Community discussions

MikroTik App
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

NetMetal ax / L23-UGSR — initial feedback from specs

Mon May 06, 2024 6:38 pm

In general I like the upcoming NetMetal ax and L23UGSR-5HaxD2HaxD, as an alternative in Mikrotik's "custom LTE router" lineup...

But some feedback based on specs...

1. Any reason it does not support USB 3.0? This kinda limits it for use with high-speed LTE networks because of the USB bus speed with USB 2.0 (since CPU looks like it can handle CAT18+). And the LtAP, wAPacR, M33 all support USB 3.0 via miniPCIe, while these newer AX model don't. Perhaps the SoC doesn't have USB3 controller, dunno. But USB 2.0 is limiting (and annoying since often modem modules like Sierra/Telit come in USB 3.0 mode, so have to be switched before use)

2. Surprised it doesn't have an M.2 slot since that's what's need for most 5G modem modules these days. There just not a lot of newer LTE/5G modems in miniPCIe FF.

3. Voltage specs is <30 volts, so yet another model without 48V PoE. Dealing with all the different PoE specs is PITA, and since can't often mix 24V passive for uplink with 48V 802.11at APs on same switch (certainly not in Mikrotik's switch products, and mixed voltage switches rare elsewhere too).

4. A 2nd SIM slot like LtAP would have been nice. Or some eUICC/eSIM hardware on board be even better...

Summary: In my LTE use cases, we use wAPacR (with 16M flash) today, so IPQ-5010 on these models still be an upgrade... but just that it's annoying replacing those we lose 48V and USB 3.0, but get more flash, better processor, and AX. (And current LtAP and RBM33 for LTE are MIPS, so no BTH/ZT/container/etc, unlike the ARM).
 
FezzFest
Member Candidate
Member Candidate
Posts: 102
Joined: Wed Jun 03, 2015 12:03 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Mon May 06, 2024 6:55 pm

Lack of 48V support is crazy, especially as the NetMetal ac2 does support 24-57V!
 
User avatar
mkx
Forum Guru
Forum Guru
Posts: 13062
Joined: Thu Mar 03, 2016 10:23 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue May 07, 2024 10:50 am

1. Any reason it does not support USB 3.0?
USB3.0 can kill 2.4GHz WiFi. USB2.0 can do up to (realistically) 400Mbps, which is not that bad either.
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue May 07, 2024 5:39 pm

1. Any reason it does not support USB 3.0?
USB3.0 can kill 2.4GHz WiFi. USB2.0 can do up to (realistically) 400Mbps, which is not that bad either.
Fair point. But annoying since always some compromise to upgrade. I still have quite a few RB953s I'd like to upgrade, and still no decent replacement board. And the wAPacR have the 16MB flash problem and slower CPU, but do support USB 3.0 miniPCI modems.

I think more just bad combo of USB2 & miniPCIe, instead of having an M.2 slot. Higher category LTE/5G miniPCIe modems ship with USB3, and theoretically need USB3 speeds. And there not a lot of miniPCIe modems that support PCIe, most use the USB lines on miniPCIe header. While M.2 LTE/5G typically DO support the PCIe bus offered NetMetal-ax / L23 – in which case the USB port speed wouldn't matter.

So perhaps my bigger grip is the lack of the M.2 slot.
 
User avatar
krafg
Forum Guru
Forum Guru
Posts: 1042
Joined: Sun Jun 28, 2015 7:36 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue May 07, 2024 7:45 pm

My only question is Wi-Fi power. I can't find if is still 30dbm.

Regards.
 
toxicfusion
Member
Member
Posts: 326
Joined: Mon Jan 14, 2013 6:02 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue May 07, 2024 8:00 pm

In my opinion, this is a TERRIBLE release by MikroTik. I've emailed them and have nearly mirrored these concerns with support. I'm vocal, and we all should be.

I've been asking them to create a "Pro" version of outdoor AX access points.
@normis
Also -- the 5Ghz radio in these appeared to be nerfed, or a typo... I've also asked MikroTik to review.

Wireless 2.4 GHz Max data rate 574 Mbit/s
Wireless 2.4 GHz number of chains 2
Wireless 2.4 GHz standards 802.11b/g/n/ax
Wireless 2.4 GHz chip model IPQ-5010
Wireless 2.4 GHz generation Wi-Fi 6
Wireless 5 GHz Max data rate 1200 Mbit/s
Wireless 5 GHz number of chains 2
Wireless 5 GHz standards 802.11a/n/ac/ax

The included Wireless chip - QCN-6102 should support bidirectional 2400Mbit/s.

Given that the previously released NetBox AX and mANTBox AX have the same QCN-6102 chip...
NetBox AX - L11UG-5HaxD-NB
@strods
and @normis
@normis
Lastly, this passive 24v-30v input makes us cry. This is not a direct replacement for existing NetMetal AC2's -- as they support 802af/at standard.

What happened to the days MikroTik allowed multiple powering options. We can pick either low voltage [24v], or just use standard 802af/at POE...

Their competitors no longer use 24volts.

This also limits network switching options, as we will need dedicated POE switching hardware to power these Access Points.

Perhaps MikroTik is looking to sell more of their own POE Switches. That is okay, however -they need to have the appropriate hardware. Chicken and egg scenario.

We are considering moving to alternative vendor for specific clients.... MikroTik fits niche market verticals. We only use MikroTik Wireless for specific use cases.
 
dfkalf
just joined
Posts: 2
Joined: Mon Mar 05, 2018 11:25 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri May 10, 2024 12:28 pm

We need a NetMetal ax with m.2 slot/USB3.0 support and af/at PoE!!! :(
 
Rox169
Member
Member
Posts: 467
Joined: Sat Sep 04, 2021 1:47 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri May 10, 2024 4:25 pm

no we need any AX outdoor client
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri May 10, 2024 6:17 pm

no we need any AX outdoor client
I'd imagine some wAPax is in the works. Technically, the NetMetal AX could be AX client FWIW.

Mainly wanted to highlight the lack of 48V PoE – that be a deal killer on some future outdoor AX client IMO, where standard 802.3at/af PoE be common.

And, if new routers are going to have some internal modem interface, in 2024, M.2 makes more sense - which avoid discussion of USB 2 vs USB 3 problem.
 
User avatar
krafg
Forum Guru
Forum Guru
Posts: 1042
Joined: Sun Jun 28, 2015 7:36 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Mon May 20, 2024 7:17 pm

And no info about radio power...
 
holvoetn
Forum Guru
Forum Guru
Posts: 6840
Joined: Tue Apr 13, 2021 2:14 am
Location: Belgium

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue May 21, 2024 5:36 pm

1. Any reason it does not support USB 3.0?
USB3.0 can kill 2.4GHz WiFi. USB2.0 can do up to (realistically) 400Mbps, which is not that bad either.
Side note: as far as L23UGSR-5HaxD2HaxD is concerned, this is a non-issue.
Only 5GHz wifi.
 
User avatar
floaty
Member
Member
Posts: 357
Joined: Sat Oct 20, 2018 1:24 am
Location: 52°08'32.34"N 14°39'05.0"E

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed May 22, 2024 7:14 am

2. Surprised it doesn't have an M.2 slot since that's what's need for most 5G modem modules these days. There just not a lot of newer LTE/5G modems in miniPCIe FF.
.
#metoo
 
User avatar
floaty
Member
Member
Posts: 357
Joined: Sat Oct 20, 2018 1:24 am
Location: 52°08'32.34"N 14°39'05.0"E

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed May 22, 2024 7:44 am

... maybe ... a tremendous "MiniPCI-e-circuit-board-connector-leftover"
... or ... ... policy ... ¯\_(ツ)_/¯
.
MiniPCI-e_ M.2.png
You do not have the required permissions to view the files attached to this post.
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu May 23, 2024 6:55 pm

... maybe ... a tremendous "MiniPCI-e-circuit-board-connector-leftover"
... or ... ... policy ... ¯\_(ツ)_/¯
.
Good find, that seems like good news. Perhaps they just compromised, for now(?), so they could get the AX part of a newer NetMetal / "RouterBoard" out. Placing a connector is different problem, than designing a whole board. And M.2 might require more testing on software side... Hopefully that's the story.

While my needs are LTE, the miniPCIe does mean it work with the LoRaWAN modules. And those modules are only miniPCI, so M.2 kinda screw you if you wanted LoRaWAN. Case for both, I guess.
 
User avatar
krafg
Forum Guru
Forum Guru
Posts: 1042
Joined: Sun Jun 28, 2015 7:36 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Sat Jun 15, 2024 6:34 pm

Finally output power published. 27dBm for 2.4Ghz and 28dBm for 5Ghz. I expected 30dBm.

Regards.
 
ilnicchio
Frequent Visitor
Frequent Visitor
Posts: 52
Joined: Sun May 28, 2006 6:11 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue Jul 02, 2024 8:23 pm

How can you use this radio? simply seems not working.
Where is country=debug?
why web connected with a good signal it not work ad more than 6Mbps?
the new interface/wifi instead of classic interface/wireless it's terrifing
 
User avatar
krafg
Forum Guru
Forum Guru
Posts: 1042
Joined: Sun Jun 28, 2015 7:36 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed Jul 03, 2024 6:04 am

Seriously? So... I'm starting to consider Ubiquiti for all Wi-Fi solutions instead of MikroTik, and for LTE, NR specifically maybe I will go back to Teltonika.

Thanks for the feedback.

Regards.
 
dulasau
just joined
Posts: 7
Joined: Mon Dec 18, 2023 4:18 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed Aug 14, 2024 9:13 pm

I've been playing with NetMetal AX for about 3 weeks, here are my thoughts:
* WIFI speed: I see around 700-800Mbps with Apple desktop clients (~ -55dBm) with one of two CPU cores maxing out at 100% load.
Disclaimer I'm using Multi connection mode on speedtest.net as a "speed testing tool".
* MiniPCIe vs m.2 - it has only USB2.0 MiniPCIe seems to be the right call since m.2 supports USB3.0 and MiniPCIe usually supports USB2.0
* MiniPCIe to m.2 converter work fine, I'm testing R11ml-RG502Q-EA 5G module right now, but you need an adapter with sim slot, I assume m.2/MiniPCIe connectors have different pinout for sim card slots
* USB2.0 vs USB3.0 - I don't think it's a big deal, running. speed test with 5G (WIFI to 5G) modem I see ~300Mbps (again multi connection speedtest.net) with close to 95% one CPU core load, so my guess that CPU will be the limiting factor here since USB2.0 can realistically do round 350-400Mbps
 
User avatar
floaty
Member
Member
Posts: 357
Joined: Sat Oct 20, 2018 1:24 am
Location: 52°08'32.34"N 14°39'05.0"E

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed Aug 14, 2024 10:03 pm

.
* MiniPCIe vs m.2 - it has only USB2.0 MiniPCIe seems to be the right call since m.2 supports USB3.0 and MiniPCIe supports USB2.0
.
guess ... this isn't a fully regular term ... there are pictures on the mikrotik support site, showing to 'isolate' the USB3 related pins on a mini-PCIe-card to allow correct initializationn of the modem ( resident on this very mini-PCIe-[form-factor]-card ) ... so why 'define' usb3-related pins on such a card ... and when - on top - not supported ... why isolate them ?
.
https://wiki.mikrotik.com/wiki/Manual:Peripherals
.
For some modems with USB3.0 support in some cases USB3.0 pins need to be isolated to ensure correct initialization:
.
m.2b and mini-PCIe are just physical interface-standards commonly used with gsm, gprs, umts and lteX-modems ( along with storage- and wifi-stuff )
.
https://www.delock.de/infothek/M.2/M.2_e.html
.
it seems to me ... that industry is adopting usb3 rather than PCIe as a datacom-protocol, for LTE-speeds > 300MBit/s (DL) ... cat6, cat16, cat18 ... wse ... it's just a 'feeling'
.
¯\_(ツ)_/¯ ... in biology some specimen are certainly over-categorized ... but for modern datacom-interfaces ...the-big-of-it-all-whats-what-vademecum (physically and protocol-wise) seems to be still in the cooking
.
there is information about everything ... on the web ... but ... subversioned ... fragmented and iffy phrased
.
[ my opinion: the extra-tiny rf-connectors on m2.b formfactor cards really give me the creeps ... the ones on the mini-PCIe-cards aren't fun ... ... but the m2.b-size is for guys with ablutomania-grade issues ]
 
dulasau
just joined
Posts: 7
Joined: Mon Dec 18, 2023 4:18 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed Aug 14, 2024 10:23 pm

I believe that the original standard included only USB2.0 for MiniPCIe, but later (probably around 2015) PCI-SIG added USB3.0, that's why there is some incompatibility there, I have MC7455 which doesn't work with NetMetal AX (i didn't try to "isolate" USB3.0 pins, but I'm sure it will work that way), but it's working without issue with RB33G/LtAP :)
Standards, right? :-D
 
User avatar
floaty
Member
Member
Posts: 357
Joined: Sat Oct 20, 2018 1:24 am
Location: 52°08'32.34"N 14°39'05.0"E

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Wed Aug 14, 2024 11:08 pm

... for ... an "experiment" ... I extracted the mini-PCIe modem from an LtAP to use it outside ...
... there was hot-glue over the rf-pluggies of the mini-PCIe-card ... ... now I know why ... !!
.
... furthermore I did the usb3-on-mini-PCIe isolation stunt myself ... accidentally on the m2.b modem (which I intended to bring to life) ... not on the mini-PCIe-to-m2.b-adapter, like described in the wiki ( I just drifted for a moment)
.
... and working on really small things not intended to be worked on ... with tools, insufficent to work on really small things ... makes you ... aggressive ... not passive-aggessive
.
... real-aggressive !
.
: )
.
tinkery-tonk.png
You do not have the required permissions to view the files attached to this post.
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Aug 15, 2024 1:08 am

I have MC7455 which doesn't work with NetMetal AX (i didn't try to "isolate" USB3.0 pins, but I'm sure it will work that way), but it's working without issue with RB33G/LtAP :)
Standards, right? :-D
FWIW the Sierra MC7455 allow you set the USB speed to USB 2.0 speeds, via AT command. See viewtopic.php?t=186137&hilit=mc7455#p953940 - no taping pins required. Some Telit and Quectel modems have similar AT commands to enable USB 2.0 speed over miniPCIe bus. These approach should work on NetMetalAX/L23, at least for Sierra MC7455 and Telit LM960s.
 
dulasau
just joined
Posts: 7
Joined: Mon Dec 18, 2023 4:18 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Aug 15, 2024 6:21 pm

I have MC7455 which doesn't work with NetMetal AX (i didn't try to "isolate" USB3.0 pins, but I'm sure it will work that way), but it's working without issue with RB33G/LtAP :)
Standards, right? :-D
FWIW the Sierra MC7455 allow you set the USB speed to USB 2.0 speeds, via AT command. See viewtopic.php?t=186137&hilit=mc7455#p953940 - no taping pins required. Some Telit and Quectel modems have similar AT commands to enable USB 2.0 speed over miniPCIe bus. These approach should work on NetMetalAX/L23, at least for Sierra MC7455 and Telit LM960s.
Yep I was going to try this, the strange thing is that my netmetal goes into constant reboot loop when the MC7455 is installed, not being able to initialize the modem is ok, but reboot....
 
User avatar
Amm0
Forum Guru
Forum Guru
Topic Author
Posts: 4406
Joined: Sun May 01, 2016 7:12 pm
Location: California
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Aug 15, 2024 6:30 pm

FWIW the Sierra MC7455 allow you set the USB speed to USB 2.0 speeds, via AT command.
Yep I was going to try this, the strange thing is that my netmetal goes into constant reboot loop when the MC7455 is installed, not being able to initialize the modem is ok, but reboot....
Well you have to send the AT commands on a PC using USB adapter first. The modem defaults to USB 3.0 mode, so the PIN are wrong on L23/NetMetal... and thus it won't come up to enter said AT command to switch it to USB 2.0. It also does require someone recent firmware on the MC7455 too.

Also, it's critical the modem is in MBIM mode. This can be done via AT commands, specifically
AT!USBCOMP=1,1,1009 after a AT!ENTERCND=A710 to enable config mode
Or if you want GPS support, the command is "AT!USBCOMP=1,1,100D" instead which enabled NMEA port for the gps.npk to use (same post linked above with USB 2.0 for MC7455 has details on GPS).
 
dulasau
just joined
Posts: 7
Joined: Mon Dec 18, 2023 4:18 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Aug 15, 2024 11:52 pm

Well you have to send the AT commands on a PC using USB adapter first. The modem defaults to USB 3.0 mode, so the PIN are wrong on L23/NetMetal... and thus it won't come up to enter said AT command to switch it to USB 2.0. It also does require someone recent firmware on the MC7455 too.

Also, it's critical the modem is in MBIM mode. This can be done via AT commands, specifically
AT!USBCOMP=1,1,1009 after a AT!ENTERCND=A710 to enable config mode
Or if you want GPS support, the command is "AT!USBCOMP=1,1,100D" instead which enabled NMEA port for the gps.npk to use (same post linked above with USB 2.0 for MC7455 has details on GPS).
Yeah I've been using MC7455 with LtAP for quite some time, it's just USB3.0 pins have never been a problem just never been a problem for LtAP
 
mada3k
Forum Veteran
Forum Veteran
Posts: 751
Joined: Mon Jul 13, 2015 10:53 am
Location: Sweden

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Aug 16, 2024 10:11 am

Mikrotik just should standardnize on 24-57V and nothing else.
 
User avatar
mjoksimovic
newbie
Posts: 48
Joined: Thu Jun 19, 2008 1:19 pm
Location: Serbia
Contact:

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Aug 29, 2024 8:00 pm

Lack of superchannel support and radio power adjustability is big minus for this MikroTik device. Correct me if I am wrong.
Returning to the dealer and big goodbye for serious mikrotik outdoor usage.
 
inray
just joined
Posts: 13
Joined: Tue Jan 16, 2024 8:58 am
Location: EU

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Aug 30, 2024 2:58 am

Lack of superchannel support and radio power adjustability is big minus for this MikroTik device. Correct me if I am wrong.

That's a big unfortunate truth. Mikrotik new devices are no longer what they used to be.
 
igorr29
Frequent Visitor
Frequent Visitor
Posts: 86
Joined: Tue Jan 02, 2024 12:53 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Aug 30, 2024 9:24 am

mine are still in box...
can the frequency be selected with 5mhz step like in non-ax chips, or must follow strict "wifi channels" frequencies? for example can i set it on 5515mhz? or must go on 5500, 5520, 5540....?
how does it work in noisy areas with channel overlap?
 
raulico
just joined
Posts: 16
Joined: Sat Oct 17, 2020 11:28 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Aug 30, 2024 10:59 am

in mikrotik, strategies on new devices are strange...
after a couple of years with a great chateau12, i had to switch to a new device, so i was curious on chateau5gax at 550€.... well i bought a hap ax3 with a m.2 adapter and a quectel rm520 for 300€ (120€hap+120€rm520+40€adapter+10€ usb3cable) and i can upgrade with a new modem in future....
i hope Mikrotik will produce a 4x4 mimo 5g/lte outdoor routerboard, with 1 or 2 m.2 full speed ports....
 
User avatar
mkx
Forum Guru
Forum Guru
Posts: 13062
Joined: Thu Mar 03, 2016 10:23 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Aug 30, 2024 11:09 pm

can the frequency be selected with 5mhz step like in non-ax chips, or must follow strict "wifi channels" frequencies?

It seems that it requires using proper channel frequencies. Just tried to set frequency to 5205 on my Audience (runing wifi-qcom-ac). CLI does allow to set it, but interface state says it doesn't run:
/interface/wifi> print
Flags: M - MASTER; B - BOUND; X - DISABLED, I - INACTIVE, R - RUNNING
Columns: NAME, MASTER-INTERFACE, CONFIGURATION.MODE, CONFIGURATION.SSID, CHANNEL.FREQUENCY, CHANNEL.WIDTH
#     NAME                   MASTER-INTERFACE    CONFIGURATION.MODE  CONFIGURATION.SSID   CHANNEL.FREQUENCY  CHANNEL.WIDTH
[snip]
;;; no supported channels
3 MBI wifi-5Glow-vlan41                          ap                  IE29DA4MikroTik                   5205  20mhz

And this probably helps (a lot) with channel overlap problems ...
 
tmichaud
just joined
Posts: 8
Joined: Fri Oct 14, 2016 10:06 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Mon Sep 09, 2024 11:09 pm

Lack of 48V support is crazy, especially as the NetMetal ac2 does support 24-57V!

We are a heavy user of the NetMetal AC2 due to the 24-57V PoE range. Any chance the L23-UGSR will be changed to handle the 24-57VDC range?
 
igorr29
Frequent Visitor
Frequent Visitor
Posts: 86
Joined: Tue Jan 02, 2024 12:53 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Sep 12, 2024 5:09 pm

can the frequency be selected with 5mhz step like in non-ax chips, or must follow strict "wifi channels" frequencies?

It seems that it requires using proper channel frequencies. Just tried to set frequency to 5205 on my Audience (runing wifi-qcom-ac). CLI does allow to set it, but interface state says it doesn't run:
/interface/wifi> print
Flags: M - MASTER; B - BOUND; X - DISABLED, I - INACTIVE, R - RUNNING
Columns: NAME, MASTER-INTERFACE, CONFIGURATION.MODE, CONFIGURATION.SSID, CHANNEL.FREQUENCY, CHANNEL.WIDTH
#     NAME                   MASTER-INTERFACE    CONFIGURATION.MODE  CONFIGURATION.SSID   CHANNEL.FREQUENCY  CHANNEL.WIDTH
[snip]
;;; no supported channels
3 MBI wifi-5Glow-vlan41                          ap                  IE29DA4MikroTik                   5205  20mhz

And this probably helps (a lot) with channel overlap problems ...
i tried now (i finally got the mmcx pigtails) - the freq selection is a joke. i had to try about 20 different random countries to be able to use 5500 mhz 160mhz wide channel. finally succeded with ghana. not yet sure how much is output power. it's written 24 but...
but anyway - when i finally do connect to the network (my phone supports ax), speedtest show about 80/80mbit , while my wired pc shows 300. (that's my isp speed). when i connected the laptop - it shows connection speed (in properties) as 195/195mbit. and real life speedtest shows only 50mbit.
the board is literally 2m away from my laptop using a pair of 5ghz omni antennas (that i removed from an old 3com dual band AP - so i'm 100%sure they are dualband).
as of now im very disappointed. no free selection of frequencies, very low output power, and every change of settings you must wait for dfs scanner. it's a joke.
 
User avatar
mkx
Forum Guru
Forum Guru
Posts: 13062
Joined: Thu Mar 03, 2016 10:23 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Thu Sep 12, 2024 5:36 pm

Well, wifi in 5GHz sucks (almost as much as 2.4GHz). According to wifi channel allocations, there are only 3 160MHz channels available:
  • channel 50 ranging from 5170 MHz to 5330 MHz (in ROS parlance that's center frequency 5180 with Ceeeeeee channels)
  • channel 114 ranging from 5490 MHz to 5650 MHz (that's center frequency 5520 and Ceeeeeee channels)
  • channel 163 ranging from 5735 MHz to 5895 MHz (center frequency 5745 for Ceeeeeee channels)
The last one is not actually avaialble everywhere (wikipedia table says it's only available in US) and is indoors only (which likely means lower Tx power allowed). The first two are available more universally, but are bound to DFS (and TPC) regulations in many regions.

Also mind the Tx power dispersion when using wide channels: specs of L23UGSR tell that transmitter should be able to transmit at 20dBm (when transmittng in fastest modes). But mind that that's "calibrated" to 20MHz channel width. When using 160MHz wide, Tx power is spread over 8-times wider channel, making 20MHz channel Tx power by 9dB lower, so it's comparable to Tx power of 11 dBm over 20MHz channel. Which greatly reduces usable range of this device (even if you compare it to itself ... when using 20MHz channels, at can do 28 dBm at low speeds).

Don't forget about Rx sensitivity ... which drops very much with increasing transmission speeds (by around 30dB) ... which really affects direction from station towards AP. Considering that stations will likely suffer from same "Tx power backoff" effect as described in previous paragraph ... and that stations likely suffer from comparable Rx sensitivity degradation (for directions from AP to stations) ... it becomes logical that the whole "gigabit over thin air" concept has weak foundations. For it to become useful, wifi chipsets would have to be much better (better linearity of Tx power amplifiers, better Rs sensitivity) ... but that comes with a cost. There's a reason for professional gear (LTE and 5G base stations) being much more expensive than WiFi APs (and economy of volumes is not the complete answer).
 
igorr29
Frequent Visitor
Frequent Visitor
Posts: 86
Joined: Tue Jan 02, 2024 12:53 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Sep 13, 2024 10:54 am

not sure i really follow that much. just considering a cheapo isp modem (speedport plus) that probably costs a few euro is giving me all 300mbit on wifi on default setup (80mhz channel, it's ac device). same spot, same setting, same client devices on same location.

not sure if to sell the mikrotik boards to get back the wasted money or wait 3-4 years to see if mikrotik will make their own wifi drivers that will support all the functions we have on an and ac boards.
 
User avatar
mkx
Forum Guru
Forum Guru
Posts: 13062
Joined: Thu Mar 03, 2016 10:23 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Sep 13, 2024 11:27 am

not sure i really follow that much. just considering a cheapo isp modem (speedport plus) that probably costs a few euro is giving me all 300mbit on wifi on default setup (80mhz channel, it's ac device).

What I tried to explain (with perhaps too many words) is that many times less is more ... due to limited resources. In this particular case L23UGSR performance might actually be much better if you configured it smilarly to ISP's modem: 80MHz channel width, etc. A decent ac AP will offer up to 866Mbps channel rate (when using 2x2 MIMO and when using channels not plagued with DFS and/or TPC) with realistic throughputs exceeding 650Mbps. And I certainly hope that L23UGSR can do it as well. But not necessarily when using 160MHz channels (with all of its gotchas).
 
holvoetn
Forum Guru
Forum Guru
Posts: 6840
Joined: Tue Apr 13, 2021 2:14 am
Location: Belgium

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Sep 13, 2024 12:31 pm

Mikrotik HW is capable enough.
The problem is there are way too many settings where you can goof up and then performance goes down the drain.
That's the difference with cheapo devices which have a "good for most, set and forget but you can not change a lot" config.
The strength and weakness of ROS is it's ability to change quite a bit of parameters.
And default on MT is mostly "maybe good enough but can be better".

Not to forget quite a bit of people also have client devices causing troubles, while they only look at the access point or forget to look at other APs blasting away the wireless spectrum.

On AC APs I can easily get passed 400Mb
On AX APs I get to 700-800 (with 80MHz channel width)
(I can even reach 400Mb using AX on 2.4GHz !)

All tested with internal iperf server to eliminate ISP issues.
 
igorr29
Frequent Visitor
Frequent Visitor
Posts: 86
Joined: Tue Jan 02, 2024 12:53 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Sep 13, 2024 12:34 pm

if i want to make a ptp link with these (let's forget the home AP idea, i don't have time to mess with that, i have aruba aps at home that work flawlessly) how to set them up, i mean - how to pick the frequencies and how will i know if those frequencies will work?
my current link is using 80mhz and passing about 220-250mbit of traffic.
winbox64_Q6jK5n2ejy.png
sync rate is 650mbit.
if i replace the boards with the ax boards and use 80+80mhz channel - how to set that up? how to know if the board will want to use that channels?
for start - there is no frequency selection , you have to manually input the number in mhz (which is a huge drawback, even old mikrotik from 2005 had freq list which you could select) - but then how to select what i want? here i would like the board to use first block 5500-5580 and second block 5660-5740. but it's not letting me.
i have no idea what is the logic, because 5500-5805 are legally open frequencies.
winbox64_FCpMVZ5e7S.png
edit: if i change the freq like this - it will work but select a random freq that i didn't allow it to use - 5775 which is nowhere in the range i selected.
winbox64_SygKgR6oCa.png
it selected it like this instead of 5500
winbox64_Ytqml15Bl2.png
You do not have the required permissions to view the files attached to this post.
 
igorr29
Frequent Visitor
Frequent Visitor
Posts: 86
Joined: Tue Jan 02, 2024 12:53 pm

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Fri Sep 13, 2024 1:06 pm

the 2nd board does not even see the wifi network from the 1st board.
where is the scanlist in the new ax driver? it shows wifi networks in the 5180-5300 range even when i selected frequency 5500-5900 so that setting obviously is not for scanlist.
how to even force the board to scan the whole 5ghz band?
winbox64_6NiGwdGNca.png
You do not have the required permissions to view the files attached to this post.
 
helipos
Member Candidate
Member Candidate
Posts: 142
Joined: Sat Jun 25, 2016 11:32 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue Dec 03, 2024 1:03 am

Mikrotik just should standardnize on 24-57V and nothing else.
I would push for either a really wide ass range input 11-57V. or two variants of the same board. One that can handle low voltages so it can work in an automotive application. say 11-30V. The second higher voltage so it can run on PoE or a Telecomms site which typically has 48V, maybe 20-57.

Just as I've written that, I can recall a radio we currently use that can run off 10-72V, so a really wide range is achievable.

The other thing I would have liked to see was multiple (say 3) gigabit ports. I was about to pull the trigger on a RBM33G because it has a lower input voltage spec until i found I need to give up the USB port to make dual Wifi cards work..
 
helipos
Member Candidate
Member Candidate
Posts: 142
Joined: Sat Jun 25, 2016 11:32 am

Re: NetMetal ax / L23-UGSR — initial feedback from specs

Tue Dec 03, 2024 11:29 pm

On the topic of input power.

HAP ac
DC jack input Voltage 11-57 V
PoE in Passive PoE
PoE in input Voltage 11-57 V

So yes, very much doable.