Jump to content

Photo

Modifying 810 drive


16 replies to this topic

#1 doctorclu OFFLINE  

doctorclu

    ******Blue Max****** *****Class 4*****

  • 5,705 posts
  • (Bubsy Bobcat fan)
  • Location:Dallas, TX - U.S.A.

Posted Fri Apr 20, 2012 7:27 AM

I was looking to get a 815 drive partly because it is double density. However, I think this might be hard to find.

But got me wondering looking at the front of the two... is it possible to change out the drive mechanism is the standard 810 to make double density?

Thanks!

#2 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Fri Apr 20, 2012 7:40 AM

You'd need to change the controller chip to one that supports double density (MFM) mode but you could use the same mechanism. Basically, you'd need to make it more like a 1050.

I've used an old Tandon Full-Height 5.25" drive with the XF551 circuit board, so I know it's possible to hack together a 360K 810. Someday I'd like to build it as a simple upgrade kit.

#3 mimo OFFLINE  

mimo

    Preppie!

  • 6,778 posts
  • It's easy living in a bubble

Posted Fri Apr 20, 2012 8:06 AM

I asked this question a while ago, I believe that it is not possible, at least a simple solution does not currently exist. Unless you can find the 810T Turbo mod :(

Third-party upgrades for the 810:
810 Archiver -- copy many copy-protected programs
810 Fast Chip by Binary 10%-40% faster
810 Turbo (810T) by Neanderthal Computer Things (NCT) -- double-density,
track buffering, speed comparable to Happy 810 Warp Speed
Happy 810 -- Happy Backup, Warp Speed 52Kbps, 18 sector buffer

#4 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Fri Apr 20, 2012 9:08 AM

The easiest path would be to make a plug-in board that simulates a USDoubler 1050. The board would need a controller, some SRAM and a new ROM.

#5 mimo OFFLINE  

mimo

    Preppie!

  • 6,778 posts
  • It's easy living in a bubble

Posted Fri Apr 20, 2012 12:06 PM

The easiest path would be to make a plug-in board that simulates a USDoubler 1050. The board would need a controller, some SRAM and a new ROM.


I'll buy one ;)

#6 Defender II OFFLINE  

Defender II

    Dragonstomper

  • 951 posts
  • Location:Traveling through space & time

Posted Mon Apr 23, 2012 4:44 AM


The easiest path would be to make a plug-in board that simulates a USDoubler 1050. The board would need a controller, some SRAM and a new ROM.


I'll buy one ;)

Me too!

#7 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Mon Apr 23, 2012 10:52 AM

I looked at this some more... The 1771 and the 2793 have very similar pin designations. If would be great if the upgrade could be as simple as a new ROM and data separator board. The biggest remaining problem would be the half-speed 6507 CPU (500KHz vs. 1MHz). This would shift a lot of work to the software side of re-writing the US Doubler ROM for a slower CPU (if that's possible at all).

Wasn't there another thread where someone mentioned an effort to support DD using the 1050's existing 256 bytes of RAM?

#8 31336haxx0r OFFLINE  

31336haxx0r

    Dragonstomper

  • 515 posts
  • Location:Germany

Posted Mon Apr 23, 2012 12:32 PM

You could implement a frequency doubler circuit.

#9 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Mon Apr 23, 2012 1:00 PM

Well, I was thinking of ways to avoid a soldering requirement and keep the upgrade as simple as possible. There's a 1MHz clock that's already fed to the 1771 chip, and it's divided by 2 for the CPU. So, it can probably be done with what's already there if you're willing to hack the board a bit.

#10 31336haxx0r OFFLINE  

31336haxx0r

    Dragonstomper

  • 515 posts
  • Location:Germany

Posted Mon Apr 23, 2012 1:41 PM

Hmmmmm a simple wire might suffice.

#11 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Mon Apr 23, 2012 5:15 PM

Turns out it's not going to be much of a problem. On data separator equipped 810's, the clock comes from the DS itself so it may be fairly easy to feed a faster clock back into the board.... still investigating...

EDIT: Nevermind, it seems the feature isn't used. A202 on the DS is for a clock divider for the 4MHz DS clock so the clock circuitry can be omitted from the sideboard, but on my 810's it looks like the 1MHz clock circuit is installed.

EDIT AGAIN: It looks like the board should work if it plugs into the CPU socket and the DS/1771 socket. The on-board clocking can be rerouted and a 4K EPROM can be supported.

EDIT #3: Also, an upgrade to the 810 should take care of the disk-glitch problem. Perhaps it would be better to offer it as a option for any 810 since it will most likely go on the analog board. It looks like the glitch problem comes from the way the 4052 head driver is used. It's INHIBIT circuit doesn't protect much on power-down.

#12 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Wed Apr 25, 2012 6:26 PM

I've started disassembling the USD ROM. Once I get it commented, I'll see about patching it to work with an 810 mech (no READY switch, no T00 sensor, etc...). Then, I'll build a proto plug-in board for the WD2793 FDC.

If anyone's already done this (the disassembly work), please let me know!! Once I get it far enough along, I'll post what I have and maybe others can help.

#13 31336haxx0r OFFLINE  

31336haxx0r

    Dragonstomper

  • 515 posts
  • Location:Germany

Posted Thu Apr 26, 2012 1:06 AM

Sorry for the stupid question, but the 810 has no track 0 sensor? :-o

#14 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Thu Apr 26, 2012 5:25 AM

Many early disk drives would simply step the head backwards 40 times to guarantee the head had arrived at T0. This causes these drives to make some weird noises as they banged the head against an end-of-travel barrier (usually an adjustable screw). They usually do this (1) at Power Up to guarantee the starting position of the head and (2) after a read error to make sure the head is still on the right track.

The down side to not having a T0 sensor (other than the occasional grinding noise the drives would produce) is that pushing the head against a barrier over and over could cause it to go out of alignment over time. Commodore 1541 drives were notorious for this.

#15 31336haxx0r OFFLINE  

31336haxx0r

    Dragonstomper

  • 515 posts
  • Location:Germany

Posted Thu Apr 26, 2012 10:13 AM

Yes I know, that's why I was asking. I've never seen let alone operated a 810, so I didn't know it did that. I remember that awful staccato of my 1541 very well...

#16 Fox-1 / mnx OFFLINE  

Fox-1 / mnx

    Stargunner

  • 1,188 posts
  • What is your Alternate Reality?
  • Location:NL, Earth 2.0

Posted Thu Apr 26, 2012 11:57 AM

If anyone's already done this (the disassembly work), please let me know!! Once I get it far enough along, I'll post what I have and maybe others can help.


I don't know if it's useful for you but I've attached what I have. It includes a source (MAC/65 format) with some named labels. (I'm not sure if the assembled ROM file in the ZIP is actually made using this source)

USDBLR.ROM = U.S. Doubler 1050 ROM
Copyright 1994 Fte
© 1984 ICD

CRC32: 605B7153



Attached File  US Doubler.zip   9.65KB   71 downloads

Edited by Fox-1 / mnx, Thu Apr 26, 2012 12:02 PM.


#17 Bryan OFFLINE  

Bryan

    Quadrunner

  • 8,614 posts
  • Cruise Elroy = 4DB7
  • Location:Costa Rica

Posted Thu Apr 26, 2012 2:38 PM

Cool thanks. I'll try to merge them and make a MADS assembly file and anyone can contribute comments to the code.

I'd like to see if I can make a modified version that can do DD sectors without the additional SRAM. This will make the USD very easy to install and simplify the 810 upgrade board. It'll all depend on if I can find enough places to stash the essential variables when RAM is full. Right now I know I'll have use of A,X,Y,SP and partial use of the Proc. Status (I,V,C), RIOT1024T if you read it back before it changes, part of PCH (jump to a mirror address to save the high order bits, and then JSR to get them back), and whatever else I can find (maybe a RIOT DDR or something). I figure the XF551 DD sector code might help here.




0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users