kenneth.tam@canremote.uucp (KENNETH TAM) (12/15/89)
After muddling around on my own for 2 months or so trying to fix this problem, I've come over to .netland hoping for some help: . Setup: B2000, USR 14.4 HST, Maxtor 360MB HD and a HardFrame 2000 v1.54 . . Problem: -------- . All the components have been verified on other machines as fully functional (BTW, still running 1.2 KS ROM -- <sigh>). Hardframe performs marvelously, modem runs fine, HD has no unaccounted for errors. Modem transfers to RAM: or floppy present no problem, but the moment I attempt an ASCII capture or binary download to the HD, the terminal program gives me CRC, Subpacket and other diverse errors. If I access the HD while the modem is recieving or sending information, said information becomes corrupt (e.g. I try a LS of the HD while reading an editor in the background and the file transfer gives me errors). Interestingly, uploading FROM the HD seems to work -- it access the drive every 1 K worth, so I suspect the access is small enough not to interfere. . Bottom Line: My HD is useless while using the modem. . I tried various solutions proposed by friends: Changing to lower baud rates does not make a difference -- the HD access still ruins the data when the serial is in use. RF shielding has been taken care of with shielded cables all over the place. Processor being to slow to keep up with both DMA HD access and serial access doesn't seem reasonable considering the data is ruined regardless of baud rate -- I haven't explicitly tested the HD + controller on a faster machine like a 2500, but I strongly doubt that is the problem. . The most reasonable proposal I've had is that it's a DMA conflict problem -- the controller is trying to take the DMA channels that the serial is using. It sounds good (bad?) to me, but there isn't much I could do about it. A call or two down to Microbotics and some talk with the techs revealed that they were aware of this problem and that it existed with any modem over 1200 baud. Apparently the Hardframe takes up all the DMA channels even if they happen to be in use? I pointed out that other DMA controllers didn't seem to have this problem, to no avail. They seemed to suggest that it was COMMODORE's fault, and that there was going to be no attempt to fix this MAJOR (IMNSHO) bug. . This is UNACCEPTABLE -- I'm very tempted to get myself a Xetec or Kronos, however, I really do like the HardFrame 2000 as it is remarkable _fast_ and performs superbly in all other respects. . Are there any HardFrame owners here that could possibly give me some assistance? Anyone else experiencing similar problems? . Any representatives from Microbotics that might be able to help? . Can some knowledgable hardware guru confirm or disprove the above 'facts'? . . Many thanks, . - Ken --- * Via ProDoor 3.2aR