a follow-up and a tip for preventing slow downs while connected Connection speed Last time, we covered the latest in the debate as to whether or not the internal modem that ships with the PowerBook 3400 really performs as fast as its specs imply. Many readers complain of an inability to get a 28.8 or 33.6 connection with this modem. However, not everyone shares this complaint. After posting this item, I received several emails from readers who said they routinely get connections in this range. For example, Gene Steinberg wrote: "I do get consistent 28,800 bps connections with such services as AOL and as high as 33,600 bps with my ISP." I am still not sure why some succeed and others (including myself!) do not. But it is at least nice to know that the potential is there.
However, even if you get connected at the desired speed, there are at least two other less-documented speed problems that you may confront:
Multiple actions slow down This slow down occurs when you try to do several Internet activities at once. This is described in an article about these modems by Craig Hunter on the PowerBook Source web site (coincidentally also posted this past week-end). He writes: "For example, with the 5300 and GV modem, I can check my e-mail, browse the web, and run a telnet window while downloading a file with Fetch or Netscape. Try the same thing on the 3400, and it doesn't always work - the download can bring everything else to a screeching halt."
Progressive slow down over time (and a possible work-around) The other problem (which is one that has been particularly frustrating to me) is that the overall transfer speed often declines the longer you maintain a connection. For example, after being online for about a half hour, I often find that web pages start to load at a rate far below 1K. Disconnecting and reconnecting fixes this problem - until it happens again. Attilio Comotto reported this same problem. Interestingly, he claimed that if he took the phone off the cradle for a second and put it back again, this too would fix the problem. He claimed that this worked because it caused a "retrain" on the line. As it turns out, this can be controlled by an AT command setting. The factory default is likely "?," which is "Enables line quality monitor and fallback/fallforward." If you add a "?" to your modem init string, this changes the setting to "Enables line quality monitor and auto-retrain." Attilio claims that when he made this change (e.g., AT&F?...) to his modem init string, the problem disappeared. I am currently testing this myself. So far, so good. I am optimistic. I will report back in a few days with more definitive results.
[By the way, if you use FreePPP, changing the modem init string is quite easy (it's located in Accounts:Connection; then make sure that the "Use init string..." option is selected from Modem Setup). For OT/PPP users, the process is a bit more involved (you have to open the modem script in a word processor). If making this change proves beneficial, I will post more detailed instructions. When I do this, I will likely combine these 3400 modem items into a new MacFixIt Report.]