Dueling with Microwave Ovens
December 04, 2003
There's been a lot of talk about microwave ovens causing interference with wireless LANs. Take a look at what was found when testing Wi-Fi in their presence.
Radio frequency (RF) interference causes wireless clients and access points to hold off transmitting, which causes delay and lower throughput. This resulting decrease in performance can make browsing websites and downloading files sluggish. In cases where interfering signals are strong enough, the wireless clients may not be able to access the LAN at all for an indefinite period of time. This is rare, but possible.
As a result, you need to be aware of potential sources of RF interference, such as cordless phones, other WLANs -- and microwave ovens. In this tutorial, well focus on interference that microwave ovens create. Most microwaves emit signals that fall within the same 2.4GHz frequency band that 802.11b WLANs utilize. It's something to think about when deploying.
Real world testing uncovers some issues
The access point was a Cisco Aironet 350 Series, which was initially set to its maximum transmit power of 100mW. As a wireless client, I used a laptop with AirMagnet monitoring software. AirMagnet is able to continually transmit packets of a specified size for a period of time while measuring signal strength, noise levels, packet retransmission rates, and throughput.
The microwave included in the testing is made by GoldStar and resides in a break room. The label on the back of the microwave indicates that it consumes 1,200 watts of power and operates at 2,450 MHz, which is close to 802.11b channel 9.
Before turning on the microwave, I set the access point to channel 9 (a worst case situation), and took some measurements within the break room to use as a baseline. The access point signal level resulting from the beacons within the break room was -63dBm, sufficient for solid 11Mbps associations. Throughput tests indicated 667 packets per second (pps) while sending 1,532 byte frames.
While holding the wireless client within one foot of the microwave, I recorded some measurements with AirMagnet while the microwave was set to high and heating up a large bowel of water. The throughput fell to 90pps, a drop of more than 85 percent.
This is a substantial reduction in performance, but its the worst-case situation. The access point was set to the same frequency of the microwave, and its unlikely that someone would use a wireless client so close to the microwave.
A more realistic distance from the microwave is from one of the break tables, which is about eight feet away from the microwave. At this range, I reran the throughput tests, resulting in 178pps. This still equates to around a 75 percent decrease, something that would likely make users unhappy.
In order to see what a 75 percent decrease in throughput feels like, I tried surfing to a web site having a few graphics. With the microwave running, the pictures would come in painfully slow. I also surfed around a bit to other pages, and sometimes the pages would freeze.
After turning off the microwave, I cleared the browser cache, and found no problems surfing the same pages. The pages loaded lightening fast.
I also repeated the tests down the hall about 20 feet away with the microwave running and still experienced fairly sluggish responses. In fact, throughput from there was still only 260pps. Obviously, the microwave was making the WLAN crawl at surprisingly great distances from the microwave.
Something to consider is that these tests were run with only one active wireless client. The results would have been much worse if there were more users on the network.
Curious to know what channels the microwave would affect the most, I ran the throughput tests again with the access point set to different channels. On channels 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11, the throughput was 660pps, 658pps, 655pps, 651pps, 643pps, 574pps, 434pps, 258pps, 178pps, 191pps, and 210pps, respectively. Based on these numbers, the microwave was most critically impacting channels 8, 9, 10, and 11.
I also found that the impact is more severe near edges of the range boundary of the access point where signal levels are lower. To test this, I lowered the transmit power of the access point to 1mW, which caused the signal power to decrease to -77dBm. This level is close to the range boundary of the WLAN.
As one would expect, the hit on throughput was even more with the weaker signals. The resulting throughput was about 10 percent lower than the case where the signal levels from the access point were higher at -63dBm.
What do you do about interference from microwave ovens?
Consider the following countermeasures:
- Change access point channels. The microwave in use with this testing didnt severely degrade channels 1 through 6. As a result, avoid the use of these channels in areas of the building where microwave ovens operate. In fact, web browsing was very fast with the microwave running and the access point set to channel 1 and channel 6. Keep in mind, however, that your microwave ovens may operate at different frequencies within the 2.4GHz band. Check the label on the back of the microwave, which should provide the center operating frequency.
- Avoid using the WLAN near microwave ovens. Keep at least ten feet away from operating microwaves while actively using WLAN applications. This eliminates working from most company break rooms when the microwave is in use. The actual impacts would only be intermittent, though, when someone runs the microwave for a few minutes while heating up a burrito or cup of soup.
As you can see, microwave ovens wont completely bring down your WLAN. Just be aware of the situation, and of the applicable countermeasures.
Jim Geier provides independent consulting services to companies developing and deploying wireless network solutions. He is the author of the book, Wireless LANs and offers training focusing on WLANs.