|Forums | Other (Public) | Disconnect On Call Waiting (+PCW=1)||Watch|
|Posted - 3/14/2006 11:42:24 PM |
Hello Richard and Hi everyone.
I have a strange problem regarding "Disconnect on Call Waiting" for which I need to consult the global brains trust.
Up until mid last year I had this feature working fine on a number of modems and computers (using Win XP) here in Australia, so that incoming phone calls immediately disrupted the dial-up connection and rang the phone. This required the appropriate command (+PCW=1) to be placed in modem properties "Extra Settings". Without this command in place, incoming phone calls were generally ignored. Of course I always make sure Call Waiting is switched on via the telephone and the computer configured so as not to disable Call Waiting for this to work; but it certainly did, for over a year, on three different computers at three locations (at that time, each had exactly the same D-Link 562I modem = Intel 536 EP).
Anyway, after realising that incoming phone calls were being ignored on my computer, and those of some people I knew (limited statistics unfortunately, as most people switch call waiting off and redirect calls to their mobile), I've spent months this year trying to get a modem, any modem on anyones computer, to disconnect on Call Waiting, with only slight success (see below). I even bought some new modems, including a US Robotics 3CP 5610, firmware upgraded to V.92, best modem I've ever used, but even it won't disconnect with a call waiting signal. The only modem I could get to work this year was a cheap connexant software modem (FM-56PCI-HSFI, connexant CX11252-11 chipset), which disconnected perfectly on first ring and held on like a limpet otherwise, including when the +PCW=1 string was not present, which proves that "disconnect on incoming call waiting" can be implemented. But this modem has nothing else to recommend it and it stopped functioning recently, so that was the end of that.
All the V.92 capable modems I've tried acknowledge the +PCW=1 command and respond with "OK" during initialisation, as recorded in the Windows XP modemlog, but don't respond to incoming phone calls (as tested by ringing in using a mobile phone). Acc. to their manuals each should work with +PCW=1.
I did, however, find that with all the modems I tested, the incoming call waiting beeps do cause repeated modem retrains, which I can hear because I use M2 to keep the modem speaker on all the time. In fact, at one point, I thought that this was the solution!! Listen for retrains!!
I can barely hear the call waiting beeps under the retrain noise, so it's not surprising the modems have trouble too. But I can hear the repeated retrains, which I found never occur unless there's an incoming call. However after several hours listening to the modem speaker and recovering from the ringing in my ears, I concluded that this is not really a practical solution.
How widespread this problem is I don't know, so I also contacted the Telecom Co. here to find out if they knew of a problem and they said they were actually doing everything to stop the reverse problem, ie CW tones causing modems to disconnect, which seemed rather ignorant, since in Australia all it takes is *44 in front of the dial-up phone number; but anyway, that was their story. So I asked that person (and later two other phone company representatives of progressively increasing rank in the company) if "doing everything" included decreasing the Call Waiting signal level, which those three representatives denied (of course), but later a more senior tech said they had altered the Call Waiting tones a number of times in response to modem disconnection complaints and may well have decreased the level, though I haven't confirmed that this is so as yet. Just a working hypothesis.
I've tried various software call waiting programs (PhoneTray etc), but they all rely on the modem detecting the signal and if that were the case it would just disconnect itself anyway, so the software would be redundant to me.
I also tried a hardware solution, namely an early model analog "Catch-a-Call" unit, which I altered internally (as per the manufacturers explicit instructions, which they emailed me) to work with the Australian 425 Hz call waiting signal, all to no avail. It occasionally responded to incoming calls but mostly not. In all fairness this is a superbly constructed and sophisticated unit that works well in North America with the NA call waiting tones (440 Hz, 0.3 sec); so I really think the problem lies elsewhere, as I've suggested.
So what I need to find out, if anyone is still reading, is:
1: Is there a way to make a modem drop out after two or three successive retrains? Can't find anything in my USR manual or anywhere else about that. Thought you might know.
2: How does a V.92 modem detect a Call Waiting tone? ie where does the frequency discrimination take place? (eg for the 440 Hz call waiting tone used in North America, France, Germany, Ireland, Hong Kong, Puerto Rico, etc).
3: Can the call waiting response frequency in a modem be altered in any way, eg to suit the local frequency and cadence in the country of use? Because Call Waiting tones vary from 334 Hz in Bangladesh to 1400 Hz in Norway. 425 Hz alone is used by many South American countries, Australia, Netherlands, Singapore, Sweden, Spain, Mexico, Saudi Arabia, etc. I can't believe modem manufacturers would not take steps to ensure their modems would be competitive in these countries, but maybe that is the case. Anyone know?
|Posted - 3/16/2006 10:20:27 AM |
As far as I know, any of the V.92 modems based on Agere Systems chipsets should respond properly to +PCW=1: disconnect when a call waiting signal is detected. The only thing that would thwart this is if V.92 has been disabled, which (as you can read elsewhere in modemsite.com) some PC-OEMs have done in some of their PCs.
The modem country must be set to Australia for this to work in Australia. This is +GCI=09, or set through Phone and Modem properties.
|Posted - 3/17/2006 7:11:50 PM |
Thanks, Ed, I appreciate your comments.
With the modem I am currently using (Intel 536EP chipset), +GCI is set to 09, as indicated in modem diagnostics and as required for Australia. Also, the Modem log indicates the modem accepts the +PCW=1 command ("OK"), as it should according to the manual for the modem.
Nevertheless, this modem, along with many others tested, doesn't respond to a call waiting signal. It USED TO, perfectly, on the first ring of an incoming call, that's the thing; and on my computer and on other people's, using the same model modem and also other V.90/V.92 modems. But mid last year something changed outside my computer (and those of my contacts) and I'm trying to find out what.
If I knew how a modem detects the call waiting signal, it would help. Any info appreciated.
In Australia, the call waiting signal is 425 Hz, not 440 Hz like in North America, so the modem has to take that into account somehow. But that hasn't changed.
However I have found by measurement that the volume level of the call waiting signals in Australia is 12 dB lower than the dial tone, whereas it should be the same volume as the dial tone, according to the Cisco Virtual Central Office manuals for the VCO's used in Australia.
When this change occurred I don't know. Maybe it was there from the start of call waiting, but I seem to remember the tones being a lot louder in the past (subjective memory unfortunately). But a 12 dB drop would make call waiting a lot harder for a modem to detect, especially as it repeatedly retrains when the Call waiting tones disturb it.
In the past that could have been a good reason for the Telco to drop the volume: fewer complaints about modem's dropping out due to call waiting, or maybe the tones was just too loud and annoying. But since modern modems with error correction mostly hang on tight during an incoming call waiting signal, dropping the volume wouldn't reduce dropouts, in most cases. Anyway, people can just put *44 in front of the dial-up number (*70 in the US, I believe) to switch off call waiting.
The above is just speculation. I really don't know why +PCW=1 doesn't do what it used to do here in Australia.
So any info is potentially useful.
|Posted - 3/19/2006 4:38:43 PM |
The CW signal must be loud enough for a human to hear and identify. Our ears are amazing signal classifiers, but modem tone detectors should be as good.
|Posted - 3/24/2006 1:39:50 AM |
Actually I have now solved this problem. I installed an inexpensive software modem (Motorola SM56) with reasonably up-to-date drivers (6/10/2005 = 184.108.40.206).
Now incoming calls cause a little pop-up box to appear, on the very first ring (tested using mobile phone) and I can choose "ignore call" or "disconnect modem and answer call".
If I choose "disconnect", it takes about five rings (15 seconds) to disconnect the modem and ring the phone, but it works! On every call.
I have no idea what other modems will work on the Australian phone system, with its low call waiting volume (-22 dBm, compared to -10 dBm for the standard dial tone), but at least I found one that does. So, problem solved.
Thanks anyway for your support. Otherwise I might have given up. Because everyone here, including the Telco "experts" kept telling me "No way. It can't be done".
Didn't I read somewhere that an "Expert" (ex-spurt) is just a former drip under pressure?
|Posted - 4/3/2006 8:07:22 AM |
Further to the above, I think the mystery is solved. I changed ISP and immediately found all my modems, including good quality hardware modems, now work quite well with +PCW=1, as good as they did a year ago. It was not long after that my previous ISP changed wholesalers and started using a different dial-in number and that's when the problem began. So I think I've discovered the cause, in general terms. Hope this helps someone else, though as to which ISP to try, have to ask other people for their experiences, I guess.
|Posted - 4/11/2006 5:26:19 AM |
Follow-up to above: I took up your suggestion, Ed, and got hold of an Agere-chipped modem (Netcomm Inmodem IN5699_4 with Agere 1648C chipset). Found this works perfectly with +PCW=1, much better than any other modem I've tried. I tried it on my old ISP as well and it works perfectly, so ISP set-up is just one influential factor, specific modem type (and driver, of course) is another. Probably makes more sense to change modems than ISP, I expect.
Anyway, now incoming calls always ring through to my home phone on the first ring of the calling phone. But without the +PCW=1 string, incoming calls are mostly ignored, ie they sometimes drop the connection, but don't ring the phone, unless the caller rings a second time after the connection has dropped.
This proves that the +PCW=1 command is working as it was designed to. I'm suitably impressed.
In addition, as a modem, the Inmodem works very well indeed, much better than various full soft modems I've tried and better than another controllerless hardware modem I tried (Intel 536EP chip) that doesn't work as well with +PCW=1 on my new ISP and doesn't work at all with +PCW=1 on my old ISP (this month I still have accounts with both ISP's so I can do these experiments).
There may be lots of other modems that work well with +PCW=1 and the Australian telephone Call Waiting signals, but the Inmodem is one at least that I've found that does exactly what I want on my system. So thanks for the tip.
Click Here To Close Thread, Administrators & Moderators Only.
Show All Forums |