9mi Bulletm2 Shot

[sbyrd]

We just deployed a 9mi BulletM2 shot. I am using 24dbi grids in h-pol. I am getting a 65/65 modulation rate with a tcp rate of 20/15. I get a signal of -59dbm and a ccq of 98-100%. Once the webinterface bug http://www.ubnt.com/forum/viewtopic.php?t=13919is sorted out these will be perfect.

This is replacing an old Canopy BH shot that only is capable of 3.5/3.5 and would cost about $2200 to replace w/10-20meg Canopy BHs. Will try to show pictures later.
http://img266.imageshack.us/img266/4338/58639164.jpg

[Shockware1]

ummm...
That link will only work for you since your c:\ is much different than my c:\


Go to www.imageshack.com and upload the photo, then link that here. No signups or email or anything.

[UBNT-Mike.Ford]

We just deployed a 9mi BulletM2 shot. I am using 24dbi grids in h-pol. I am getting a 65/65 modulation rate with a tcp rate of 20/15. I get a signal of -59dbm and a ccq of 98-100%. Once the webinterface bug http://www.ubnt.com/forum/viewtopic.php?t=13919is sorted out these will be perfect.

This is replacing an old Canopy BH shot that only is capable of 3.5/3.5 and would cost about $2200 to replace w/10-20meg Canopy BHs. Will try to show pictures later.
http://img266.imageshack.us/img266/4338/58639164.jpg

Thanks for the info.


Are you using AutoACK or static ACK timing?

Thanks,

Mike

[sbyrd]

Right now I am using an Static ack timing of 120% the distance of the link. I tried testing w/Airmax enabled when testing in the shop and got less throughput then with it disabled. Would love to know what that is cause so far the only thing it does for me is lower my throughput.http://img19.imageshack.us/img19/7339/11591076.jpg

[UBNT-Mike.Ford]

Right now I am using an Static ack timing of 120% the distance of the link. I tried testing w/Airmax enabled when testing in the shop and got less throughput then with it disabled. Would love to know what that is cause so far the only thing it does for me is lower my throughput.

Hello,

AirMax itself will decrease the throughput in most Point to Point links due to the added overhead (except in heavy noise enviroments where it helps).

Can you test the link with AutoACK on both sides. You should see an increase in throughput as 11n is VERY VERY sensitive to the ACK settings in outdoor enviroments. Please note, it can take 30-60 seconds for the ACK settings to normalize. It normalize when data is pushed over the link. This normalization process will only occur the first time.

Thanks,

Mike

[sbyrd]

I won't be able to test the throughput on both sides since I would a have to climb a 165ft Water tower to do so, but I will try setting it to AutoAck. Is there any downside to using Auto vs static besides possible increase in throughput on AutoAck.

[UBNT-Mike.Ford]

I won't be able to test the throughput on both sides since I would a have to climb a 165ft Water tower to do so, but I will try setting it to AutoAck. Is there any downside to using Auto vs static besides possible increase in throughput on AutoAck.

Nope, on some links though, it will take some serious playing with a static ACK to get it to work right. I mean literally trying one number at a time untile you find the right value. It can be THAT sensitive.

Thanks,

Mike

[sbyrd]

I tried it and the AP has an Ack of 9.5 miles and the Station has an Ack of 11.3 miles on Auto will that affect my link more than having it static at 10.8 miles on both the AP and Station.

[UBNT-Mike.Ford]

I tried it and the AP has an Ack of 9.5 miles and the Station has an Ack of 11.3 miles on Auto will that affect my link more than having it static at 10.8 miles on both the AP and Station.

Yes it can.

Thanks,

[WHT]

it will take some serious playing with a static ACK to get it to work right. I mean literally trying one number at a time untile you find the right value. It can be THAT sensitive.
No sh*t, Sherlock! :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:

[Shockware1]

Perhaps some sort of embedded ACK tuning applet in the radios is in order? Something that does tests for you for each ACK value and displays the results?

Or is auto ACK just that, but without showing you what is better or worse, just picking the best?

[UBNT-Mike.Ford]

Perhaps some sort of embedded ACK tuning applet in the radios is in order? Something that does tests for you for each ACK value and displays the results?

Or is auto ACK just that, but without showing you what is better or worse, just picking the best?

AutoACK does just this.

[sbyrd]

I tried it and the AP has an Ack of 9.5 miles and the Station has an Ack of 11.3 miles on Auto will that affect my link more than having it static at 10.8 miles on both the AP and Station.

Yes it can.

Thanks,

Yes it can, make the link worse when one side is 9.5 and the other 11.3, or make it better?

[Shockware1]

For a PtP link the ACK should be the same at both ends I would assume. Any difference between them should make things worse.

[WHT]

I don't know why, but when I use Auto ACK, I get two different distances.

Mike had me try setting it to auto and wait at least a minutes for the distances to settle down.I set both ends for AutoACK and did several minutes of pinging from both ends.

TX end ACK/Distance: 24.0 miles with AutoACK
RX end ACK/Distance: 25.2 miles with AutoACK

[sbyrd]

Did it make you link worse, better, or no change? What happens if Auto settles at one distance, but during the course of operation drastically differs on one end. Will my link get worse, or will it be better because the unit has adjusted to a better setting because of interference or what not?

[WHT]

We tested this a few weeks ago, so I don't know f this was with the XM.v5.0-rc5.2188.090618.1223.bin or more recent XM.v5.0-devel.2401.090722.1132.bin, but what we sw was the Auto ACK dropped the Station side to one mile when there was no data being passed. Once data started passing, it took a few seconds to adjust up to about 22 miles.

I just tried duplicating it, but can't; so it might have been an earlier firmware issue. Right now the link is stable with 65/65 TX/RX at both sides regardless if I used fixed 26 miles at both ends or Auto for TX 24 and RX 21 miles.

<shrugz>

[UBNT-Mike.Ford]

I don't know why, but when I use Auto ACK, I get two different distances.

Mike had me try setting it to auto and wait at least a minutes for the distances to settle down.I set both ends for AutoACK and did several minutes of pinging from both ends.

TX end ACK/Distance: 24.0 miles with AutoACK
RX end ACK/Distance: 25.2 miles with AutoACK

Hey Guys,

ACK is not just a factor of distance, there are some other mechanisms that help with interference. Also when the ACK varies between two sides, its usually an indicator of interference on one side. When a packet is dropped, the ACK will bump up. So high error counts can push the ACK timing up. I have found in ALL links that AutoACK is the best way to go due to the sensitivity of 11n ACK timing from teh Atheros chip.

Thanks,

Mike

[toshik]

why 20 MHz? up to 40?