wjc@llex.ll.mit.edu ( Bill Chiarchiaro) (05/22/91)
Can anyone tell me how to get in touch with the author(s) of HP's PDL software? I am having a problem (I'll describe it below for the curious) which is beyond the expertise of the people at HP's Calculator Support line (503-757-2004). When I received my HP-48SX (rev. E) and PDL package I loaded PDL onto my PC at home and hooked up the 48. PDL seemed to run correctly, except I couldn't effect any transfers between the PC and the 48. I started up the PDL kermit and put it into connect mode. Using the 48's XMIT and SRECV commands, I tried to exchange character strings between the two machines. I was able to get strings from the PC to the 48, but not the other way around. The PC just didn't seem to see the characters coming from the 48. I wondered if the PDL programs were looking for certain modem control signals. I made a null modem which looped the PC's DTR back into its DSR and CD. This didn't help. In fact, I was unable to get echo on the "connected" kermit with pins 2 and 3 of the PC's RS-232 line shorted together (with or without the null modem). The PC in question is an old Leading Edge (with a 4-digit model #, something like 1610) 8088-based machine running MS-DOS 3.10. It has one async port on the motherboard and another on a third-party expansion card. The PDL behaviour is the same with both ports. I tried two other versions of kermit on this PC with the same uni-directional results. With the PC I have in my office (a Standard 286), PDL and the kermits all work fine -- I can move characters in either direction between it and the 48. My test with kermit in connect mode and XMIT and SRECV on the 48 also works. I began to think that the PC at home had a BIOS bug. I wrote some Turbo-C programs to talk directly to the motherboard's async chip. I was able to use these to send characters to and from the 48. At this point I was sure that PDL and the kermits were using BIOS calls to do their IO and that the BIOS was messing up. After all, my test programs which ingored the BIOS were working. Also, other communication software such as Phil Karn's IP/TCP code works properly with the async ports on my home PC. As another test of this theory, I wrote some more C programs which did use the BIOS (actually, DOS interrupt 14H -- I hope I have the terminology correct). Unfortunately, these worked! I am now quite confused. I want to find out just how PDL talks to the PC's async ports. I think I need to find someone familiar with the internals of PDL. The people at the Calculator Support line didn't even understand the concept of shorting RS-232 pins 2 and 3 to do an echo check with kermit in connect mode. Thanks for any help, Bill Chiarchiaro