cancel
Showing results for 
Search instead for 
Did you mean: 

WiFi speed problems – g is faster than n!

Community Veteran
Posts: 6,111
Thanks: 1
Registered: 05-04-2007

WiFi speed problems – g is faster than n!

Hi all,
I was wondering if anyone who's particularly knowledgeable about the various 802.11 standards might be able to shed some light please on strange behaviour I'm witnessing with my WiFi speeds. I've had my router (an Apple Time Capsule – don't laugh, it genuinely is a very solid router) set to 802.11n 2.4GHz with b/g compatibility for some time now, but recently I've wanted to do lots of data transfers across our home network and I noticed that the speeds I was getting were very odd. Namely, they were fluctuating a great deal during any single file transfer; the speed would be constantly zigzagging between about 1 and 4MBps when sending data from the Time Capsule to my computer, whereas in the other direction the speed would be much much more constant.
To try and get to the bottom of the issue, therefore, I did some testing. First I tested the time it took for my mother's computer to send or receive 250MB to/from the Time Capsule set to use various wireless standards exclusively, utilising the clearest channel available in each case. No other WiFi devices were enabled whilst this test was carried out, and the results are thus:
Sent
Received
n 2.4GHz
34s, steady-ish82s, fluctuating
n 5GHz
30s, steady-ish82s, fluctuating
g
80s, steady68s, steady
a
78s, steady66s, steady

I then enabled WiFi on my computer and did the same test, but this time with my mother's computer sending or receiving the same data to/from my computer:
Sent
Received
n 2.4GHz
70s, fluctuating78s, steady-ish
n 5GHz
77s, fluctuating92s, fluctuating
g
181s, steady-ish170s, steady
a
191s, steady-ish250s, steady-ish

(The words after the numbers of seconds denote how steady the data transfer rate was during the file transfer, having observed the Mac OS X Activity Monitor's network activity graph. ‘Steady-ish’ meant there were some fluctuations in the graph of perhaps 10-15%, whilst ‘fluctuating’ obviously means somewhat worse than that.)
Those results have left me even more confused than before I started testing the speeds, I think! Specifically, these are the things I'm wondering:

  • Why is it faster for the computer to receive data from the router using g than it is with n, yet n is faster in the other direction?

  • Why does the speed fluctuate wildly whilst receiving data from the router over n, yet it's perfectly steady over g?

  • Why does the disparity in the sent/received rates when transferring between the router and computer over n disappear when transferring between two computers instead?

  • Why do I not see any improvement when using n at 5GHz, considering that there are loads of 2.4GHz networks around here?


I understand that 802.11n uses multiple antennas, and prior to doing the testing above it was suggested to me that perhaps the cause of my issues was one or more of the n-only antennas in the router being broken. That would have made sense, except for the fact that it's possible to send data to the router over n at a much higher speed than over g, which implies that the n antennas are working correctly after all. Unless there are different antennas for sending and receiving data?
It's also worth noting that 68 seconds to receive 250MB of data over g – about 29Mbps – is, as far as I'm aware, a pretty fast real-world speed to achieve using that wireless protocol. So it's not as if the channel in question was chock-full of interference or that the router's generally rubbish and not capable of good WiFi performance.
So would anyone be able to shed some light on the strange wireless performance please? Undecided
Edit: I should also add that we have no known sources of interference at home (other than the microwave when it's on), although the walls here are thin so it's possible that we could be experiencing interference caused by neighbours' devices. But then that'd raise the question of why does only n become affected by the interference and not g…