Image Map Image Map
Page 2 of 5 FirstFirst 12345 LastLast
Results 11 to 20 of 47

Thread: IDE hard drive on an IBM PC Convertible

  1. #11

    Default

    I have totally forgotten about this thread. The idea of an IDE hard drive on a Convertible is exciting!

    The technical reference has a good description of the I/O channel:

    • Multiplexed low order address and data lines
    • High order address lines
    • Six IRQ lines
    • Memory and I/O read and write control lines
    • Clock
    • DMA


    At the falling edge of ALE the 8 shared address/data lines as address lines. When the memory read/write or I/O read/write lines are low the shared address/data lines contain data.

    Address lines 16 - 19 are also latched and are treated the same way. Their other signal is "status bits" but I can't find what those status bits are. There is nothing in the timing diagrams that indicates they are used for anything other than addressing, so I'm wondering if the book has an error where it says those lines are multiplexed with status bits.

    I need to work on my electrical engineering skills. As a software person I feel so isolated by all of this hardware.

  2. #12
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    33,719
    Blog Entries
    18

    Default

    There were a few ATA drives that could be jumpered as "8 bit", but I'm not sure exactly what was meant by it. Dig back in the archives here and there have been several discussions about XTA vs. ATA hard drives and how it isn't just a matter of swapping a few signals.

    If the item has been discussed several times before, I'm not going looking for it again.

  3. #13
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    2,062

    Default

    Quote Originally Posted by Chuck(G) View Post
    There were a few ATA drives that could be jumpered as "8 bit", but I'm not sure exactly what was meant by it. Dig back in the archives here and there have been several discussions about XTA vs. ATA hard drives and how it isn't just a matter of swapping a few signals.
    Here's a thread where someone did, with minor hardware modifications, turn an XTA port into a CF port using the XTIDE universal BIOS. Broadly speaking it looks like the only major hardware gotchya is the XTA port only has two address lines, while for a CF you need three. The I/O base is x320, which is no problem for the XTIDE BIOS out of the box.

    In any case, that supports the idea I was getting to, which is if the OP were willing to trace the schematic of this card it may take only minimal hacking to turn it into an XT-CF-Lite that would work on an IBM Convertible.
    My Retro-computing YouTube Channel (updates... eventually?): Paleozoic PCs

  4. #14
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    2,062

    Default

    Quote Originally Posted by mbbrutman View Post
    Address lines 16 - 19 are also latched and are treated the same way. Their other signal is "status bits" but I can't find what those status bits are. There is nothing in the timing diagrams that indicates they are used for anything other than addressing, so I'm wondering if the book has an error where it says those lines are multiplexed with status bits.
    That part left me scratching my head as well. There's no indication of what "status" they're meant to relay.
    My Retro-computing YouTube Channel (updates... eventually?): Paleozoic PCs

  5. #15
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    33,719
    Blog Entries
    18

    Default

    Quote Originally Posted by Eudimorphodon View Post
    Here's a thread where someone did, with minor hardware modifications, turn an XTA port into a CF port using the XTIDE universal BIOS. Broadly speaking it looks like the only major hardware gotchya is the XTA port only has two address lines, while for a CF you need three. The I/O base is x320, which is no problem for the XTIDE BIOS out of the box.

    In any case, that supports the idea I was getting to, which is if the OP were willing to trace the schematic of this card it may take only minimal hacking to turn it into an XT-CF-Lite that would work on an IBM Convertible.
    Read the whole thread, if you haven't already. This involved changing out the BIOS routines for accessing the disk. CF is capable of 8 bit transfers, but not so ATA IDE hard drives.

    Again, there is no such animal as an IDE ATA or XTA "controller", though same call it that.. What most call a "controller" is a bus adapter. Given that, (as I suggested 5 years ago), you can repurpose the signal lines on an XTA interface to talk to a CF card capable of 8 bit transfers--but it'll take a reworking of the BIOS code.

  6. #16
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    2,062

    Default

    I think we have been talking past each other.

    I am well aware that from a software driver standpoint an XTA hard drive is totally different from an ATA drive. If you notice I have been referring to the thing it hangs off of as a port, not a controller, and I have been pretty explicit that I dont think the original BIOS code on this adapter is going to be useful *. What I was interested in is if there was some critical electrical difference that would rule out repurposing the schematic of this device as the basis of a new-build XT-CF-Lite for this computer. (It seems to have a remarkably low chip count considering it includes latches but still must decode both a port address and a ROM chip.) And that thread seem to confirm its a viable plan with only very minor changes to the original circuit, assuming the XTIDE Universal BIOS code works on a Convertible?

    (Yes, most computers with XTA are stuck with the driver code in their system BIOS chips. We dont have that problem here, there is clearly a distinct ROM chip on that adapter card.)

    I was *not* talking about some kind of general purpose fudge to put CFs on existing XTA ports with XTA BIOS code.

    Edit: * Well, I tried to be clear since I started talking specially about the XTIDE bios, but Ill endeavor to put a finer point on it when touching third rails.
    My Retro-computing YouTube Channel (updates... eventually?): Paleozoic PCs

  7. #17
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    33,719
    Blog Entries
    18

    Default

    Okay, I think we have been talking past one another. To summarize:

    The critical thing here is can the device (whatever it might be) perform using an 8-bit data path exclusively. On an XTA drive, the answer is of course, "yes" as it is on the CF drive (part of the spec). So long as you have a sufficient number of address lines, you're good--the rest, as they say, is just software.

    Almost every ATA drive made since about 1990 does not have the ability to operate on an 8-bit data path exclusively, which is what makes the XTIDE adapter (multiplexing the data bus) necessary with rotating rust. Of course, it's possible to transfer only half-sectors, ignoring the other half, but then you've got the problem of dealing with 256 byte sectors and having to deal with blocking/deblocking to synthesize 512 byte sectors.

  8. #18
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    2,062

    Default

    Quote Originally Posted by Chuck(G) View Post
    Almost every ATA drive made since about 1990 does not have the ability to operate on an 8-bit data path exclusively, which is what makes the XTIDE adapter (multiplexing the data bus) necessary with rotating rust.
    That's what makes targeting the XT-CF-Lite support of the XTIDE BIOS so great, the driver work is already done to drive a CF without multiplexing. Forgetting spinning rust support seems like a reasonable trade-off, especially since we're talking about a portable machine here. The XT-CF-Lite boards I churned out for my Plus-bus Tandy 1000s run at over 300KB/s transfer rates (when limited to 8088 code, over 500KB/s with a V20) and work great with a 44 pin laptop IDE->SD card adapter that sells for about $8, so, yeah, who wants a real hard disk?

    Sigh. Now I wish I had an IBM Convertible to play with.
    My Retro-computing YouTube Channel (updates... eventually?): Paleozoic PCs

  9. #19
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    33,719
    Blog Entries
    18

    Default

    Related question: Do ATA-interface DOMs support 8-bit mode?

  10. #20
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    2,062

    Default

    Quote Originally Posted by Chuck(G) View Post
    Related question: Do ATA-interface DOMs support 8-bit mode?
    I've been wondering about that myself. I was going through the junk boxes in the garage the other day and I found a laptop 44-pin form-factor DOM that has a weird pass-through style header, I was thinking of trying it on one of my XT-CFs. (And then seeing if I could put a CF card adapter jumpered for slave after it if it worked.) But any results from that may not match up to the ones that come in desktop form factor packages.
    My Retro-computing YouTube Channel (updates... eventually?): Paleozoic PCs

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •