WiFi Connection Quality

snipped-for-privacy@coop.radagast.org (Dave Platt) wrote in news:2gni1f-jn2.ln1 @coop.radagast.org:

Yer jus' spittin' out words to see where they splatter...

Horseshit presumption. The WRT guys are as good as it gets.

OMG they're everywhere... NOT.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

This should not be true. Care to name some?

Rob wrote in news: snipped-for-privacy@xs9.xsall.nl:

That "tweaking" does not require an expert. A mere installation of WRT- DD OS will do. It turns it into the pro level settings registry for that board, and max power can be tweaked as well as several other operational elements. No expert required.

Show us your actually measured transfer speed!

Why should it not be true? Are you suggesting that the only devices operating on the 2.4GHz band are 802.11 and Wi-Fi based?

Spread Spectrum is not completely jam free. When the Wi-Fi direct sequence spread spectrum (DSSS) signal meets a frequency hopping spread spectrum signal (FHSS) with even a partial overlap of occupied spectrum, the frequency hopper wins, and the direct sequence link drops out.

The article you cut trimmed includes a short list of interference sources that I've seen or found. Since you apparently missed them, I'll repeat them here for your benefit.

If I dig through my ancient service records, I could probably find some more.

Also: Checklist of possible interference sources. [ * Non Wi-Fi or non 802.11 based ] [ - Wi-Fi or 802.11 based ]

• Microwave oven

• 2.4 Ghz video or security camera link (X10).

- Municipal wireless network.

• Bluetooth devices (mouse, phone, PDA, headset, cell phone, etc.)

• Portable wireless TV camera used at sports events.

• Spread spectrum Frequency hopping cordless phones (e.g., Panasonic Gigarange) 2.4 GHz phones are not an issue in the UK(Eur), as DECT phones operate from 1.88 GHz to 1.9 GHz.

- Wi-Fi (IEEE 802.11b/g) wireless keyboard, PDA, and/or cell phone.

• 2.4 GHz wireless gamepad and/or controller.

• RF Excited Lighting (e.g., Fusion Lighting). See Part 18 RF Lighting, A Potential "Extinction Level Event" For Communications Users Of The 2.4 GHz Band.

- 2.4 GHz wireless baby monitor.

• 2.4 GHz ham radio operation.

• WISP (Wireless Internet Service Providers) which may be using non-Wi-Fi type of modulation (e.g., WiMAX).

• Breezecom/Alvarion/Symbol/Raylink spread spectrum frequency hopping networks.

• Western Multiplex or Proxim non-Wi-Fi wireless links (e.g., Lynx).

• HomeRF spread spectrum frequency hopping network.

• ZigBee IEEE 802.15.4 sensor wireless network.

• Microwave fruit drying oven, plastic mold preheater.

• Unstable high power Wi-Fi power amplifier spewing RF "garbage".

• 2.4GHz radio control

- 802.11 Wireless security camera and video

- Media player (Roku, Apple, WD, etc).

- DVR, Hifi, and TV internet backhaul via Hi-Fi.

- Illegal power amplifiers.

"Potential sources of Wi-Fi and Bluetooth interference"

"7 WiFI Killers that may surprise you"

More links:

"Make a WiFi Jammer With ESP8266 - Hack Any WiFi Without Password!"

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

snipped-for-privacy@gmail.com prodded the keyboard with:

Some time ago, a few months, I had exactly this problem. It turned out to be two bad capacitors inside the router, on the output of the switching power supply. The five volt rail was more spikes than a cactus, and periodically floated up and down by nearly a volt when looked at on a scope. Apart from the barely visible bung starting to push out, those two caps got very hot. Two new 220uf 16v caps cured the problem.

--
Best Regards: 
                      Baron.

ld. Any suggestions as to model?

f good reviews. I believe this unit is long out of

try opening it up to see if I the PSU is working ok.

LOL! No, I am not in the least worried about "drive by" attacks. To get w ithin range they would have to come a quarter mile up my driveway or my nei ghbor's driveway perhaps after getting past his locked gate. Yeah, complet ely possible, but also completely unlikely. If this were to happen the mos t likely culprit would be the NSA or other US gov agency.... so quite unlik ely indeed.

Ok, thanks

Rick C.

If you keep ignoring my description of the environment I might just.

THERE'S NEARLY NOTHING HERE!!! My home has no microwave running (the one I have has a twist dial mechanical timer and hasn't been used in weeks) and I have already indicated all the possible interferers in the house. I've e xplained how only one other house is close enough to cause any problems and they are typically not here and have no routers or cameras or other gear t hat might cause interference.

Where would those signals come from? One of the houses that are at least a tenth of a mile away through dense woods?

Ok, forget I said Ted Kazinski. Think Gilligan's Island! No phone not lig hts no motorcar... well, I have lights. The car is off and has nothing in it that will pollute other than the water. If you think a neighbor from mo re than 500 feet away can interfere with my wifi connection when I can't ev en begin to connect to their signal, then I will listen. But you need to r ead this and tell me their weak signal is the problem.

Do CFLs or LED lights pollute the 2.4 GHz band? Even then, none have been added unless one has gone bad. I have a car radio from the 70's that I lis ten to. I don't even have an Raspberry Pi fired up these days. It the sin gle PC and my flip cell phone... that's it!

I found this router at Microcenter and on Amazon.

_Router

It has 1 Gb Ethernet as well as other useful (potentially) features. At $4

4 the price is right. You seem to like ASUS a lot. If there is a differen ce I'll spend the extra $4 and get the ASUS Certified RT-AC1200G assuming t hat is one you are recommending.

Rick C.

.com:

ect directly to the WISP modem and get a good connection. Connecting to th e router by cable seems to get a good connection. By "good" I mean the pin g times to 8.8.8.8 are consistently in the 20's or 30's of ms. When on the wifi connection lately the ping times are up to 500 ms and beyond 20 to 30 % of the tests.

ion with much around. The nearest neighbor is next door ~100 feet but no o ne is there (weekend home) and they don't have a router. They use their ce ll phones for Internet access. The next closest neighbor is 300 to 400 fee t away and has a typical router like I do. Everyone else is over 500 feet away and I can only see their wifi show up sporadically at levels of 1 or 2 bars if any.

PC is only 8 feet away so the signal strength is full scale. The laptop i s a Dell Precision M6800 which seems to work ok at other locations.

eb sites I get a noticeable lag compared to the direct connection.

)

I'm not unwilling to try things people intelligently suggest. I have alrea dy said I am going to get a dual band unit that will support 5 GHz to get a round interference at 2.4 GHz. I

I have already said I tried that. I went from 11 to 1.

I don't have cables that long. No one is home so I would have to break in. They don't have Internet through anything other than their cell phones, s o I'd have to wait for them to come down and then steal their phones. When Abbie Hoffman wrote "Steal This Book", I did. But I'm drawing the line th ere.

As I've already posted, I have tried what is possible to try.

Or perhaps you could read what I post.

Thanks anyway, I appreciate the effort.

Rick C.

It is the same router I was using a couple of weeks ago before the current slowdown. Are you suggesting the router has developed clogged arteries?

Rick C.

My understanding is that the recent problem with the problematic wifi proto col is fundamental to the protocol, not any specific implementation. So WR T would not be in any better position than anyone else I think. It's just not an issue for me here where only one neighbor could possibly pick off my signal without using a parabolic dish.

How far do you think the range could be extended with a top end antenna on just one end of a typical router link?

Rick C.

Rick C.

That could well be. Look at the hint about dried capacitors. However, in general slowdown is always relative to the currently accepted normal situation. In 1995 you could surf the internet using a 33k6 modem with good results. Today that is no longer possible. The modem has not slowed down, but the expectations have gone up.

Do the LED lights have to be on? If so, not my problem.

alarm clocks,

Ok, I have one. If I let it run down and don't wind it, will it still be a problem? Ok, just kidding. So an LED alarm clock can interfere with 2.4 GHz Wifi? That's pretty amazing. I'll try it sometime with the clock unpl ugged and the backup battery out.

and switching power supplies

In the laptop? Nothing else is running. NOTHING ELSE IS RUNNING! My 1970 's car radio is being powered by a linear regulator hanging off a transform er.

No landline, corded or cordless.

Does it have to be running? Mine is completely off.

None of my neighbors are around. As I've said many time, only one is withi n 100 feet and the others are at least 300 feet away, likely more like 500 feet. If I can't even see their routers on my laptop, I don't think it can be causing interference.

I returned my drone long ago. I was only $80 and it sucked.

I am on WISP. But they are running 900 MHz and it is always up when I am u sing the router. The problem is just the last couple of weeks.

Huh? I don't see any of those in my kitchen parts drawer.

I repeat... Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Huh? Ain't got none.

Am I correct in assuming all interferers would have to be electronic in nat ure? Ain't got none. I do have a 30 year old ceiling fan running.

Rick C.

Ok, thanks for the info. I'm going to buy a new router anyway. But once I do I will be free to rip open the old one and see what might be wrong. He ck, we are out in the boonies where power surges are more likely. It could easily be a power problem even if it's just in the wall wart.

Rick C.

I'm not sure why you posted. Yes, I am getting a new router to see if the old one is defective. But otherwise over a two week period I don't think expectations of the router are the issue rather than an objective measurement of the router performance.

Rick C.

The KRACK vulnerability (which is the one I was referring to) is fundamental to the WPA2 security protocol - or, rather, to the way that the protocol is actually implemented. As I understand it, the protocol itself has been proven to be secure... but the protocol definition doesn't explicitly cover certain corner cases as to how the keys are used, which can be exploited if the implementations don't take extra precautions.

The specific vulnerability in KRACK is that if an attacker monitors a WPA2 handshake, and then "re-plays" some of the packets and sends them to the WPA2 client, it can cause the client to re-install the same key that was exchanged during the handshake... and, as a side effect, reset the encryption "initialization vector" to its initial value as well. Once this happens, it's almost very likely that the client will transmit new packets using the same key and IV that it used earlier... and this is a no-no, because it allows the underlying encryption protocol to be attacked, and (potentially) allows packets to be decrypted by the attacker. Other aspects of WPA2 are also vulnerable to this handshake-replay attack.

It's even worse if the client uses an older session-encryption protocol (e.g. TKIP) since the attack will then allow the attacker to forge and inject packets, not just decrypt them.

That's the bad news - the vulnerability is in the protocol.

The good news is this: there *are* effective mitigations for it. The obvious one is for the client to detect the fact that it's being asked to reinstall the same key it's already using, and refuse to do so. This "hardening" of the protocol is what has been rolled out in (e.g.) Linux, Windows, OpenBSD, and the recent security-firmware updates for many routers and APs.

This is the concern I have with the WRT54GL. It's an older router, with limited flash and RAM space, and it's in the class of devices that the OpenWRT people say have "limited functionality" - they're too small to take many of the more recent firmware updates.

Looking at the OpenWRT web site, it appears that the most recent "officially supported" release for the WRT54GL is a 10.03.1 "Backfire" release, which dates back to December 2011... long before KRACK was discovered. Hence, the "official" OpenWRT build for that hardware is almost certainly at (some) risk. Whether that's enough risk for you to be concerned, in your situation, is your call.

It's possible that one of the newer LEDE-based OpenWRT builds for the WRT54G might work on the WRT54GL, at least in the "stripped down, no GUI" mode. I did succeed in installing a LEDE 17 release on one router of roughly this vintage and capacity, but there simply wasn't enough flash available to include the Luci graphical interface... the only way I could configure it was to log in via ssh and use the uci command-line configuration tools. It was functional, but maintaining it was a hassle... and when I got the chance to grab a couple of the current-generation Asus routers on sale (with a rebate, no less) I jumped at it, and retired my older gear.

I have a great deal of respect for the OpenWRT/LEDE/DD-WRT/Tomato communities... lots of very bright developers out there. However, as volunteers, they can't be expected to provide usable turn-key firmware solutions for every bit of hardware in the universe... and at some point, as the code size grows and the older low-capacity devices start dying of old age (e.g. capacitor dry-out) there just may not be anyone sufficiently interested to gin up working firmware builds for the old devices.

If nothing else, this issue has been educational.

Rick C.

Yep. If the router supports alternative open source firmware, I can usually fix things without waiting for new firmware from the vendor. I also try to recommend hardware that I've worked with or own. I switched from Linksys to Asus after Belkin bought Linksys. The Asus hardware seems better, but more expensive.

The Tenda AC9 AC1200 look ok to me. I've never used any Tenda hardware so I can't tell how well it works. It's been sold since early 2016 so you'll find reviews everywhere. Skimming a few, it seems to be ok.

Interestingly, it has a "WISP Mode" which might be of some use if you have to login to your WISP's router:

That's Asus "recertified" which seems to be the same as "refurbished". That sometimes means nothing more than dusting off customer returns and stuffing them in a brown box. I've done well with refurbished equipment, but every once in a while, get a lemon.

Watch out for the Asus RT-AC1200 (without the G). I bought one and discovered that it has a MediaTek chipset, which I don't like. I was looking for Broadcom but screwed up by not checkingf first. The "G", "G+", "GPlus", and "GP" use a Broadcom BCM47189 chip: while the non "G" uses a MediaTek MT7628AN Avoid the non "G".

Also, none of these are supported by AsusWRT Merlin firmware, which will probably not be a problem for you (but would for me):

That's the "G" suffix, so it's probably acceptable. Your decision on the $4. Basically, you're not trying to buy a feature infested router, or squeeze the last bit of performance out of it. You're trying to determine if the WRT54GL router is the cause of your ummm... erratic performance. So, basically ANY decent router will suffice as a replacement.

Good luck.

Drivel: After lunch today, I was volunteered to accompany a friend to one of his nearby customers that was having an "interference" problem. After a bit of thrashing around, and some hardware substitutions, I found a miswired CAT5 cable between the cable modem and the wireless router. During this exercise, my cell phone rang with a rather complicated Linux update problem from a very knowledgable users. The problem was that he hadn't rebooted his machine in at least 2 weeks. I must admit that I didn't figure it out, my friend did.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Ted: Mine: Mine is bigger. I think the redwood forest is denser than what might be Ted's pine forest.

It's difficult to obtain decent MAC layer 2 Wi-Fi packet loss statistics because most drivers don't allow direct access to those numbers. You can get TCP/IP Layer 3 numbers, but those tend to hide wireless retransmissions and retries. The only evidence of those are extended ping times. So, ping something continuously through your Wi-Fi link and let it run for a while. Use something that displays sequence numbers so you can see if anything disappears. For Windoze, hrping works well: If you see a stable string of ping times, then I'll agree with your claim that you're not getting any interference. However, if you see occasional bursts of much longer ping times, usually clustered together, then you're picking up interference from somewhere.

Another good test is iPerf3. I did a nice writeup on how to use it, but it's a home right now. If you're running speeds under about

30Mbits/sec, Iperf2 or Jperf will work, producing graphs like these: If the graph is a straight line, with little variation in speed between data points, then you don't have any interference. However, if it look like the 2nd URL, you have interference, which in this case was caused by a store and forward "range extender".

Thank you. I was going to mention that but you beat me to it. Yeah, it's a big problem, especially when I sometimes feel like instead of being asked to explain my advice, I'm getting cross examined instead.

Yep. Worse, it is possible to actually produce a likely solution, but the customer will dance around the problem, trying everything possible EXCEPT the easiest and most likely solution. Fear of success is a very real problem.

I think he already tried that. With 11ea 5MHz channels on 2.4GHz in the US, and a Wi-Fi occupied bandwidth of about 22Mhz, there are only

3 non-interfering channels available 1, 6, 11. However, a feature of most wireless routers is to have the router select the "best" channel. Of course, it can select channels other than 1, 6, and 11, which now produces interference on two of the non-overlapping channels.

That which is most obviously correct, beyond any need of checking, is usually problem. My money is on a sick WRT54GL. (Incidentally, I donated about a dozen WRT54G routers to a group that builds mesh networks. They eventually gave up and are buying Ubiquiti hardware).

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558