I know this is an old thread but it was a decent discussion, and I'd been considering upgrading 200Mbps plugs to 500Mbps after a broadband upgrade. I decided to measure things first to see what was really being achieved so sharing what I found here.
The summary if you don't care about the details is that my 200-rated TP-Link devices have worked solidly for over 2 years now (Jan 2011, so quite old but really very reliable). They may be based on the same chip set as the Zyxel PLA-407 so possibly Atheros INT6400 from 2010 based on firmware name but I couldn't verify this for certain.
In favourable conditions the link quality is detected as 172 out of 200 and in my real world setup the connection is more like 70 or 80 out of 200. So on the real world 70-80 measure I actually get about 18.5Mbps or at the going rate of 8bits to a byte, then a little under 2.5 megabytes per second on real network transfers over my less than trivial power connection.
Achieving one quarter of the quality rating as real Mbps on the network between Windows machines seems to be very consistent. Whether that holds up for the more recent 500Mbps rated devices with newer chip sets (i.e. giving 125Mbps or 15 megabytes per second transfer rates with a 200 out of 200 quality link) I really cannot say.
I never really noticed this up until recently as the broadband connection itself maxxed out at 10Mbps so that was the weakest link in the chain. Now my next question is that since my aging EoP devices are limited by the actual physical power connections between them, would I get anything better out of 500Mbps devices?
Anyway, here are some basics on the setup - I have an outside home office with no practical Ethernet cabling option, 11-inch brick walls killing wireless, but with a secondary power circuit fed off the same mains supply as the main house. I bought a couple of cheap '200-rated' EoP devices with the plug-through, and set up the EoP connection from the broadband supply to the office into a small Netgear router supplying 100Mbit and 1000Mbit connections to devices in the office. Link over the EoP connection is always detected as 100Mbit by the router. Cat5e cabling everywhere, nothing over 5m.
Using speedtest.net to check broadband capacity, I get 38Mbps download and 9.5 Mbps upload at the modem, up to 36Mbps/9.5Mbps on wireless depending on signal strength, and typically no more than 18.5Mbps/9.5Mbps on the EoP connection.
Using iPerf, the link between the office and the house reports itself as supporting 18.9Mbps on network transfers (TCP), connections inside the home office on the cables reports 99.2Mbps (so near the theoretical maximum on the 100Mbps network). Using the EoP connections directly next to each other (separate mains plugs and with other devices such as main computer running) the EoP throughput is 40Mbps.
So iPerf and the observed limits on the broadband provider match almost exactly, indicating that my EoP link will support 18.9Mbps with this limit being due to the physical environment rather than the EoP devices. The limit in a near optimal working environment for the EoP devices (quality 172 of 200) is not much higher than 40Mbps.
Regarding ping times, I can't detect any differences (too small to notice). Also the throughput holds up even if someone is using a vacuum cleaner off an adjacent socket (so line noise isn't obviously a factor and I think these devices include some filtering).
The devices themselves have no obvious markings (bought them from eBuyer, product code 179960 discontinued) but the TP-Link diagnostic tool works with them and reports the following:
Vendor: TP-LINK
Firmware: INT6400-MAC-4-0-4011-00-3430-20090501-FINAL-C
I couldn't track down the actual chip sets (and to be honest didn't try very hard as these are quite old) but Atheros INT64000 looks like a good bet based on the firmware.
The diagnostic tool also shows a low-high rate range of 66-79 'widgets' (not sure that this is actually supposed to be Mbps but it should relate to the 200 'widgets' the device is rated for i.e. the line is providing about a third of the maximum capacity/quality). With the EoP Devices close to each other giving 40Mbps on Windows to Windows, the diagnostic utility reports the rate at 172 'widgets' out of a possible 200. The actual achieved computer-to-computer throughput for network exchanges appears to be about one quarter the number on the EoP devices (measured rating of 79 gave 18.5Mbps, measured rating of 172 gave 40.5Mbps).
One other point is that without configuring the security, these devices will connect to someone else's house down the street (and presumably vice versa) so make sure you configure them properly if you care at all about someone freely snooping on your home network.
Facebook
Twitter