[ath9k-devel] Question about antenna throughput.

[ath9k-devel] Question about antenna throughput.

[Ben Greear]

We have a 3x3 NIC (wpea-127n).  A user reports that when using
a single antenna cable to directly connect to an AP (with attenuator),
that one of the antenna can do about 30Mbps download from AP, the
other about 25Mbps, and the third cable can only do about 1Mbps.

Is that expected results?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Question about antenna throughput.

[Adrian Chadd]

On 12 July 2012 07:11, Ben Greear <greearb@candelatech.com> wrote:
> We have a 3x3 NIC (wpea-127n).  A user reports that when using
> a single antenna cable to directly connect to an AP (with attenuator),
> that one of the antenna can do about 30Mbps download from AP, the
> other about 25Mbps, and the third cable can only do about 1Mbps.
>
> Is that expected results?

If I recall what's going on, the actual signal sent out the antennas
isn't necessarily a mirrored copy from each antenna, even for 1 stream
and legacy signals. Definitely for 2 and 3 stream signals that is
absolutely to be expected.

The baseband may decide to "scramble" it so part of the wireless
transmission goes out each antenna. Don't ask me about the details, I
don't know them.

There's also STBC for 1 stream 11n rates.

If you want to connect a single cable to an AP, you have to put it
into a single chain TX/RX mode and use the correct chain. On that NIC
you can configure it to be a single chain on any of chain 0, 1 or 2.
(It doesn't have to be 0.) But you have to choose a single chain.

So, yes. I think that's totally understandable behaviour.



Adrian

[ath9k-devel] Question about antenna throughput.

[Ben Greear]

On 07/15/2012 02:39 AM, Adrian Chadd wrote:
> On 12 July 2012 07:11, Ben Greear <greearb@candelatech.com> wrote:
>> We have a 3x3 NIC (wpea-127n).  A user reports that when using
>> a single antenna cable to directly connect to an AP (with attenuator),
>> that one of the antenna can do about 30Mbps download from AP, the
>> other about 25Mbps, and the third cable can only do about 1Mbps.
>>
>> Is that expected results?
>
> If I recall what's going on, the actual signal sent out the antennas
> isn't necessarily a mirrored copy from each antenna, even for 1 stream
> and legacy signals. Definitely for 2 and 3 stream signals that is
> absolutely to be expected.
>
> The baseband may decide to "scramble" it so part of the wireless
> transmission goes out each antenna. Don't ask me about the details, I
> don't know them.
>
> There's also STBC for 1 stream 11n rates.
>
> If you want to connect a single cable to an AP, you have to put it
> into a single chain TX/RX mode and use the correct chain. On that NIC
> you can configure it to be a single chain on any of chain 0, 1 or 2.
> (It doesn't have to be 0.) But you have to choose a single chain.
>
> So, yes. I think that's totally understandable behaviour.

Ok, I'm going to look into this chain configuration.

This user is also reporting that our box can only communicate
down to about -72dBm, whereas some off-the-shelf USB dongles
are providing around 100kbps at -78 to -80dBm.

Are there any published rate v/s range results for ath9k available?

Perhaps the chaining issue is also involved in this?

Thanks for the detailed answers on this....this is new
territory for me!

Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Question about antenna throughput.

[Sujith Manoharan]

Ben Greear wrote:
> Ok, I'm going to look into this chain configuration.
> 
> This user is also reporting that our box can only communicate
> down to about -72dBm, whereas some off-the-shelf USB dongles
> are providing around 100kbps at -78 to -80dBm.
> 
> Are there any published rate v/s range results for ath9k available?

Not published, but we do have RvR numbers for ath9k, unfortunately
they are really old.

I can get new numbers based on current -wt, maybe by this weekend,
and we can add them to the wiki or something.

Sujith

[ath9k-devel] Question about antenna throughput.

[Ben Greear]

On 07/16/2012 10:28 AM, Sujith Manoharan wrote:
> Ben Greear wrote:
>> Ok, I'm going to look into this chain configuration.
>>
>> This user is also reporting that our box can only communicate
>> down to about -72dBm, whereas some off-the-shelf USB dongles
>> are providing around 100kbps at -78 to -80dBm.
>>
>> Are there any published rate v/s range results for ath9k available?
>
> Not published, but we do have RvR numbers for ath9k, unfortunately
> they are really old.
>
> I can get new numbers based on current -wt, maybe by this weekend,
> and we can add them to the wiki or something.

That would be great!  Even brief details on how the tests were
done would help too..we'd like to run similar tests and verify
that our systems are performing at expected levels...

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Question about antenna throughput.

[Adrian Chadd]

On 16 July 2012 10:19, Ben Greear <greearb@candelatech.com> wrote:

> Ok, I'm going to look into this chain configuration.
>
> This user is also reporting that our box can only communicate
> down to about -72dBm, whereas some off-the-shelf USB dongles
> are providing around 100kbps at -78 to -80dBm.
>
> Are there any published rate v/s range results for ath9k available?

That's odd. You should be able to receive low rates at a much lower
signal than that.

I'll work with Sujith and see if we can pull up the expected
sensitivity tables for each rate. Maybe the RX gain tables are all
wrong, or the NICs you're using have poor(ish) RX EVM?



Adrian

[ath9k-devel] Question about antenna throughput.

[Ben Greear]

On 07/18/2012 09:44 AM, Adrian Chadd wrote:
> On 16 July 2012 10:19, Ben Greear <greearb@candelatech.com> wrote:
>
>> Ok, I'm going to look into this chain configuration.
>>
>> This user is also reporting that our box can only communicate
>> down to about -72dBm, whereas some off-the-shelf USB dongles
>> are providing around 100kbps at -78 to -80dBm.
>>
>> Are there any published rate v/s range results for ath9k available?
>
> That's odd. You should be able to receive low rates at a much lower
> signal than that.
>
> I'll work with Sujith and see if we can pull up the expected
> sensitivity tables for each rate. Maybe the RX gain tables are all
> wrong, or the NICs you're using have poor(ish) RX EVM?

No idea.  It could be mis-configuration at any level..but if there
are published results, we can run similar tests and then pay attention
if we see any significant differences.

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Question about antenna throughput.

[Ben Greear]

On 07/18/2012 09:44 AM, Adrian Chadd wrote:
> On 16 July 2012 10:19, Ben Greear <greearb@candelatech.com> wrote:
>
>> Ok, I'm going to look into this chain configuration.
>>
>> This user is also reporting that our box can only communicate
>> down to about -72dBm, whereas some off-the-shelf USB dongles
>> are providing around 100kbps at -78 to -80dBm.
>>
>> Are there any published rate v/s range results for ath9k available?
>
> That's odd. You should be able to receive low rates at a much lower
> signal than that.
>
> I'll work with Sujith and see if we can pull up the expected
> sensitivity tables for each rate. Maybe the RX gain tables are all
> wrong, or the NICs you're using have poor(ish) RX EVM?

Any luck on this?

Thanks,
Ben

>
>
>
> Adrian
>


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com