]

Yeah, AX at close range, trying to figure out if the router's the problem (fixable via settings), the environment or it's damaged hardware.

@elp, I tried both with enabled and disabled. Which do you suggest I use?

 

[Ripshod]

Windows 11 with AX210, from 3m. Router as per sig.

PING 10.0.0.1 (10.0.0.1) 56(84) bytes of data.
64 bytes from 10.0.0.1: icmp_seq=1 ttl=64 time=1.33 ms
64 bytes from 10.0.0.1: icmp_seq=2 ttl=64 time=1.22 ms
64 bytes from 10.0.0.1: icmp_seq=3 ttl=64 time=3.96 ms
64 bytes from 10.0.0.1: icmp_seq=4 ttl=64 time=3.67 ms
64 bytes from 10.0.0.1: icmp_seq=5 ttl=64 time=3.97 ms
64 bytes from 10.0.0.1: icmp_seq=6 ttl=64 time=3.66 ms
64 bytes from 10.0.0.1: icmp_seq=7 ttl=64 time=3.65 ms
64 bytes from 10.0.0.1: icmp_seq=8 ttl=64 time=4.41 ms
64 bytes from 10.0.0.1: icmp_seq=9 ttl=64 time=5.43 ms
64 bytes from 10.0.0.1: icmp_seq=10 ttl=64 time=3.69 ms
64 bytes from 10.0.0.1: icmp_seq=11 ttl=64 time=4.56 ms
64 bytes from 10.0.0.1: icmp_seq=12 ttl=64 time=3.95 ms
64 bytes from 10.0.0.1: icmp_seq=13 ttl=64 time=3.51 ms
64 bytes from 10.0.0.1: icmp_seq=14 ttl=64 time=3.66 ms
64 bytes from 10.0.0.1: icmp_seq=15 ttl=64 time=4.14 ms
64 bytes from 10.0.0.1: icmp_seq=16 ttl=64 time=3.77 ms
64 bytes from 10.0.0.1: icmp_seq=17 ttl=64 time=3.70 ms
64 bytes from 10.0.0.1: icmp_seq=18 ttl=64 time=4.05 ms
64 bytes from 10.0.0.1: icmp_seq=19 ttl=64 time=2.55 ms
64 bytes from 10.0.0.1: icmp_seq=20 ttl=64 time=3.75 ms
^C
--- 10.0.0.1 ping statistics ---
20 packets transmitted, 20 received, 0% packet loss, time 19034ms
rtt min/avg/max/mdev = 1.222/3.630/5.429/0.942 ms

 

[elp]

I tried both with enabled and disabled. Which do you suggest I use?

Regarding WMM No-Acknowledgement my opinion is very negative. Although i have 0 interference in my house it introduced problems when i enabled it so my suggestion is to keep it disabled.

 

[tashcz]

What problems did you encounter with no ACK? I usually kept that off, and on AX6000 it's on by default.

 

[SallyBrown]

There may be issues with some AX6000's. I had one, two months ago. It worked, only on ac for a day or so, ax was very unreliable. I rebooted and it died. Reset switch wouldn't do anything, so it was returned. Three days ago I bought a AX88U-Pro and it's working well. Wifi pings on ax about a meter from the router (on a shelf right above me) 3-4 ms. Using laptop with intel AX200 card. Etherent is 1ms, as expected.

Interestingly, I checked the laptop power plan as mentioned above and wifi was on max power save, should have been max performance. The power plan was bungled up, not sure how it got that way.

Before fixing the power plan, ax pings were erratic, 3-6 ms. RSSI before change was -35 at best, now is -25. Varies a bit of course. Signal quality went from 92 to 99. wifiinfoview from nirsoft is my utility. Old values kinda go along with ping variation and signal quality. I don't think there should be much ping variation with a decent signal.

So, my router is OK; you may want to consider returning yours for an AX88U-Pro. They're the same price now and the bizarre red gui is replaced by the normal Asus blue/gray one. JMO but in default config, your AX6000 should be excellent.

 

[tashcz]

Thanks for the response Sally.

Regarding power saving, I think I disabled all there is, but still run my tests connected to AC power or plug in my PD battery bank to avoid any power saving by any link. I think my AX6000 is a dud but I'm still doing tests since returns are a PITA. I'm getting really, really mixed results and must normalize them, but overall, I can't get the average ping times to the ruter under 4ms. It's constantly 4-7ms at least with spikes again.

Worse of all is I get like 6ms, 6ms, 6ms, followed by one 2ms and it makes it average to 4-5ms. Same like when you get a high spike that messes up your avg.

That something's wrong is backed up I often see a YT loading circle when I click on an unbuffered section of the video, and instagram/facebook scrolling is acting like lazy-load on web pages sometimes. Again to note, I'm using 12th gen i7 laptop with a AX211 NIC on mains power/PD battery bank, can't get better than that.

I've again tried to set up the router to work on 80MHz AX with the channels that aren't utilized by anyone in the neighbourhood. It's mostly DFS, the airport is 15-20kms away. I've tried using high and low channels too, not making ANY difference at all.

Basically, any setting I change, beamforming, MUMIMO, NOACK, doesn't make any, any difference at all. Channels make no difference.

What makes me sad that in spots of my 75m^2 appartment I get full signal everywhere and consistent ping times, I get those 4-7ms in all spots, even those where with AC58U I'd get "half" of the signal strength reported by Windows.

What's weird... pinging from my home server to the router I get sub 1ms, but pinging the HGW of the ISP I get 2-3ms, meaning I get 1ms per each "hop" added for no reason even on wired.

To sum it all up, today I'll reinstall the old network as it was, and check my ping times compared to new setup with AX6000. If it's a 2x increase in any pings to home network devices, I'd call it a dud. There's no chance a router 10x more powerful to add aditional ping with all extra settings disabled.

Yes to note, QoS is disabled, Ai protection is disabled, anything that adds strain to the router is disabled. It's literally a router, switch and an AP doing nothing else fancy.

 

[SallyBrown]

Seems like you've tried everything, I read the other posts.
Plan B: return it and go for a AX88U-Pro.
If you can. If not, maybe Asus can help you via. RMA.

 

[Volt]

BTW, did you try disconnecting all other clients when pinging the router? Theoretically, other clients, especially using older WiFi generations, can mess with WiFi a bit, leading to increased pings.

 

[tashcz]

Plan B will be the only plan if my full old testing checks out to be different than with the new router. I haven't done those ping checks in a couple of years for sure, but that's the whole point - set those router things and forget about them, enjoy using the network. But that's gonna take a lot of moving stuff around since I keep my server, printer, NAS , HGW and other stuff in the cabinet, where on top I put the router above. I even have tin foil on the top of the cabinet to prevent any EMF dangling with the WIFI. as much as possible.

Hey @Volt,

Yeah, all other *wireless* clients were disconnected. I left my home server attached via ethernet and depending on what I was testing, disabled and enabled my ISP's HGW. So there was only one client to the AP to prevent any other device "stealing" time from the network. One config was only 20MHz 5GHz enabled, with all legacy devices being blocked (even though they were not connected) on a 100% free channel. That provided me with 3-4ms ping times but only ~100mbps bandwidth since I even reverted to AC.

Both of my desktops have 6E connectivity too, so that's one thing I haven't tried, but that's even overkill since they're always on cable... can only try that since one is Ryzen and the other one is Intel. But a concrete reinforced wall is between them and the router.

This is getting annoying Been a week without any progress, if I count out the 20MHz AC testing where there was minor latency compared to something actually usable in real world.

 

[glens]

Here's some data for your consumption.

 

[tashcz]

Thanks a lot!

This is a bunch of useful data. Should I treat all ping data as consistent?

Also, is dale connected via wire or WiFi to Zenmaster? I can see it has AX and 100mbit wired, so I guess it's wifi to it?

 

[glens]

Yes, wireless connection to dale. Like I'd suggested, it's such an ancient Dell (hence "dale") laptop that wireless is its best possible connectivity (especially with the retrofitted-to-HMC-format AX-200 card).

This is in a semi-rural locale with several clients using various wireless connectivity throughout the process. Most traffic otherwise was dale wirelessly to zenmaster, then wired to roger, then wirelessly to the stereo integrated amp, with an android tablet, also wireless to roger, doing the co-ordination, all 5 GHz on roger. But who can say what all else may have been causing insignificant traffic.

dale was (is) the only client on zenmaster, and I believe mimis is the only client on zenslave 5-2 during that capture.

Near the end, I tried to show how each client doesn't "speak" particularly well with (toward) it's /own/ AP-as-an-endpoint for some reason. I'm unconcerned about that, though.

I've been quite content with the pair of XT-8s for a bit over a year now. Best $350 I've spent in some time. Having noted how they couldn't "iperf3" any better than 1.5 Gb/s to (within) themselves, and not being able to wire 2.5 Gb/s betwixt them, I recently acquired a GT-AX6000 to stir things up. The final impetus was $229 at microcenter(.com) (which also now has that price on the AX88U Pro, but didn't at the time).

You'll see I iperf'd roger to itself twice. The initial burst was something I'd not seen before, and it did it twice! Typically it's 5x what the XT-8s can do to themselves.

As it turns out, I'd be /best/ served by a pair of XT-12s, but I'm still in under that cost, and I have a "spare" XT-8 to play with.

I'd had nearly a year of uptime on the main XT-8 (zenmaster) running gnuton's Merlin 386.07, before a regional power outage outlasted my UPS. How's that for "set and forget"?

Except for the appearance of the antennae themselves, I quite like the look of the GT-AX6000. And I believe it takes up a slightly less amount of shelf space than does the AX-88U Pro once you include the side-mounted antennas it uses. And I have a hunch that the four-corner location of the GT's antennas can garner a bit better beamforming &c. in all directions. In any event, it was $50 cheaper at the time.

I have no clue whether any or all of my units tend toward "dud"liness individually. I do know that they're meeting my expectations, and I really hope you can achieve that same peace in your endeavors.

Ciao

 

[drinkingbird]

Thanks a lot!

This is a bunch of useful data. Should I treat all ping data as consistent?

Also, is dale connected via wire or WiFi to Zenmaster? I can see it has AX and 100mbit wired, so I guess it's wifi to it?

I think rather than pulling your hair out and wondering you may just need to try another router (same model or a better one) just to see if you have some sort of hardware issue or if it is just your environment.

I started with 802.11b, back then your ping times would have been a dream come true

 

[tashcz]

@glens,

I went over all the data and as far as I can see, your network is doing excellent. I too hope I will bring this down to the level "set it and forget it" Appreciate the time you took to do all those tests, they will benefit not only me but others too I hope.

@drinkingbird,

Yeah, seems like an option. I managed to squeeze out the max possible bandwidth and speed out of this, I'm "okay" with the current setup but still wonder if there's something going on. I've kicked out all the neighbours from the 100+ channel range (oh my what 2 routers can do ) so the whole spectrum is clear just for me, no overlapping on any channel. On mains power, I'm getting 3.5-4.5ms average ping from my router now. But there's another issue - the damn thing is adding a 1ms for each wired connection and 2ms for my HGW. I don't have sub 1ms pings from my wired devices, I get ~1ms wired. So if it's a switching issue, that 1 or 2 ms of added latency does in fact account for what I see on WIFI. That'd give me 1.5-3.5ms on wireless which is probably as much as you can get.

And hell yeah does battery power destroy my latency! I've done all my tests with the power bank attached since I don't have a setting for power saving on my AX211 in W11. It adds 5-10ms!

As you've said, I think it's enough of testing and time to return the router & see if I get better luck with a different unit or a model. For some reason I can't find any AX-88U Pro's in stock here, and the rest of better models are AXE. They are pretty expensive here, AX6000 was around 320EUR, while the AXE models are 500EUR+. And the first AXE11000 as far as I remember doesn't have dual 2.5G which is one of my requirements.

So, next week I'll try to replace my unit with another AX6000. Bummer to do it for 2-3ms added latency BUT it is noticable - I see spinners on websites I haven't seen yet. If it really adds that much, 2 switches that I plan on using (one 1Gbps and one 2.5Gbps) will probably do a lot more damage to the network.

This was supposed to be the start of the whole home network + services renewal and it's a shame I got stuck at the very beginning. I'm yet to get a switch that can aggregate 2x1Gb, one 2.5G 8-port switch, a newer NAS with 2.5G connectivity along with higher capacity drives and a NUC to replace my home server from 2015.

Thank you guys for all the support, I'll give more details on what's going on as things start to move forward

Cheers!

 

[drinkingbird]

- the damn thing is adding a 1ms for each wired connection and 2ms for my HGW. I don't have sub 1ms pings from my wired devices, I get ~1ms wired. So if it's a switching issue, that 1 or 2 ms of added latency does in fact account for what I see on WIFI. That'd give me 1.5-3.5ms on wireless which is probably as much as you can get.

Not sure I'm totally following, are you saying pinging between two wired devices is 2msec where pinging between wired and the router is 1msec? The switches in these units are not great but that still seems a bit high. Though I have tested on my older AC1900 and seen the internal switch gets around 5 to 10% less throughput than a cheap external switch, and a bit of latency could explain that.

An extra msec to the ISP gateway device may not be a problem, could just be not great at processing ICMP responses, but still seems odd, any modern wired devices should easily be sub millisecond if there are no problems.

And hell yeah does battery power destroy my latency! I've done all my tests with the power bank attached since I don't have a setting for power saving on my AX211 in W11. It adds 5-10ms!

Yeah power saving mode is good for one thing, and it isn't performance. I haven't gone to W11 yet but there must be somewhere you can adjust the power level when on battery, but mine works fine on th medium setting when on battery. Haven't checked the difference in ping times.

BUT it is noticable - I see spinners on websites I haven't seen yet. If it really adds that much,

That seems a bit extreme for a couple msec and would imply to me there may really be a problem with the router, when processing actual traffic the latency may be going much higher than what your pings are showing or it is dropping packets or something. 2-3 extra msec should not be at all noticeable for surfing the web (some web pages with tons of embedded images/advertisements etc might have a measurable difference but not a spinning timer). Are you sure it isn't just possibly on your mind now and you're noticing it more - the same sites are faster with no spinning on your old router?

2 switches that I plan on using (one 1Gbps and one 2.5Gbps) will probably do a lot more damage to the network.

Modern switches should add only a small fraction of a millisecond each so should not be a big concern on that part. Just about anything you buy will be better than the cheap switch built into these. I have a $25 TP-Link smart 8 port switch that outperforms the built in asus a bit.

But that does bring another question to mind (which you may have already covered, don't recall), have you tried both WAN ports (2.5 and 1G) to see if there is any difference? And on the wired to wired tests, try different LAN ports - i.e. one is a "gaming port" and one is a 2.5G port I think - so try different ones just out of curiosity.

Mostly just a shrug from me at this point, perhaps that's just what that router is like, perhaps it is a dud. The best would be to test another model but if not available, then another AX6000 would at least confirm "operating as designed" or "hardware issue".

 

[glens]

But there's another issue - the damn thing is adding a 1ms for each wired connection and 2ms for my HGW. I don't have sub 1ms pings from my wired devices, I get ~1ms wired. So if it's a switching issue, that 1 or 2 ms of added latency does in fact account for what I see on WIFI. That'd give me 1.5-3.5ms on wireless which is probably as much as you can get.

Bummer to do it for 2-3ms added latency BUT it is noticable - I see spinners on websites I haven't seen yet.

I'm rather shrugging my shoulders over your "dilemma," to be honest.

Maybe if it was multiplicative instead of additive would a couple ms lost in-network become noticeable. Most certainly no cause for spinners IMO.

I reviewed my attachment and don't recall pinging any AP's clients /from/ the AP. I did throw in such an iperf at the end.

As recorded, I can "speak" (xfer of data, and, most likely, ping) much better /from/ an AP to its client than the other direction. I guess I'd prefer that be not the case, but that's got to be a good 4 decimal places in for tolerance level. Especially at a residential non-commercial (depending, maybe) locale.

See again how [edit: add here "wireless"] pings (if/where recorded) /and/(/ or just) data xfer /to/ an |adjacent| AP/router outperforms same to ones /own/ AP/router, in every case.

So far as I recall, my ping times are right in line with what you were initially fussin' over, which is likely why I recorded the interaction: to allay your concerns.

I've never had such a responsive network as that which I now have. SSH (terminal and graphical) connection to fellow clients feels near-indistinguishable from native.

 

[glens]

I've never had such a responsive network as that which I now have.

Have to confess that I've been using GNU-Linux at least 99.999% since about the time the 2.0 Linux kernel and discussion of a.out vs. ELF was relevant topic (likely Windows 95 was fixin' to debut). Also, that pretty much the last time I'd set up a network at home it was 100Mb and dialup, pretty much, until a bit over a year ago when I got the XT-8s to replace the failed fiber-provider-provided (1/2G symmetrical, CGNAT'd feed) Netgear AC 1700 (or so) "router."

I cringe when I read comments alluding to the in-necessity of >Gb connectivity within the network if the Internet connection ain't that high.

In final summation: I truly hope you can achieve your goals within budget. Please do follow up with a report of the substitute router, if that's the route you take.