Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!portal!cup.portal.com!thad From: thad@cup.portal.com (Thad P Floryan) Newsgroups: comp.sys.amiga Subject: Re: Supra Controllers and SCSI ID's Message-ID: <20392@cup.portal.com> Date: 13 Jul 89 11:31:40 GMT References: <20240@cup.portal.com> Organization: The Portal System (TM) Lines: 159 Re: Doug's problem with the Maxtor XT3280S and the Supra ... I believe Doug when he says he tried my steps and "it didn't work." There are several possiblities: 1) he's not booting TWICE on a cold start. For some reason I've been unable to fathom, the Maxtor drive (and some other exotic drive I tried last year) requires a second reset ONLY AFTER A COLD BOOT before it will talk properly with most SCSI host adaptors/interfaces on the Amiga. Works fine on a Sun, but that's not the issue here. In other words: - Cold-boot your Amiga (gee, sounds kinky! :-) - After you notice the drive has NOT been "mounted" (by SupraMount), reboot leaving everything powerered up. In my startup-sequence I have an "echo" followed by a "wait 2" which provides a convenient (and safe) point at which to do the second boot; since I boot only once every several months or so, this is not a big deal (the systems are operated continously). - Notice the drive THIS time being properly mounted. 2) "someone" is setting the LUN (Local Unit Number) in software instead of the SCSI ID. Don't laugh! I've seen one person doing this last year and wondering why their drive didn't work. Perhaps now is the time to briefly describe a SCSI bus. Consider the following (rough) illustration of a fully-configured (and loaded!) bus: Interface Devices whose LUNs are: +----------++-----------+ | Your || SCSI host | SCSI +-----+ +---+ +---+ | Computer || Adaptor |==========::=|SID=6|-----| 0 |--- ... ---| 7 | | (Amiga!) ||SID=7,LUN=0| BUS || +-----+ +---+ +---+ +----------++-----------+ || || +-----+ +---+ +---+ ::=|SID=5|-----| 0 |--- ... ---| 7 | || +-----+ +---+ +---+ where: .. .. SID = SCSI ID [0..7] .. LUN = Local Unit Number [0..7] || +-----+ +---+ +---+ ::=|SID=0|-----| 0 |--- ... ---| 7 | || +-----+ +---+ +---+ || -/\/\/\- Bus Termination Given the proper interface at each SCSI ID, one could place up to, say, 8 disk drives on it, and with 7 interfaces, one could have, say, 56 disk drives on one's bus. This is what the SCSI bus and its specs permit. In reality, however, a interface typically will be designed to operate only one or two devices. For example: a) An embedded-SCSI disk drive (such as Seagate ST138N, Maxtor XT3380S, Quantum 80S, etc.) This is an integrated SCSI interface with disk drive controller. The SID should be set to a unique value on the bus, and the LUN is hard-coded to be zero (0). b) A (multi) disk-drive adaptor/translator card (such as Adaptec 4000 and 5000 series, Omti, etc.). This is a stand-alone SCSI interface card which is usually bolted by the end-user or system-integrator to the back of a disk drive (for convenience) and will adapt a non-SCSI disk drive (such as ST251, Maxtor XT1140 or XT2190, Miniscribe 6085, etc.) to the SCSI bus; such cards usually have provision to handle up to 2 drives whose LUNs will be 0 and 1 as determined by the connector on the interface to which one plugs the drive's cabling. c) Other embedded-SCSI devices, such as tape drives, laser printers, modems, page scanners, etc. Such devices permit setting a unique SCSI ID and will have a hardcoded LUN=0. d) Other adapter cards (such as NCR, Fujitsu, etc.) designed to adapt specific devices (such as floppy and/or tape drives, scanners, etc.) to the SCSI bus; these cards permit setting a unique SCSI ID and "may" provide multiple devices with LUNs ranging from 0 to 7. e) "Host adapter" cards. Such cards are the interface between a computer and the SCSI bus. One CAN have more than one "computer" on the bus, each with its own SCSI ID. Typically the LUN is hard-coded at zero. In a full-SCSI, arbitrating [de]-select setup (which I have NOT seen yet for the Amiga), the choice of a SCSI ID takes on additional meaning because it will determine the device's PRIORITY. The highest SCSI ID, 7, will have the highest priority on the bus. With such a setup, one would want to have the host adapter at 7, the primary disk drive at 6, other devices at 5 thru 1, and your tape backup system at SCSI ID 0. It is possible for devices whose IDs are, say, 5 and 4 to be "talking" to one another and exchanging data while the host is talking to, say, a tape drive at ID 0. Note that only ONE Initiator-Target pair has its data on the bus at any given instant, though. SCSI setups for the Amiga are (usually) non-arbitrating systems, and all bus control is under direction of the host adaptor (and its software), so the choice of SCSI ID isn't as important (except that all interfaces must still have a unique ID! :-). What all this is leading up to is that the address of a specfic device on a SCSI bus is determined by the combined SCSI ID and LUN. A device will only be selected and activated when it sees its "address" on the bus. If, for some reason, one is attempting to address the typical embedded-SCSI hard disk with a non-zero LUN, the drive will simply NOT respond. This is the situation I saw with one person's setup last year, and it "may" be the problem Doug is having. 3) Another set of potential problems concerns an improperly terminated SCSI bus, miswired cabling, and/or mis- or non-use of the SCSI terminator power line on the bus. SCSI bus wire 26 has been perverted by many manufacturers and users, and others have "done" non-standard things with other "undefined" SCSI bus wires. In a previous posting, I've detailed the SCSI bus lines that one should be SURE are not connected (improperly). I think this was also reprinted in one of the Amiga-related magazines, but I haven't seen it. In any event, among the mis-wiring offenders I've seen are included Supra, Seagate, Apple, etc. And a lot of "Apple" 3rd party disk subsystem assembly houses really screw the bus by having too long a "stub" off the bus (spec is 10cm MAX, and I've seen up to 40cm; sheesh!). 4) And a BIG problem with using Maxtor drives is the use of an inadequate power supply. These mongo drives require tremendous startup-surge capability AND very clean power during operation. Doug should look at both the +5 and +12 VDC lines on a 'scope while powering-up and running the drive. 5) And a final suggestion/idea concerns the specific Supra host adapter. For the A1000, there are TWO different 4x4 host adaptors; one has the dual 25-pin connector, and the other has the DB-25 (a la Apple). The DB-25 unit is the more recent, and I've noticed it has some minor circuitry changes (mostly enhanced or more bypass capacitors, etc.). He should be calling Mark White or Clark Stevens at Supra for help in this regards; Bill Seymour may also be able to provide some assistance since he's been putting a lot of Maxtor drives on Supra setups. All the above should provide some clues as to where to look for problems. When everything is assembled correctly, operation will be faultless. The SCSI busses on my Amiga systems extend to the full 20 feet (6 m) permitted by the SCSI spec, "talk" to 4 or 5 interfaces on each bus (which includes several Maxtors), and are rock-solid and operated 24 hours/day, 7 days/week. There are ALL Amiga A1000 systems using the Supra 4x4 Host Adaptor, and presently running Supra's version 5.3 software. Hmmm, Clark Stevens promised to send me the Series II software after DevCon, and I haven't received it yet ... :-( The interfaces I've personally verified as operating fine on the Supra-based SCSI busses on my systems include: 1) ALL embedded-SCSI drives I could get my hands on. This includes Fujitsu, Seagate, Maxtor, Imprimis, Connor, Quantum, CAST, etc. 2) Adaptec interfaces/translators. This includes the 4000A and 4070 used to operate Seagates, Maxtors, LaPines, Miniscribes, Microscience, Micropolis, Priam, etc. 3) Fujitsu M10008A (for driving ESDI tape drives) Thad Floryan [ thad@cup.portal.com (OR) ..!sun!portal!cup.portal.com!thad ]