Re: Hardware Guru for Hire?

I was not as clear as I should have been in my previous post.

A standard NEC 765 can do the 4 sector sizes that have been widely used (128, 256, 512, 1024 bytes). But the USB-to-FDC chip in question doesn't, taken as a whole, necessarily support everything that the 765 could do. PCs have only used the 512 byte sector size. The USB-to-FDC chip is a whole microprocessor PLUS an FDC chip WITH A FIRMWARE OPERATING SYSTEM. Even if the hardware component of this chip can do all 4 sizes (likely, but not an absolute given), the firmware in the stock chip might only do 512 bytes. There may be versions of this chip that support external firmware, and on a volume basis, custom firmware can be implemented and ordered. But if the discussion relates to the most common mask firmware stock version(s) of the chip, it may not be possible to implement sector sizes other than 512 bytes. I'm not saying absolutely that it isn't, I'm saying it's an area for concern/investigation.

Also, in the context of my response that certain things might not be "programmable", I was defining "programmable" as "programmable from the PC via the USB connection", and I was referring to the "stock" version of the chip, not a version with custom external firmware.

There is no doubt that what we want can be done (with this part or other parts), but there is very great doubt that it can be done with the stock, standard version of this chip and it's stock, standard internal firmware.


no.spam@xxxxxxxxxxxxxxxxxxxxxxx wrote:
On Thu, 19 Jul 2007 10:11:55 -0400, Barry Watzman
<WatzmanNOSPAM@xxxxxxxxxx> wrote:

Neither data rates nor sector sizes are programming issues. They both require hardware support in the FDC chip and associated circuitry and if it's not present, there is no software workaround.

Ok then. Last time I used a FDC I had to program it to do things.
Things like what sector size and how many just minor programming thats
all. That doesnt mean the FDC on board that beast can do it all but,
to get it to do anything requires programming. So variable sector
sizes, number of sectors, number of tracks, sides are programming
issues within the limits of the hardware. The chip being hardware and dumb as a stump without programming and host support programming
on the other end of the USB.

Allison


no.spam@xxxxxxxxxxxxxxxxxxxxxxx wrote:
On Wed, 18 Jul 2007 21:46:19 -0400, Barry Watzman
<WatzmanNOSPAM@xxxxxxxxxx> wrote:

That's a CHIP, not a finished product. There are a number of chips that
Yep ther anre many bot most are similar.

could form the basis of a finished product. However, while I'm not familiar with that particular one (and there are others), these were intended, for the most part, for use in PCs with 3.5" drives. So there are a number of issues to investigate:

-Is FM encoding supported (necessary for SSSD 8")
Maybe depending on data rate.

-Are various data rates supported (necessary for all formats other then 3.5" 1.44MB)
Programming issue.

-Are all 4 commonly used sector sizes supported (128, 256, 512, 1,024 bytes)? PCs ONLY use 512 bytes
Programming issue but may also be limited by ram buffer.

-Is the chip in some way operating system specific (to PC formats)?
Thats the USB driver issue, technically a programming issue but
constipated by the fact taht USB and the Miclyspooge OS are a PITA.
Likely worse if the OS is win2k or XP.

-How flexible is the chip? For example, will it support media that has totally different formats on different tracks (many early 8" double density disks had some tracks recorded in single density)
Again that is a programming issue.
-Will it support formats that the Western Digital chips (17xx series) could create but that the 765 series controllers are incapable of dealing with?
There lies the dirty dawg. Most use a 765 core for the floppy
controler and therefor it has no special properties over the average
PC chipset that uses the 765 core.

FYI: some of the 1771 formats can't be read even with the 1793!

Generally those chips are a cpu, ram, memory (eeprom/ram) and of
course the FDC and USB IO and hopefully DMA (for speed). What they do and how they do it is programable but... After programming the
chip you also have to have a driver that knows what the device can do
to exploit it. Programming a PC at the driver level for XP is likely
the least fun thing I can think of.

I don't know the answers to these (and other) questions, but not all of the chip products will work for a general purpose non-PC disk drive of possibly 5.25" 360k, 5.25" 1.2MB and 8" (various formats, many "oddball"). It's all part of getting from a chip to a finished product.

It's 765 based so if it doesn't work on most PCs over the last 10
years the USB thingie isn't likely to do it either.

The easiest and cheapest solution is to pick up and older PC and use that as the FDC in those supports two floppy drives and still has a drive bay that can accept a 5.25" drive. FYI:
the older the better (486 works too). Generally those machine are found on the curb working and for free. It will not solve 100%
of the problem but the cases that don't fly are things like some 8",
some 5.25" oddballs and of course any hard sectored stuff.
For an OS load up w98se as it does networking and is still close to
DOS and will fit easily on a 500Mb or larger disk.

Allison

craigm wrote:
Barry Watzman wrote:

I want one too.....

But as a hardware person (somewhat, but this is probably beyond me),
what you want is a LOT more than a cable.

Unfortunately, the 3.5" USB drives are all made as integrated devices
... they are not simply USB to floppy converters with a "standard" (e.g.
34 pin) 3.5" drive.

I want a USB to "standard floppy" controller, ideally it should support
5.25" drives (both 360k and 1.2M) and even 8" drives. I think there is
a market for it (not huge, but large enough to pay for development and
production). But no such device currently exists.


How close is this?
http://www.smsc.com/main/catalog/usb97cfdc2_01.html

.

Relevant Pages

  • Re: Hardware Guru for Hire?
    ... Neither data rates nor sector sizes are programming issues. ... They both require hardware support in the FDC chip and associated circuitry and if it's not present, ... constipated by the fact taht USB and the Miclyspooge OS are a PITA. ... The easiest and cheapest solution is to pick up and older PC and use that as the FDC in those supports two floppy drives and still has a drive bay that can accept a 5.25" drive. ...
    (comp.os.cpm)
  • Re: Hardware Guru for Hire?
    ... intended, for the most part, for use in PCs with 3.5" drives. ... Programming issue but may also be limited by ram buffer. ... constipated by the fact taht USB and the Miclyspooge OS are a PITA. ... chip you also have to have a driver that knows what the device can do ...
    (comp.os.cpm)
  • Re: Hardware Guru for Hire?
    ... Last time I used a FDC I had to program it to do things. ... Things like what sector size and how many just minor programming thats ... The chip being hardware ... on the other end of the USB. ...
    (comp.os.cpm)
  • Re: USB Cartridge ?
    ... That chip isn't really needed anymore since the USB pics came out. ... programming port connected to the programming pins on the the flash chip. ...
    (comp.sys.cbm)
  • Re: Structured Programming using Forth
    ... don't attempt to automate optimization at all. ... In your last reply to me, you disagreed with my statement that programming a SEAforth processor was going to be much more like assembly language programming than it would be Forth programming. ... Intellasys might take an different view-- that market acceptance of the SEAforth chip will require that they make it as easy as possible for engineers to use their chips. ...
    (comp.lang.forth)