Stuart Baker Software offers high quality consulting services and products to users of IBM PCs and compatibles (DOS and Windows)
    as well as Interdata, Perkin-Elmer and Concurrent Computer Corporation 3200 series machines (OS/32).

The Pitfalls of Progress


 

The Pitfalls of Progress


by Stuart Baker
Stuart Baker Software
9370 Golondrina Drive
La Mesa, CA 91941-5654
(619) 466-8811

Stuart Baker Software develops telecommunications software for OS/32 and IBM PC's. To stay abreast of the current technology I have been experimenting with the Hayes V-series Smartmodem 9600. This article outlines the problems I had getting the modems to function on an OS/32 system.

The first problem is the clock rate dilemma. One of CCUR's 8-line mux design features is that you must live with pre-defined clock groups. The groups do not address the choices needed to handle today's modems. Observe that the Hayes modem supports 300, 1200, 2400 and 19200 bps. Unfortunately neither the 8-line mux or MPC is flexible enough to allow you to group these speeds together. If you have a CCUR 8-line mux the only solution to the problem is foil cut and jumper or replace the interface. If you have a MPC, you can purchase a custom ROM chip to solve your problem. Page 3-39 of the MPC installation manual instructs you to replace A24 (9B3) to regroup the baud rates. I have never tried to obtain a new ROM so I can not unveil the mysteries of that procedure. If you are fortunate enough to own a Macrolink 8-line mux, simply set the baud rate grouping switches. You don't even have to remove the board or shut the system down. Let's hear it for instant reconfiguration.

After you solve the clock rate dilemma, the next problem is auto baud adjust. After using PASSPORT to dial a number at 2400 bps, the modems would not sync up at 19200 bps. This occurs because the speed of the last AT commands determines the maximum speed for the modem. There is a way around this problem. You must configure the modems at 19200 bps and include the &D3, S37=1 and &W0 commands in the configuration. This forces the modem to restore the configuration of user profile zero whenever DTR drops. Therefore, at the completion of every call the modem is reset because BIOC drops DTR for 3 seconds.

After auto baud is working, the next problem is flow control. You noticed I was using a top speed of 19200 on a modem rated at 9600 bps. This modem uses data compression and can achieve speeds that exceed the rated line speed. The CPU must supply data to the modem fast enough to allow buffering for the compression to work properly. Obviously, since the modem does not have infinite data storage, eventually the CPU must stop sending to prevent data loss. The Hayes modem offers three basic methods of flow control, RTS/CTS, XON/XOFF or local flow control disabled.

I chose RTS/CTS flow control because I knew that XON/XOFF would not work with XMODEM file transfers. I could not get the modems to function using the existing modem cables. Investigation revealed I needed to add a wire to the modem cable. I added a wire from pin 13 of the 15 pin plug to pin 4 of the 25 pin plug. This supplied the RTS signal to the modem. Once this is accomplished, the modem communicates properly under the BIOC driver. I tested XMODEM and it worked, but very slowly. The speed of the XMODEM file transfer with these new modems is a subject for another article. I next tested RELIANCE, it did not work. If CTS drops during a RELIANCE screen transfer, ITAM generates an I/O error. If there is a way for ITAM not to report loss of CTS as an error, I am not aware of it. If there is, will someone please let me know.

I next disabled local flow control. RELIANCE seemed to work, but I have a suspicion that some complex screens or line error conditions might exceed the modem buffer limits. MTM worked if you limited displays to about one screen. A display files with no page pause failed miserably. The XMODEM protocol worked because each packet must receive an acknowledge before sending the next. This is definitely not a functional configuration.

I assume that XON/XOFF works with the ITAM driver, but I have not tested that combination. The BIOC driver works fine with XON/XOFF, but remember that this will prevent running file transfer protocols like XMODEM. If you need to run only RELIANCE, give XON/XOFF a try. If you are running with BIOC on MTM, and don't plan on adding RELIANCE, use RTS/CTS. If you need both MTM/XMODEM and RELIANCE you will need to disable local flow control and take your chances with data loss. I find that totally unacceptable and hope there is a method of making ITAM work with RTS/CTS.

In conclusion, try out these new modems with caution, there may be surprises along the way. Unless you can justify writing off the complete cost in less than one year, delay your purchase for a while. With the current incompatibility between various manufacturers modems, you will likely be an island onto your self.

Return to Index
Article Copyright 1988 Stuart J. Baker

This SBSW.COM page has been optimized for printing.
Web Content Copyright © 1997 - 2010 Stuart Baker Software. All rights reserved.

Please use our Feedback form to submit questions or comments about this web site.
Web Content Copyright © 1997 - 2010 Stuart Baker Software. All rights reserved.
This site was last modified: Wednesday, December 29, 2010