How do I force a Macbook in OSX to prefer 802.11a over 802.11g? I cannot shut off 802.11g on the WAP because it is used by other devices. Both use the same SSID which also cannot be changed.
Asked
Active
Viewed 2,699 times
2 Answers
3
Apple doesn't provide a way to force bands on the client. So the only way to do what you want is to give the 5GHz network a different network name (SSID) and only put that network name in your preferred networks list.
For best results, make sure the 2.4GHz network name is not in your preferred networks list; if you have to have both in your list, make sure the 5GHz network outranks the 2.4GHz network. However, 2.4GHz networks are much easier/quicker to find in scans than 5GHz networks are, so even if you rank 5GHz over 2.4GHz, if you have the 2.4GHz network in your list, there may be times when the 5GHz network gets missed but the 2.4GHz network gets found, so your machine joins the 2.4GHz network after all.
Spiff
- 104,423
-
+1 to using two separate networks. It's a built-in feature in Apple's Airport Extreme, tho: http://www.apple.com/airportextreme/specs.html – cregox Mar 26 '10 at 17:51
0
edit: reference to the comments.
Adding to Spiff's answer, it's better to keep the networks as separated as possible. And it's not really that Apple don't do it (tho it really doesn't offer any way to choose, neither does windows or ubuntu, by OS default)...
From experience, if there's one device connected to the slowest protocol, that's the one every other device will have to deal with. So if you have a b/g/n network, and all the devices are "n", then they all will reach "n" speeds. But the moment you introduce a "b" or "g" device to the network, they all drop to the slowest speed.
This also used to happen even in wired networks. If you have a switcher that can handle 100mbps but there's just 1 computer plugged in that's 10mbps, most switchers would drop down the whole network velocity to 10mbps. That's sure to happen on a hub, but some more advanced switches can actually isolate the 10mbps and make the rest of the network communicate at the fastest speed.
Things are not so simple with wireless routers, tho. In the cable router it can be physically isolated and it's just a matter of a board architecture. In wireless it's all broadcast. Picture a radar. So for a multi-band being able to cast more than 1 band, it would have to do it in very small intervals (very unlikely) or have two transmitters and two receivers (such as apple airport extreme).
As a side-note, since this ended up being about wireless quality, it's interesting to note each frequency have their own set of interferences. The 2.4 GHz can suffer lots of interference for instance. So, depending on your environment, it could be better or worst using either one.
cregox
- 5,894
-
2Please don't spread that myth that all devices on an 802.11 network drop to the speed of the slowest device. It is manifestly untrue.
It was even less true of wired Ethernet networks.
– Spiff Mar 24 '10 at 20:08 -
@Spiff: it was true about thicknet/thinnet ethernets (mostly). what were those, 10base-5? 5base-2? i can't remember anymore. – quack quixote Mar 25 '10 at 09:03
-
@Spiff Ok, sorry. I think I'll just delete this then. What you suggest? – cregox Mar 25 '10 at 17:37
-
@Cawas, yes, I think it would make sense to delete it.
@quack, I love your sense of humor. (For those who didn't get it, those PHYs only supported 10mbps signaling, so plugging in a 10mbit device would limit all the other 10mbit devices to 10mbit because they were all 10mbit.)
– Spiff Mar 25 '10 at 20:01 -
@Spiff, sorry again for my bad research. But everything I said makes sense, and thought I understand you're saying that doesn't happen, it did happen to me. If not that idea, what could make the faster device reduce speed on that situation? And about hubs, I'm 100% positive it's connecting slower devices that drops the whole network down. So it does happen on hubs. Could you point me to learn better about all that? (PS: I un-deleted this just so I could send you this "msg"/question, spiff, since I could find no better way to do it) – cregox Mar 26 '10 at 17:29
-
@Spiff I will just delete this and verify if you still can read the question. Then you can send me an e-mail or just edit your answer or add a comment to it. – cregox Mar 26 '10 at 17:49
-
-
2@Cawas, A standard part of Wi-Fi certification testing requires that your newer higher-rate clients and APs still use their higher rates even when older lower-rate clients are active on the same AP. The higher-rate devices usually have to make a few small adjustments to accommodate the presence of the legacy devices (to that they don't interfere with each other too badly), but that should only have a minor effect, like a 10% slowdown, not a complete drop from 300mbps to 54mbps (N to A or G) or from 54mbps to 11mbps (G to B). – Spiff Mar 26 '10 at 19:26
-
1@Cawas, regarding wired Ethernet, hubs are single-speed devices by definition, you can't plug a 10BASE-T-only device into a 100BASE-TX hub. So-called "dual-speed hubs" are a misnomer, and even they won't slow down 100BASE-TX devices when 10BASE-T devices are plugged in.
If you can show me a crappy "dual-speed hub" or switch that cycles all ports to renegotiate link speed for all clients down from 100BASE-TX to 10BASE-T just because you plugged in a 10BASE-T client, I'll eat my hat.
– Spiff Mar 26 '10 at 19:32 -
@spiff: actually i was referring more to the fact that in thicknet/thinnet all network nodes were on the same wire. no switches; this was pre-star topology, so nodes could only transmit one at a time. (but these were homogenous in regards to link speeds so you wouldn't have a mix of different speeds.) – quack quixote Mar 26 '10 at 19:33
-
@quack so you weren't really responding to my contention at all. I was only speaking to the notion that plugging in devices of a slower link speed drops the link speed for all other devices. As you say, that doesn't apply to 10BASE-5 or 10BASE-2. – Spiff Mar 26 '10 at 19:37
-
More on so-called "dual-speed hubs" being a misnomer:
http://superuser.com/questions/123308/under-what-circumstances-can-will-a-10base-t-device-slow-down-a-100base-t-or-fast/123375#123375
– Spiff Mar 26 '10 at 19:38 -
@Spiff yeah, I've seem that and maybe all your posts so far. Thanks for the the clarification and even repeating yourself. I still can't imagine what else could cause the problem, tho. – cregox Mar 26 '10 at 19:52
-
For the wired part, I've had "slow-down" experiences with some "dual-speed hubs" and with one 1gbps huge switcher as well. I will do a proper examination on them the next time I bump into one. For the wireless part, that's nice to know. Again, eventually I'll force myself to properly test the issue the next time. This can happen a lot sooner than with the wired tests. – cregox Mar 26 '10 at 19:53
-
@spiff: right. i believe you are correct on the technical aspects, but i honestly can't confirm the lower-level details without digging into some external references. (for wired or wifi, tho i'm more familiar with wired.) the misnomer-of-dual-speed-hubs link is a good read, thanks. – quack quixote Mar 26 '10 at 20:27