Flashback2 Michele CPU Discussion & Observations

[BigO]

.Pin...Name.........Description.....................Notes

...1...A1...........Address Line 1..................20 address lines = direct access to 1 Meg RAM instead of 2600's 4K

...2...A2...........Address Line 2..................

...3...A3...........Address Line 3..................

...4...A4...........Address Line 4..................

...5...A5...........Address Line 5..................

...6...A6...........Address Line 6..................

...7...A7...........Address Line 7..................

...8...A12..........Address Line 12.................

...9...A15..........Address Line 15.................

..10...A16..........Address Line 16.................

..11...A18..........Address Line 18.................

..12...VCC..........+5VDC...........................

..13...A17..........Address Line 17.................

..14...A14..........Address Line 14.................

..15...A13..........Address Line 13.................

..16...A8...........Address Line 8..................

..17...A9...........Address Line 9..................

..18...A11..........Address Line 11.................

..19...GND..........Ground..........................

..20...A10..........Address Line 10.................

..21...D7...........Data 7..........................

..22...D6...........Data 6..........................

..23...D5...........Data 5..........................

..24...D4...........Data 4..........................

..25...D3...........Data 3..........................

..26...A19..........Address 19......................

..27...VCC..........+5V.............................

..28...NRESET.......?...............................

..29...GND..........Ground..........................

..30...NHS..........Horizontal Sync.................

..31...NCS/NVS......Composite Sync / Vertical Sync..

..32...CHROMA4(R3)..Chroma4 Composite / Red3 RGB....

..33...CHROMA3(R2)..Chroma3 Composite / Red2 RGB....

..34...CHROMA2(R1)..Chroma2 Composite / Red1 RGB....

..35...CHROMA1(R0)..Chroma1 Composite / Red0 RGB....

..36...CHROMA0(G3)..Chroma0 Composite / Green3 RGB..

..37...LUM2(G2).....LUMA2 Composite / Green2 RGB....

..38...LUM1(G1).....LUMA1 Composite / Green1 RGB....

..39...LUM0(G0).....LUMA0 Composite / Green0 RGB....

..40...VCC..........+5V.............................

..41...I0...........?...............................

..42...I1...........?...............................

..43...PA7..........?...............................

..44...PA6..........?...............................

..45...PA5..........?...............................

..46...FIRE2_N......?...............................

..47...PA4..........?...............................

..48...GND..........Ground..........................

..49...AUD04........?...............................

..50...AUD03(B3)....Audio3 Composite / Blue3 RGB....

..51...AUD02(B2)....Audio2 Composite / Blue2 RGB....

..52...AUD01(B1)....Audio1 Composite / Blue1 RGB....

..53...AUD00(B0)....Audio0 Composite / Blue0 RGB....

..54...GND..........Ground..........................

..55...VCC..........+5V.............................

..56...I2...........?...............................

..57...I3...........?...............................

..58...PA3..........?...............................

..59...PA2..........?...............................

..60...PA1..........?...............................

..61...FIRE1_N......?...............................

..62...PA0..........?...............................

..63...GND..........Ground..........................

..64...SOFTNRST.....?...............................

..65...PB0..........?...............................

..66...PB1..........?...............................

..67...PB7..........?...............................

..68...PB6..........?...............................

..69...PB3..........?...............................

..70...PB5..........?...............................

..71...PB4..........?...............................

..72...PB2..........?...............................

..73...VCC..........+5V.............................

..74...NTEST........?...............................

..75...MGNODE.......?...............................

..76...YCNRGB.......?...............................

..77...CLKIN........?...............................

..78...CLKOUT.......?...............................

..79...GND..........?...............................

..80...NWR..........?...............................

..81...D2...........Data 2..........................

..82...D1...........Data 1..........................

..83...D0...........Data 0..........................

..84...A0...........Address Line 0..................

 

 

 

 

General Notes

 

[

http://www.atariage.com/forums/index.php?s...st&p=888131 (Hack to bring 5v to controller port)

Questions to be answered

Edited August 18, 2007 by BigO

[+bcombee]

18 address lines A0 - A17 means access to 256k instead of the 8k of the original 2600

 

AUD0 - AUD4: assuming AUD = audio, what might be the functions of 5 audio related pins?

Is it significant that 0 thru 3 are labeled B0-B3, but AUD04 has no B4 designation? Maybe just an oversight when the document was created.

 

When I saw the five audio pins, my thought was you connect them all up to different values of resistors, then combine the output to produce the analog signal, similar to how the old Disney Sound Source made audio from a PC's parallel port. I've not see the schematic of the Flashback 2 to know if that's the plan or not.

[kenfused]

18 address lines A0 - A17 means access to 256k instead of the 8k of the original 2600

 

AUD0 - AUD4: assuming AUD = audio, what might be the functions of 5 audio related pins?

Is it significant that 0 thru 3 are labeled B0-B3, but AUD04 has no B4 designation? Maybe just an oversight when the document was created.

My guess is the B0 to B3 are 4 bits of Blue when it is switched to RGB mode? (there are R0-R3 and G0-G3 in the chroma/luma pins.

Edited July 2, 2007 by kenfused

[antron]

18 address lines A0 - A17 means access to 256k instead of the 8k of the original 2600

 

AUD0 - AUD4: assuming AUD = audio, what might be the functions of 5 audio related pins?

Is it significant that 0 thru 3 are labeled B0-B3, but AUD04 has no B4 designation? Maybe just an oversight when the document was created.

My guess is the B0 to B3 are 4 bits of Blue when it is switched to RGB mode? (there are R0-R3 and G0-G3 in the chroma/luma pins.

 

so no audio in RGB mode?

[BigO]

Oops...make that 20 address lines: 1 Meg

 

It seems fairly easy to correlate the CPU pinout physically to the circuit board. It should just be a matter of massive amounts of tedium to produce a schematic of the FB2.

Edited July 2, 2007 by BigO

[+karri]

Once the thing is out I would love to have a look inside to get some idea of the real usage of "MGNODE (Mega Gyration Node) ".

 

Also the lack of USB signals is a bit surprising. Perhaps the USB connection is handled by a different chip? The LCD seems to be operating on some kind of video interface as there is HS and VS signals and no LCD address pins.

 

It looks as it would have 5-bit audio in YCN mode and only 1 bit in RGB-mode.

--

Karri

[mimo]

This is for the FB2, not the portable

[+karri]

This is for the FB2, not the portable

 

Of course! How stupid of me...

[BigO]

More specifically on the audio, I saw what appeared to be a hint that the two audio channels as existed in the 2600 are present somewhere in the FB2. I expect the audio in the FB2 test mode might give me a clue when I look at those 5 pins with an oscilloscope.

[antron]

NCS/NVS - composite or vertical sync

NHS - horizontal sync

Edited July 2, 2007 by antron

[antron]

It looks as it would have 5-bit audio in YCN mode and only 1 bit in RGB-mode.

 

which would basically leave you with no audio. why would they go through the trouble of including RGB mode like that?

[werejag]

ok how do we get stereo out on this?

[antron]

ok how do we get stereo out on this?

 

see if the audio bits feed to a Covox

 

if one of the pins are unused then perhaps it is a strobe signal to split the channels.

 

edit: or perhaps we have to get our own strobe from the clock? would leave 5 bit sound.

Edited July 4, 2007 by antron

[BigO]

Can anyone point me to a good reference that will help me understand RGB displays to the extent that I would be able to understand how these 4 bit R,G,B signals would be applied in the context of an LCD display?

 

Also, I assume that the i0, i1, i2, i3 pins are related to the video so expect that I'd gain an understanding of that.

 

I've searched around a bit, but I don't yet know enough to decipher what I've found. Hoping someone will be able to point out something easy to digest.

 

Thanks.

Edited August 2, 2007 by BigO

[supercat]

which would basically leave you with no audio. why would they go through the trouble of including RGB mode like that?

 

The chip could quite easily generate pulse-width-modulated audio out of a single pin that would, with a simple RC filter on the back end, sound perfectly decent. I don't know what the highest clock used on chip is, but I'd guess 14.31818MHz, which is 456 times the audio sampling rate. That's high enough that even a crude filter should be adequate (the PWM frequency would be about 447KHz) and the highest frequency even remotely of interest would be about 1/28 of that.

[supercat]

More specifically on the audio, I saw what appeared to be a hint that the two audio channels as existed in the 2600 are present somewhere in the FB2. I expect the audio in the FB2 test mode might give me a clue when I look at those 5 pins with an oscilloscope.

 

I would guess that those five pins probably output in binary the sum of (generator0 * AUDV0) + (generator1 * AUDV1), where generator0 and generator1 each go between 0 and 1 as directed by AUDCx and AUDFx. If that's the case, no stereo unless there's some alternate mode.

[supercat]

Can anyone point me to a good reference that will help me understand RGB displays to the extent that I would be able to understand how these 4 bit R,G,B signals would be applied in the context of an LCD display?

 

Also, I assume that the i0, i1, i2, i3 pins are related to the video so expect that I'd gain an understanding of that.

 

I've searched around a bit, but I don't yet know enough to decipher what I've found. Hoping someone will be able to point out something easy to digest.

 

If the chip can indeed be put into a useful RGB mode, then converting that to an analog RGB display is easy. For the green you simply do:

G0--2K--+---2K--Gnd

. 1K

G1--2K--+

. 1K

G2--2K--+

. 1K

G3--2K--+

. +------Green Out

. 220

. Gnd

[code]

Red and blue are similar; you may have to tweak the 220 ohm resistor to scale the output voltage appropriately.

 

I'm curious, though, what indication there is that the FB2 can actually produce RGB?

 

Even if the FB2 can't produce RGB directly, I would guess that its output could probably be coaxed into pixel-perfect RGB pretty easily using a PLD along with some resistors. I'd have to see what's actually output on the chroma pins to really know, though.

[antron]

I'm curious, though, what indication there is that the FB2 can actually produce RGB?

 

the RGB mode select and the Red Green Blue bits on the Michele CPU (according to Curt)

[BigO]

I'm curious, though, what indication there is that the FB2 can actually produce RGB?

 

the RGB mode select and the Red Green Blue bits on the Michele CPU (according to Curt)

 

See the initial post in this thread. I've copied over the Michele CPU pinout that Curt provided.

Specifically, see pins 76, 50-53, 30-39, (41-42, 56-57?) for indications of RGB functionality.

 

Based on the fact that RGB output appears to be integrated, I imagine that the shelved FB portable would have highly leveraged the FB2 design.

 

Curt provided the information originally here: http://www.atariage.com/forums/index.php?showtopic=109488

and stated "This info is provided as is, I will not answer detailed specifics on particular pin functions, have some fun, tinker and explore, you'll find some very interesting and fun exploits if you know how". (emphasis mine)

 

It looks to be fairly easy to map the CPU pinout to the connections around the blob on the FB2 board.

 

I had two goals in that motivated me to pick through this information and tinker. The secondary goal related to stereo output. The primary goal was to build my own portable. NTSC-composite-video-friendly LCD's are about as scarce as hen's teeth. The one I did acquire worked perfectly with every other video signal I threw at it but didn't play well at all with Atari 2600 or FB2. (Though, it did ultimately net me over $400.00 profit! )

 

As can now be seen in the initial post, I have started trying to compile the information that people are providing/confirming/discovering and/or speculating about.

[Hornpipe2]

The existence of separate CHROMA and LUM outputs gives me continued hope of an S-Video mod. I really need to open my FB2 up.

 

EDIT: What about removing the existing 1M rom-containing chip and replacing with our own?

Edited September 20, 2007 by Hornpipe2

[BigO]

The existence of separate CHROMA and LUM outputs gives me continued hope of an S-Video mod. I really need to open my FB2 up.

 

EDIT: What about removing the existing 1M rom-containing chip and replacing with our own?

 

I was just thinking about that 1M opportunity recently. I'd think this could allow for direct addressing of a good amount memory (both ROM and RAM probably would be useful) beyond what was accessible in the 2600. It would certainly be a niche market, but it seems like you could build a bigger non-bankswitched game...whatever that's worth. I guess it'd be like putting a full 6502 CPU into the 2600 architecture similar to a "super atari" project I saw that gave access to 64K, but with more addressing space.

 

...I wonder if there is any additional RAM native in the FB2? I suppose if I'd get around to doing a cart slot mod or similar such thing to one of my FB2's, I could eventually figure that out.