republican-creole
site Search:


 
    All Forums Hot Topics Gallery






how-to block ads


 
Search Topic:
Uniqs:
996
Share Topic
Posting?
Post a:
Post a:
Links: ·WISP Forum FAQ ·WISP Directory ·Radio Mobile
AuthorAll Replies

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON

Anyone using 3x3 antennas outdoors?

The 3x3 radio cards work great indoors even using good old rubber ducks. But when I tried outdoors using 3 small patch antennas on each side, throughput is not much better than 2x2 using same antennas beyond a few hundred meters.

WHT

join:2010-03-26
Rosston, TX
kudos:5

I came across three different sources for case studies of 4x4 MIMO LTE.

They reported only one optimum path polarity consistently supported a 75 Mbps stream and the other polarities provided clean up to pick up the load when the optimum path degrades.

They also reported where there were two each less than optimum paths, the other two paths provided clean up.

They didn't' see more than two optimum paths at any given time, and they also saw four very less optimum paths combine to proved 75 Mbps throughput.


lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
Reviews:
·TekSavvy DSL

I see some 3x3 panels out there, probably just an array of patches inside. My thought was that 3 separate 8dBi patches with more space between would be better and cheaper than those ... but it does not seem to help much over 2x2 at long range.

»Thinking out of the box ... Gigabit PTMP?
Quantenna 4x4 discussed long ago. Performance indoors was like the new 802.11ac radios but less spectrum usage. I tried outdoor 4x4 setups a few times but does not scale in distance.

»[Equipment] Tiniest triple radio setup
This old thread shows a very rare miniPCI with 3 completely independent 802.11a/b/g radios. I could even do tricks like running 5.3;5.8 and 2.4 simultaneously to reach 100mbps+30mbps TCP full duplex. That radio paralleling method actually performed well at very long range using very low TX power and narrow beam antennas. But they don't make them any more.

I also had brief access to an FPGA-based 8x2 MIMO platform and that actually scaled well in distance even with 7-9dBi rubber ducks at the AP side and rubber duck or patch antenna at CPE. But those platforms are so expensive that only cellular companies can afford them.


Chele

join:2003-07-23
kudos:1

reply to lutful
I was told by a manufacturer that in their case the radio only used the two of the 3x3 antennas that had better links.


WHT

join:2010-03-26
Rosston, TX
kudos:5

Appears consistent with what I was saying above with 4x4 MIMO LTE.


Chele

join:2003-07-23
kudos:1

Correct


rconaway8

join:2005-11-10
Phoenix, AZ

reply to lutful
MIMO is based on a 20dB isolation factor which is only created with a 90 degree polarity differential. That means a single antenna can only handle up to a 2x2 feed. Others claim a 45 degree offset but then you can't use either of the other polarities. The only way to make it work is to use two 2x2 antennas and they have to be separated by a certain minimum distance, depending on how far the link is. The farther the link, the more the antennas have to be separated. Although Quantenna can do 4x4, that's only for very, very short ranges relative to what a WISP needs.


lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
Reviews:
·TekSavvy DSL

said by rconaway8:

The only way to make it work is to use two 2x2 antennas and they have to be separated by a certain minimum distance, depending on how far the link is.

Based on a few tests, you have to space them more than 10 wavelengths apart to go even 1000 wavelengths.

WHT

join:2010-03-26
Rosston, TX
kudos:5

reply to lutful
As i posted in the UBNT forum,
Case studies of 4x4 MIMO LTE shows only one chain carries the burden of data and other three chains for error correcting redundancy.

A reply pointed out that LTE is different.

I was going to reply to that, but the new forum editor deleted my entire post when the preview crashed with an HTML error.

I was trying to say the salient point was not because it was LTE, but rather the efficacy of a 4x4 MIMO regardless if LTE, HPSDA, WiMax, or 802.11ac is used.


rconaway8

join:2005-11-10
Phoenix, AZ

With 802.11ac in 2x2 mini, its really no different. Both streams carry traffic on both channels. With dual antennas, all 4 channels will carry traffic.

In terms of separating the antennas, I remember that that 1 mile needed the antennas at least 5 feet apart.


WHT

join:2010-03-26
Rosston, TX
kudos:5

said by rconaway8:

With 802.11ac in 2x2 mini, its really no different. Both streams carry traffic on both channels. With dual antennas, all 4 channels will carry traffic.

As I pointed out, the case studies of 2x2 and 4x4 showed (I'm going from best memory at the moment) for data streams, i.e. 4 MIMO polarity diversity chains:
One will have 100% if path is optimum.
If not optimum, the second path picks up the slack.

Hold on..let me look for the info....

said by rconaway8:

In terms of separating the antennas, I remember that that 1 mile needed the antennas at least 5 feet apart.

A lot more than that....

I don't have my long range chart on this laptop, just the 3,000 foot range chart, and it shows
900 MHz NLOS of 40 feet
900 MHz LOS of 32 feet
2.4 GHz NLOS of 26 feet
2.4 GHz LOS of 16 feet

This is based on the scattering equation at longer distances. When you add Cambium's PT600 multi-beam,space-time coding (MBSTC) that adds RF diversity to the mix, you have to consider the minimum frequency diversity.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
Reviews:
·TekSavvy DSL

1 edit

reply to rconaway8

said by rconaway8:

In terms of separating the antennas, I remember that that 1 mile needed the antennas at least 5 feet apart.

I will appreciate any info/links for physical separation versus range. I have a small project where they really want to use Atheros cards in 3x3 mode for higher throughput.

WHT See Profile same request for you too. I need to determine required separation for a reliable 5.8Ghz link just under 2 miles. Smaller building is about 30ft wide on the link side but I could ask them to attach reasonable length booms on either side.

rconaway8

join:2005-11-10
Phoenix, AZ

reply to lutful
The only other number that comes to my fading memory is 100' at 100 miles. That at least gives you a scale to work from.


WHT

join:2010-03-26
Rosston, TX
kudos:5

reply to lutful

Click for full size
I originally used Bell System's formulas for 10 GHz PtM links for my 48 MHz and 450 MHz links back in the mid-'70s. I updated the formulas several years ago for 900 MHz and 2.4 GHz.

These are the minimum spacing lengths for two antenna diversity.

/edit
5.8 GHz would scale very close to half the 2.4 GHz spacing.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
Reviews:
·TekSavvy DSL

Based on tests so far, I am guessing that stronger side lobes and weaker polarization of panel antennas are compromising 3x3 operation. However, I have to learn a lot more MIMO theory to prove that hunch.

Based on that chart, I will suggest 3 circular polarization dish antennas with 3m separation between them and 3 cables coming from the radio as equal length as possible. Just to use both circular polarizations in the link, middle dish will be LHCP and the outer two dishes RHCP.

Now if 3x3 does not work out as expected with all that effort, it is back to 2x2 using LHCP+RHCP feed on a single dish.


Saturday, 18-May 17:47:52 Terms of Use & Privacy | feedback | contact | Hosting by nac.net - DSL,Hosting & Co-lo
over 13.5 years online © 1999-2013 dslreports.com.
Most commented news this week
Hot Topics