Any recommends on a wireless setup for the new house?

It's also why I presented the Ethernet over Coax option. Odds are the house already had runs of coax to every room. Add a few adapters and you're looking at a nice 135-270 Mbps range depending on equipment used.

 

Why would it have that ?

 

Many newer homes are pre-wired for cable tv in most rooms because Americans like to have tvs in every room. Sadly my house is older so I have cat5 in 3 rooms that I ran myself poor me.

I'm so old I remember when all Ethernet was coax.

 

Davros, that's like asking why would all the rooms have electricity run to them.

Exactly what Mize stated. All the newer homes in the US have all the rooms wired for coax. That was even the case in my sister's neighborhood 8 years ago. They have a central drop in the basement with 1 feed coming in from the outside. It was all nicely done by the developers. All I had to do was add in an active 8-split signal booster to ensure perfect quality to all their rooms.

I too remember the days of all ethernet being coax with a terminator going bad or someone removing a PC and not placing a terminator on it and having the entire network go to crap.

 

Or someone simply T-ing off an exiting line not realizing coax is essentially a waveguide. Then you have to get out the TDR and bounce the lines for reflections to find where they screwed things up because of the rat's nest of wires in the drop ceiling.

Coax is fast though.

 

Are you talking about a wired or wireless device connected to the router? About connecting to the internet, to another device on the network, or to a HD/NAS attached to the router? Smallnetbuilder sees upwards of 60Mbps (7.5 MBps) at 6' away (Location A), but different wireless adapters, interference, etc.

Any interference from neighbors' wifi networks (d/l inSSIDer to check) or older cordless phones? Are you constantly microwaving things?

 

I'm guessing you mean 1.3GBytes, not bits right? Just making sure. 1.3GB * 1000MB/GB / 9.5m / 60s/m ~= 2.3MB/s, which is about what you were getting before.

Smallnetbuilder's review says they saw ~9MB/s to a NAS and ~50MB/s WAN-to-LAN, so you probably can do better. Is your router set to 'n' only, or 'n' and 'g' (like Mize asked)? And did you try running inSSIDer to see what's competing on your channel?

 

You bought a dual band router, right? If you need b/g put it on one and keep N alone on the other. Running in a multi-mode (b/g/n) will slow things down. Also, metal ducts (in the walls) and plumbing, etc. will mess with signals. One of my wifi computers is in the kitchen area nowhere near my CAT5 and even with a dual band n router I have enough drywall and metal between my router and this computer that I'm running at or below your 10 minutes for 1.3 GB speed. Fortunately the only thing that streams to it is Youtube.

 

So what kind of throughput (aka wrongly used term "bandwidth") are you getting now?

 

3 MB/sec = 24 Mb/sec = ALMOST enough for HD streaming.
Pretty impressive.

 

I have about double that (1.3 GB in ~5min) to the bedroom (~10m), with 12 other visible access points around. I didn't turn off g, because I need it for the EEE PC 900, which means that it tops out at 64 Mb/s.

It seems to depend a lot on the firmware, as the routers that play by the official rules (switch back to a single channel and/or drop down to g when a g channel is detected, instead of only skipping the channel that overlaps when it is actively used) suffer quite a bit when there are other 2.4 Ghz sources or g devices nearby.

Perhaps you can configure that?


Edit: you can buy 2.4 Ghz signal detectors for less than $10, some are USB. Most of those only show signal strength, but you can get around that by turning off the 802.11n devices. And they nicely show why you tend to lose your connections when you or your neighbours turn on the microwave

Signal strength (2.4 Ghz) is like this:

Microwave oven leakage: up to 0.005W/cm2, or about 5W total for a big, leaky one.
WiFi: 0.3W
Bluetooth: 0.0025-0.1W
Zigbee: 0.001W

 

I just got ~6MB/s WAN-to-LAN (laptop to desktop), ~2.4MB/s WAN-to-WAN (same computers) moving a ~500MB file about 6' away from my cheapo Asus RT-N13U. No metal ducts for central air here. I have a bunch of competing APs around, but only one weak competitor on my channel. WPA2, 20MHz, 'auto' mode (b/g/n) with "b/g protection." I don't even know if I'm connected at n speeds or if I'm always at g b/c of my other g devices. And I do have an old 2.4GHz phone that can mess with my signal when it's in use, but the base is far enough away that it doesn't seem to affect things when it's not in use.

I'm sure your Netgear will be better at handling multiple connections, digi, but it still seems weird that you can't see better speeds in the same room. It's also funny that WAN-to-WAN is so much slower than -to-LAN, but I'm totally clueless about the overheads involved.

 

Do you mean LAN to LAN?

Cheap routers don't have a hardware switch, but rely on the microcontroller to do all the work, and they only have a single channel to the LAN ports and/or a single buffer there. Essentially, they're half-duplex (read OR write).

For the WAN connection, they do have a dedicated channel, as it's completely different than the LAN ones.

So, in that case, your LAN-LAN speed is at most half your WAN-LAN speed, and the maximum speed (throughput) is determined solely by the speed at which the microcontroller can read and write the data.

And, of course, that bandwidth has to be shared by all devices and connections.


If you're using the "b/g protection", that means that it will only use one channel (out of three, or less if you have less antenna's on the router) if it detects a single b/g source, and will drop back to g if there are multiple and/or there is no room for a single channel that isn't interfering.

The 2.4GHz phone is pretty bad, in that it likely uses a big piece out of the available bandwidth, so with that and a single, external g channel, you're likely only using g.

You can try to unplug the phone, and/or disabling the b/g protection, but the latter might make g channels impossible, depending on the implementation.

On the other hand, if they did implement a time sharing workaround (as on my router), you would get the maximum throughput possible without interfering with the g channel(s). That would work best.

If you have a cheap router with only a single antenna (802.11n 120 Mb/s), that won't work, and there is a minimal distance you need to be from the router for it to work optimally. Reflections are a problem as well.

You could measure that with smart WiFi software (Toshiba had a nice tool for that, perhaps there are OpenSource ones?) or with a signal strength detector.


EDIT: found a nice graph:



As you can see, there's only room for 3 different n channels at the same time, which will use up all the bandwidth available. The g channels are much smaller, but still overlap with both channels next to them. And the phone is very likely using a pretty wide channel as well, up to the whole bandwidth available in the worst case (if you have an old, cheap one), but that is unlikely.

EDIT2: If you add a channel marker before and after the ones in the graph, you have the 13 g channels, which are all about 2 markers wide.

EDIT3: well, the above is actually a bit of a simplification, as the bandwidth for n is wider, but it gives a good picture.

 

Good info, Frank. Thanks. I might try disabling b/g protection (my router's supposedly 2T2R, though its max transfer rate is only 143Mb/s).

When I said WAN I meant wifi, and LAN meant CAT5. I didn't test LAN-LAN. I think the minimum distance got me, as the desktop was 2 feet from the router.

 

Ah. Wifi = WLAN