Xref: utzoo comp.sys.att:10233 unix-pc.general:5934 Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!wuarchive!zaphod.mps.ohio-state.edu!sdd.hp.com!decwrl!fernwood!portal!cup.portal.com!chad From: chad@cup.portal.com (Chad The-Walrus Netzer) Newsgroups: comp.sys.att,unix-pc.general Subject: Re: Unix PC and WD2010B Message-ID: <32953@cup.portal.com> Date: 18 Aug 90 08:41:06 GMT References: <1476@madnix.UUCP> <14873@csli.Stanford.EDU> <32819@cup.portal.com> Organization: The Portal System (TM) Lines: 170 In a previous article (Thad Floryan) writes: )If he hasn't already (postings are slow to return back to PORTAL), Chad Netzer )will post some preliminary findings re: WD2010, a Miniscribe 3085, and several )revisions of the UNIXPC motherboard. In summary, all of the 3085, the WD2010, )and the motherboard(s), are good and OK in and of themselves, but some combos )simply do NOT work. The highest motherboard rev level tested so far is ``C''. Here are the preliminary results of my testing: Equipment: WD2010B disk controller chip WD1010A disk controller chip ATT 7300 w/ motherboard upgrade to 2 megs (Rev B. Motherboard) ATT 7300 (Rev C.) motherboard Miniscribe 3085 hard disk Miniscribe 10 meg drive (model number unknown, came standard with certain machines) UNIX 3.51m operating system Enhanced diagnostics disk with support for larger drives Basic Problem: WD2010 should allow formatting and use of drives with greater than 1024 cylinders. Should replace WD1010 on motherboard with a simple chip swap. However, WD2010 does not work correctly in all situations (detailed below). Description of tests and results: After the failure of my WD2010 to work properly, I did some simple tests. By swapping the WD2010 and the 3085 drive under all possible configuarions (with available hardware), I was able to chart which combinations worked, and which didn't. This chart is detailed below. WD1010A | WD2010B ------- | ------- Rev B Motherboard | w/ Stock 10 meg WORKED | WORKED drive and 2 meg | ram. | ----------------------------------------------------------------------------- Rev C Motherboard | w/ Stock 10 meg WORKED | WORKED drive and 512 K | ram. | ----------------------------------------------------------------------------- Rev B Motherboard | w/ 3085 drive WORKED | * FAILED * and 2 meg ram. | ----------------------------------------------------------------------------- Rev C Motherboard | w/ 3085 drive WORKED | * FAILED * and 512 K ram. | ----------------------------------------------------------------------------- Symptoms of failure: The symptoms of failure remain consistant across motherboard revisions. Upon booting up on systems with both the WD2010 and Miniscribe 3085 installed, when trying to boot off the hard disk, which prints out the "Searching for Hard disk" message followed by printing of successive asterisks, it will silently fail and simply start to boot off the floppy disk again. It doesn't reset the machine or clear the screen. It simply skips down the screen and tries to boot off the floppy again, then the hard disk, then the floppy, etc. over and over. The failure when booting off the hard disk occurs at different times and is not particularly consistent. Occasionally it even gets past the asterisks stage, but fails later. Note that this is version 3.51m of the operating system, and it does correctly print out that a WD2010 is installed when started up. If I proceed by booting up the diagnostics disk, and doing a hard disk subsystem test, an initialize, or any disk operation, I get (variations on) this error message: "Subtest: format" "Error: Winchester: Can't Recal: Response = 10" or: "Subtest: read all the disk" "Error: Winchester: Error on check read: Response = 10, startblock = 10" The "Response = 10" is pretty consistent. Also, when formatting the drive, it actually appears to go through all the cylinders and initialize them, but after the last cylinder, the above type error messages appear when attempting to do a surface test. But the format itself seems to work. Summary of tests: The tests show that the two cases where the combinations didn't work were with both the WD2010B and Miniscribe 3085. This may lead one to believe that these two items are incompatible, but I know of cases where the combination of WD2010 and Miniscribe 3085 do work perfectly together. (As Thad Floryan will attest to.) Also, I have tried with two separate WD2010B chips, and they both exibited the exact same symptoms. I have not tried it with either the WD2010A or WD2010C chips. Speculations and general notes: Because I know that the WD2010 and 3085 are not inherently incompatible, I suspect the problem to be due to an obscure motherboard bug that was fixed in later revisions. I would therefore like to know what motherboards are being used in those machines that have been able to get the WD2010 and 3085 to work. It may be due to an incompatability between the "B" model of the WD2010, which may not be present in other moodels (ie. the "A" or "C") It may also be due to improper setup of 3085 somehow. It came with only one jumper, the drive select, which didn't work when set to the (shipped) default of "2". Setting to one fixed the problem (before installing WD2010). Do any of the other jumper selections on the drive need to be set (highly unlikely I suspect). Perhaps it is a problem with the power supply not giving enough output. But since the 3085 (should) use less power than the older drives, I would suspect the failures to be reversed if that were the case, and it should also fail with the WD1010A. Request for help and experiences: If any of you have working, (or non-working) combinations of the WD2010 and the Miniscribe 3085, please E-mail me your motherboard revision if you know it, and anything else you might consider helpful. I initially want to try and get as wide a range of data on which motherboards the combination will or won't work. If there is a pattern, it may help us find a fix for ALL cases. Also indicate what rev. level of the WD2010 chip you are using, ie. "A", "B", or "C". This letter should appear after the number "2010" on the chip itself. I have tested on Rev. B and Rev. C motherboards, with only the "B" chip so far. In the future, I may be able test with the "A" chip, on higher Rev. level motherboards, once Thad Floryan has more time to spare, and his Lab is restored from the fire damage. This may yet be quite a ways off if ever, however, and the sooner we start tabulating data, the better. So please send in any facts, or suggestions you might have. If the problem ever gets fixed, it WILL be made available for all. Send all correspondence to: chad@cup.portal.com (until the end of September) chad@slugmail.ucsc.edu (From October onwards) PS. As an aside, I noticed someone offering to sponser a group buy for the WD2010 chips, for about $50 apiece. Just to let you know, I got my first WD2010 off an IBM XT hard disk controller which cost about $40 list (that is for the whole controller board, from which you can desolder the chip like I did.), although I got the board for free. I got the second chip at Ace Electronics in Santa Clara, Ca. for $11, and they had about 20 or more at the time. The stock was rapidly depleted by a person who will remain nameless :-), But the point is that they should NOT be that hard or expensive to get if you just shop around, even if you have to avoid the "mainstream" outlets... Just some friendly advice for prospective buyers. If you are interested, the number of Ace Electronics is: (408) 730-4660 (408) 730-2145 FAX ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Chad Netzer -> chad@cup.portal.com