jaclaz

Really I cannot understand what is the difficult part in this. As said n times this Sparkfun thingy is a peculiar board with a peculiar behaviour. Another board (including yours) may behave in a completely different way, but this one, from the posted reports, behaves EXACTLY as advertised. And again with all due respect, you are missing some logic. Let's see, we are now talking of your self built interface: You have in the above THREE statements: MAX232 powered with +5V: MAX232 transmits @5V HDD transmits @3.3V #1 is a "real" statement, as I preeume you know what you power your interface with #2 is EITHER an assumption OR you measured the signal on the board TX wire #3 is - by your own words - an assumption based on your trust in what I said Now, logic tells that if I say two things, you cannot trust my word for only one of the two, you have to choose if EITHER believe BOTH or NONE, i.e. I'm EITHER reliable or I am NOT. On the other hand the same logic says that if you actually measured the signal on the board TX wire (and thus #2 is NOT an assumption) you have the means (oscilloscope or signal analyzer) to measure the signal levels, if it is so , you can well measure the signal on the HD TX pin and thus need not to trust me, and need not to assume #3. I have NOT the faintest idea at what level works what, I simply draw sums out of data. From this, which remains probably the only thing I can understand : http://www.interfacebus.com/voltage_threshold.html A "proper" TTL (@3.3V) will understand a signal at anything over 3.3 V as "a suffusion of yellow". It is possible that the whatever is on the stoopid seagate disk understands something more than this let's say 4V and your particular interface sends data at a lower level than the maximum 4.7 V that the (@5V) TTL/CMOS allows, say 3.99999 whilst the Sparkfun thingy, if powered at 5 V actually sends signals at 4.6999999 V. Or it is possible that for any reason your interface actually produces 3.3V level signals and thus works allright. Without actual measurement of signals of BOTH your circuit and of a sample of the Sparkfun thingy it's hard to say. UPDATE: After all maridadi is NOT using the RS232 SMD shifter, but rather a Prolific USB to TTL adapter: jaclaz

Why? Loop-back #1 has a (5v) signal exiting a (5v) board and looping back, thus entering a board capable of understanding a (5v) signal. OR: Loop-back #2 has a (3.3v) signal exiting a (3.3V) board and looping back, thus entering a board capable of understanding a (3.3v) Nothing strange. In loop-back test: Board transmits @5v->loop-back connection->loop-back connection->Board understands @5v OK Board transmits @3.3v->loop-back connection->loop-back connection->Board understands @3.3V OK In real life: Board transmit @5v->hard disk DOESN'T understand @5V->a suffusion of yellow -/- Hard disk trasmits @3.3V->Board DOESN'T understand @3.3v-> a suffusion of yellow NOT OK Board transmits @3.3v->hard disk understands @3.3V->hard disk transmits @3.3V->Board understands #3.3V OK After all this is called TTL where the L is for "Logic" . jaclaz

Just to avoid further misunderstanding , this: Actually means: The GND of the shifter to BOTH the GND of the harddrive AND the black wire (ground) of the power source (be it a 3V battery or wires coming from a power supply, Orange +3,3V and Black 0V or ground) Please review point #7 of the read-me-first jaclaz

But, did you use the 5V with your home made (officially "super-deluxe" ) thingy: or with this specific board "Shifter" from Sparkfun?: http://www.sparkfun.com/commerce/product_info.php?products_id=449 Maybe the$#!Za. (at least from the name he could be) is an undercover agent form SPECTRE attempting to upset world order by underpowering the stoopid shifter : and, later: Let's see if I can clear it. There are two different "TLL" levels standard, a "5V" and a a "3.3V" (sometimes called "2.8V") one. http://www.interfacebus.com/voltage_threshold.html The hard disk "wants" or "likes" the 3.3V ones. A converter can be: powered at 5V and output a 5V TTL level signals <-this WON'T work powered at 2.8V÷3.3 and output 3.3/2.8V TTL level signals <-this WILL work powered at 5V and ouput 3.3/2.8V TTL level signals, or more generally powered at ANY voltage and output a 3.3/2.8V TTL level signals<-this WILL work powered at 5V and ouput BOTH 5V and 3.3/2.8V TTL level signals, or more generally powered at ANY voltage and output BOTH 5V and 3.3/2.8V TTL level signals<-this WILL work if the HD is connected to the "right" TX/RX" of the converter That particular converter you used/got "autoswitches" output TTL level based on the voltage it is powered with. In other words it has an "advanced" feature that turned out as a problem in your case. With all due respect for your "deluxe thingy", it is perfectly possible that it works at 5V, or even at 4/9*SQ(PI) , but I presume that the Sparkfun board being an industrial product, ALL single boards of this model behave the same and "as advertised". Right now we have: Gradius2 (you know the peep you initially thanked as his guide saved your a**) using this adapter with a 3V battery Brad Garcia (you know, one of the other peeps that mostly contributed to save a**es around) using a plain adapter with a 3.3V supply (orange cable) the$#!Za. reporting specifically that this adapter did NOT work for him when powered at 5V BUT the same (this) worked when powered at 3.3V Everyone remains of course perfectly free to draw his/her conclusions, and either act accordingly or supply any voltage into any TTL converter, actual results may vary , but I won't buy that this adapter will ever work if powered at 5V. jaclaz

VideoRipper might have been a bit too laconic. I am pretty sure he meant without anything, but with the "shifter SMD" actually powered. Or is it supposed to get 3.3 V (NOT 5V) from thin air? Please note that theoretically the loopback test carried with: NO POWER -> will fail +5V -> will (hopefully) succeed BUT won't work with the hard disk +3.3V -> will (hopefully) succeed, AND will (hopefully) work with the hard disk Anyway, I give up, I see it's of no use. jaclaz

The shifter SMD NEEDS to be powered at 3.3 V, as I already pointed out to you on the "main thread", here: AND also (I hope clearly) stated on the read-me-first, point #11: WHAT is the difficult part in it? WHY starting a new thread? jaclaz

@lesmond74 Just for the record the given link is OK: http://web.archive.org/web/*/http://www.v72735.f2s.com/LetAssig/ It leads to a page on which you can choose between several snapshots of the page, just click on last one: http://web.archive.org/web/20080619125702/http://www.v72735.f2s.com/LetAssig/ and see if the file is available (it is ): http://web.archive.org/web/20080401085803/http://www.v72735.f2s.com/LetAssig/LetAssig.zip or try earlier snapshots. To disambiguate possible mis-reporting of the actual "placement" of the partitions, though it seems like clear enough, and you should wait for rloew's patch, can you in the meantime get from here grub4dos: http://nufans.net/grub4dos/ http://nufans.net/grub4dos/current_release/ http://nufans.net/grub4dos/current_release/grub4dos-0.4.4-2009-10-16.zip From the zip get just grub.exe, boot to plain DOS and run it. At the grub prompt issue command: geometry (hd0) [ENTER] and confirm that the output corresponds to: (hd0,0) C:\ - Win98 SE, FAT32, 25GB, Primary (hd0,1) F:\ - Win XP, FAT32, 15GB, Primary (hd0,2) Linux, Ext3, 10GB, Primary Extended partition containing the following logical volumes - (hd0,4) Linux, Swap, 512MB (hd0,5) Linux, Ext3, 35GB (hd0,6) E:\ - FAT32, 25GB @dencorso The dlmanip page is online : http://sta.c64.org/dlmanip.html no need in this case for the wayback machine... jaclaz

I have nothing much to add to what I wrote in points #6 and #7 of the read-me-first, but I can add the following data collection from a statistical viewpoint. (actual data for the "those that did NOT use ground is obviously faked , but anyway representative of actual values ): Group A (those that did connect properly the ground): Percentage of success 100% Problems with connection 0% Whining about the guide/tutorial 0% Group B (those that did NOT connect properly the ground): Percentage of success 56.65% Problems with connection 34,35% (of which solved by connecting ground 99.99% +/- 0.01%) Whining about the guide/tutorials 34,35% In other words, we have NO reports of people having problems with the ground connection, and a significative number of reports about people that had problems with the connection that were INVARIABLY solved once ground connection was established. Since you have that connector available, and there is NO evidence that it creates a problem, why NOT connecting it? Or, let's put it in yet another way, try without the ground connected, if it works, good , if it doesn't try again after connecting the ground, and hope that the problem was just the ground . Grounding is, as said STRONGLY advised, which doesn't mean that is COMPULSORY, not that the stoopid hard disk cannot be revived without a ground connection, it simply means that if I were you I would connect the ground too, with a total surcharge in costs of 0.00 US$ AND no harm done to any living being in the making of. jaclaz

You never had any issues with it integrating something into NT 4.00? http://www.nliteos.com/nlite.html Integrating NT 4.00 Service Packs is: NOT possible with nlite NOT possible (actually possible but NOT "easy") Here : http://www.msfn.org/board/index.php?showtopic=109439 jaclaz

Yep, the FTDI PL303 chip is said to be very reliable. You won't need an external power source, NO. It is STRONGLY recommended to use GROUND, ALWAYS! READ points #6 and #7 in the read-me-first: Don't you wonder WHY that adapter has THREE leads (Tx/Rx/GND)? Or do you think they haven't thought of saving 0.01 US$ and give you just Tx and Rx? Should you want TWO of them and with also a +5V line: http://cgi.ebay.it/2x-USB-auf-Seriell-Adapter-Kabel-Konverter-TTL-RS232-PC-/320550268639?cmd=ViewItem&pt=DE_Technik_Computerzubeh%C3%B6r_Kabel_Adapter&hash=item4aa248eedf Seriously, only to confirm that the PL2303 is a known "good work horse". From the photo on your link it seems not like a 2.54 mm plug, but rather single connectors, that you can use allright, all you will need is (maybe) a pair of pliers to "tighten" them a bit... This is a Dupont connector housing 2.54: http://cgi.ebay.it/50pcs-Dupont-Connector-Housing-Female-2-54mm-2x5P-/140368742795 i.e. it's a "housing" for the single wire connectors. If you want an actual 2 mm pitch plug, you can have it with a *small* surcharge : http://cgi.ebay.it/INTERFACCIA-USB-3V-TTL-per-HD-Seagate-7200-11-no-rs232-/190387498311?cmd=ViewItem&pt=Accessori_per_Radioamatori&hash=item2c53faad47 Sometimes I cannot believe it, the Italian text: means: jaclaz

Naah, that's as said - a "standard" setup, compatible with *everything*. Not at all to put down the nice app by rloew, but it repesents an "advanced tool" to do "advanced things". If you decide to go "standard", you can use "standard" (FDISK) or "super" (RPM) and need not "hyper" (RFDISK), just like to commute every day to work you may use your current car or (maybe ) a Porsche 911 , but don't *need* a Koenigsegg CCR or CCX, (mainly because you wouldn't be able to spell it's name ) jaclaz

How many "Rs232 shifter SMD" are there around? http://www.sparkfun.com/commerce/product_info.php?products_id=449 @maridadi Check ATTENTIVELY these (about 15 posts starting from here) : http://www.msfn.org/board/index.php?showtopic=128807&st=2326 AND points 6 to 10 in the "reame_first": jaclaz

Just as a recent prove that Murphy's Laws do apply to non-standard setups : http://www.msfn.org/board/index.php?showtopic=144947 jaclaz

Is the internal drive SATA? Have you integrated the proper driver? Have you a setting in BIOS for "IDE emulation mode" or something like that? Try removing the flash card, most probably you will have a BSOD 0x000007b. Compare with: http://www.boot-land.net/forums/index.php?showtopic=8758 jaclaz

Basically you created a Primary partition on which you used a filesystem that is called "Extended filesystem" and you created a second primary partition with a filesystem readable by Windows 98. It is highly probable that those partition create the issue. Workarounds are (cannot say WHICH one may work for you): Hide the 2nd primary with XP on FAT32 and the third primary Ext3FS when booting DOS/Win98 Convert the XP partition to NTFS (thus making it "not understandable" by Win98) AND hide the third primary Convert the XP partition to NTFS only Try Letter Assigner: http://web.archive.org/web/*/http://www.v72735.f2s.com/LetAssig/ Try the patched IO.SYS: http://www.msfn.org/board/index.php?showtopic=118119 Try the above in reverse order, if the first partition is "big enough", it is possible that the partition with XP falls in the cases where #5 will work. More generally, to avoid this kind of problems, the rule of thumb is to use ONLY one primary partition with DOS/Win9x and all the rest as logical volumes inside extended, both NT based systems and Linux OS were designed to be installed on logical volumes. jaclaz

For the record, having two primaries is mostly pointless, as a logical volume inside extended is also safer (from dumb virus or most partitioning programs errors): http://www.msfn.org/board/index.php?showtopic=33964 And with some tweaks you can even install Windows 9x on a logical volume. (no don't do it) The point as I see it is: the most basic "standard" is having a SINGLE primary partition (formatted with a filesystem DOS 7.x recognizes, i.e. FAT16 or FAT32) the next "standard" is having a SINGLE primary partition and an Extended partition with how many logical volumes (within the limit below) you want/need in it, optionally having other two Primaries formatted with filesystems NOT recognized by DOS 7.x/Win9x/Me ANYTHING different from the above is NON-standard, and although it may cause NO problems whatsoever in all the life of that hard disk/win 9x install, I don't see the utility of doing something non-standard UNLESS you actually *need* some features that the "standard" does NOT allow. Just for the record, any Windows NT based system AND any Linux based systems are designed to be installed on Logical Volumes inside extended. I personally have been partitioning along standard #2 (standard #1 is simply inadequate) on "production machines" since more than 15 years, and had no problems whatsoever. Of course if the scope is to experiment, you are very welcome to try anything, though if this is the scope, "two primaries" is so common that it cannot be classified in the "new test" section. Of which roughly 976 NOT accessible with a drive letter... Norton 360? http://www.msfn.org/board/index.php?showtopic=140499 http://www.911cd.net/forums//index.php?showtopic=23902 jaclaz

@jorgeA FORGET (temporarily) whatever you have read. Read THIS: http://web.archive.org/web/20080726065134/http://www.ranish.com/part/primer.htm (I am linking to the Wayback Machine as the curremnt page seems like not having images anymore ) It seems to me the clearest expalanation around. About the historical part, when it all began, there were NO partitions. Then the good guys found a way to fit 4 partition entries in the MBR (first 512 bytes sector of the disk/device). Then they saw that 4 partitions could have been too little a number and invented the Extended partition and Logical Volumes inside it. The difference is simple: MBR contains a Partition Table where each entry points to a given address on the disk. In the case of a Primary partition this address contains the actual partition (i.e. "something" that starts with the bootsector). In the case of an Extended this address contains a pointer to ANOTHER pointer, this latter pointing to the address where the actual Logical Volume or parition (i.e. "something" that starts with the bootsector) resides. Primary: MBR->Entry in PT->Primary partition Logical Volume: MBR->Entry in PT->Extended partition->EPBR->entry in PT->Logical Volume @Mijzelf JFYI, there are ways to boot from Logical Volumes inside Extended (by correcting the "sectors before" data in the bootsectors) and with Letter Assigner you can change pretty much everything related to default drive lettering. There is not an actual "standard" for partitioning and formatting, most of the work has been done based of a certain set of assumptions derived by the way the original FDISK program and the various IBM/MS originated filesystems worked. The problem just like any non-standard setup (i.e. the same reasons I pointed out for Partition Magic or System Commander) is that you never know if and when you will run some other program that have a different set of assumptions about how the drive should have been set. For the record, even MS itself has not very clear ideas, compare with: http://www.boot-land.net/forums/index.php?showtopic=9897 jaclaz

Almost, but not quite : For option 1 you need nothing but the FDISK you already have (though RPM is a very good tool that can be used INSTEAD of FDISK) For option 2 you need nothing but the FDISK you already have (though RPM is a very good tool that can be used INSTEAD of FDISK) For option 3 you need any of the freewares capable of resizing (RPM and Partition resizer are examples) For option 4 you need any of the Commercial tools capable of resizing (Partition Magic Acronis and Partition Doctor are examples) RFDISK s Commercial/Shareware, cannot say if it is usable as option 4, it may have been option 0 (as it can partition regardless of BIOS settings ) but since your BIOS proved to be allright, you don't need option 0. I.e. you choose between 1 or 2 or (3 or 4). 1 is "how things should be" 2 is "how things can be" 3 is "how things should be if everything works" 4 is "how things should be if everything works after having parted from a few bucks" Up to you the choice. jaclaz

@jorgeA Still, I guess we are talking about two different things. There is NO doubt that the original Toshiba geometry is the one rloew gave: 29,264*16*63=29,498,112 You can enter the above AND re-partition the disk with FDISK AND re-format and re-install allright. (OPTION 1). OR (if you wish NOT to re-partition the disk and re-format and re-install) you can try using the 31,214*15*63=29,497,230 and adding an extended partition and one or more logical volumens in it. (OPTION 2) Since 29,497,230 is LESS than 29,498,112, there should be NO problems except losing exactly 29,498,112-29,497,230=882 882*512=451,584 bytes of capacity. Nonetheless, checking that the last sector is accessible is advised, and it seems from your edit like allright. Now you only have to choose between the available options 1 or 2, but also option 3 (using one of the suggested freeware apps) and option 4 (needlessly using a commercial software you don't own a licence for), the latter two being deprecated, obviously #4 being MUCH MORE deprecated than #3. jaclaz

Cleaning the lens and using a pad usually fixes it: http://www.ehow.com/about_5366556_mouse-pointer-move-itself.html There are reports about the actual LED colour/wavelength, but I wouldn't go as far as that: http://reviews.ebay.com/Microsoft-Basic-Optical-Mouse-Drift-Fix_W0QQugidZ10000000004119253 From the mouth of the wolf: http://www.evoluent.com/support.htm jaclaz

I guess this must be cleared once and for all. Partition Magic ("good" versions are 2.0, 3.0 -but these are limited to the 8 Gb HD barrier - and 8.x - if you have 4,5,6 leave them alone) and Partition Table Doctor and Acronis are COMMERCIAL apps. I wouldn't spend the money for a license for any of them (and yes, I am notoriously cheap , to do something that: 1) is NOT *needed* 2) can be done anyay with freeware apps This said, BOTH Partition Magic AND Acronis apps have some UNDOCUMENTED very little quirks (like writing in some cases "non-standard CODE in the MBR/bootsector AND writing WITHOUT any warning to some hidden sectors of the HD). @JorgeA The main problem that we still have to solve is whether 31214 (or, say, 31213 or 31215 ... you get the idea) is an accurate value or not. If the highest accessible cylinder is 31215 having the BIOS set to 31214 won't do any harm and you'll simply lose a very little part of the overall capacity of the drive. On the other hand if highest accessible cylinder is 31213 you have potentially a system that may crash ANY time. Most probably your best bet is to set in BIOS a higher value, say 31250, then boot to DOS (NOT Win9x!) and try accessing latest sectors with PTS Disk Editor: http://thestarman.pcministry.com/tool/FreeTools.html Stop when you have an error. jaclaz

Good. Now I am completely lost again. The fact that chainloader /io.sys worked and and chainloader +1 only means that you either failed to establish root properly or your bootsector is NOT a Win9x one or it is NOT properly set. It is very likely that you need something "more" to have Win898 working from second drive, but from a quick look you performed such a number of changes/experiment ALL TOGETHER and without properly detailing them that I have NO idea what is causing the problem and even cannot say how it can be solved. If you are happy with WinME, I would consider the thing closed, otherwise please post again how you have setip the thingy with ALL details, including menu.lst and the ususal MBR's/bootsectors. It is very possible that the problem is originated by the primary FAT on 1st drive, if this is the case, try making it a logical volume inside extended and try again. jaclaz

Since the hard disk is "designed" with a 16/63 geometry, if it works on that BIOS, it would be better. All roads lead to Rome, but some are straighter/shorter! Try with the 29264x16x63. Just in order to say "I had told you" : The actual point now is to understand exactly how many cylinders you have actually available, I suspect that if you manually give a "higher-than-real" number of cylinders, everything will work UNTIL you get to actually write access one of those non-existing sectors. (This usually happens when you type in the "D" of "THE END" of your great literary work - the one that took you 15 years to write and of which you have NO backups - usually it happens on Saturdays, around 11:59 PM ) About the partitions, personally I would use FDISK to create an Extended partition and in it a Logical Volume (or more than one) that I would label "DATA" leaving your current primary as-is. I am personally not much a fan of ANY partition resizing software. Unless I am mistaken this nice thingy here: http://partitionlogic.org.uk/about/preview.html does not (yet) allow for FAT16/32 resizing. Good ol' RPM or Partition Resizer http://www.thefreecountry.com/utilities/partitioneditors.shtml may work, though. jaclaz

ACPI.SYS is a driver for particular (now pretty much standard) ACPI hardware. Is it possible that your PC mainboard is NOT ACPI compatible? Or that it needs some additional drivers slipstreamed? http://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface Which settings have you tried in BIOS? Some BIOS do have a setting for "Plug and play OS" or "ACPI" that may affect the installation. Unfortunately often BSOD's messages are not accurate and point to something whilst the actual cause is something completely different. Try using F7 to disable ACPI while booting: http://www.theeldergeek.com/clean_installation_of_windows_xp.htm You can later re-enable ACPI, sometimes XP is "picky" during installation, but after it works allright. jaclaz

Those are relative to the partitioned/formatted part. Boot to plain dos and run FDISK. Check what you see in option 4. You can then try 31214x15x63. jaclaz