Jump to content

  2. View New Content
  3.  
  2. → Viewing Profile: Posts: MrMartian


MrMartian's Photo

  • Overview
  • Profile Feed
  • Likes
  • Friends
  • Topics
  • Recent Activity
  • Gallery
  • Blog
  • Issues
  • Stats

MrMartian

Member Since 5 Jul 2007
OFFLINE Last Active Jan 14 2012 7:53 AM

Posts I've Made

In Topic: Understanding Atari floppy drive timing

Thu Jan 12, 2012 9:13 AM

Pretty much every drive steps differently. AFAIK, the 810, 1050, Indus and (some?) Rana drives are CPU controlled steppers, whereas the XF551 and Trak are FDC controlled. The track skew from formatting is also very variable between drives, from the difference between step times, delays on index pulse generation, and code effenciency.

In Topic: Looking for MP3 of Atari Booting Sound? Anyone have?

Fri Nov 11, 2011 5:13 PM

View PostHiassofT, on Fri Nov 11, 2011 3:58 PM, said:

I made 2 recordings, one at divisor 0 (~125kbit/sec) and another one at divisor 8 (standard PC "highspeed SIO", 57.6kbit/sec). Be careful, the latter contains a really nasty high-pitch tone!

Somewhat off topic, but I wanted to know how reliable you've found running at 57.6k, where the Atari is expecting around 59k at that divisor.. If 57.6k is close enough, I could fairly trivially do US in the Trak drives..

In Topic: Looking for MP3 of Atari Booting Sound? Anyone have?

Fri Nov 11, 2011 9:40 AM

View Postphaeron, on Thu Nov 10, 2011 11:45 PM, said:

I disassembled the 1050 rev. B ROM and found the transmit routine. It takes 51 cycles per data bit and 549 cycles/byte at 1MHz, which comes out to 19608 baud and 1821.5 bytes/second. I don't understand why the routines were written with the timing off, since it has a lot of time wasted in delay loops and could have been fixed. The 1050's CPU is twice as fast as the 810's (1MHz vs. 500KHz) and the only thing extra the routine does is compute checksum, and yet the 810's transmit routine is actually closer timing-wise (19230 baud and 1886.8 bytes/sec).

I remember optimizing my code when I did the Indus GT Ultraspeed hack, to minimize the inter-byte and end of sector overhead to make the transfer as fast as possible.. I wonder how much overall time could be saved by doing this for any of the drives...

In Topic: New Indus GT firmware

Fri Nov 11, 2011 9:25 AM

View Posttregare, on Mon Nov 7, 2011 6:53 PM, said:

View PostMrMartian, on Mon Nov 7, 2011 9:07 AM, said:

I doubt it.. Even if I did, it wouldn't have the same level of comments in it as I would have added most of that as I was hacking away... I will look to see what I kept, though..

Thank you for looking.

No, it seems I didn't keep any other version of the disassembly, sorry.. At some point I should probably do this and do my hacks as .ifdefs..

In Topic: New Indus GT firmware

Mon Nov 7, 2011 9:07 AM

View Posttregare, on Mon Nov 7, 2011 2:03 AM, said:

View PostMrMartian, on Sun Sep 4, 2011 6:25 AM, said:

View PostMrMartian, on Sun Sep 4, 2011 6:10 AM, said:

Many apologies, I've actually not been on this site in a very long time... I will dig up the source and post it as soon as I can!

This is it! The only parts fairly well commented are the SIO routines...

I presume that this is the modified version? do you have the pre-modified source?

I doubt it.. Even if I did, it wouldn't have the same level of comments in it as I would have added most of that as I was hacking away... I will look to see what I kept, though..

  1. AtariAge Forums
  2. → Viewing Profile: Posts: MrMartian
  3. Guidelines