VoIP Box Review & Setup Guide (SPA-2102)

[ElectronGuru]

Unfortunately my isp seems to have set a password. I spent about 15min on hold tonight trying to get through to their helpline before giving up. None of the default codes I googled seemed to work.

Egad, it sounds like your 2102 is locked! There are guides for unlocking such boxes but they come filled with disclaimers about the lock coming back if you do a, b, or c. If you want to anything more than what the phone service provider says you can, I suggest returning it and getting an unlocked model: SPA-2102-NA.

 

[js]

. . .

Everything I've read focused exclusively on codecs ending in letters and most describe 729a as the best sounding but the most bandwidth intensive. My tests showed this as well. Let us know what you find out. Keep in mind, if Call Centric doesn't support a codec, it will be a short test.

. . .

I think you mean "most describe 711u as the best sounding but the most bandwidth intensive" right?

Have you messed with 711u vs. 711a? There is a subtle difference between the two, apparently, although I doubt it's important enough to worry about too much.

 

[js]

There is contradictory information on the "true meaning" of Maximum Uplink Speed. My best understanding is that is tells the 2102 the total possible upload of both voice and data that is available. I don't see why they wouldn't design it to exceed this number, so think of it as Expected Total Upload Speed.

It won't exceed the uplink maximum total. So if you set this too low, you will be unduly limiting your uplink speed. If you set it too high, however, your QoS won't really work. The tutorial I linked to spells this out nicely, but it's for a WRT54GL running some open-source firmware--not Tomato and not DD-WRT--so it's entirely possible that the SPA2102 works differently. But everything I've seen so far makes me suspect that it works the same.

We shall see.

 

[js]

Oh, and I also read that the SPA2102 FXS ports are "hot" as shipped/configured, which can cause echo problems and clipping problems, and that you should lower the gains. The guy I read said that he was happy with -6 and -6 for in / out. I'm currently using -5 / -5, but I was using -6 /-6 before that. I may go back to that, I think. Better to be a little too low than risk clipping.

 

[ElectronGuru]

I think you mean "most describe 711u as the best sounding but the most bandwidth intensive" right?

Have you messed with 711u vs. 711a? There is a subtle difference between the two, apparently, although I doubt it's important enough to worry about too much.

Oops, yes, U. Here's a quick table

ITU G.711 - 64 Kbps, sample-based Also known as alaw/ulaw
ITU G.722 - 48/56/64 Kbps ADPCM 7Khz audio bandwidth
ITU G.722.1 - 24/32 Kbps 7Khz audio bandwidth (based on Polycom's SIREN codec)
ITU G.722.1C - 32 Kbps, a Polycom extension, 14Khz audio bandwidth
ITU G.722.2 - 6.6Kbps to 23.85Kbps. Also known as AMR-WB. CELP 7Khz audio bandwidth
ITU G.723.1 - 5.3/6.3 Kbps, 30ms frame size
ITU G.726 - 16/24/32/40 Kbps
ITU G.728 - 16 Kbps
ITU G.729 - 8 Kbps, 10ms frame size​

and a page showing pros and cons.

http://ozvoip.com/codecs.php

It won't exceed the uplink maximum total. So if you set this too low, you will be unduly limiting your uplink speed. If you set it too high, however, your QoS won't really work. The tutorial I linked to spells this out nicely, but it's for a WRT54GL running some open-source firmware--not Tomato and not DD-WRT--so it's entirely possible that the SPA2102 works differently. But everything I've seen so far makes me suspect that it works the same.

[throwing out the assumptions] So the original 80% of possible upload still applies, provided that possible does not exceed about 7000k.

Better to be a little too low than risk clipping.

So your clipping is all gone now? Another way to adjust this would be setting your handset volume to medium, then adjusting the gains until most calls just right. That way you've got each adjustment either way.

Something else to keep in mind with volume: Normal ports are designed to power an entire house worth of wiring and phones. If you're running a single 6ft wire from the box to the wireless base (the ideal config), the 'extra' energy doesn't have a place to go.

 

[ElectronGuru]

Here's a tech document comparing 711u with 711a:

http://www.voipforo.com/en/codec/codecs-g711-alaw.php

G.711 A Law(a-law) and µ Law (u-law) encoding scheme

The two main encoding laws used nowadays are A law (a-law) and µ law (u-law), that are also known as g.711 codec. A Law (a-law) is used mainly in European PCM systems , and the µ law (u-law) is used in American PCM systems.​

 

[js]

So, I'm now having other issues and it's really making me crazy.

My landlord also uses my wireless network, and my old WRT54GL with DD-WRT and the power cranked to 240 mW/m worked just fine to cover his side of the house and out to the barns.

But when I got the AEB, it wouldn't reach as far and he was having issues. So, I got another AEB for him and used the 2nd one to extend the wireless network. Which it did.

So far so good. But since then, I've noticed that the performance of the connection using wireless is CRAZY variable. Like sometimes nearly the full 2 mbps down and 7 or 8k up, and other times like 350kbps down and 20kbps up. Meanwhile, the desktop Mac Pro, which is connected via internet cable to the LAN port of the primary AEB, always gets pretty good performance, even when other wireless clients are on the LAN.

WTF is going on? I've googled to see if people have had trouble doing WDS with two AEB's, and I get zero reports of problems. I doubt that the wireless bridge is the problem, but it could be. I mean, I understand that I cut the throughput of that second AEB in half, but that's still way more capability than my non-stellar internet connection, right? Half of 54mbps is >> than 2 mbps.

I suspect that the problem is with the linksys/sipuria router and DHCP server. It seems like it is dividing up things this way:

ethernet port --> Mac Pro
all wireless clients -->my MBP (sometimes)
_______________ -->2nd AEB-->landlords laptop 1
_________________________ -->landlords laptop 2
_________________________ -->landlords desktop in barn
_________________________ -->my MBP (sometimes)

So, the 2nd AEB gets half service, let's say. Then it divides that half among maybe up to three or four wireless clients. So, if I'm downstairs and connect to the 2nd AEB, which is in my landlords house, but just on the other side of the wall, then I get absolute crap for bandwidth.

I'm going to try reconfiguring things for a test, so that the AEB1 is not bridged to the 2102, but is rather the first in line downstream of the modem. Then I will plug the Mac Pro and 2102 into the AEB1, and will keep the AEB2 wirelessly bridged to the AEB1, to extend the network.

So, I'll need to forward port 80 to the 2102, right? I assume this is straightforward, but I haven't read up on it.

Is this what you do with your TimeCapsule? And if so, do you have a permanent DHCP lease to the 2102, or is it a static IP? Or what?

Sorry to ask for your help again, but this wireless CRAP PERFORMANCE issue is going to drive me effing INSANE. I can't live with it. It was NEVER this way before, and that was with a 1x1 mbps connection and not a 2mbps connection.

I can't believe that Apple would implement such a crap routing scheme, so right now I'm assuming its the 2102. The box does get too hot for my taste, that's for sure. I wonder if the router part of the 2102 is less than great.

 

[ElectronGuru]

There's a lot going on there. First a few points, then we'll get to some options.

I can't believe that Apple would implement such a crap routing scheme, so right now I'm assuming its the 2102.

While it is an Apple box and everything inside the box can be built the way Apple wants to do it, most of its wireless features require interoperability and so are based on wireless standards, including WDS. So while they can work their magic with putting USB harddrives ON the network, the wireless network itself must still conform to the standard.

The box does get too hot for my taste, that's for sure. I wonder if the router part of the 2102 is less than great.

Boxes that just do routing/QoS often cost more than the 2102 does with all the VoIP. The 2102's router is not high performance and has limited memory, with a particular weakness when it comes large numbers of transactions (as opposed to volume of data). As long as its not hot enough to auto-restart, its not *to* hot, but do what you can keep it cool, like mounting it vertically or laying it on a metal plate.


Okay, getting down to business. Your biggest challenge is that two big variables changed at the same time that a problem started, so there is not an obvious culprit. Take the voip box out of the network for an hour and see if the problem goes away.

My landlord also uses my wireless network, and my old WRT54GL with DD-WRT and the power cranked to 240 mW/m worked just fine to cover his side of the house and out to the barns.

AEBs have an option to turn down the power, is it possible its not maxed out? Otherwise the AEB's antenna is not as strong. A definite drawback to the cleaner design, no external antenna. Option #1, add one:

http://www.quickertek.com/products/80211N_upgrade_BS.php

But when I got the AEB, it wouldn't reach as far and he was having issues. So, I got another AEB for him and used the 2nd one to extend the wireless network. Which it did.

So far so good. But since then, I've noticed that the performance of the connection using wireless is CRAZY variable. Like sometimes nearly the full 2 mbps down and 7 or 8k up, and other times like 350kbps down and 20kbps up. Meanwhile, the desktop Mac Pro, which is connected via internet cable to the LAN port of the primary AEB, always gets pretty good performance, even when other wireless clients are on the LAN.

WTF is going on? I've googled to see if people have had trouble doing WDS with two AEB's, and I get zero reports of problems.

Back in the b/g days, I tried to make WDS work and was so :hairpull: that I gave up. N is supposed to have an improved version, but 1) any non-N devices connected to either AEB may require b/g use and 2) N WDS may not be as good as advertised.

So, the 2nd AEB gets half service, let's say. Then it divides that half among maybe up to three or four wireless clients. So, if I'm downstairs and connect to the 2nd AEB, which is in my landlords house, but just on the other side of the wall, then I get absolute crap for bandwidth.

This is my issue with WDS. You've got an antenna in a box. Any time that antenna is doing one thing, it can't also do another. So thats fine if a relay is only doing relay. But when a relay is also providing service for one or more clients, thats a lot to do. Bit 01 comes in from client, bit 02 has to wait for bit 01 to be retransmitted before bit 02 can come in. Then bit 03 has to wait for bit 02 to do its thing. The solution would be dual antennas, which the new AEB has, so maybe...

I'm going to try reconfiguring things for a test, so that the AEB1 is not bridged to the 2102, but is rather the first in line downstream of the modem. Then I will plug the Mac Pro and 2102 into the AEB1, and will keep the AEB2 wirelessly bridged to the AEB1, to extend the network.

So, I'll need to forward port 80 to the 2102, right? I assume this is straightforward, but I haven't read up on it.

Let us know how it goes. Again, I suggest pulling the 2102 off the network for this test. Its a lot faster to reverse and completely removes it as a cause if the issue continues.

If you don't want to access the web interface of the 2102 during the test, all it needs is an internet connection and a phone. If it works when you reconnect, don't bother changing anything unless you 1) need to or 2) want to leave the new configuration long enough to change settings with it.

To answer your question, don't bother with port forwarding. All it should need is WAN access to the web interface.

Is this what you do with your TimeCapsule? And if so, do you have a permanent DHCP lease to the 2102, or is it a static IP? Or what?

With Airport Admin, log into AEB1. Under Internet > DHCP > DHCP Reservations, click +. Type in something like 2102 > Mac Address > Continue. Then just type in the MAC ID of the 2102 and an IP you want it to have.


The best way (#2) to expand the range of a wireless network is to connect AEB2 to AEB1 via an ethernet cable, then configure them with the same network name. Are we talking separate buildings here? If so (#3), is AEB1 against the wall closest to house2?

 

[js]

Great post, ElectronGuru!

Indeed. Hmmm. So, starting at the end, no, the two AEB's are in essentially the same building. I could run an ethernet cable from AEB1 to AEB2. I'd have to drill some holes through walls, and it would be maybe a 100 foot run (or less), but it's the option I'm currently favoring, regardless of whether or not I put the 2102 directly behind the modem, or plugged into a LAN port of AEB1.

As for the 2102, it's mounted vertically, with the vent openings top and bottom for best air flow. It's not too hot. It's never restarted. And, yes, you're right: there is a compounding of probems here that I will have to deconvolve. But, yes, the power levels on the AEB's are at 100 percent.

As for Qos, I honestly don't know if it's worth it. I mean, it only matters for the uplink side of things. You have yours in a configuration where QoS is impossible, and you've been happy.

I remember reading in a review of the 2102 that the router part of it wasn't an all around great router--good at just one or two things, but weak in other areas. Apparently, having five or six DHCP clients at once is confusing it--added together with the WDS thing--and it's just not acceptable.

Ah, well, this is what you go in for when you try to do something like this! I try to keep focusing on the FUN side of it.

Anyway, thank you so much for your help, ElectronGuru.

In terms of the guide--i.e. the first post in this thread--maybe you might want to

1. link to that QoS tutorial to explain how it works and why we're setting the max uplink. I actually like the 80 percent of max reading rule of thumb, at this point, as the QoS is only in effect when the phone is on, and you're better to be conservative than optimistic, so as to ensure QoS will actually work.

2. mention turning on the token bucket filter (TBF) as well as setting the policy to "On when phone in use".

3. Give a glossary of terms, like DTMF, QoS, FXS, POTS, PSTN.

4. 729a sucks for sound quality. I would say that ANYONE who isn't on a dialup connection (which would be stupid here anyway) will want 711u.

5. Add the neato admin web access phone code--what a lifesaver!

 

[ElectronGuru]

I'd have to drill some holes through walls, and it would be maybe a 100 foot run (or less)

Who said anything about drilling?

http://www.netgear.com/Products/PowerlineNetworking/PowerlineEthernetAdapters.aspx

* Power outlets and electrical wiring must in the same electrical system​

As for Qos, I honestly don't know if it's worth it. I mean, it only matters for the uplink side of things. You have yours in a configuration where QoS is impossible, and you've been happy.

The best part about this box is it can go back and forth as needed. Put the AEB1 first for a while and see what happens. You may find that its only an issue during certain online activities - that you can curtail during phone calls. Some torrent software, for example, has scheduled bandwidth limits.

In terms of the guide--i.e. the first post in this thread--maybe you might want to

Thanks, I'll get these added/changed

 

[js]

ElectronGuru,

I don't like the idea of sending information over power lines, especially not the ones in this house! So, I went to a lot of trouble to map out a route, drill holes, and pull cat5e ethernet cable, so that I have my two AEB's connected via an ethernet cable, from LAN port to LAN port, as per the diagram in the manual. They also talk about setting channels for such a roaming network, saying that you want at least 5 channels of separation for two routers that are within 150 feet of each other. They talked about channels 1, 6, and 11. These are for the 2.4 GHz band, right? So I set AEB1 to channel 1 and AEB2 to channel 11. But for the 5 GHz band, there were choices like 36-54, and 149-162 or something along those lines. In automatic mode, previously, AEB1 had chosen 149 for the 5GHz band, so I picked that, and the one farthest from it, but in that same range--i.e. 162 (or whatever the largest one was). Also, of course, I named both wireless setups the same name, same security, same password--creating one big single network from the two base stations.

And everything is now smoking fast, no matter where I am connecting from: wired, wireless, upstairs, downstairs, over at the landlords house--it's all good!

But I'm curious about those 5GHz choices and have done a bit of searching, and don't seem to get a clear answer. Why that strange triple range of choices for 5GHz. There's the old 1-11, then something like 36-54 (or whatever), then 149 through whatever. What's the deal with that?

Also, using wide channels--is that the same as channel bonding? In any case, I turned that feature off, just in case it might possibly be interfering with the 5.8 GHz signal coming in to my rooftop antenna. Maybe I'm being paranoid, but I don't need the extra throughput, as I rarely do computer to computer file transfers. 95+ percent of network use is LAN-WAN or vice versa.

On another note, my VoIP phone service crapped out today big time. Kept dropping calls, breaking up, etc. I'm presuming that it was a problem with my internet connection, but either way, if this continues, I'll have to go back to a POTS landline.

Have you ever had any such problems? Is it possible that the fault was with callcentric (i.e. did you experience any trouble today?)

Finally, I forgot to mention yesterday that it's probably best to say outright in the guide that the very first thing you should do after getting your 2102 connected to the internet is to upgrade the firmware. You can mention that I have the .bin file on my web server space and will gladly reply to any PM with the exact URL so that others can easily upgrade via the upgrade rule field.

 

[ElectronGuru]

I went to a lot of trouble to map out a route, drill holes, and pull cat5e ethernet cable, so that I have my two AEB's connected via an ethernet cable, from LAN port to LAN port, as per the diagram in the manual. They also talk about setting channels for such a roaming network, saying that you want at least 5 channels of separation for two routers that are within 150 feet of each other.

Yea you did. What you've setup is called active roaming or a roaming network. Way better than WDS, its the ideal way to run multiple base stations over a large area. Like a cell phone in a moving car, it allows you to move from base to base (and back) with no loss of signal as you change and with maximum signal strength at every location. Here's a diagram from the pdf below:

They talked about channels 1, 6, and 11. These are for the 2.4 GHz band, right? So I set AEB1 to channel 1 and AEB2 to channel 11. But for the 5 GHz band, there were choices like 36-54, and 149-162 or something along those lines. In automatic mode, previously, AEB1 had chosen 149 for the 5GHz band, so I picked that, and the one farthest from it, but in that same range--i.e. 162 (or whatever the largest one was). Also, of course, I named both wireless setups the same name, same security, same password--creating one big single network from the two base stations.

But I'm curious about those 5GHz choices and have done a bit of searching, and don't seem to get a clear answer. Why that strange triple range of choices for 5GHz. There's the old 1-11, then something like 36-54 (or whatever), then 149 through whatever. What's the deal with that?

Reading several Apple wireless guides, the newer they are, the less they mention channel separation. The latest (below) doesn't show it at all. This feels to me like Apple found a way to remove the requirement for channel separation in their recent designs and rather than saying "hay, look at this cool new fix", they simply pulled the old "fix" from their documentation. The newer stations are even removing channels as an option all together unless you hold down the option key when clicking the menu. Overtime, I expect channels to become less like channels as more like radio frequency identifiers, such that it won't mater what the setting, provided they coordinate.

Also, using wide channels--is that the same as channel bonding? In any case, I turned that feature off, just in case it might possibly be interfering with the 5.8 GHz signal coming in to my rooftop antenna. Maybe I'm being paranoid, but I don't need the extra throughput, as I rarely do computer to computer file transfers. 95+ percent of network use is LAN-WAN or vice versa.

I believe wide channels is not a bonding of 2+ pipes, but is rather a wider pipe. Its not on by default because it reduces range and 5.8 already has less range than 2.4. Yes, it would only apply if you has massive LAN-LAN transfers, such a local file server. I use mine for audio/video.

The cool part for you is since both your AEBs are 2009 models, you basically have 4 antenna's. Assuming they don't interfere with other equipment, any client device can pick - automatically from moment to moment - which offers the best signal.

On another note, my VoIP phone service crapped out today big time. Kept dropping calls, breaking up, etc. I'm presuming that it was a problem with my internet connection, but either way, if this continues, I'll have to go back to a POTS landline. Have you ever had any such problems? Is it possible that the fault was with callcentric (i.e. did you experience any trouble today?)

If I understand correctly, this is with 2102 not upstream, so the QoS no longer functions, yes? If so, this gives you 3 options

1. Leave QoS off (yuck)
2. Put the 2102 back upstream and just let it be warm
3. Get a more powerful router and off load the QoS to a dedicated box

BTW, I placed two calls that day and both were perfect. For the record, I've yet to have a VoIP issue that was not correctible on my end.

And here's some good reference material:

http://db.tidbits.com/article/10125

http://manuals.info.apple.com/en_US/Apple_AirPort_Networks_Early2009.pdf

You can mention that I have the .bin file on my web server space and will gladly reply to any PM with the exact URL so that others can easily upgrade via the upgrade rule field.

I've added a note with links.

 

[js]

EG,

My 2102 is still upstream of the AEB's. I also found the stuff in both the links above. The latest manual (in the link) does actually mention channel separation, just not in the exact same place as how to set up a roaming network, but they only mention the 2.4 GHz channels, 1-11, which are overlapping unless you separate by 5 (i.e. 1,6,11 are non-overlapping).

As for my VoIP quality, it's taken a SERIOUS dive for the worse. Having lots of cut-outs, both ways, and it's dropping calls very frequently. Pisses me off, because it was working perfectly for a week or two before this. Nothing changed between the time it was working good and now. My creation of the roaming network happened after the VoIP badness, and plus, it's all downstream of the 2102, so that can't be it. And I hadn't changed any settings on the 2102. It's continued now for two days.

So, I'm punting. I reinstated my POTS since I could get my same number back and they waived the re-instatement fee. It seems I can't get away from the $50 - $60 monthly charge. Which sucks, but there it is. Marginally worse sound quality I'd be fine with. But dropping calls, cut-outs, and calls that just plain don't ring the 2102--that's another story.

Maybe some day, I'll try again, but not today, and not tomorrow. I've had enough. If anyone wants my 2102, send me a PM, otherwise I'm going to list it on ebay, I guess.

 

[js]

I did some research and found the deal with 5GHz 802.11n channels. It is surprisingly difficult to find ANY information on the subject!

So . . . check out this link for the channels:

5Ghz channels & bonding

Note that if you count up the number of channels you'll find more than 24, but there are only 24 non-overlapping 20MHz channels, due to 34-48 being only 10 MHz apart. If you go with the "use wide channels option" this will reduce you from 24 to 12 non-overlapping channels, which is still loads better than 2.4GHz's 3 20 MHz wide non-overlapping channels, or TWO 40MHz channels (with bonding).

Now, the story doesn't end there! Turns out that the confusing numbering system refers to different power limitations that are imposed on the different bands. Check out this link and in particular on the following from it:

There are now four bands in 5 GHz channelized for 802.11 in the US, although they're numbered somewhat strangely. In brief, there is total of 555 MHz across 23 channels in 802.11a/n. The lower four are indoor only; the higher 19 are indoor/outdoor. The lowest four (5.15 to 5.25 GHz) can have 50 mW of output power, the next four (5.25 GHz to 5.35 GHz), 250 mW; the next 11 (5.47 to 5.725 GHz), 250 mW; and the top four (5.725 to 5.825 GHz) up to 1 W. (There are further restrictions on 5.25 GHz to 5.725 GHz in terms of detecting and avoiding stepping on military radar transmissions, which share those bands. And the 802.11a spec specifies 40 mW/200 mW/800 mW instead of 50, 250, and 1,000, just to make it even more complicated.)

There are an enormous number of details about effective output, antenna gain, and so forth, but most of that affects the use of directional antennas and point-to-multipoint outdoor connections, not the use of interior omnidirectional antennas.

Because 5 GHz signals have shorter wavelengths than 2.4 GHz signals, at the same amount of power, they propagate shorter distances. They're also worse at penetrating solid objects. This means that even if you use the top four channels for 802.11a or single-wide channels for 802.11n in 5 GHz, you will only be able to send data less than half as far if that. There are only two double-wide channels possible in that top band.

In the 250 mW restricted range of 5 GHz, you could achieve the same range by using higher power in 5 GHz than in 2.4 GHz. But many of the devices that offer 2.4 GHz and 5 GHz radios don't compensate in 5 GHz by having higher-powered signal output. Thus a device that gives you 100 interior feet in any direction in 2.4 GHz could span less than 50 feet for this reason in 5 GHz. The lack of interference from competing networks could compensate for the shorter distance, however.

So, in terms of setting the 5GHz channels manually, it looks to me like you're best bet is to chose 149-161 (or 165--although, this seems to be missing from some attributions/hardware). However, note that if you're using 20MHz channels, that all of these are non-overlapping. If you're using wide channels, you'd go with 149, 157, 165. The reason to chose the uppermost ones is that they can go up to 1000 mW/m of output power and have dynamic allocation (or something like that--not sure what that means right now). Plus, on its own, in automatic mode, the AEBS seems to go for 149 or one of those uppermost channels.

Personally, I'm thinking that putting all basestations in automatic mode would probably work out just fine. I may test this theory out later tonight.

 

[js]

What an ordeal this VoIP thing has turned into lately. Of course, my POTS line isn't activated, as they promised, as the technician called me while I was out in the tunnel of the accelerator, and the phone I reached to pick up my call was dead. I asked the phone operator to take a message. His message? "Verizon technician tried to reach you". OK. WTF? THAT is a message? I don't think so. No call back number, no reason why he was trying to reach me. Nada. So, I call customer service and spend forever on hold listening to country music non-stop, then get transfered to the installation CS, then wait some more, still listening to country music, then I get put on hold while the woman checks with the local supervisor. THEN, she tells me that the technician is on his way back to my apartment, but not why he needs to talk to me or what's going on. Presumably, it has to be that he needs to get inside. OK. Fine. I call using the VoIP phone to tell my wife to wait around for him. It works just long enough for me to tell her this before it cuts out and drops the call.

Well, I get home at 5:30PM, and no dial tone, --the guy never showed. I'm pissed. You'd think that a freaking PHONE COMPANY would be able to directly call its installation technicians using something fancy schmancy known as a CELL PHONE. But no. Instead, I talk to a woman in CS god knows where, she talks to a local supervisor in an office not too far away, and he somehow talks to the technician--or no, he probably just issues an order via a text or something. Or who knows. But in any case, the CS woman can't work out an arrangement, an appointment, a specific time. No reasons, no communication.

Insane. Fastest way to get me to see red and become really angry. Crap like this. Ah, I should have known. This is what always happens with service technician calls like this. Not that I understand WHY they need a technician to come to the apartment! But, whatever.

And I still don't know why my VoIP isn't working any more. I've thought it through and checked most or all of the variables and settings, and all I can think at this point is that my SPA2102 is defective, and was working OK for a time, but then just up and died or something.

Right now, it's TERRIBLE. Like, no call will last more than a minute or two before it gets dropped, and there will be massive cutting out even then. It's crazy! What the heck happened? And why did it have to happen two days after I canceled my POTS? It was working just fine for well over a week, if I remember correctly.

My advice to anyone doing this: after you think you've got everything tweaked just right and working perfectly and are very happy, continue to test it out for a full month before canceling your regular phone service.

Oh, well, two good things came out of this. My new Apple Airport Extreme Simultaneous Dual-Band Base Stations, and the ethernet cable that I ran between them to create a roaming network. If it weren't for me trying out this VoIP thing, I would probably still have the WRT54GL, and I like the AEBS's so much better. You win some, you lose some, I guess.

 

[ElectronGuru]

As for my VoIP quality, it's taken a SERIOUS dive for the worse. Having lots of cut-outs, both ways, and it's dropping calls very frequently. Pisses me off, because it was working perfectly for a week or two before this. Nothing changed between the time it was working good and now. My creation of the roaming network happened after the VoIP badness, and plus, it's all downstream of the 2102, so that can't be it. And I hadn't changed any settings on the 2102. It's continued now for two days.

If this were happening to me, I would

1. Unplug everything
2. Connect only the DSL modem and the 2102 and test
3. Retest the DSL connection with the speakeasy and VoIP web sites
4. Connect only a wired phone, and test
5. Reload the firmware, and test
6. Reset the 2102 to defaults, and test
7. Reenter the settings from post 1, and test
8. And if all that yields the same result, throw the 2102 out the nearest window

I did some research and found the deal with 5GHz 802.11n channels. It is surprisingly difficult to find ANY information on the subject! So . . . check out this link for the channels:

5Ghz channels & bonding

Note that if you count up the number of channels you'll find more than 24, but there are only 24 non-overlapping 20MHz channels, due to 34-48 being only 10 MHz apart. If you go with the "use wide channels option" this will reduce you from 24 to 12 non-overlapping channels, which is still loads better than 2.4GHz's 3 20 MHz wide non-overlapping channels, or TWO 40MHz channels (with bonding).

Great info. Its also useful to read that the 5G range is better at this because there is more headroom

And I still don't know why my VoIP isn't working any more. I've thought it through and checked most or all of the variables and settings, and all I can think at this point is that my SPA2102 is defective, and was working OK for a time, but then just up and died or something.

Right now, it's TERRIBLE. Like, no call will last more than a minute or two before it gets dropped, and there will be massive cutting out even then. It's crazy! What the heck happened? And why did it have to happen two days after I canceled my POTS? It was working just fine for well over a week, if I remember correctly.

My advice to anyone doing this: after you think you've got everything tweaked just right and working perfectly and are very happy, continue to test it out for a full month before canceling your regular phone service.

The first rule in trouble shooting is variable control. But there are so many here, the only way to control them is to remove all of them and slowly add them back, one at a time. If you do want to try again, give the list above a try.

 

[js]

If this were happening to me, I would

1. Unplug everything
2. Connect only the DSL modem and the 2102 and test
3. Retest the DSL connection with the speakeasy and VoIP web sites
4. Connect only a wired phone, and test
5. Reload the firmware, and test
6. Reset the 2102 to defaults, and test
7. Reenter the settings from post 1, and test
8. And if all that yields the same result, throw the 2102 out the nearest window

Indeed. I pretty much have done everything except that last step. I didn't reload the firmware, however. That is good idea. But after a certain point, I had to punt, as I said. I absolutely must have a reliable phone, as accelerator operators call me at all hours in certain types of emergencies. I got called Tuesday morning (as in just after midnight Monday) and the connection was so bad, I had to tell her to email me!

I'll probably reload the firmware and reset and re-enter and plug the 2102 into the AEBS, which is now my router and DHCP server. I'll see what sort of results I get from that configuration. My strong suspicion is that it's the 2102 itself that's gone bad. I also thought about testing out VoIP via Skype, which bypasses the 2102, but would test the connection in actual use. It's possible that my connection is stable for between 15 seconds to 3 minutes, and then goes unstable, causing calls to get dropped. If I do the test, which only lasts 5 seconds or maybe less, it's unlikely to bridge one of those instabilities, and so could test good, despite my connection having taken a turn for the worse. In that case, Skype calls would also be bad. But if the 2102 is the culprit, then skype calls would be fine.

But, honestly, I've reached my limit. Out of curiosity I will eventually trouble shoot this situation, but right now, I've got more important things to do--like build USL's and do survey work at the accelerator!

It was fun, nonetheless, however! And I want to thank you, ElectronGuru, for all of your help and advice.

 

[DarrellBray]

New to the site. I read all that, and then no resolution?!

Way to leave it hanging . I suppose you do, as you pointed out, have more important things to do than continue the 2102 odyssey.

Associate

Busienss Voip Telephone Systems

 

[js]

DarrellBray,

It was the 2102 itself that failed on the hardware side of things. Or at least that is the only possibility that was left that I could think of in my troubleshooting. I could have sent it back for repair or replacement, and maybe I still will, but I decided to just stick with POTS for now. I may revisit the hole VoIP thing at some point in the future, though. Sorry for not updating my story in this thread!

 

[ElectronGuru]

Googling "spa 2102 review" brings this thread up in 4th place, a high ranking. Best if we can clear this up.

If you wanna send it to me, I'll put my config in your box and swap it out for mine. This should tell us what's up.