Improved Sector Layout CX8111

[+Nezgar]

Well look at what has serendipitously turned up on the AtariAge Marketplace group on Facebook... https://www.facebook.com/groups/atariagemarketplace/

 

Rev B 810 ROMs! and lots of them amongst other things.

 

I'm going to get a couple, but I'm hoping there's someone else interested to acquire one that has the means to dump it.

Cheers!

[+Nezgar]

In this post by Kevin Savetz http://atariage.com/forums/topic/234310-new-atari-8-bit-scans-and-video/?p=3978608 is the document "Atari 810 Disk Peripheral Device Description dated 12/15/80" by Harry Stewart.

 

Two things I found interesting in this early document about the 810:

 

The rotation is documented here as 300RPM, not the production 288RPM standard for the 810 (815?) and later 1050. Maybe the 12 sector interleave in the 810 Rev B ROM was better suited to 300RPM and a holdover from pre-production before Rev C replaced it with the 9 sector interleave.

 

He claims this documented interleave has a skew of 7, Not sure how he came up with '7' - its clearly 12, and this is the same interleave I have evidence of on a few of my older disks.

 

But he does note the sector 11/18 GAP here, which is documented as between sector 1/12 in the 810 hardware service manual.

[phaeron]

Different definition of interleave. Notice that the sector numbers increase by 7 in physical order.

 

The sector layout described is exactly the same as that used by the rev. C ROM, which implies that the only difference in formatting was to the sector order table. Furthermore, they may have had to slow down the drive due to 300 RPM not working. The track layout specified produces 3379 encoded bytes, of which about 277 are lead-in/out. This makes the critical portion of the track 3102 bytes long. At 125Kbits/sec and 300 RPM, the track holds 3125 bytes, leaving a margin for speed variation of 0.8%. That's below the 1-1.5% specced for various spindle speed variation parameters on the drive mechanism.

 

300 RPM to 288 RPM is a difference of 4%, which is not enough to explain a change from 12:1 to 9:1. DOS 1 vs. 2 might have something to do with it -- DOS 1 can sort of read at a reasonable speed with a 12:1 interleave, but it blows revs constantly at 9:1.

[+Nezgar]

Well lookey what I got my hands on, just in from Best. Last one they (he) had in stock apparently!

 

The two chips for comparison, before I 'downgraded'

 

The bottom of the 2 chips of mine:

 

I formatted a disk with it, and checked the results using the Archiver Software, confirming the suspected 810 Rev B sector layout of:
18 7 14 3 10 17 6 13 2 9 16 5 12 1 8 15 4 11 (12:1)

 

And as seen by the Speedy Disk-Mapper: (It does not seem to care about index mark)

 

A 'mythical unicorn' no more!

 

In my brief run with it this evening, it otherwise seems to operate identical to Rev C. Disks read the same speed according to formatted skew, idle timeout , seek to track 39 on idle, and loud track 0 buzz on read failure seems the same. I wonder if the formatting sector layout really is the only change in the 810 code from Rev B to C.

 

Interestingly, the audible time of sector write I/O is just about the same as the pause between each write. When writing, because of the 'clicky' nature of the POKEY SIO audio, it's difficult to determine which periods of time are actual data transmission and waiting for the drive between sectors.

 

Once I'm done with my own testing, anyone out there willing to attempt to dump it? I can mail it over. Old threads on this topic seem to indicate that the 2KB 810 ROM is very difficult to dump or wire up an adapter with most EPROM burners... I presume a lot of re-wiring of an adapter would be needed if it was possible with my TL866.

Edited April 28, 2018 by Nezgar

[Rybags]

Maybe the inter-sector gaps are different?

 

What's needed is a cryoflux run against the various formats.

[joeventura]

On 12/18/2017 at 12:05 AM, tschak909 said:

and to think, Percom was selling data seperators at $45 a pop.

 

-Thom

I actually have a Percom data seperator for my 810 the pins broke on it in a move

 

What value does it provide?

 

 

[tschak909]

On 3/6/2022 at 12:37 PM, joeventura said:

I actually have a Percom data seperator for my 810 the pins broke on it in a move

 

What value does it provide?

 

 

much improved reliability.

[joeventura]

41 minutes ago, tschak909 said:

much improved reliability.

Time to get out the soldering iron

[+Nezgar]

On 3/6/2022 at 12:37 PM, joeventura said:

I actually have a Percom data seperator for my 810 the pins broke on it in a move

 

What value does it provide?

The original 810's, like the TRS-80, used a WDC1771 controller, but did not heed the advice in the documentation from WDC to not use the built-in data separator or read reliability would suffer. See this from Page 1 of the Percom TRS-80 install guide:

 

https://archive.org/details/Percom_Data_Separator_Installation_Manual_19xx_Percom_Data_Co./mode/2up

 

For more detail on exactly *WHY* WDC recommended an external data separator, see this from "FD 1771-01 Application Notes" https://archive.org/details/rearc_fd1771-01-application-notes-198x/mode/2up

 

 

 

[tschak909]

19 minutes ago, Nezgar said:

The original 810's, like the TRS-80, using a WDC1771 controller, and did not heed the advice in the documentation from WDC to not use the built-in data separator for read reliability. See this from Page 1 of the Percom TRS-80 install guide:

 

https://archive.org/details/Percom_Data_Separator_Installation_Manual_19xx_Percom_Data_Co./mode/2up

 

For more detail on exactly *WHY* WDC recommended an external data separator, see this from "FD 1771-01 Application Notes" https://archive.org/details/rearc_fd1771-01-application-notes-198x/mode/2up

 

 

 

boy, that's a queasy note.

-Thom