Jump to content
IGNORED

Atari 800 Engineering Serial 26


ClausB

Recommended Posts

In 1984 I worked for a small company creating educational software for the Apple II, Atari 800/400, and (ugh) Commodore 64. One of our development machines was an old 800 labeled "Engineering Serial 26". It looked and acted pretty much like a standard 800 but a peek under the lid revealed bare RAM and ROM cards. Both ROM cards held EPROMs with paper stickers and three RAM cards totalled 36K. The OS card had an extra chip socket bolted to the back with wires flying down to the motherboard, so it could not come out.

 

A look at the RAM cards showed 4K in slot 1 and 16K in each of 2 and 3. Both sizes used the same circuit board, dated 1978. They contained standard 4K or 16K DRAM chips and had jumpers to select the size. This was unlike the later 8K and 16K cards, which used different circuit boards and selected RAM in units of 8K. There must have been differences on the motherboard too.

 

When the company failed, the owner gave me old #26. (He also had an Apple I but, wisely, he kept that!) I immediately tore it open and traced the circuits. The motherboard was labeled "Rev. 2" and it selected RAM and ROM in 4K blocks. I found that RAM slot 1 could only address 4K but 2 and 3 could hold either 16K or 4K. I also saw cuts and jumpers on the motherboard and OS card. These mods made the prototype act like a production unit. To see the prototype design, I traced the original circuit and discovered the following memory map:

 

Original #26................................................Production

$0000 - $8FFF..............Full RAM..................$0000 - $BFFF

$9000 - $AFFF.............Right ROM................$8000 - $9FFF

$B000 - $CFFF.............Left ROM.................$A000 - $BFFF

$D000 - $D7FF.........CTIA & ANTIC.............$D0xx & $D4xx

$D800 - $DFFF..........POKEY & PIA..............$D2xx & $D3xx

$E000 - $FFFF..............OS ROM..................$D800 - $FFFF

 

Notice the changes from the prototype to the production version. Maximum RAM went from 36K to 48K. Cartridge ROM moved down to overlap (and disable) the top of RAM, leaving a 4K hole at $Cxxx. POKEY and PIA moved down to make room for the 2K Floating Point ROM. The 8K Operating System ROM remained. One can speculate on the reasons for these changes.

 

There were probably two main reasons, one firmware and one hardware. Bill Wilkinson has written that, late in 1978, Atari changed from an 8K Microsoft BASIC to the 10K Shepardson BASIC including the 2K BCD floating point package. The #26 OS card originally held 4 2K ROM chips. It was modified to house the OS in 2 4K chips and the FP ROM in a 2K chip.

 

The hardware reason was likely that, by mid-1978, DRAM makers were labeling defective 16K chips as 8K and dumping them onto the market. These worked as long as you addressed only the half which had no bad bits. They sold for little more than the older 4K chips, especially in high volumes. So, Atari developed an 8K card and reworked the 16K card and the motherboard to select RAM and ROM in 8K chunks.

 

The rest you know.

Edited by ClausB
  • Like 4
Link to comment
Share on other sites

Well, I dug out old number 26 and looked her over. A little yellow, but the Start key doesn't stick like my newer 800's did. I dare not turn her on without a thorough internal inspection but I did open the lid. The foam tape had disintegrated and the cartridge shield fell off! I pulled out all 5 EPROMs and checked them out with an old Shooter. Each one read and verified three times - after 29 years! The binary looked OK in the few places I checked. Of course, it only takes one bad bit to crash the system.

 

The EPROMs are labeled "OS3 6-15-79" and "BASIC 4-18" (presumably also 1979). Does anyone know if these would be the production versions?

 

Should I post the binaries or are they still covered by copyrights?

Link to comment
Share on other sites

AFAIK, the 8-bit firmware is in the public domain.

 

Would be interesting to get a look at them.

 

If the memory map of the system is as you've specified, then it's a 100%er that the ROMs would be significantly different.

 

Although, maybe not so different from OS Rev A - remembering that Atari had the habit of jumping the gun and committing to enormous ROM production runs with half-baked products.

Link to comment
Share on other sites

Well, I dug out old number 26 and looked her over. A little yellow, but the Start key doesn't stick like my newer 800's did. I dare not turn her on without a thorough internal inspection but I did open the lid. The foam tape had disintegrated and the cartridge shield fell off!

 

Very common! Later 800's lack this shield.

Link to comment
Share on other sites

If the memory map of the system is as you've specified, then it's a 100%er that the ROMs would be significantly different.

No, they modified the hardware of this prototype to be compatible with the later production machines, so the EPROMs were made for the Production memory map. The Original memory map would have held true before those modifications. Unfortunately, I don't have the original prototype EPROMs.

Link to comment
Share on other sites

You know, it kind of makes sense you had one of the early models as the company you worked for was one of the very first software publishers on the Atari :)

Yes, IIRC, PDI was originally an Apple house and Atari gave them this unit to foster early software development. The label underneath also says "Software #4" and I'm sure PDI did not have 4 Ataris. Could be it was used for in-house programming at Atari. ???

Link to comment
Share on other sites

It's not quite a prototype, but it's not the same as the production unit either.

Does anyone know where the prototypes are?

 

I think Curt may have a pre-production prototype Atari 800 somewhere. I know he had some internal developer units that had cool ports on them for connecting to mainframes.

 

Tempest

Link to comment
Share on other sites

Label is interesting... I have a 400 with the same styling, so this is the only other unit I've seen with such a label...

 

 

The internal boards must've been swapped out during the course of things, because I've got a pre-production unit and everything is non-green coated boards... The cartridge door also looking like it may have still been undergoing some final changes as it looks off, like its not a good seating fit.

 

 

Thanks for sharing Claus.

 

 

Curt

 

Look, it's real.
Link to comment
Share on other sites

Label is interesting... I have a 400 with the same styling, so this is the only other unit I've seen with such a label...

 

 

The internal boards must've been swapped out during the course of things, because I've got a pre-production unit and everything is non-green coated boards... The cartridge door also looking like it may have still been undergoing some final changes as it looks off, like its not a good seating fit.

Curt, are there pics somewhere of your units?

Link to comment
Share on other sites

  • 2 weeks later...
A look at the RAM cards showed 4K in slot 1 and 16K in each of 2 and 3. Both sizes used the same circuit board, dated 1978. They contained standard 4K or 16K DRAM chips and had jumpers to select the size. This was unlike the later 8K and 16K cards, which used different circuit boards and selected RAM in units of 8K.

Looking at the Atari 800 OS Manual and Listing, I noticed that the routine that figures out total RAM size does so by testing memory in 4K blocks. Shades of the prototype hardware?

Link to comment
Share on other sites

I opened up #26 to show off the innards better. Check out the ANTIC and CTIA ICs - aren't they just beautiful? And then there's the hack job on the OS board and mobo! Those are the mods that made this former prototype act more or less like a production unit. How many of you are old enough to remember wire-wrapping? One of the pink wires had broken off its connection over the years. Can you spot which one? After photgraphing, I repaired it. Enjoy.

post-18605-1206919988_thumb.jpg

post-18605-1206920010_thumb.jpg

post-18605-1206920113_thumb.jpg

post-18605-1206920135_thumb.jpg

  • Like 1
Link to comment
Share on other sites

I opened up #26 to show off the innards better. Check out the ANTIC and CTIA ICs - aren't they just beautiful? And then there's the hack job on the OS board and mobo! Those are the mods that made this former prototype act more or less like a production unit. How many of you are old enough to remember wire-wrapping? One of the pink wires had broken off its connection over the years. Can you spot which one? After photgraphing, I repaired it. Enjoy.

 

Are the three ICs on the bottom of the OS board the same as the production?

Edited by warerat
Link to comment
Share on other sites

I opened up #26 to show off the innards better. Check out the ANTIC and CTIA ICs - aren't they just beautiful? And then there's the hack job on the OS board and mobo! Those are the mods that made this former prototype act more or less like a production unit. How many of you are old enough to remember wire-wrapping? One of the pink wires had broken off its connection over the years. Can you spot which one? After photgraphing, I repaired it. Enjoy.

 

Are the three ICs on the bottom of the OS board the same as the production?

Not quite. The OS board has no Atari CO# part number nor a date. It originally held 4 2716 EPROMS (8K total) but was modified later to hold 2 2532s and 1 2716 (10K total). The three chips are NAND and NOR gates that do similar jobs as those on the production board but they also decoded the original I/O memory map shown in the first post. The bolted-on, wire-wrapped chip, added later, is the same decoder used on the production OS board and it decodes the production I/O memory map.

post-18605-1206927571_thumb.jpg

  • Like 1
Link to comment
Share on other sites

I opened up #26 to show off the innards better. Check out the ANTIC and CTIA ICs - aren't they just beautiful? And then there's the hack job on the OS board and mobo! Those are the mods that made this former prototype act more or less like a production unit. How many of you are old enough to remember wire-wrapping? One of the pink wires had broken off its connection over the years. Can you spot which one? After photgraphing, I repaired it. Enjoy.

 

Are the three ICs on the bottom of the OS board the same as the production?

Not quite. The OS board has no Atari CO# part number nor a date. It originally held 4 2716 EPROMS (8K total) but was modified later to hold 2 2532s and 1 2716 (10K total). The three chips are NAND and NOR gates that do similar jobs as those on the production board but they also decoded the original I/O memory map shown in the first post. The bolted-on, wire-wrapped chip, added later, is the same decoder used on the production OS board and it decodes the production I/O memory map.

 

Still an interesting motherboard. I'm assuming that 'LS245 on your motherboard is for the data bus to isolate the memory slots from the OS board and the cart slots? If so, I don't get why they didn't keep that simpler design instead of going to the pair of 'LS243's on the production machine.

Link to comment
Share on other sites

Still an interesting motherboard. I'm assuming that 'LS245 on your motherboard is for the data bus to isolate the memory slots from the OS board and the cart slots? If so, I don't get why they didn't keep that simpler design instead of going to the pair of 'LS243's on the production machine.

Yes, and the LS30 gate enables the LS245 buffer. I can see no logical reason for changing to LS243s. The two LS138s decode the address space into 16 4K blocks. The gates on the OS board generate RASTIME, WRITIME, RAM buffer enables, ROM selects, and POKEY, PIA, and CTIA selects. There is no circuitry to deselect RAM since cart ROM did not overlap. After the mods, the right cart slot could only take 4K, since the RAM totals 36K. Notice too that the cart slots are smaller.

Link to comment
Share on other sites

  • 3 weeks later...

IT WORKS!

 

After hours of tracing circuits, scoping signals, and replacing the OS EPROMs with a simple RAM-less rainbow display loop, I found that D1 of the data bus was corrupted somewhere between the EPROM and CPU. Replacing the data bus buffer did not help and checking for shorts and opens revealed nothing. Finally I replaced the CPU with the one from my 400 and it works.

 

You can see from the photos above how beat-up the original CPU chip is. Also, it is stamped "A" (it looks like an afterthought) and my 400's is marked "6502B". And the paper sticker says "Slow Micro", so I wonder if it just had marginal timing.

 

I then put the OS ROMs back and added the RAM boards in one-by-one and Memo Pad came up each time. Next I'll reassemble the whole thing and try BASIC...

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...