W8SDZ@SIMTEL20.ARPA (Keith Petersen) (03/31/85)
The following review was not written by me. It was downloaded from a
Remote CP/M system. Unfortunately there is no way to reach the author
because it is unsigned. It is presented here for its possible
informational value. Please address discussions/comments to the
mailing list, not me. I don't own a 2400 baud modem.
--Keith
10 Mar 85
*** DISCLAIMER ***
The products described here were repeatedly tested for a specific
application only. No value was placed on advanced features not
directly related to their intended use. The opinion expressed herein
is that of the reviewer and may, in fact WILL differ considerably from
other reviewers' opinions.
This is an unsolicited review. Anyone able to disprove the reviewes
claims is welcome to do so. This review is about as unobjective as it
can get.
In a data processing environment, chances are the employee
with a terminal and a modem (or computer) and access to the business
computer via dialup will be more productive and is more likely to put
in a few hours' worth of unsolicited overtime per week than the
employee who has access to the same computer only during working
hours. Therefore, it was decided that the office computer be set up
with at least one high-speed dialup line and the most economical
choice was that of a 1200/2400 baud modem.
The modems were going to be used for two purposes:
1. unattended autoanswer
2. occasional use for dialout
Testing was started as soon as modems became available through
a local distributor. Due to this factor only two brands were
evaluated. Here is the story on both of them:
1. PENRIL 2024
The Penril 2024 seems to be the first widely available
1200/2400 baud modem, with the exception of the prohibitively priced
VADIC 4400 series. The 2024's list price is somewhere around $900.
The 2024 offers two baud rates, 1200 and 2400. The 1200 baud
protocol can be switched from 212A to V.22 at configuration time. By
today's standards, the 2024 cannot be considered a "smart" modem in
that its smartness is limited to the ability to dial a phone number.
Placing the modem in autoanswer mode is accomplished by simply
configuring the internal and external switches according to the
manual, releasing all front panel switches and plugging it in. The
2024 does not have a power switch (a definite plus in this
application).
Originating a phone call with the 2024 is a cumbersome
procedure, especially in an application where the modem may be 100
feet away from the terminal. First, the modem must be taken out of
autoanswer mode by pressing a front panel switch. Next, the originate
baud rate must be selected by locking the HI/LO switch IN or OUT.
Note that if you set up the modem for 2400 baud, you can call a 1200
baud number because of the "fallback" feature. You just have to
adjust your terminal baud rate after connect. The 2024 has no abort
provisions. While dialing, the TALK/DATA switch can be used to abort.
When connected, you must either cause the remote computer to drop
carrier or you again have to hit the switch. An alternative is
dropping the DTR line low, but in some instances that's a bit hard to
do. Dialing a phone number is very awkward. The sequence is
"CRNnnnnnnn<CR><LF>" so to dial 555-1212, you type CRN5551212^M^J.
Fine if the ENTER key on your keyboard generates a CR-LF sequence;
with most terminals you have to hit two keys. Sorry, no redial
capability.
The Penril worked fine calling the local TYMNET 2400 baud
access number, but no connection was established to any long distance
modem at 2400 baud.
The modem is superbly suited for unattended autoanswer mode.
the 2024 can be turned on and left alone and if something goes wrong
it's the software but not the modem. The continuous high-pitched
noise coming out of the built-in speaker may be objectionable to
some -- it picks up the strongest local AM radio station. the
speaker can be turned low or off via an internal jumper block.
One 2024 modem was tested initially in early November 1984,
and two were again tested in late February 1985. No difference was
found between the three modems, even though the early onemay have been
a preproduction unit.
2. USR COURIER 2400
The USR Courier seems to be the first smart low-priced modem
to be released, probably due to the fact that, unlike many
manufacturers, it does not use the Rockwell chip set. It features
Hayes 2400 compatibility. I will briefly summarize the positive
aspects of the modem:
- externally accessible, well-labeled configuration switches
- external switch to reverse pins 2 & 3, thus eliminating the
need for a null modem
- result codes can be completely turned off via switch
- volume control for internal speaker
After setting the configuration switches (an easy task for
anyone who has ever set up a modem) the Courier is ready for
operation. With the appropriate switch setting, it can be used both
in originate and answer mode without any hardware changes.
Originating a call can be accomplished with the now-famous
ATDT sequence, except that command letters no longer have to be in
caps. As with other smart modems, any character typed while dialing or
waiting for carrier aborts the action and hangs up the line. The
"escape" character can be used to either return the modem to command
mode (like the Hayes) or to hang up (like other USR modems) depending
on a configuration switch setting.
The Courier was used to successfully connect to the local
TYMNET number. A later model also was able to talk to a VADIC 2400
baud unit over long distance (Wayne Masters' RCPM). The Courier was
also able to call and be called by a Penril 2024 and another Courier.
As to autoanswer mode, the modem was a complete washout to put
it mildly. Surely, hard- and software are partly to blame but the
fact that other modems (including USR Password and AD212A) work with
the same setup indicates a serious flaw in the Courier.
The hardware used, for whatever reason, drops DTR while
changing baud rates. The duration is so short that all other modems
tested on the hardware, EXCEPT the Courier, are totally unaffected.
The Courier will, upon carrier lock and receipt of the first character
typed, drop the carrier 3 out of 4 times at 1200 and 2400 baud. While
no considerations were given to 300 baud performance, it was noted
that those problems only exist at 1200 and 2400 baud. Placing a 5MFD
capacitor from the DTR line to ground totally fixed this problem.
In autoanswer mode, the modems were used as follows:
- all result codes are inhibited
- on carrier loss, computer reboots, cycles DTR, then waits for a
character typed by constantly polling the data input port
- on receipt of character, baud rate is tested and, if necessary,
changed
I must again stress that this method works with all modems
tested. The USR Courier, however, would simply refuse to answer any
more phone calls after answering a few. No set pattern was
discovered. Sometimes, the modem would work properly for 5-10 calls
then refuse to answer, at other times it would only allow 1 or 2
calls. When it refused to answer, no outside indication was given
as to the problem. The appropriate LEDs on the front panel were lit
yet the modem did not respond to the ring. Surely, this is a most
serious deficiency and I have decided that the Courier is unfit for
use in this particular application. I am convinced the problem lies
solely with the Courier, particularyly sonce both the Auto Dial 212A
and the Password 1200 work in the exact same environment (except for
the much-needed 2400 capa- bility, of course).
In closing I must again stress that the USR certainly is a
superb modem and very well suited to originate applications.
Considering that 99 out of 100 modems sold will never be used for pure
autoanswer purposes, the Courier is not at all a failure.
One modem was tested in late November, two in late February
(both were preproduction units) and thre PRODUCTION units were tested
in March. All five units exhibited the same problems.
...may those who have the power to change things do so, may
those whose toes I stepped on test for themselves before stepping on
mine, may those who want to buy a Courier not be discouraged.
If YOU intend to use a USR Courier in an autoanswer-only
environment, please by all means give it a try, it may sure work for
you. If it doesn't, you have been warned, and if it does, either
"they" fixed it or the problem is installation-dependent...rick@cadtec.UUCP (Rick Auricchio) (04/04/85)
given to 300 baud performance, it was noted >that those problems only exist at 1200 and 2400 baud. Placing a 5MFD >capacitor from the DTR line to ground totally fixed this problem. I expect the DTR drop to be due to the design of the serial chips; 8251(A)'s, for example, must be reset before new modes can be stored. A reset would cause DTR to drop momentarily, causing the problem you mention. Though a reset is not always necessary, a UN*X driver would probably always do one because a change in baudrate is done via TIOCSETP, which also allows the caller to change other params (e.g. parity and bits/character). My UN*X system does this; luckily my Hayes 1200 doesn't get upset. ============================================================================== Opinions expressed have been generated solely by line-noise. {cbosgd,decwrl,hplabs,ihnp4,seismo}!nsc!cadtec!rick N1150G (408) 942-1535 "The sooner you fall behind, the more time you'll have to catch up!"