magicJack and MagicJack Plus Support, Reviews, FAQs and Hacks

corchard · MagicJack Newbie Joined: 26 Jan 2008 Posts: 2

I'm new to this whole technology, but I am technical.

I live in Canada, and am getting .5-1 second delays when speaking. There is no echo, so you end up talking over each other.

I am on a high bandwidth cable (Shaw Internet) but have several other systems and applications that sporadically use the pipe and wonder if a bandwidth manager will help. Moreover if anyone knows of a suitable manager if at all for use with the MagicJack.

your thoughts are appreciated.

Chris

Google · AdSense

HolmanGT · Posted: Tue Jan 29, 2008 2:19 am Post subject:

corchard,

I would think a manager would NOT help much and the delay is most likely not on your end but some server you are going thru. Do you have this problem all the time?

The MJ uses such a small amount of bandwidth ... I think about 80 kHz which is like nothing. The only way in my humble opinion that it could be on your end if whatever it is that you say kicks on from time to time is a real bandwidth hog.

Frankly I have done things like burn CDs, play music and surf the WEB with no ill effects. I have also used Remote assistance to my Daughter and MJ to provide Voice communications while we trouble-shot whatever she has recently broken. Now that has to be a heavy load on my system which only has 512 up and have never experienced any noticeable delay.

I don't know what the problem could be because even a slow server is not going to be in your path all the time, unless your ISP has chosen to intentionally do this like some of them actually block VOIP for whatever reasons known only to them. You could ask them... who knows maybe they will tell you the truth. Confused

_________________
- George -

HolmanGT - St. George, UT MJ-Area/Prefix 435-275
ooma-Area/Prefix 435-579
Baja Broadband, up-1mb dn-10mb, on days with a good tail wind.
MJ on HP T5730 2GBF/2GBR Thin Client XPe SP2 Router Dlink Dir-655

SpamBox

testing123 · Dan Should Pay Me Joined: 13 Nov 2007 Posts: 703

SpamBox · Posted: Tue Jan 29, 2008 8:05 pm Post subject:

WOW great post testing123!

Make this a STICKY!!!

testing123 · Dan Should Pay Me Joined: 13 Nov 2007 Posts: 703

mufon · Posted: Wed Feb 06, 2008 7:29 pm Post subject:

Don't forget to set max mtu in Windows, or you will still frag packets. It's usually more effective here since it's an endpoint. Windows supposedly has a detect path mtu feature implemented in the ip stack, and has to be set in the registry. Even so it doesn't work (remember we're talking about Windows) , so set the mtu manually in Windows after you determine the correct value, and should be the same or less than your router's mtu. Ping from your router if you can, then do the same from windows to get the correct values for both. Usually they (mtu) will be the same for both but not always.
_________________
"Looking for a new job. I have worked for Adelphia, Enron, Health South, Worldcom, and most recently British Petroleum."

testing123 · Dan Should Pay Me Joined: 13 Nov 2007 Posts: 703

LikeMagic · Posted: Wed Feb 06, 2008 11:35 pm Post subject:

Routers and PCs which must "talk" TCP/IP must have their "own" MTU value defined.

The router could have defined its own value (such as 1492 or 1500) because it must "talk" to the ISP with optimum MTU value to minimize overhead thus more efficient data transfer. Most DSL/PPPoE ISPs prefer MTU=1492 and most cable ISPs prefer MTU=1500.

You should also define the optimum MTU value for each of your PC's network connections (built-in LAN connection, built-in WLAN connection, PC card WLAN connection, etc) to match the MTU setting of your router to get the best optimized throughput. Windows' TCP/IP process will then use the network connection's defined MTU setting to "fill up" each TCP/IP packet for transmission. If you don't define the MTU value for each network connection that your PC has, then the OS may assign a lower default value such as 576 which may not be efficient for TCP/IP transmission.

Basically, if you define the optimum MTU value for your router, then the router will have optimized transmission to/from the ISP. If you define the optimum MTU value for each of your PC's network connection, then the PC will have optimized transmission to/from the router. Make sense? Yes? Laughing

testing123 · Dan Should Pay Me Joined: 13 Nov 2007 Posts: 703

Gerry_52 · Posted: Thu Feb 07, 2008 11:36 am Post subject:

Using logic...which according to Mr. Spock is essential...if you change the MTU to say 200 from what is possibly now in the area of 1500 to reduce stutter, wouldn't you now need to send roughly 7 times more packets to accomplish the same thing. Over time, epecially if your upload speeds are marginal couldn't this increase lag times? Since I use a Sprint wireless card for my internet I at times do have problems with stutter. I average about 950kbps down and 225kbps up, however during peak times the upload speeds can drop to 75kbps, which is at best marginal for voice communications. The MJ on a laptop with a wireless card is an elegant solution when it works and for a RF engineer is just geeky enough. I am working on educating myself on terrestrial based landline communications, but the more I read the more I realize I have a lot to learn. I am thoroughly enjoying learning however, and the communications periodicals I digest every month are telling me that the money is going to be in engineering, installing and servicing commercial VOIP services. Anyways, I'll help where I can, and the rest of the time I'll listen and learn from you guys.
_________________
Gerry
kc0cat

Windows XP Home and Pro

testing123 · Dan Should Pay Me Joined: 13 Nov 2007 Posts: 703

Gerry_52 · Posted: Thu Feb 07, 2008 12:11 pm Post subject:

I'm just wondering how much effect setting the MTU on your machine is really going to matter in the big picture. Won't the packet size be dynamically changed several times downstream before it reaches it's destination anyways, or is preferable if not important to set the packet size prior to introduction to the internet to prevent excessive fragmentation once the packet hits the IP layer?
_________________
Gerry
kc0cat

Windows XP Home and Pro

HolmanGT · Posted: Thu Feb 07, 2008 2:27 pm Post subject:

testing123,

And with the hypothetical 200 for an MTU you might actually be Transmitting more packet overhead than data. I don't remember the exact figures but I read on DSLreports that IP packets at 1472 - 1500 are about 12% overhead (none data). At an MTU of 200 the data to packet overhead would be insane.
_________________
- George -

HolmanGT - St. George, UT MJ-Area/Prefix 435-275
ooma-Area/Prefix 435-579
Baja Broadband, up-1mb dn-10mb, on days with a good tail wind.
MJ on HP T5730 2GBF/2GBR Thin Client XPe SP2 Router Dlink Dir-655

Gerry_52 · Posted: Thu Feb 07, 2008 3:35 pm Post subject:

At that point packet loss becomes catastrophic because I would imagine a fair portion of that information keeps the call in session. The other problem, if you want to call it that is I am running a Kyocera KR1 router. It's primary feature is that it accepts the Sprint PCMCIA wireless card and gives me both WAN and WLAN (G)...however it is not a very feature rich router as settings go. There is no way for instance to set QoS, but still...it is just way cool for what it does do. I however have not found much of a difference in QoS between running the MJ plugged into my desktop which runs on a home network thru the Kyocera with the internet accessed via the Sprint card and the Sprint card plugged directly into my laptop via PCMCIA and the MJ plugged into a USB port on the laptop. The results of side by speed measurements are also very similar taking into consideration the constantly changing RF environment wireless internet lives within.
_________________
Gerry
kc0cat

Windows XP Home and Pro