Re: 56K Modems - Do I really want mine?

Karl Denninger (karl@Mcs.Net)
Thu, 21 Aug 1997 11:39:28 -0500

On Thu, Aug 21, 1997 at 11:14:03AM -0500, dtrucken@zsassociates.com wrote:
>
> I'm using a PM3 with ComOS 3.7 and the old modem cards. I have an IBM 760e
> with a MWave modem. When I boot the laptop in NT 4.0, I MUST use *TH8 in
> my init string to connect. (Or I get random gibberish on the screen instead
> of a Login prompt.) According to the MWave help page, this increases the
> Signal to Noise Ratio required for a given speed.
>
> When I boot the same machine in Windows 95, I don't need the init string.
> Go Figure.
>
> The Mwave version I loaded is 2.2 (first available from IBM in May97). So
> far, I haven't found a way to identify the version of MWave running on a
> machine, I had to reload it just to make sure it was 2.2.
>
> Is there any word about the seemingly broken V.42 implementations that hang
> up after 12 resyncs? Is this a Livingston problem or a problem with all
> (or most) modems?

That's not a bug! The end which rejects six packets (requesting
retransmission) is supposed to ask for a retrain and/or renegotiation after
the six. If you get to *12* and haven't renegotiated the connection, you
get dropped. This is what is *supposed* to happen.

What's happening is that some modems don't request the renegotiation after
the errors occur. Thus, the count piles up, and eventually the Lucent
chipsets hang up with a V.42 retry limit exceeded error. This is *proper*
behavior.

> Is there a fix?

Lucent is reported to know about this and be putting in code to deal with
those modems which do the "wrong thing".

> Should I try and get MNP to work?

That's a temporary workaround for the problem.

> Is
> error control and modem compression really not necessary since I'm using
> PPP and PPP compression (what ever that is) anyway? What do people think?

If you have STAC cards in the box, you can shut off error control entirely
and use PPP's error control for this (it works), and MS Windows compression
(check the "software compression" box in the profile) for compression. This
is a STAC derivitive.

That's a *very* viable solution to the problem.

In fact, doing either that or using V.42 (but not V.42bis) and the Microsoft
Windoews compression *IS* faster than V.42bis if your processor is a Pentium
90 or better.

If you're running K56Flex, there is another reason to do this -- with even
moderately compressible data (ie: Usenet news postings, email, etc) you'll
overrun the 115.2kbps port rate on most PCs. Unless you have a 230kbps
serial card, that's the ONLY WAY to get the full K56Flex transfer potential.

We are *strongly* recommending that users on K56Flex do exactly this for the
best possible performance.

--

-- 
Karl Denninger (karl@MCS.Net)| MCSNet - Serving Chicagoland and Wisconsin
http://www.mcs.net/~karl     | T1's from $600 monthly to FULL DS-3 Service
			     | NEW! K56Flex modem support is now available
Voice: [+1 312 803-MCS1 x219]| 56kbps DIGITAL ISDN DOV on analog lines!
Fax:   [+1 312 803-4929]     | 2 FULL DS-3 Internet links; 400Mbps B/W Internal