no v.90 with some networks
Bob Young
3.210-V90_2M_DLS.
The problem: I have used AT&T as my ISP for about a year, and have always
connected at 50666, 52000, or 53333 (52000 most common). Now I need to
use a different ISP. All the reasonably-priced ISP's are resellers of
access to one of the following 5 networks -
Qwest; Broadwing; Genuity; Uunet; Aleron
I have tested the local access numbers of each of those 5
networks, and they all connect at 33600, using V.34 (I can tell by the
handshake it is not even trying V.90). These numbers are all reported by
the network owners as supporting V.90. Under the exact same test
conditions I get 52000 (typically) with the AT&T number, and with one
other local number I found, owned by IBM. I have verified that other
modems on the same phone line can connect to at least 2 of those networks
at V.90 speeds.
My modem has the Rockwell ACF chipset. It accepts the +MS=V90 command.
I have tried +MS=V90,0,300,33600,44000,56000 in the hope of forcing a V.90
connection, but it makes no difference. I also tried S202=64, which is
supposed to increase the signal level for V.90 from -12dBm to -6dBm
(according to a web page I found), but that does not help. It seems to be
impossible to force a V.90 conenction with this modem. Does anyone know
how? Why is V.90 failing with so many networks? Any help appreciated.
Bob Young
Hooda Gest
"Bob Young" <news1.jum...@spamgourmet.com> wrote in message
news:20020416143854.2d2b3d...@spamgourmet.com...
Not all telephone routes are equal. That is, the way the telco routes the
call to the new ISP numbers may be the main factor in preventing your V.90
connection. I would try two things...
1. Use Hyperterminal, or whatever terminal program you prefer, to dial up
the new POP numbers for testing purposes (don't try to log in).
2. Dial through an LD service to the new POP numbers to see if that results
in a V.90 connection.
If you can get a V.90 connect in Hyperterminal on a local call then the
problem is in how you've configured the connectoid. If you can only get a
V.90 connect through LD calling then the problem is the telco routing (and
you are out of luck). If you can't get a V.90 connect no matter what you do
then the problem is incompatibility with the POP equipment or you are
dialing into banks of V.34 modems.
--
Hooda Gest
"The only thing I do immediately is procrastinate."
ted.f...@virgin.net
Im in the UK and have just had BT's Home highway fitted. This uses the 2
coppers to bring the digital link direct to ones home so this gives 2
analogue lines with different phone nos and one digital or isdn
guaranteed 64k a sec line. You can use any 2 of the 3 at any time.
Whats interesting is as yet ive not tried the digital line with my set
up as Ive not been able to get card and socket services to work on my
laptop with win 3.1. Thats a seperate issue. However ive always used my
usr external dual standard courier V34 fax V32 bis via the serial port
and this never connects faster than 33600. HOWEVER !!! as the analoge
section of the line from the courier to the Home highway box is now only
a few feet, then its digital to the telcos exchange my download speeds
have doubled. I can get a consistent 5.5k on text files without data
compression. So the 33600 connection speed doesnt seem to have anything
to do with the download speed. Does this make sense? Anyway it works so
thats all that matters. Ive also tried the usr professional voice modem
thich connects at 45666 but had very large nos of overun errors, and
doesnt download any faster. I guess the courier is the better product.
Also im using it powered by a 19v dc 2amp power supply from a laptop as
the 19v ac standard one wont last more than a couple of months without
butning out. I use AT&F1 as my init string.
Ted Frater, Dorset UK..
Bob Young
"Hooda Gest" <Be_@One_With.Calm> wrote:
> If you can get a V.90 connect in Hyperterminal on a local call then the
> problem is in how you've configured the connectoid. If you can only get
> a V.90 connect through LD calling then the problem is the telco routing
> (and you are out of luck). If you can't get a V.90 connect no matter
> what you do then the problem is incompatibility with the POP equipment
> or you are dialing into banks of V.34 modems.
>
I do not use Windows, but Minicom (similar to Hyperterminal) running under
Linux is how I already tested the access numbers for the 5 networks I
mentioned. All of them would connect only at 33600. They are all
advertised by both the network owners and the ISP's selling access as
V.90; one might be lying, perhaps, but I doubt that all 5 are - and I have
connected to at least one of them at V.90 speed using different hardware.
I am not clear how I can connect using LD without traveling outside the
local calling area - and taking my computer with me. In any case, this
would not really help as I am not going to pay LD charges for internet
access whatever happens - although it would be interesting to find out the
result. Incompatibilty with the POP modems would seem to mean nobody with
a Rockwell/Conexant ACF modem could use any of the 5 biggest networks and
get V.90 connections - possible, I guess.
Anyway, I seem to out of luck. My choice is to pay double what I need to
for internet access, or buy another modem. But, thanks for your help.
Bob
Bob Young
This web page - http://808hi.com/56k/rockacf.asp - gives information on
the settings for the S202 register in the Conexant ACF modems. It implies
that a setting of S202=32 raises the signal level by 6 dBm for K56Flex,
and S202=64 does the same thing for V.90 - but the info is wrong! The
settings are reversed. I tried S202=32 and immediately get a V.90
connection at 52000, instead of the V.34 at 3600 I always got without this
setting. Needles to say I am very happy.....
...but, what makes me angry, is that I have wasted untold hours and
bothered this group unnecessarily, when a simple document accurately
listing all of the AT command set for this modem would have have given me
the answer in minutes. There is no such listing in the printed manual,
nor on the CD that came with it, nor anywhere to be found on the
creative.com website. The printed manual is particularly ridiculous, as
it has page after page of screenshots of Windows dialog boxes for
"installing" the modem (as a Linux user I find this doubly pointless) -
but absolutely nothing on the all-important command set. The cheap
unbranded HCF winmodem I used to use came with a full technical manual in
PDF on the CD. OK - end of rant.....
Bob Young
Franc Zabkar
<news1.jum...@spamgourmet.com> put finger to keyboard and
composed:
>To close out this thread - I seem to have solved it!
>This web page - http://808hi.com/56k/rockacf.asp - gives information on
>the settings for the S202 register in the Conexant ACF modems. It implies
>that a setting of S202=32 raises the signal level by 6 dBm for K56Flex,
>and S202=64 does the same thing for V.90 - but the info is wrong! The
>settings are reversed. I tried S202=32 and immediately get a V.90
>connection at 52000, instead of the V.34 at 3600 I always got without this
>setting. Needles to say I am very happy.....
Hmmm, it appears that, using a "magic" init string, one _can_ force a
modem to connect at speeds which it thinks are not possible. ;-) Have
you checked your throughput to see if the initial connect speed is
sustained?
If your modem supports the AT#UD and AT&V2 diagnostic commands, you
may also want to check the post-call stats for details about your
dial-up session.
See http://modemsite.com/56k/diag.asp#rock
BTW, the latest versions of Stanislav Mekhanoshin's decoders are here:
http://www.geocities.com/rampitec/files/s1-1201.zip
http://www.geocities.com/rampitec/files/ud-1100.zip
>...but, what makes me angry, is that I have wasted untold hours and
>bothered this group unnecessarily, when a simple document accurately
>listing all of the AT command set for this modem would have have given me
>the answer in minutes. There is no such listing in the printed manual,
>nor on the CD that came with it, nor anywhere to be found on the
>creative.com website. The printed manual is particularly ridiculous, as
>it has page after page of screenshots of Windows dialog boxes for
>"installing" the modem (as a Linux user I find this doubly pointless) -
>but absolutely nothing on the all-important command set. The cheap
>unbranded HCF winmodem I used to use came with a full technical manual in
>PDF on the CD. OK - end of rant.....
I believe the information on Richard Gamberg's site is based on the
information I posted to Usenet in February 2002:
http://groups.google.com/groups?selm=3c6596f5.2092035%40news.dingoblue.net.au&output=gplain
This information comes from a Rockwell document, number 100426A,
entitled "CX802xx/CX803xx AnyPort Multi-Service Access Processor AT
Commands Reference Manual":
Admittedly the CX802/803 device is not a client modem, but my initial
experimentation suggested that several bit settings were applicable to
my own ACF2 chipped Rockwelloid.
BTW, in my original post I did say that I was "still experimenting
with bit 6 (V.90 Tx level)". I suggest you submit your own results to
Richard for inclusion on his site.
As for your rant against poor documentation, I'd have to agree that
most OEM vendors supply only basic command manuals, if any at all.
However, Rockwell/Conexant documentation is very good in general ...
if you can find it. I've managed to find a lot of good stuff, but it
involved a good deal of scrounging.
-- Franc Zabkar
Please remove one 'g' from my address when replying by email.
Hooda Gest
> On Wed, 17 Apr 2002 06:06:05 GMT
> "Hooda Gest" <Be_@One_With.Calm> wrote:
>
> > If you can get a V.90 connect in Hyperterminal on a local call then the
> > problem is in how you've configured the connectoid. If you can only get
> > a V.90 connect through LD calling then the problem is the telco routing
> > (and you are out of luck). If you can't get a V.90 connect no matter
> > what you do then the problem is incompatibility with the POP equipment
> > or you are dialing into banks of V.34 modems.
> >
> I do not use Windows, but Minicom (similar to Hyperterminal) running under
> Linux is how I already tested the access numbers for the 5 networks I
> mentioned. All of them would connect only at 33600. They are all
> advertised by both the network owners and the ISP's selling access as
> V.90; one might be lying, perhaps, but I doubt that all 5 are - and I have
> connected to at least one of them at V.90 speed using different hardware.
I saw your other post with your solution. I do not understand why one would
have to modify the signal level manually to get a V.90 connection.
Especially with an undocumented register setting.
> I am not clear how I can connect using LD without traveling outside the
> local calling area - and taking my computer with me. In any case, this
> would not really help as I am not going to pay LD charges for internet
> access whatever happens - although it would be interesting to find out the
> result. Incompatibilty with the POP modems would seem to mean nobody with
> a Rockwell/Conexant ACF modem could use any of the 5 biggest networks and
> get V.90 connections - possible, I guess.
The purpose of using LD is to determine if the problem is due to local telco
routing or compatibility with the POP's equipment. It is not for regular
usage of the POP. To do an LD call without leaving town, you just use an LD
access code which puts you into the LD carrier's network.
--
Hooda Gest
"In a New York minute, everything can change..."
Bob Young
franc...@dinggoblue.net.au (Franc Zabkar) wrote:
> Hmmm, it appears that, using a "magic" init string, one _can_ force a
> modem to connect at speeds which it thinks are not possible. ;-) Have
> you checked your throughput to see if the initial connect speed is
> sustained?
Yes. I downloaded a 5 MB compressed file at a rate of over 5 kB/s.
> I believe the information on Richard Gamberg's site is based on the
> information I posted to Usenet in February 2002:
>
> http://groups.google.com/groups?selm=3c6596f5.2092035%40news.dingoblue.net.au&output=gplain
Sure looks like it.... I searched google groups early in the
troubleshooting process, but failed to find your post; it would have
pointed me in the right direction and saved some time.
> This information comes from a Rockwell document, number 100426A,
> entitled "CX802xx/CX803xx AnyPort Multi-Service Access Processor AT
> Commands Reference Manual":
> http://www.unitronic.de/Hersteller/conexant/Netzwerk-DSL-Optische%20ICs/100426aCX802xx-CX803xxAnyPort-TM-MultiServiceAccessProcessorATCommands.pdf
Great! I just downloaded a copy, and will burn it on a CD so I can't lose
it. Thanks!
> BTW, in my original post I did say that I was "still experimenting
> with bit 6 (V.90 Tx level)". I suggest you submit your own results to
> Richard for inclusion on his site.
Yes - I already did that. I see from the Conexant document that bit 6 of
S202 is specified as the Tx signal level setting for V.90, but I achieved
my result using bit 5 instead. Bit 6 had no effect. Strange....
> As for your rant against poor documentation, I'd have to agree that
> most OEM vendors supply only basic command manuals, if any at all.
> However, Rockwell/Conexant documentation is very good in general ...
> if you can find it. I've managed to find a lot of good stuff, but it
> involved a good deal of scrounging.
>
In the case of Creative (Modem Blaster), none at all. I don't see how
this helps them. I'm very happy with my modem right now, but a few hours
ago I wanted to smash the thing, entirely as a result of missing
documentation.
Thanks for your valuable information.
Bob Young
Franc Zabkar
>On Wed, 17 Apr 2002 23:07:52 GMT
>franc...@dinggoblue.net.au (Franc Zabkar) wrote:
>> BTW, in my original post I did say that I was "still experimenting
>> with bit 6 (V.90 Tx level)". I suggest you submit your own results to
>> Richard for inclusion on his site.
>
>Yes - I already did that. I see from the Conexant document that bit 6 of
>S202 is specified as the Tx signal level setting for V.90, but I achieved
>my result using bit 5 instead. Bit 6 had no effect. Strange....
I wonder if experimenting with register S91 (PSTN Transmit Attenuation
Level) would have achieved the same outcome. Note that the C802xx
documentation suggests that S91 applies only to analog modulations,
but other Rockwell/Conexant literature makes no such distinction.
Another thing to note is that 52000bps is also a K56flex speed. Are
you sure you connected with V90?
My modem's post-call stats (AT#UD and AT&V2) show the following signal
level info:
AT#UD
TX/RX signal power level, -dBm : 10/19
AT&V2
Signal Level (TX/Power Drop), -dB : 10/0
RX Signal Level (Last/Min/Max), -dB : 18/18/18
Remote power drop support : ON
I notice that occasionally the Tx signal level will drop by one or two
dBm, eg from "10/0" to "10/1". I believe this is because the remote
modem commands a "power drop" of 1dBm. It would be interesting to
compare your own data.
In the past I have experimented with S91:
http://groups.google.com/groups?hl=en&selm=399e3fb9.444318%40news.ozemail.com.au
http://groups.google.com/groups?hl=en&selm=3c08a19a.14580396%40news.dingoblue.net.au
Bob Young
franc...@dinggoblue.net.au (Franc Zabkar) wrote:
> Another thing to note is that 52000bps is also a K56flex speed. Are
> you sure you connected with V90?
Oh yes. I also sometimes get 50667 or 53333 to the same POP. And, my
modem firmware supports V.90 only. It has to be flashed with a different
version for Flex. ATI3 reports the firmware revision as -
V3.210-V90_2M_DLS
> My modem's post-call stats (AT#UD and AT&V2) show the following signal
> level info:
>
> AT#UD
> TX/RX signal power level, -dBm : 10/19
>
> AT&V2
> Signal Level (TX/Power Drop), -dB : 10/0
> RX Signal Level (Last/Min/Max), -dB : 18/18/18
> Remote power drop support : ON
>
> I notice that occasionally the Tx signal level will drop by one or two
> dBm, eg from "10/0" to "10/1". I believe this is because the remote
> modem commands a "power drop" of 1dBm. It would be interesting to
> compare your own data.
OK - I'll try to get some and post it later.
Bob
Franc Zabkar
>On Wed, 17 Apr 2002 23:07:52 GMT
>franc...@dinggoblue.net.au (Franc Zabkar) wrote:
>> BTW, in my original post I did say that I was "still experimenting
>> with bit 6 (V.90 Tx level)". I suggest you submit your own results to
>> Richard for inclusion on his site.
>
>Yes - I already did that. I see from the Conexant document that bit 6 of
>S202 is specified as the Tx signal level setting for V.90, but I achieved
>my result using bit 5 instead. Bit 6 had no effect. Strange....
Well, I finally tested S202 bits 5 and 6. Based on my results, I've
come to the conclusion that the documentation for S202 is only partly
applicable to my client modem. While several bits performed as
expected, bit 5 (K56flex) had no obvious effect, and bit 6 (V90)
produced the attached hex dump. FWIW, notice that the data appear to
be fairly regularly spaced and are grouped in several distinct blocks
(are these the results of a line probe?). BTW, all the data lines have
been wrapped by my news reader.
====================================================================
ats202=64
OK
atdtxxxxxxxx
0001 0001 0001 1000 0002 0001 0001 1000 0004 0001 0001 1000 0008 0001
0001
1000 0010 0001 0001 1000 0020 0001 0001 1000 0FE8 0FE7 0FE6 0FE5 0FE4
0FE3 0FE2
0FE1 0FE0 0FDF 0FDE 0FDD 0FDC 0FDB 0FDA 0FD9 0FD8 0FD7 0FD6 0FD5 0FD4
0FD3 0FD2
0FD1 0FD0 0FCF 0FCE 0FCD 0FCC 0FCB 0FCA 0FC9 0FC8 0FC7 0FC6 0FC5 0FC4
0FC3 0FC2
0FC1 0FC0 0FBF 0FBE 0FBD 0FBC 0FBB 0FBA 0FB9 0FB8 0FB7 0FB6 0FB5 0FB4
0FB3 0FB2
0FB1 0FB0 0FAF 0FAE 0FAD 0FAC 0FAB 0FAA 0FA9 0FA8 0FA7 0FA6 0FA5 0FA4
0FA3 0FA2
0FA1 0FA0 0F9F 0F9E 0F9D 0F9C 0F9B 0F9A 0F99 0F98 0F97 0F96 0F95 0F94
0F93 0F92
0F91 0F90 0F8F 0F8E 0F8D 0F8C 0F8B 0F8A 0F89 0F88 0F87 0F86 0F85 0F84
0F83 0F82
0F81 0F80 0F7F 0F7E 0F7D 0F7C 0F7B 0F7A 0F79 0F78 0F77 0F76 0F75 0F74
0F73 0F72
0F71 0F70 0F6F 0F6E 0F6D 0F6C 0F6B 0F6A 0F69 0F68 0F67 0F66 0F65 0F64
0F63 0F62
0F61 0F60 0F5F 0F5E 0F5D 0F5C 0F5B 0F5A 0F59 0F58 0F57 0F56 0F55 0F54
0F53 0F52
0F51 0F50 0F4F 0F4E 0F4D 0F4C 0F4B 0F4A 0F49 0F48 0F47 0F46 0F45 1FFF
0000 0006
0001 C854 01E1 01FE 0004 0004 0006 803F 0004 0000 0000 0000 0000 0000
0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0026 5B2B 0674
003F 0000
0000
0000 0000 0200 0000 0008 0081 0000 0000 0000 0000 0000 0000 0000 0000
0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000
0000 0000 001A 0029 003D 004B 005B 006C 007A 008C 009A 00AC 00BC 00CA
00DE 00ED
00FE 010B 011E 0130 0141 014F 0160 0172 0180 0191 01A0 01B3 01BF 01D1
01E5 01F4
0202 021D 0239 025C 027E 029D 02BC 02DF 02FF 0322 0340 0362 0381 039E
03C1 03E1
0402 0437 0477 04B7 04FA 0539 057A 05BF 05FE 063E 067E 06C4 0701 0745
0784 07C6
0806 086A 08EF 0972 09EE 0A76 0AF6 0B78 0BFE 0C7F 0CFE 0D81 0E05 0E89
0F0B 0F8D
1000 10D3 11D7 12DB 13E1 14E4 15E9 16EF 17F2 18F4 19F9 1AFD 1C00 1D04
1E02 1F0C
200D 219F 23A9 25AF 27BB 29BE 2BCC 2DD1 2FD7 31E4 33EF 35F1 37F9 39FE
3C0A 3E27
4037 432D 473A 4B4B 4F55 536B 575B 5B2B 803F 0000 0011 002D 003F 0049
005E 006F
0077 008B 009F 00B2 00C6 00C9 00E2 00EE 0102 010C 011C 0131 0140 014C
015E 0175
0185 0195 019F 01AE 01BF 01D0 01EF 01F5 0207 021E 0239 0259 0283 029E
02C0 02DC
02FF 0326 0342 0363 0384 039E 03BD 03E3 0404 0435 0479 04B7 04FD 0538
0579 05BD
05FE 063A 0682 06C5 06FD 0745 077E 07C5 07FE 0868 08E8 0977 09E9 0A70
0AF2 0B75
0BFF 0C81 0CFF 0D7D 0E01 0E85 0F04 0F8B 1000 10CB 11D4 12D4 13E3 14DE
15DB 16EC
17ED
18EA 19EE 1AF7 1BF7 1CF9 1DF7 1F02 2006 219A 23A5 259F 27B4 29B4 2BB7
2DCB
2FC4 31E0 33EA 35FD 37EF 3A13 3BFC 3DF6 3FF8 4333 4742 4B5A 4F5E 537B
5777 5B8F
0000 0000 0016 0029 0042 0046 005A 006B 0079 0090 009A 00A7 00BA 00CC
00E0 00F3
00F6 0107 011E 012A 0143 014E 0160 0177 017B 0195 01A5 01B2 01BE 01D0
01DE 01ED
0202 021B 023E 0258 0282 02A0 02BC 02E0 0302 031D 0344 0362 037D 039E
03C2 03DC
03FD 043C 046F 04B8 04F6 0532 057D 05BC 05F9 063E 0678 06C0 06FD 073F
0781 07C2
07FC 0868 08EF 096C 09ED 0A74 0AF1 0B78 0BF9 0C7C 0CFD 0D82 0E04 0E8B
0F09 0F8A
1000 10D0 11D5 12D8 13DD 14DE 15EF 16EF 17F6 18F6 19F7 1B03 1C00 1D00
1E02 1F11
200F 21B2 23AE 25B7 27C5 29CB 2BE3 2DE3 2FF3 31F4 33DB 35D5 37E7 3A0B
3BFF 3E0D
4022 431A 472A 4B55 4F3E 5368 5770 5B9B 0000 0000 0028 0025 0043 0050
0063 006F
007E 008B 009C 00B0 00BB 00CC 00DB 00F1 0100 0109 0121 0135 0146 0156
0162 0170
0184 018C 019F 01B6 01BA 01D4 01E9 01FC 0205 0222 023F 0266 027D 02A0
02C1 02DF
0301 0323 033F 0366 0382 039C 03B8 03E8 0404 043E 047D 04B8 04F8 053D
057D 05C5
0602 063B 0684 06C3 0702 074C 0786 07CA 080E 0863 08F2 0975 09F0 0A77
0AF6 0B75
0BFE 0C82 0CFB 0D82 0E04 0E89 0F0F 0F85 1000 10D9 11D8 12DE 13E2 14DA
15E3 16F3
17EC 18F2 19F9 1AFE 1C03 1D05 1E02 1F0C 2008 21A1 2396 25A7 27BD 29B9
2BB8 2DCC
2FC6
31DA 33DA 35CB 37E1 39DA 3BED 3DEF 3FFB 4326 472B 4B36 4F3B 5359 56CD
592C
0000 0000 0018 0027 003A 0046 0053 0067 006F 008A 0096 00A8 00B5 00C4
00D6 00E7
0102 0107 011C 012E 0142 014C 0160 016F 0177 018A 0198 01A7 01B9 01CA
01E0 01F2
01FC 0218 022C 0258 027F 0299 02B1 02D9 02FE 031D 0337 035F 037E 039E
03C6 03D9
0403 0431 0475 04B2 04F2 052F 0572 05BD 05FB 0643 067B 06C2 0701 0744
077B 07C2
0803 086A 08EF 0969 09E5 0A7A 0AF5 0B75 0BFE 0C78 0D01 0D83 0E04 0E84
0F0B 0F92
1000 10D2 11D1 12D8 13D8 14EE 15EC 16F2 17EE 18FC 19FE 1AFA 1BFD 1CFE
1DFA 1F0D
200A 2187 239A 25A9 27AA 29B3 2BCC 2DBE 2FD2 31D6 33DD 35E8 37E5 39E2
3C0C 3E28
4029 431A 472A 4B28 4F3F 5350 576F 5B7E 0000 0000 0023 0028 0033 0048
0054 0069
007B 0088 0090 00A6 00B3 00C3 00DA 00E6 00FA 010E 011B 0129 013A 0149
0159 016D
0176 018A 019E 01B5 01C3 01CE 01D9 01ED 01FF 021A 0235 025A 0273 0296
02B2 02DC
02F6 031F 0337 035A 0380 039A 03C0 03D6 03FC 042C 0475 04AF 04F8 053A
0575 05B8
05FA 063A 0672 06C2 0700 0740 0784 07BF 0805 086D 08E9 0972 09EC 0A6F
0AF2 0B7A
0BFA 0C7B 0CFC 0D79 0DFF 0E87 0F08 0F8D 1000 10D7 11D9 12E5 13DD 14EC
15ED 16EC
17F7 18F5 19F8 1B00 1C05 1D0E 1E0C 1F08 2017 21AB 23B0 25BD 27CC 29CC
2BE2 2DD4
2FE6 31F8 3409 3608 381E 39FD 3C4B 3EA0 40E6 4343 4753 4B5D 4F73 5385
5783 5B8E
0000
CONNECT 230400
====================================================================
Franc Zabkar
>To close out this thread - I seem to have solved it!
>This web page - http://808hi.com/56k/rockacf.asp - gives information on
>the settings for the S202 register in the Conexant ACF modems. It implies
>that a setting of S202=32 raises the signal level by 6 dBm for K56Flex,
>and S202=64 does the same thing for V.90 - but the info is wrong! The
>settings are reversed. I tried S202=32 and immediately get a V.90
>connection at 52000, instead of the V.34 at 3600 I always got without this
>setting.
I've been researching the Net for more info on S202 but haven't found
any information that is very detailed. However, there are numerous
references to an A/D detection bug in early versions of Rockwell
K56flex firmware (rev 1.00x) which caused problems with certain ISP
equipment.
Here are some references:
======================================================================
http://homepages.force9.net/paulmac/modemk56.htm
"S202=32 -is an obscure command which should not affect the operation
of the device from day-to-day. If you MUST know, it disables part of
the negotiation sequence, where the modem attempts to detect more than
one digital-to-analogue conversion along the path from the ISP to your
location."
======================================================================
http://www.56k.com/trouble/known.shtml
http://www.wmis.net/home/setup/56k.htm
PROBLEM
Poor connects with 1.00x K56flex firmware
Symptom: When you issue an ATI3 command in a terminal program, the
modem reports 1.00x firmware (1.003, 1.009, etc.). The modem fails to
connect at high speeds. Prior to the firmware update, the modem
connected at higher speeds.
SOLUTION
Try adding S202=32 to your init string. This has been very effective
for many readers. Once this string has been added, many K56flex modems
show much improved performance.
What does this command do? According to Tom Hanson of Zoom technical
support:
The command disables dual pcm detection. i.e. it won't look for how
many analog to digital connections are on your phone line. This is a
bug in 1.0x code. It was fixed in 1.1 maybe 1.09 but I'm not sure.
Note: using S202=32 on higher- numbered firmware may prevent K56flflex
connects. Keep this in mind when you upgrade your firmware.
======================================================================