The datasheet  for the 3600 states that max transmit power and max power setting are 23dbm.
However, when setting up 2 different devices using the GUI, the highest setting available for 2.4G is 22dbm, and for 5G is only 14dbm. Why the discrepancy? Do I need to use the console to achieve the specified max for the 3600?
There is a comment in the datasheet: "The maximum power setting will vary by channel and according to individual country regulations." AFAIK these are North America devices, and 23dbm should fall well within regulations.
I'm noting also that I set up a 3700 and the GUI provided options for 23dbm on both channels.
The transmit power level depends on: the AP model, the AP regulatory domain, the number of antennas in use, the configured gain, the modulation being used, and the specific channel in use.
You can see all the gory details in the Channels and Maximum Power spreadsheet - https://www.cisco.com/c/dam/en/us/td/docs/wireless/access_point/channels/lwapp/reference/guide/3600e_3600i_pwr_chn.xlsx .
Thanks Aaron, the spreadsheet was helpful.
So what I gather from that is that the total output power is the sum of the power at the antenna ports. It would seem that this could be a little misleading in some cases, if one is thinking of the total radiated power and thus achievable range based on power output.
If the device is beamforming, then yes, it would make sense that the total radiated power (in a certain direction) would be a result of the sum of the power at each antenna port, and thus would be reflected in maximum achievable range.
However, if the device is optimizing mimo spatial characteristics in order to achieve maximum data throughput by distributing data equally across all four outputs, it doesn't seem in this case that the output of each antenna port would be additive, and therefore would offer no better range than a single port would.
This also raises another question, does the 3600 automatically choose between beamforming and optimized data spatial mode, or a mix of the two (however the purest form of each of these is mutually exclusive) presumably based on machine learning? And is there a way to hard-configure one or the other? (Perusing through the config doesn't seem to indicate a setting for that, but I could be missing something.)
(WLC2504-2) >config 802.11a beamforming ?
global Configures all 802.11a Cisco APs.
ap Configures one 802.11a Cisco APs.
If autonomous IOS:
Enter configuration commands, one per line. End with CNTL/Z.
Jul 18 13:33:10.636 MST: %PARSER-5-CFGLOG_LOGGEDCMD: User:console logged command:interface Dot11Radio0 amform ?
ofdm enable beamforming at OFDM rates
rssi signal strength threshold to apply beamforming
Do you have a -A or -B model? That can also make a difference.
By -A or -B you mean regulatory domain?
Also your other devices must support that output power, otherwise you have one way communication, or in other words an outage.
I understand your point if you are talking about the limits, however output power also has an affect on data throughput as the range is increased. Ie, as the client device gets farther away from the AP, throughput diminishes. If the output power of the AP is increased, the throughput from the AP to the client (ie, download) will diminish at a slower rate, ie, for a given range, an increase in AP power will increase throughput from the AP to the client.
Since the vast majority of data transfer for the average user is download, there is something to be gained by increasing AP output power, even if the client device has lower (even much lower) power.
I have wondered why in the US, the federal regulation is 1 watt maximum, but APs are never near this this limit, but perhaps as you point out, the discrepancy between a 1watt AP and the rather small output of many client devices would be so great there would be no point, since, even considering the scenario I just described there is a dimishment in the effectiveness of this scenario as the discrepancy between the power of the AP versus the client device becomes very large.