The 5.0 Band is weaker than 2.4 Band but faster. What gives?

[jborchel]

I have a new AC router, an Asus RT-AC66U. I have a laptop with an N antenna built in and an AC antenna in one of the USB slots. I have setup two wireless connections, one for each band.

My 5.0 band says it communicating at 352mbps while the 2.4 says 72mbps. All well and good as I guess those are the speeds that AC and N are supposed to be. But the signal strength for the 5.0 band says 3 out of 5 bars while the 2.4 says 5 out of 5 bars. I thought that the AC tech was supposed to give me a stronger signal for the 5.0 band. Have I done something wrong in the setup?

 

[Phillip Corcoran]

http://kb.netgear.com/app/answers/detail/a_id/19433/~/why-dont-i-get-a-strong-signal-with-my-5-ghz-wireless-connection%3F

 

[bill001g]

That is to be expected. If you were to test 802.11n on the 5g band it too would get a weaker signal but I would bet you would see 150m because it is more likely to get 2 channels. The 5g band does not penetrate as far is the only reason it is weaker. As to the speed it depends on lots of things. The connection speed is only part of the equation the actual throughput tends to be much less depending on how much interference there is. This is where signal level makes a difference since a stronger signal will not get interfered with as much.

You can't really read a lot into any of these numbers you need to just see how much data you can actually get though.

You likely can get your 2.4g 802.11n to go to 150m if you were to force it to 40mhz channels. This does not mean it would actually transfer data any faster since you have no double the chance you have to get interference....but it would likely connect at 150m.

 

[jborchel]

Theoretically, what would be better response time 72 mbps at 5 out of 5 bars, or 352 mbps at 3 out of 5 bars?

 

[bill001g]

The throughput speed only slightly affects the response time. If you get a perfect transmission with a maximum packet size at most you are going to see a couple ms difference. The large difference will be how many times the packet gets corrupted and must be retransmitted. That is impossible to predict but is the common cause of delays in wireless networks.

 

[jborchel]

So, if packet corruption is the main cause of delayed response then it seems that 72 mbps at 5 out of 5 bars would be the best choice. Is there an app that measures the average response time over a session time? 30 minutes for example? I know there is one that measures signal strength.

 

[bill001g]

Some device drivers will tell you how many errors you are receiving. Some call this quality and other just call it errors...and some totally ignore it. There really is no easy way to measure a session and really be sure it is just your wireless unless you are only talking to servers in your house. There are a couple of tools that capture the traffic and generate reports. You could try wireshark and run its report on delays but this tends to be a overkill way to get the information. Off the top of my head I don't know a free tool that will do this and I am too lazy to search.

I have been looking for a very long time for a way to get the number of times a packet is retransmitted in a wireless session due to packet damage. The router and the nic card have to know this since they are the ones doing the retranmission. I am starting to suspect this is actually done by this wireless chips themselves without the route/computer having to worry about it.

 

[jborchel]

Some device drivers will tell you how many errors you are receiving. Some call this quality and other just call it errors...and some totally ignore it. There really is no easy way to measure a session and really be sure it is just your wireless unless you are only talking to servers in your house. There are a couple of tools that capture the traffic and generate reports. You could try wireshark and run its report on delays but this tends to be a overkill way to get the information. Off the top of my head I don't know a free tool that will do this and I am too lazy to search.

I have been looking for a very long time for a way to get the number of times a packet is retransmitted in a wireless session due to packet damage. The router and the nic card have to know this since they are the ones doing the retranmission. I am starting to suspect this is actually done by this wireless chips themselves without the route/computer having to worry about it.

 

[jborchel]

I have an app named "insider" or something like that that tells me signal strength. It seems like an app that not only measures the retransmission rate but also tunes the specifications for maximum performance would be a hot seller. If a tech who knows wireless can advise you on the various tuning parameters, like optimim packet length for instance, than a well crafted program should be able to analyze your wireless situation and real time adjust your parameters to give you the best response time, if that what you are after. I would think there are only a small number of objectives that one might have and they could be ranked in importance by the user. I can think of two; response time and throughput. Depending on what was most important to you I would think the parameters would be different.

 

[bill001g]

INSSIDER is mostly smoke and mirrors. It is a good program and displays really nice charts but it is really stupid. All it is doing is listening for SSID broadcast and keeping track of how strong the signal is. These are just very short messages sent by the routers. They in no way indicate the actual utilization of any channel. It also only sees the very simple signal level. This really is just a pretty representation of the same list of networks you get when you setup a network connection. The main problem it has it assumes that a SSID somehow indicates traffic. I could have a single router on a channel sending data at maximum rate and a second channel that had 10 SSID doing nothing and it would pick the channel with the least SSID. Many routers can transmit multiple SSID per radio so the raw number of SSID means nothing about the actual traffic.

The only way to really do this is to actually capture the traffic on each channel and the graph amount of actual data seen per channel. Unfortunately microsoft has decided we don't need programs like that because of all the evil hackers. The have disabled the ability to put a nic card in promiscuous or monitor mode. So any good tool would have to run under unix.

Still even if you capture packets under unix there is no way to get damaged ones. I don't know why it does this but everyone says there is no way to get the nic to give you a packet that is not correct. This means the nic is some way is hiding all the retransmitted packets.