158 replies to this topic

[atariksi]

Quote

The 6502C is the custom version Atari used as I mentioned and I see no cost advantage of using another 6502 derivative instead. (I'm not sure what atarisky was getting at since the A8's used the same CPU) Custom chips vs off the shelf (or licensed) components had a number of tradeoffs though: accepting design costs of custom chips and one time licensing fees you've got volume production of mass market components vs smaller numbers of custom components, one you pay more to a distributor for their profit, the other will only be cheaper if the custom chips are produced in huge volumes (granted, that's the case with a popular console). The big advantage with custom chips comes when you have your own fabrication company (or close associate), ie vertical integration as with CBM and MOS.
...

Well, if Atari PIA is also some generic chip like 6502C w/slight modifications by Atari, then weren't both part of one big deal so why would they drop the PIA. It should have been cheap like 6502C.

[kool kitty89]

atariksi, on Thu Mar 25, 2010 10:11 PM, said:

Okay, PIA is different from the one currently available from WDC which they claim to be backward compatible; current one in Atari does input in output mode and is faster at switching between input and output modes. So I thought perhaps Atari had a hand in it. 6502C meanwhile I believe they told them to make it with HALT line.

For Halt, do you mean an internal mechanism to directly halt the CPU with a single signal rather than dealing with the ready line before halting?

I'm not sure about current offerings, but I've seen PIA listed under the standard accessory chips for the 650x family (along with 6532 RIOT, 6522 VIA, etc). Wiki's page also lists similar "PIAs" used for the 680x architecture. (it seems several chips were named "peripheral interface adaptor" all with similar functionality and some intercompatibility including the MC6821/6822 of the CoCo)

atariksi, on Thu Mar 25, 2010 10:16 PM, said:

Well, if Atari PIA is also some generic chip like 6502C w/slight modifications by Atari, then weren't both part of one big deal so why would they drop the PIA. It should have been cheap like 6502C.

I'm not sure PIA was custom at all, not even as custom as the 6502C (or 6510), it seems to have been as standard as RIOT in that sense from the information I see online at least. (for some reason Atari referred to the 6532 as "PIA" on the Sylvia/3200 plans seen on Atarimuseum too)

The only reason I mentioned replacing PIA with RIOT was for VCS compatibility. Looking at the planned Sylvia design though, they seemed quite interested in going for as consolidated and (VCS) compatible a system as possible, yet the 5200 didn't fit that as much. It seems the simplest "quick fix" after the 3200 was dropped would have been simply including both TIA and CTIA/GTIA instead of the new STIA, less consolidated, but probably the closest option. (using SRAM still doesn't make a whole lot of sense)
Otherwise, the most straightforward choice is to take the A400/800 chipset directly without modifications (internally, at least, incompatible cartridges and such are another matter) onto a consolidated board, but they didn't do that either.

Edited by kool kitty89, Fri Mar 26, 2010 2:33 AM.

[atariksi]

kool kitty89, on Fri Mar 26, 2010 2:25 AM, said:

atariksi, on Thu Mar 25, 2010 10:11 PM, said:

Okay, PIA is different from the one currently available from WDC which they claim to be backward compatible; current one in Atari does input in output mode and is faster at switching between input and output modes. So I thought perhaps Atari had a hand in it. 6502C meanwhile I believe they told them to make it with HALT line.

For Halt, do you mean an internal mechanism to directly halt the CPU with a single signal rather than dealing with the ready line before halting?
...

Yeah, halting with single signal. A400/A800 do it via external circuitry whereas XL/XE/XEGS/5200 do it with a pin on the chip.

Quote

I'm not sure about current offerings, but I've seen PIA listed under the standard accessory chips for the 650x family (along with 6532 RIOT, 6522 VIA, etc). Wiki's page also lists similar "PIAs" used for the 680x architecture. (it seems several chips were named "peripheral interface adaptor" all with similar functionality and some intercompatibility including the MC6821/6822 of the CoCo)

atariksi, on Thu Mar 25, 2010 10:16 PM, said:

Well, if Atari PIA is also some generic chip like 6502C w/slight modifications by Atari, then weren't both part of one big deal so why would they drop the PIA. It should have been cheap like 6502C.

I'm not sure PIA was custom at all, not even as custom as the 6502C (or 6510), it seems to have been as standard as RIOT in that sense from the information I see online at least. (for some reason Atari referred to the 6532 as "PIA" on the Sylvia/3200 plans seen on Atarimuseum too)

The only reason I mentioned replacing PIA with RIOT was for VCS compatibility. Looking at the planned Sylvia design though, they seemed quite interested in going for as consolidated and (VCS) compatible a system as possible, yet the 5200 didn't fit that as much. It seems the simplest "quick fix" after the 3200 was dropped would have been simply including both TIA and CTIA/GTIA instead of the new STIA, less consolidated, but probably the closest option. (using SRAM still doesn't make a whole lot of sense)
Otherwise, the most straightforward choice is to take the A400/800 chipset directly without modifications (internally, at least, incompatible cartridges and such are another matter) onto a consolidated board, but they didn't do that either.

Either way you look at it, PIA dropping from 5200 doesn't make sense. If it was an in-house modified chip, it would be cheap to keep it compared to buying from another party. If it was part of a standard chip along with 6502, then they already had a good deal in place to get them cheap together. From my tests, the original PIA in Atari computers is better than the one they sell now. Remember that Atari computers needed faster chips back then since 1.79Mhz was considered top of the line compared to competitors all doing 1Mhz (like Apple, Vic-20, C64, TRS-80s, etc.).

[kool kitty89]

atariksi, on Fri Mar 26, 2010 3:10 PM, said:

Either way you look at it, PIA dropping from 5200 doesn't make sense. If it was an in-house modified chip, it would be cheap to keep it compared to buying from another party. If it was part of a standard chip along with 6502, then they already had a good deal in place to get them cheap together. From my tests, the original PIA in Atari computers is better than the one they sell now. Remember that Atari computers needed faster chips back then since 1.79Mhz was considered top of the line compared to competitors all doing 1Mhz (like Apple, Vic-20, C64, TRS-80s, etc.).

The only sense it makes to remove PIA is if Atari/Warner was really gung-ho about using analog joysticks as the main control as a key feature like Mattel's touted 16-direction disc (which was horrible), which in all honestly is probably true; or at least the marketing department might have gotten hold of the idea and let loose with it. -It would have made a lot more sense to simply mock the shortcomings of the IV controller (and I'm not a particular fan of negative advertising), but seriously, they could have pointed out how uncomfortable the disc (and controller as a whole) was and how the 16-direction was not only of limited utility, but often detrimental to accurate control. An analog joystick would have made a good accessory though, especially for games like Star Wars. (would have been nice to have that on the 800 too -which should have been quite possible)

All they really needed was more buttons to match arcade games of the time (pause is really nice too).

There's no reason they couldn't have made the cartridge and expansion interfaces the only things changed compared to the computers (if onboard VCS compatibility was no longer a concern), of course there's curt's previous post on this with a neat example concept for the console:

Curt Vendel, on Sun Nov 1, 2009 12:59 PM, said:

You have to balance this from a marketing and consumer standpoint... you need to keep the cost of the unit down to get consumers to buy it. You also want to dangle carrots in front of buyers to get them to buy and upgrade - hence selling a separate keyboard and cassette option. If you include the keyboard then its more of an entry level computer - we are trying from the company standpoint to deliver a high end game console system to the public which could be upgraded to a computer by consumers, not to just directly sell an entry level computer - that defeats the purpose of the conversation...

So, lets see the "re-evolution" of the 400 into the new "5200-Phase II" design...

Stage 1: Remove top cartridge door, cpu/ram card and go with single board design
stage1.jpg

Stage 2: Remove keyboard from design
stage2.jpg

Stage 3: Add 5200 cartridge port, venting style and silver label
stage3.jpg

Stage 4: Adjust color to console black
stage4.jpg

Stage 5: Remove original SIO port, here is our new 5200 Design...
stage5.jpg

Thorsten Günther, on Sun Nov 1, 2009 11:17 AM, said:

I would like to disagree on some of your proposals:

IMHO it would have made no sense to change the cartridge port or to entirely leave out the keyboard when the Odyssey had one to fake "educational" value. I would have simply aimed at a cost-cut 600XL with a 400-like membrane (if it were identical, upgrades would already have been available - so that'd be a plus) or chiclet keyboard. 16 KB RAM is OK, but I'd have kept the PBI and SIO bus to advertise it as "able to grow to a full 64K computer - fully compatible with all the peripherals".

And - heck - why no orange case?

Analog joysticks would have been nice for some games - considering the fact that the original joystick port already has two analog inputs (for the paddles) and five digital ones, it would have been a no-brainer to develop one similar to that of the Vectrex.

The bulky size of the 5200 has come up before (both in terms of cost and convenience), but that concept design reminds me: Just how big is the A400 compared to the 5200? (or A800 for that matter)


But in the context of Sylvia, the most direct quick fix to that end would be dropping the super TIA plans and substitute that with TIA+CTIA/GTIA.

Edited by kool kitty89, Sat Mar 27, 2010 2:02 AM.

[atariksi]

kool kitty89, on Sat Mar 27, 2010 1:37 AM, said:

atariksi, on Fri Mar 26, 2010 3:10 PM, said:

Either way you look at it, PIA dropping from 5200 doesn't make sense. If it was an in-house modified chip, it would be cheap to keep it compared to buying from another party. If it was part of a standard chip along with 6502, then they already had a good deal in place to get them cheap together. From my tests, the original PIA in Atari computers is better than the one they sell now. Remember that Atari computers needed faster chips back then since 1.79Mhz was considered top of the line compared to competitors all doing 1Mhz (like Apple, Vic-20, C64, TRS-80s, etc.).

The only sense it makes to remove PIA is if Atari/Warner was really gung-ho about using analog joysticks as the main control as a key feature like Mattel's touted 16-direction disc (which was horrible), which in all honestly is probably true; or at least the marketing department might have gotten hold of the idea and let loose with it. -It would have made a lot more sense to simply mock the shortcomings of the IV controller (and I'm not a particular fan of negative advertising), but seriously, they could have pointed out how uncomfortable the disc (and controller as a whole) was and how the 16-direction was not only of limited utility, but often detrimental to accurate control. An analog joystick would have made a good accessory though, especially for games like Star Wars. (would have been nice to have that on the 800 too -which should have been quite possible)

All they really needed was more buttons to match arcade games of the time (pause is really nice too).

There's no reason they couldn't have made the cartridge and expansion interfaces the only things changed compared to the computers (if onboard VCS compatibility was no longer a concern), of course there's curt's previous post on this with a neat example concept for the console:

Yeah, on this very planet with a high mortality rate (earth) we talked about how A5200 design was trying to imitate the inferior joysticks of intellivision rather than sticking to the best that they had.

They could have just incorporated the Atari 2600 keypad into the CX-40 and used the CONSOL select lines to switch between the joystick and keypad.

[kool kitty89]

atariksi, on Sat Mar 27, 2010 12:27 PM, said:

Yeah, on this very planet with a high mortality rate (earth) we talked about how A5200 design was trying to imitate the inferior joysticks of intellivision rather than sticking to the best that they had.

At least they ended up being somewhat more useful than the IV controllers, if less reliable. (buttons mainly) That later prototype self-centering controller sound really promising though, compact spring-loaded 2-axis pot module, like modern analog sticks (and perhaps the Vectrex). Still, that would have been best as an accessory.

Quote

They could have just incorporated the Atari 2600 keypad into the CX-40 and used the CONSOL select lines to switch between the joystick and keypad.

How could they do that while maintaining the standard ports and pinout?

[atariksi]

kool kitty89, on Sat Mar 27, 2010 3:43 PM, said:

atariksi, on Sat Mar 27, 2010 12:27 PM, said:

Yeah, on this very planet with a high mortality rate (earth) we talked about how A5200 design was trying to imitate the inferior joysticks of intellivision rather than sticking to the best that they had.

At least they ended up being somewhat more useful than the IV controllers, if less reliable. (buttons mainly) That later prototype self-centering controller sound really promising though, compact spring-loaded 2-axis pot module, like modern analog sticks (and perhaps the Vectrex). Still, that would have been best as an accessory.

...

Still by copying IV, it was the poison ivy for contaminating the then currently superior controllers.

Quote

Quote

They could have just incorporated the Atari 2600 keypad into the CX-40 and used the CONSOL select lines to switch between the joystick and keypad.

How could they do that while maintaining the standard ports and pinout?

Just like Sega Genesis used the pin 7 for a select line so use pin 7 as +5V and 0V via the CONSOL select line. Joysticks normally don't use +5V at all so they function with pin 7 at 0V and +5V on pin 7 selects other controllers like touchpad. In fact, current A5200 systems use a CONSOL pin (S2) to select POT GND enable/disable (+5V/0V).

So it remains perfectly backward compatible.

[kool kitty89]

atariksi, on Sun Mar 28, 2010 1:11 AM, said:

Just like Sega Genesis used the pin 7 for a select line so use pin 7 as +5V and 0V via the CONSOL select line. Joysticks normally don't use +5V at all so they function with pin 7 at 0V and +5V on pin 7 selects other controllers like touchpad. In fact, current A5200 systems use a CONSOL pin (S2) to select POT GND enable/disable (+5V/0V).

So it remains perfectly backward compatible.

So by default, pin 7 becomes a select line and when voltage is detected on that line, a different mode is automatically switched to? (for paddles, VCD touchpads, etc)

Having a select line set-up like that would double the number of digital inputs possible, correct? (10 rather than 5) Would you need additional circuitry onboard the controller, like the genesis does, to use such a layout?
What are the select lines normally used for on the A800? (I know the 5200 uses them for selecting the controller having keys scanned by POKEY) I mean, if an integrated keypad didn't really matter, you could simply use the select line in such a configuration as a second trigger input. (rather than using an analog input for trigger)
Of course, if you opted for a 2 port system, you'd have an extra GTIA trigger input per controller too.

[atariksi]

kool kitty89, on Sun Mar 28, 2010 2:42 AM, said:

atariksi, on Sun Mar 28, 2010 1:11 AM, said:

Just like Sega Genesis used the pin 7 for a select line so use pin 7 as +5V and 0V via the CONSOL select line. Joysticks normally don't use +5V at all so they function with pin 7 at 0V and +5V on pin 7 selects other controllers like touchpad. In fact, current A5200 systems use a CONSOL pin (S2) to select POT GND enable/disable (+5V/0V).

So it remains perfectly backward compatible.

So by default, pin 7 becomes a select line and when voltage is detected on that line, a different mode is automatically switched to? (for paddles, VCD touchpads, etc)

Having a select line set-up like that would double the number of digital inputs possible, correct? (10 rather than 5) Would you need additional circuitry onboard the controller, like the genesis does, to use such a layout?
What are the select lines normally used for on the A800? (I know the 5200 uses them for selecting the controller having keys scanned by POKEY) I mean, if an integrated keypad didn't really matter, you could simply use the select line in such a configuration as a second trigger input. (rather than using an analog input for trigger)
Of course, if you opted for a 2 port system, you'd have an extra GTIA trigger input per controller too.

Yeah, you would need some circuit for an integrated keypad + joystick controller. However, for normal joystick only or paddles only or keypad only, you just have to set CONSOL output so it's +5V default which is what pin 7 is anyway. Yeah, it would allow you to read the 5 digital lines in two different modes with pin 7 = 0v or +5V if the circuit is there. I guess you could use pin 7 for input as well since CONSOL supports both input and output.

On A800, select lines are input only for START/SELECT/OPTION and S3 for speaker output.