cAP ac (wifi repeater) - issues with bandwidth

RouterOS Wireless Networking
1

Hi all,

I recently bought cAP ac device (firmware: 6.46.4 stable). I needed additional device to use as a wifi repeater to cover 2 kids room. Main router is Vodafone station (cable modem/router with wifi 2,4G and 5G). Picture of the configuration is attached. Main router has wifi configured (ssid: AP5G) and “moved” to clear frequency channel (5540). cAP ac is configured as wifi repeater(for 5G network), main interface wlan2, virtualAP wlan3. Win laptop is connected with cable directly to main router. Client (mobile phone) is connected to extended wifi network (interface wlan3). All signal details are on image.

For all bandwidth internet tests I used speedtest.net and for local testing: mikrotik bwtest app in win and bandwidth test from mikrotik router.
Tests:

  1. Bandwidth test from laptop (connected with cable to main router, 192.168.1.222 to internet):
    speedtest net: ping 11ms, 150/9 Mbps (isp connection speed): OK
  2. Bandwidth test from cAP to laptop (mikrotik bandwidth test connect to bw test app on laptop; 192.168.1.195 to 192.168.1.222; udp, direction receive)
    Rx total average: 170 Mbps
  3. Bandwidth test from client (192.168.1.131 to internet)
    speedtest.net: ping 22ms, ~20/4 Mbps

Client speed/bandwidth in test 3 is really low, comparing to the Client Tx/Rx rates (on image) and the speeds in tests 1 and 2.

Can I do some additional tweaks/config or this is the expected bandwidth for this device?
Anyone with similar issues? I also tried long term and testing firmware and results are similar.

Thanks,

Best regards,
Mladen.
IMG-2218.jpg

2

Unless we see the CAP configuration am not really sure we can help…

3

This is kinda expected for a repeater, no matter what kind of device you are using.

4

From 150Mbps down to 20 Mbps no it is not expected… It is a huge drop… Even for repeating!

5

Bandwidth test from client 192.168.1.131 to internet, but connected to wlan 5GHz Vodafone on AP place ? Do you use station pseudobridge mode !

6

It is searching in the dark, you only give part of the information. Well … it is difficult to document.everything.

Anyway , let’s try. You never know …

58/60% CCQ is low for a “clear” 5 GHz channel. This 80 MHz wide channel should be able to give you 800 - 866 Mbps interface rate (80Mhz/2S/SGI). RX power of -63 dBm is acceptable and good enough to work.
Is this 60% CCQ while you operate a repeater mode traffic or only to the Vodafone?

My suggestion is to try to lower the bandwidth to 40 MHz . (40 Mhz/2S/SGI) can give you 400 Mbps interface rate, still better than what you have now. Hope is to improve the CCQ (energy more concentrated in the smaller spectrum, lesser interference even if that comes from the cAP ac and the Vodaphone. Vodaphone signal can be disturbed by reflections on metallic surfaces and by wall’s, certainly if the TX power is at the maximum.

Its a suggestion, it might learn us something.

One other possible idea: my experience is that the 6.45.6 (stable, not in LT) gives better wifi stability than 6.45.8 (LT), 6.46.4 (stable), or 6.47beta49 for an ARM based AP.

7

@bpwl i didnt even see those info inside the diagrammm.. :laughing:
I will only comment on the CCQ, the CCQ can be really low when the WiFi is not used and as soon as there is activity it can raise to 90+ % … So you cant tell unless you know if that CCQ Value is of an active link or not !

8

“So you cant tell unless you know if that CCQ Value is of an active link or not !”

Agreed 100%., inactive is mostly some bad number.

@Mladen should follow up the CCQ in all the different scenario’s. It’s very dynamic.

Just wonder if rts/cts could bring some improvement. (“Hidden node problem” for Vadofone and client? But the CAP AP is not AP for both clients, so rts/cts will not bring the same regulation as usual)

9

Hi all,

@JohnTRIVOLTA, yes, I test the bandwidth from the client (connected to AP5G) to the internet. Speed is around 140-150Mbps. And yes, in pseudobridge mode. Config is attached.

I downgraded the firmware to stable 6.45.6, as suggested but it didn’t bring higher bandwidth. I did 2 additional tests.

  1. I connected cAP with a cable to Vodafone router and made AP on cAP. Speeds are very similar to the speeds that I get with the clients directly connected to Vodafone station.

  2. I moved cAP in the same room in which Vodafone router is. The cAP is configured as a repeater (with the attached configuration). In this case, CCQ are 95%-100%. I did a 10+ speedtest.net requests and average speed is a bit better, 30-35Mbps.

I returned cAP to the original position and speed is around 25Mbps. What I additionally noticed, is that upload speed didn’t go above 3Mbps. When I follow the tx/rx CCQ value on the client
during speedtest testing, I get 15-20%. When the test stops, value returns around 70%. Strange?

Best regards,
Mladen.
myAP.cfg.rsc (1.2 KB)

10

When I follow the tx/rx CCQ value on the client
during speedtest testing, I get 15-20%. When the test stops, value returns around 70%. Strange?

Actually this is what I suspected the two transmissions disturbing each other. The reduced CCQ is the expected outcome.in that case.

Now the more difficult part. How can we correct this? Might not be easy but lets try something.

If the 3 transmitters CAP, Vodafone and client all can “hear” each other, then this is the normal designed of a wifi cell.
If however ‘client and CAP’ and ‘CAP and Vodafone’ can communicate, but ‘client and Vodafone’ cannot understand each others packetheaders then we have an ill designed wifi cell that should coordinate transmission time but cannot.

Just wonder if rts/cts could bring some improvement. (“Hidden node problem” for Vadofone and client? But the CAP AP is not AP for both clients, so rts/cts will not bring the same regulation as usual)

rts/cts for tansmission is overhead, but sometimes this is better than just free attempts to use a channel by challenging for ‘channel clear’ moments. Its like putting traffic lights.

In Mikrotik this is called “HW protection mode” in the advanced tab. Normally it’s “none”, I suggest “rts cts”. We leave the “HW protection treshold” on “0”, to make it a strong action.

We should have the same in the Vodafone. It will be called differently.(it’s sometimes activated by lowering the packet size threshold from 2347 to 0)

11

Question: is it possible to service the client via the 2.4GHz band, or is that one overcrowded?
If we have 20 MHz free on channel 1, 6 or 11 , we can have 110-130 Mbps as interface rate (20MHz/2S/SGI)

12

Hi,

We should have the same in the Vodafone. It will be called differently.(it’s sometimes activated by lowering the packet size threshold from 2347 to 0)
I setup the option “HW protection mode” on cAP but I probably have restricted access from Unitymedia on Vodafone station. I “see” only wifi password and channel setting in Expert view mode so I can not setup anything there. Testing with only changing cAP config didn’t help a lot.

Question: is it possible to service the client via the 2.4GHz band, or is that one overcrowded?
I have over 20 AP on 2.4 in range, signal strengths range from -55 to -90. If you have an idea we can try.

Best regards,
Mladen.

13

OK. Lets try the 2.4 GHz then. Idea is to start using WLAN1 as “AP bridge” as the WLAN3 is set up. You can choose the same or a different SSID and same security profile.
In this case you will have to enter the frequency and power settings for the WLAN1 (WLAN3 copied this from WLAN2, and WLAN2 got it from the Vodafone). E.G. country , freq/channel, 20 MHz bandwidth …
You can leave WLAN3 on if you want for testing and comparing (It is still sending AP beacons however, so disturbing a little bit the 5 GHz spectrum.)
WLAN1 is already member of the wifibridge, so that’s OK.

14

Hi,

I setup wlan1 as AP as suggested. 2 pictures are attached (sorry for the lines, I made a photo of the screen :slight_smile: ).
Client_connected is the client connected to a newly created AP on wlan1.
client_connected.jpg
Client_running is the picture during speedtest. CCQ is 54% on the picture, but it goes down to 40 during the tests. From a couple of speed tests, I manage to get an average of 33-38Mbps on the client.
client_running.jpg
Is this good/expected result for these conditions?

Best regards,
Mladen.

15

Really? Cant you just take some screenshots ? :laughing:

16

@Zacharias, it’s quicker this way, when I am posting from phone :slight_smile:

17

Blurry pictures, but at least readable when zoomed in.

This is a normal “average good” image. Unfortunately the CCQ drops to 54% when loaded, but you had 20 AP’s around. Maybe setting rts/cts on this WLAN1 can improve this somewhat (less collisions)

100 Mps interface rate will give something between 50% and 80% as usable bandwidth in one direction.(it drops to 47700bps , and this number still includes some overhead).

It’s only average. Even no improvement on the 5 GHz test. Unfortunately.

The Vodafone is sending us 80 MHz bandwidth appearently. I don’t like that in this poor receiving condition, I would step down to 40 MHz. But you had no access, isn’t it? It uses channel 52 till 64 (5250-5330) . I would definitively use channel 100 upward , this has much more allowed power to transmit. (7 dB more). I know it’s listed as outdoor, but it is legal to use it indoor. Fiddling with the cAP ac does not help, it is the Vodafone dictating the channel and channel width!

[admin@MktwAPac] /interface wireless info> country-info
country: germany
ranges: 2402-2482/b,g,gn20,gn40(20dBm)
2417-2457/g-turbo(20dBm)
5170-5250/a,an20,an40,ac20,ac40,ac80,ac160,ac80+80(23dBm)/passive,indoor
5170-5330/a,an20,an40,ac20,ac40,ac80,ac160,ac80+80(20dBm)/dfs,passive,indoor
5250-5330/a,an20,an40,ac20,ac40,ac80,ac160,ac80+80**(20dBm)/dfs,passive,indoor
5490-5710/a,an20,an40,ac20,ac40,ac80,ac160,ac80+80(27dBm)**/dfs,passive
5190-5310/a-turbo(20dBm)/dfs
5180-5300/a-turbo(20dBm)/dfs
5520-5680/a-turbo(27dBm)/dfs,passive
5510-5670/a-turbo(27dBm)/dfs,passive
902-927/b,g,g-turbo,gn20,gn40(30dBm)

And also the opposite option is to be tested. The cAP ac is quite powerfull, and you have bad CCQ when transmission is on. This happens with reflections and a strong signal.
Changing the “antenna gain” from 3 to 10 on WLAN2 will lower the TXpower of the 5 GHz cAP ac.with 7 dB. This could give a better CCQ (contra intuitive, but that is often the case with wifi)
Keep Vodafone and cAP away from reflecting surfaces.

18

Hello,

I made a some progress but results are confusing…
First I manage to get a 5G signal with 40MHz bandwidth channel. Second I moved channel to 100+ (108 currently). Also after some testing I set antenna gain to 8.

Results:

  1. Connected WinPC to cAP eth port. I got 35-75 Mbps.
  2. Connected client to a repeater WiFi virtual AP. I got 15-40Mbps.
  3. Connected client to a AP (on wlan1 interface). I got 20-50Mbps.

I tested it couple of times per day, couple of days, always with the same client in the same place. During the tests I also check the internet connection on Vodafone router and I got always a full bandwidth there. I didn’t manage to get any pattern why speed varies more than double in all 3 tests.

Any thoughts?

Thanks.

19

Hi, your results will always be confusing in a complex environment. The one overall number you get (throughput) depends on too many independent things to be enough to find the bottleneck.
Measuring overall is OK for knowing if it works. To find out why it is slow, needs looking in detail on what happens, and to find whats the cause.

Systematic check of the whole path is necessary, and the wifi world is one of a shared medium. Shared between your devices and devices in the neighborhood. That shared medium is the most complex and dificult to analyze, because even non-wifi devices may interfere. However Mikrotik has quite some tools built in.
Please check your wifi registration information, the status of the wifi interface (signal strength, CCQ, retransmissions) , and use “freq usage”, “scan” and “snooper” to make things visible.

Like in http://forum.mikrotik.com/t/mikrotik-drops-wifi-frequently/138296/1

20

Using the wrong device at the capaC position. What the OP should be using is an outdoor 2 radio system.
Why flog a dead horse of crappy connectivity and dropouts and frustration and an over complicated configuration.
Spend some coin and do it right.
I really don’t get you guys giving whacked out advice…
If he was trying to dissect wings of a fly and he was using a baseball bat, you would try to tell him how to do with the bat instead of recommending a microscope and some tweezers.

Back to my alcool…