Ath9k distance

openwrt ath9k qca9531 .

Two devices are 1m apart, use IPERF test throughput ,

Set different distance.

==========1km==========
[ 3] 19.0-20.0 sec 5.25 MBytes 44.0 Mbits/sec
[ 3] 20.0-21.0 sec 5.12 MBytes 43.0 Mbits/sec
[ 3] 21.0-22.0 sec 5.00 MBytes 41.9 Mbits/sec

========-50km=========
[ 3] 2.0- 3.0 sec 1.12 MBytes 9.44 Mbits/sec
[ 3] 3.0- 4.0 sec 1.50 MBytes 12.6 Mbits/sec
[ 3] 4.0- 5.0 sec 1.25 MBytes 10.5 Mbits/sec

Is there a question here?

I don't even understand your statement... 1m apart, then showing iPerf results for what appears to be 1km and 50km??

With distance, SNR drops, which means that link speed also drops. Not sure if you're really talking about 1 kilometer and 50 kilometers -- even 1km is significantly father than standard wifi can operate (maybe you're using some dedicated long distance radio links?).

Both equipment are in my desk

When I set up 1000m:

/etc/config/wireless
   option distance '1000'

Device A: iPerf -s
Device B: iPerf -C xxxx

[ 3] 19.0-20.0 sec 5.25 MBytes 44.0 Mbits/sec
[ 3] 20.0-21.0 sec 5.12 MBytes 43.0 Mbits/sec
[ 3] 21.0-22.0 sec 5.00 MBytes 41.9 Mbits/sec

When I set up 50000m:

/etc/config/wireless
   option distance '50000'

Device A: iPerf -s
Device B: iPerf -C xxxx

[ 3] 2.0- 3.0 sec 1.12 MBytes 9.44 Mbits/sec
[ 3] 3.0- 4.0 sec 1.50 MBytes 12.6 Mbits/sec
[ 3] 4.0- 5.0 sec 1.25 MBytes 10.5 Mbits/sec

Only modify the "distance" parameters, the throughput has dropped a lot

It wasn't clear initially that you were setting a wifi parameter while keeping a constant physical distance.

I can't speak to exactly how this parameter works, but suspect that this is actually downgrading the bandwidth to accommodate the large distances. Essentially, if you know that you will have poor SNR due to physical distance, you can cap the bandwidth on the link to hopefully improve overall reliability. In a way, it is like a using a low pass filter to prevent aliasing -- for example, setting an LPF for 20KHz before a 44.1KHz analog to digital converter in an audio system.

Yes, I agree with you.
But I want to know if there is a way to set a long distance from fixed distance,
it will not affect the actual throughput ?

I'm not sure what this means.

What are you trying to accomplish by setting this parameter in the first place?

Both equipment are in my desk.

option distance '1000'
option distance '50000'

I want to implement the same environment, set different distance, but does not affect throughput.
What is your good way?

I still don't understand what you are tryin to achieve.

The distance parameter is there to describe the physical distance involved between client and AP. It is not a required parameter. If the devices are on your desk (i.e.1m apart), there is no reason to set this parameter, especially to a very large number.

As I said before, I suspect that this parameter intentionally reduces the bandwidth to make a more reliable/robust signal. So it makes sense that adjusting this parameter will affect the throughput. And this is why I don't understand why you are expecting to be able to set different distance values without affecting the throughput. That's like saying you want to adjust the volume on your TV but not have it get louder.

distance Distance between the ap and the furthest client in meters.

To set a large distance parameter (50000) on such a near distance, the throughput is very low. If I actually 50 km, the throughput will only be lower. So I want to make Distance without affecting the throughput. Test a 50km throughput

We have established that the parameter appears to reduce bandwidth. We also know that physical distance also has the effect of reducing the bandwidth. So increasing the value of the parameter and/or increasing the physical distance will both affect bandwidth. I don't think there is any way around that.

No standard wifi devices can communicate across this range. You need special dedicated radios with highly directional antennas to bridge a gap that large. You're not talking about standard wifi anymore when you talk about anything over 100-200m.

You are right, we already have these conditions.

I found the above-mentioned problem during the test, so I want to find out if there is a way to solve this problem, and then conduct a long-distance measurement.

I don't think this is classified as a problem when it is a feature that is performing by design. If you turn off your light switch, the light also goes off. That is not a problem -- it is a feature and the purpose of the switch.

1 Like

It may not be one, but is there a way to change this problem and let me do remote testing?
I just want to test if there is a higher throughput over long distances

You’ll have to just run tests over the actual physical distance you are hoping to achieve. Setting the parameter should help improve the reliability of the link, but you can start by placing your devices and then experiment with the distance value to see how it affects both throughput and link integrity (look at the packet loss rates).

I’m not sure what kind of physical distance you are hoping to achieve - 1km is well beyond normal WiFi range.

1 Like

You don't need to worry, I have done a 10km test, and the throughput is still ok.

Setting 10000m throughput drops very little (2 devices are 1 meter apart.), but setting 50000m throughput drops severely (2 devices are 1 meter apart.), which affects the test.

So it is suspected that distance affects throughput

Assuming no interference or obstacles, signal strength and snr will follow the inverse square law. The distance parameter likely calculates the target throughput based on inverse square and other factors.

You forget about power!
A standard wifi is by law about 20mW max sending power. But are we supposed to talk about 50km we need a lot more power to be able to sustain this speed. Like at least 5W or more likely 20W if we want higher throughput.

And a high tower because we talk about “in line of sight” at these frequencys.

To have a radio link setup with one meter isn’t really meaningful since radio transmitters that near each other will interfere each other just by being in the same room.

1 Like

I just want to test if there is a higher throughput over long distances