Buffalo "High Power" router not so high powered in UK

[TheLastMan]

Spurred on partly by the review on this site I recently bought a Buffalo WZR-HP-G300NH2 wireless Router and Access Point - which I use as an Access Point.

The wireless strength from this router is disappointing given that it is called "High Power". The wireless signal to my 802.11g devices (three Logitech Squeezeboxes) is reported by the Logitech Media Server software at between 50% - 70% which is no higher than with the WiFi on my Netgear DG834GT router. That lacks 802.11n so at least my two laptops and one wireless desktop PC (all of which have 802.11n compatible wireless NICs) are seeing much faster connections.

The difference between my router and the one reviewed on this site turns out to be because the UK has much stricter emissions rules for routers than the US.

Over here, for 802.11n (40 MHz) we are limited to a maximum output of 200mW (20dBm) compared to in the USA where the maximum is 1000mW (30dBm).

I wonder if flashing with the US version of the User Friendly firmware will help? I would worry about bricking it though.

Looking at other reviews, this router / access point is not much higher in power than pretty well any other on sale. Therefore the use of "High Power" in the description is seriously misleading, at least in the UK.

 

[stevech]

The difference between my router and the one reviewed on this site turns out to be because the UK has much stricter emissions rules for routers than the US.

Over here, for 802.11n (40 MHz) we are limited to a maximum output of 200mW (20dBm) compared to

Reality: most all WiFi consumer products in the $200 and less range operate with 30-60mW in the OFDM modes of 11g/n, 20MHz. Some get to 100mW or so in the lowest modulation modes (non-OFDM), e.g., 6Mbps burst rate.

I don't think you'll see any difference in real products' transmitter power between UK/US.

DD-WRT and other private firmware have a user interface to let the user "ask" for higher power than stock.. but the products' low level firmware won't do so, if the requested power exceeds what can be achieved for the selected bit rate for the WiFI Alliance's required "Rho" (quality). Higher the bit rate, lower the (RMS) power, with OFDM. That's just the way it works with OFDM. To get to 200mW or more in high burst rates of OFDM adds too much cost to the power amplifier. Gee, the best power/range is 802.11b where there's no OFDM and the burst rates are low.

Deeper dive: In OFDM the peak-to-average ratio of the TX power of the modulated signal is large. In non-OFDM modes (slow burst rates), the ratio is far less. To accomodate the ratio, all hardware reduces the RMS power in the higher rates. If this "OFDM Backoff" is not done, the transmitted waveform is too distorted and causes bit errors at the receiver, and this leads to lower throughput due to retransmissions.

And last, the recurring reminder: In WiFi, it's a two-way medium of equal speeds: A super-powerful transmitter on the access point/router doesn't help throughput or range if the client device's power is far less. The rock-band's amplifiers are the analogy; can the band hear the guy in the audience?

 

[TheLastMan]

And last, the recurring reminder: In WiFi, it's a two-way medium of equal speeds: A super-powerful transmitter on the access point/router doesn't help throughput or range if the client device's power is far less. The rock-band's amplifiers are the analogy; can the band hear the guy in the audience?

That is not my experience. A friend bought one of these rather illegal devices:
http://www.wifi-antennas.co.uk/budget-500mw-wifi-booster.html
to go on his Buffalo "G" class, single antenna router. It made a massive and real world difference to his download speeds - particularly streaming.

Like me he has a Squeezebox Duet. This has a wireless client called the Receiver with rather poor internal antennae and reception. The Duet has a network testing app that shows real-world throughput - it streams data at a selected rate and graphs / logs the performance. Without the wireless booster it was showing 40% signal strength and streaming at barely 500 Kbps, with the booster it showed 90% signal strength and was consistently streaming at over 2,500 Kbps. FLAC music files that were unplayable prior to the booster were streaming effortlessly afterwards.

I was so impressed I nearly did the same for my router, but did not really need it and it would probably have been a major source of interference to our neighbours. He lives out of town so has no near neighbours to disrupt! Unfortunately it does not work with "N" class devices, only "G" and "B" at 2.4 GHz, and anyway it needs a detachable antenna.

I am slightly puzzled by the idea that a powerful AP will somehow interfere with its less powerful client. Surely the transmit and recieve information is separated in time? In other words the AP will not be transmitting data at the same time as the client is transmitting it. As you say, the AP will transmit a packet and then wait for a return confirmation before transmitting the next packet. In which case no interference will take place.

Using your analogy the WiFi "rock band" is not playing constantly, it stops playing in order to listen for the guy in the audience (about 20,000 times a second!).

I can certainly see how it might interfere with other access points though.

 

[stevech]

good discussion... "Using your analogy the WiFi "rock band" is not playing constantly, it stops playing in order to listen for the guy in the audience (about 20,000 times a second!)."

My point is that a high powered transmitter (the rock band), vastly out-talks the unamplified guy in the audience. Of course, in 802.11 as we're talking about, the MAC protocol is half-duplex (one frequency, transmit then receive). Even when receiving as best as is (economically) possible, if the difference in "his" transmitted power (antenna gains included) is far less than yours, it's a no-go.

 

[TheLastMan]

good discussion... "Using your analogy the WiFi "rock band" is not playing constantly, it stops playing in order to listen for the guy in the audience (about 20,000 times a second!)."

My point is that a high powered transmitter (the rock band), vastly out-talks the unamplified guy in the audience. Of course, in 802.11 as we're talking about, the MAC protocol is half-duplex (one frequency, transmit then receive). Even when receiving as best as is (economically) possible, if the difference in "his" transmitted power (antenna gains included) is far less than yours, it's a no-go.

Again, I have to politely disagree. If the WAP is not transmitting while it is listening for the return signal from the Client then its transmit power at that point is irrelevant - because it is not transmitting!

The key point is the antenna / receiver sensitivity. To extend your analogy, it is not how loud the rock band plays but how good the players hearing is when it is not playing that matters. The problem with your analogy is people find it difficult to imagine a rock band that can play loudly for a 1/20th of a millisecond and then be totally silent for another 1/20th of a millisecond before playing loudly again the next 1/20th millisecond! I have no idea how long these packet bursts actually last - this is just a thought experiment.

Provided the system is truly half-duplex (i.e. it is never transmitting when it is receiving) then the output power of the WAP can, and demonstrably does, have an effect on throughput.

Your comment about receiver sensitivity is valid. However you have to take care with very long, highly sensitive antennae as they are very directional perpendicular to the axis of the antenna. In other words, if the antenna is vertical then it will create a flat "disk" shaped transmission / receive pattern around itself so the client will need to be within that plane to benefit from the increased sensitivity.

In contrast stubby 3db omni-directional antennae have a "doughnut" shaped transmission pattern which is capable of covering a much bigger area above and below the transmitter.

Unless you live in a very large single storey house or apartment a single tall 9db antenna is not terribly useful. Most people with wireless routers or WAPs live in multi-storey houses sometimes on 4 levels (basement, ground, upper floor and loft) but not much more than 30 feet or so wide/deep. The WAP therefore has to cover a rough cube shape. That is why most consumer grade WAPs and routers have short antennae and can even work effectively with internal antennae printed on the circuit board.

So I draw the conclusion from all this that the ideal WAP or wireless router for the average domestic house is 802.11g/n with a high power transmitter and short stubby 2-3db antennae. Which explains the long-standing popularity of Buffalo's "HP" range of routers in the USA such as this one where Craig Ellison concludes that:

" With peak power almost ten times that of the D-Link DIR655 (A4), you would expect better coverage. And the good news is that the WZR-HP actually does provide significantly higher throughput at weak signal locations where other draft 11n routers only squeeze out a Mbps or so. And although we didn't test for increased range, we're pretty confident that it will deliver."

A pity us in the over-regulated, paranoid, tin hat wearing EU have to do without