Forum

I guess we can't attach pictures here....

http://boostedmopar.com/temp2/ouch.jpg

-40 dBm is a very strong signal and it's on continuously. I would bet that devices on channels 1 and 6 get almost no throughput if they are anywhere close to this source.

It looks to be about 1 MHz or so wide, and its' between between Wi-Fi channels. So based only on the frequency I would guess it is a jammer, or a Class one Bluetooth device set in some FCC test mode.

Sorry I'm not up to analysing it by its shape (yet).

From the long-term waterfall on the far left, it looks to me like there's still Wi-Fi traffic on 1 and 6. I'm sure it's causing a measurable performance difference though.

Performance on 1 and 6 was bad however turning off the device didn't [b]seem[/b] to make a different. It was a really harsh Wi-Fi environment and there were some gateway issues that may have come in to play. Very limited test window also.

I am not an AP guy and I am just getting in to the analyst side of life. My question is... An 802.11 device will do a energy detection to determine if it can transmit, right? What I am wondering is how wide does it check? Does the interference have to be right in the middle of the channel it is using or will a device like this one that is right on the edge cause the same problem? I can't seem to find any documentation that says what is going to happen.

In the CWAP book page 471 it does say this.

"One final statement about Wi-Fi cards and their misrepresentation of ambient raw RF.
Sometimes if the interfering device is strong enough, when a Wi-Fi device performs a
physical clear channel assessment by detecting the RF energy (as opposed to a virtual
CCA), it can sense that the frequency is not clear and then refuse to transmit any frames.
If this occurs, there would be no frames transmitted in the air, meaning no bits flying
across the RF media to be harmed by the interfering device. In this scenario, with your
Wi-Fi card monitoring, the RF will show no interference and also no signal since the Wi-Fi
devices are not sending any frames at all, but it will display a channel utilization at or near
100 percent."

BTW what kind of device was this?

It's not a properly functioning Wi-Fi device which would show up on a channel - not between two of them.

Look in the 802.11 spec.

Here is part of it:

15.4.5.13 PMD_ED.indicate
15.4.5.13.1 Function
This optional primitive, which is generated by the PMD, shall indicate to the PLCP layer that the receiver
has detected RF energy indicated by the PMD_RSSI primitive that is above a predefined threshold.

15.4.5.13.2 Semantics of the service primitive
The PMD_ED (energy detect) primitive, along with the PMD_SQ, provides CCA status at the PLCP layer
through the PHY-CCA primitive. PMD_ED indicates a binary status of ENABLED or DISABLED.
PMD_ED shall be ENABLED when the RSSI indicated in PMD_RSSI is greater than the
ED_THRESHOLD parameter. PMD_ED shall be DISABLED when the PMD_RSSI falls below the energy
detect threshold.

15.4.5.13.3 When generated
This primitive shall be generated by the PHY when the PHY is receiving RF energy from any source that
exceeds the ED_THRESHOLD parameter.

15.4.5.13.4 Effect of receipt
This indicator shall be provided to the PLCP for forwarding to the MAC entity for information purposes
through the PMD_ED indicator. This parameter shall indicate that the RF medium may be busy with an RF
energy source that is not DSSS PHY compliant. If a DSSS PHY source is being received, the PMD_CS
function shall be enabled shortly after the PMD_ED function is enabled.

Also find this document by Steve Shellhammer for some interesting ideas:

[u]Clear Channel Assessment Energy Detection (CCA-ED) in 802.11y[/u]

Thanks Wlanman, I'll check that out.

Kevin,

Also check out 18.4.8.4 CCA.

If you look at CCA Mode 4, you will see that it specifically talks about timer durations at particular bit rates.

I'm sure there are other references too if you look in the actual spec.

here is a better picture of it without all the other noise.

http://t.co/qtXGw1Ly

I would stay with my original estimate above, as to the source.

However, just below what should be CH 14 is an interesting spike. Since that hasn't been mentioned before, is it possible that this is a harmonic (noise) from something else on the PC's PCB ? Say a poorly grounded oscillator?

Believe it or not, I have seen harmonics this high ! And it caused two problems.

The first was that overall sensitivity of the radio was lower than it should have been by about 10 dB. Secondly, there was so much noise in the unit that it failed the 802.11 test for Receiver Maximum Input Level (18.4.8.2).

A conducted sensitivity test run from say -20 to -100 dBm showed that only signals [u]between[/u] approximately -45 and -70 dBm could be seen. Really nuts.

The unit gave very strange range test results too, until we found the problem and fixed it.