RB912UAG-5HPnD and additional AR9380 - intermittent gain

[lustyffh]

Hi Folks,

I’ve bought Atheros AR5BXB112 (AR9380) off eBay, installed it on my RB912UAG-5HPnD. Then I connected it via two additional pigtails to dual-band omni 5dbi antennas.
It was recognized correctly and seems to be running fine. Logs are crystal clear as well. However, the gain seems to be sporadic or intermittent. Please refer to the screenshot attached that shows the card running in parallel with embedded Atheros interface.

Tx power is set to card rates.

I then tried to plug it in the pci-x adapter on my PC and it seems to be running fine as well.

Does anyone have any clue to this strange power output?

[laithmikrotik]

dont use rb912 with any type of omini antenna …use it just with dual chain sector antenna…

[lustyffh]

@laithmikrotik

dont use rb912 with any type of omini antenna …use it just with dual chain sector antenna…

why?

I have main card with omni antennas as well. Moreover, I’ve tried to swap first card antennas (5ghz only) with secondary card to no avail.
I use RB912 as a home AP because of its power, robustness and flexibility - couldn’t find any viable alternatives in the SOHO market for 5 Ghz.

[lustyffh]

doesn’t anyone have a clue to this?

In the meanwhile, I have also tried a test by sitting closer to the AP. However, the result is similar:

[uldis]

please describe more detailed your problem.
Also please note that tx-power-mode=card-rates are not working with AR93xx wireless chipset. If you want to lower the output power you need to use tx-power-mode=all-rates-fixed.

[lustyffh]

@uldis

please describe more detailed your problem.
Also please note that tx-power-mode=card-rates are not working with AR93xx wireless chipset. If you want to lower the output power you need to use tx-power-mode=all-rates-fixed.

Hello,

thank you for your feedback! What other details do you need?
The reason I set it to card rates is just because I thought it would be the safest option. First I set it manually, however the result was the same. I set it to 14dbi (max according to the FCC test report). Strangely, Mikrotik allowed me to set rates higher than the card can support and interface confirmed that the rate was set successfully.

[uldis]

even if you set a higher tx-power than card supports it will not use that - it will automatically use the max supported power of the wireless cards.
At what times are you monitoring the signal?
When the signal drops - did you send more traffic?
What tool are you using for monitoring the signal?

[Ivoshiee]

Sporadic gain is one thing when you are constantly monitoring the signal, but what does happen if you monitor a active link to that device? Does it have weird signal dropping as well?

Btw: Have you managed to find any radio specification to that given card? What should these gains be according to the specification and what are the receiver sensitivity numbers?

[lustyffh]

@uldis

even if you set a higher tx-power than card supports it will not use that - it will automatically use the max supported power of the wireless cards.
At what times are you monitoring the signal?
When the signal drops - did you send more traffic?
What tool are you using for monitoring the signal?

I assume it as well that the card will default to the hardware Max. But the bad thing that Mikrotik still misleadingly reports what was set manually.
Anytime tried, morning or evening.
No clients are connected to this card while testing. All clients sit on default-5 ssid.
The tool is Homedale

@Ivoshiee

Sporadic gain is one thing when you are constantly monitoring the signal, but what does happen if you monitor a active link to that device? Does it have weird signal dropping as well?

Btw: Have you managed to find any radio specification to that given card? What should these gains be according to the specification and what are the receiver sensitivity numbers?

Here are specifications from FCC test report for this wireless card(PPD-AR5BXB112):

20Hz channel

datasheet from the similar card:

Both adapters run in the UNII-3 band in the non-overlapping 20Hz channels!!

I did a few more tests and here are some interesting results.

First I switched the card from 5ghz to 2ghz w/o changing SSID. Testing is done while being connected to default-5. No clients are connected to default-51. You may see this switch in the suden TX increase for default-51:

…continued

[lustyffh]

…continued


Next, I connected one client(laptop) to default-51. Tests are performed on the PC connected to default-5 still.

Finally, I decided to run the test from the laptop connected to default-51. In parallel, PC runs it’s own test and still shows diagram as above. But notice now that on the laptop the gain is quite normal!!!

So, how can this be explained? And does the card works correctly in the end or not?

BTW, RB912G-5HPND has just no space to connect 3rd pigtail for an extension card. AUX2 almost sticks to the motherboard leaving no room for external pigtail.

[uldis]

the best way to monitor the card is to lock the wireless card to use the lowest data-rate - then connect a client and start a test where the AP sends traffic all the time.
Then you can run your tool and monitor the wifi link. As it is more accurate as you never know how good your wifi monitoring tool is working (how often it reads that channel for wireless signal and for how long time it monitors the channel till it scans the next one).