A8 vs C16/Plus4

[oky2000]

What I did find was that there is a hell of a lot of games I tried so far that don't work properly in VICE! Mostly vertical scrolling games. There is YAPE of course but the video recording does not work properly so I can't upload C16 video *meh*

[Rybags]

I just use YAPE - it's the best one and emulates the undocumented hardware features better than anything else.

[oky2000]

I use YAPE too but can't record any videos for my Youtube channel with it which is a shame.

 

Looking at some vids on youtube I see the C16 can produce some punchy sfx in the style of the VCS (and some A8 games). Digisamples are possible if not as nice sounding as the 6581.

[Rybags]

You could probably use Camtasia or maybe FRAPS to record from YAPE.

 

TED isn't so great for samples, there's only 3 bits of volume control, the high bit forces maximum volume (so I guess 9 levels of amplitude).

 

Additionally (I think) you can only have 1 digital or 2 synth channels unlike A8 and C64 where you can mix and match as well as combining voices to get more amplitude levels.

 

Really, TED sound lets the side down - it's poorer than a YM running on 2 voices. About the only positive of it is that you have 10 bits frequency control per voice.

[JamesD]

Clearly, the sound is the weakest part of the Plus/4.

When its used well it sounds alright, but when it isn't it's pretty weak.

[oky2000]

Because laptop audio drivers are never done properly I can't use things like Camtasia to capture with audio sadly.

[oky2000]

Clearly, the sound is the weakest part of the Plus/4.

When its used well it sounds alright, but when it isn't it's pretty weak.

 

For the price the plus/4 should have had SID too.

[JamesD]

Clearly, the sound is the weakest part of the Plus/4.

When its used well it sounds alright, but when it isn't it's pretty weak.

 

For the price the plus/4 should have had SID too.

That was marketing stupidity. They were convinced they could sell it as a business machine. It should have cost about half what they were asking at it's introduction.

[kool kitty89]

I think the C16 would've given the Atari 400 something to think about - but sadly it's actually a more recent machine than the C64 produced in a blind panic as a knee-jerk response to machines like the ZX spectrum. It probably could've had some great software to rival the Atari but people were already on the C64 by this point and it all seemed like a step backwards (no SID, no hardware sprites, limited memory), so it got overlooked.

If they just wanted a lower-end machine, they could/should have released a cut-down C64 (but NOT the Max). A 16k C64 in the normal breadbox form factor might have been right. (especially if some of the peripheral logic was cut out -maybe one of the CIAs- with provisions for full expansion to C64 spec)

 

By the time the C16 was actually released, 16k was just too little and the full C64 was already in the low-end price range (well in the US it was . . . in Europe the Speccy was king of the low-end -but a 16k C64 really wouldn't have solved that as 16k is just too limited for tape based games -and carts were unpopular in Europe due to cost)

Hell, the VIC itself was already better than the Speccy for games in some respects (better color, trade-offs for graphics, better sound), but it wasn't as cheap -especially if they boosted the internal RAM. (maybe a DRAM based VIC could have been used instead if it didn't cause compatibility issues with the SRAM based original)

 

The Plus/4 was a different issue . . . it wasn't such a low-end system and definitely was not a successor to the VIC or C64, but something overlapping and in direct conflict with those systems. (I don't know why they didn't just release a C64 with a software bundle comparable to the Plus/4, or have the software in expanded onboard ROM like the Plus/4)

Releasing yet another machine with similar (weaker in some areas) specs and an unattractive price point was rather pointless and wasteful. (waste of R&D, manufacturing, advertising, and watering down the C64's own market)

 

At least the C128 made some sense. (though it would have been a lot better had it actually been an efficient evolution of the C64 hardware directly -apparently hindered by lost engineering staff and management decisions) Albeit, one could argue that they could have pushed a low-end Amiga derivative instead and further cut-out unnecessary hardware lines to support. (cutting things back and focusing just on the C64 and a range of Amiga models might have been the smartest option for 1985 -and if AmigaDOS was impractical to run on the low-end model, perhaps support CP/M 68k -and allow Amiga DOS to run after RAM/ROM expansion)

 

 

 

 

 

The Plus4 is virtually unrivaled among 8-bitters for static graphic capability, but that's it's trump card. When it comes to what matters for gaming, it doesn't quite have what it takes to match the A8 and C64 - ie ability to have scrolling/sprites with little CPU usage, and proper sound/music.

If CBM had updated the VIC 2 to have an extended palette (or even some sort of hack to support 16 indexed colors -from a larger palette- to displace the default palette -but otherwise manipulate those 16 colors in a similar manner).

I wonder if that was something they could have done even without the VIC's designers.

 

Regardless, the TED's graphics capabilities were still relatively limited overall . . . the ST or Amiga would have been the ones to have for that at the time, but they were both a lot more expensive, obviously -but if you were doing low-end home graphics/art type stuff, the C64 or A8 alone weren't too bad. (the MSX2 had better graphics than the TED too, but that was '86 -as was the CoCoIII, though 16 colors indexed from 6-bit RGB is a mixed bag against the TED's atribute cells)

On the other hand, both the ST and (especially) the Amiga could have been cut down to compete closer to the price range of the Plus/4 (or a lot closer to the C128). (the Amiga obviously had a lot more room to get cut down, though even the ST probably could have been cut back a fair bit more -in the case of RAM, they both would need alternate low-memory OSs to work with -and at best, they'd still have been a good bit more expensive than the Plus/4's initial $300 pricing, but the long-term value would have been a lot better)

 

 

 

 

 

All I know is Jack Tramiel basically wanted a replacement for the VIC20 to clean up at the low end. So that was meant to be a $60-70 retail machine with the rubber keys like the Aquarius/Spectrum otherwise identical to the C16 as we know.

That's not a replacement for the VIC, it's an even lower-end machine than the VIC ever was (which at least had a full keyboard and built-in peripheral ports).

 

If they just wanted a low-end model, it would have made a hell of a lot more sense to stick with existing hardware (so as to not segment the market with incompatible machines and to consolidate production). But, again, a 16k C64 probably wouldn't have been marketable at all in '84/85. (the 600XL was seen as a complete joke in Europe due mainly to the piddly amount of memory -the Spectrum 16k had never sold very well compared to the 48k, even back in '82, but by '84 16k sales were virtually nonexistent from what I've seen -or at least very tiny compared to 48k sales)

 

Maybe, just maybe they could have gotten by with a 32k machine (the 600XL might have been OK if it was at least 32k), especially since 32k on sprite/character based hardware (A8 or C64) would stretch further than the framebuffer based Spectrum/CPC where . (aside from games using framebuffers on the A8 or pseudo framebuffers on the 64) The Speccy 48k had up to 16kB reserved for graphics memory (though for normal 256x192 double buffered games, it would only be 13.5 kB, or 6.75 kB if single buffered), but on the CPC you're definitely stuck with at least 16k for the framebuffer and 32k if you double buffer (without clipping to a window smaller than 160/320x200).

 

I think had Commodore not messed about and tried to make such a hash of the whole project it would have been a nice machine for the price.

Nice hardware, perhaps, but totally unnecessary compared to what was already on the market. (the 128 is more arguable, but there should have been better alternatives for that too)

 

Expanding a compatible range of machines makes tons of sense, but pushing out a new line that was incompatible (and in pretty direct conflict with) existing machines really doesn't make sense at all. (at least the 600XL was largely backwards compatible with the 400, and readily expandable to 800XL spec)

 

I don't think it is as powerful as the 600XL/400 models except possibly if it has a C64 style char mode for very flexible total on screen colours, but I don't know so I am hoping we can have a nice discussion between the machines.

it does have fairly C64-like color flexibilities (except 2 of the colors were fixed per screen/line rather than 1 in the C64), but no sprites.

The C16 also had a nominally slower CPU (.89 MHz in RAM), but I think the Plus/4 was able to run at full speed (or close, but with wait states) in RAM. (sort of like the BBC Micro vs Electron)

 

Compared to the A8, it had weaker sound (fewer channels, less flexibility), somewhat better color capabilities (though the A8 can be made quite nice with good optimization, especially with 5 colors per line plus careful use of sprites), lacks hardware scrolling, lacks sprites (even more rudimentary Atari-style sprites), and it obviously lacks the true bitmap/framebuffer modes as well as the low res 4bpp modes of GTIA. (and that's on the hardware end, not considering software support -or the A8's huge back library- or the lack of incentive for developers to support yet another machine when it really had little to offer over the C64, the Atari 8-bit had a hard enough time getting support in the mid/late 80s as it was, let alone a new/incompatible machine -and much of what the Plus/4 did offer could have been accomplished with a C64 compatible machine, or simply a software bundle with a stock C64)

 

You could do spectrum style rendering on it pretty well, but with much more color (and more per cell in the lower res mode) and better sound than a 48k. (worse than a 128k obviously)

But again, why bother when the C64 is already there and around $100 cheaper than the Plus/4 was at release? (and the C64 was too weak to be worth much in '84, let alone later on)

 

I have seen now it has pixel scroll so really it's the sound and the sprites that are different to A8/C64.

It has scrolling? Where did you see that? (especially horizontal scrolling, since that's the really hard one to do in software -unfortunate that the MSX2 only got the V scroll register)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Edit:

Also note, I'm not trying to bash the Plus/4 or C16 out of spite or anything, nor do I dislike it personally . . . I was just speaking from a historical observation perspective. (at least from what I know)

The Plus/4 and C16 really didn't make business sense for CBM. (more so in hindsight obviously, but even at the time it made little sense to release them or even invest in their development -even the C128 seems a bit odd in a number of areas, but it at least made a lot more sense than either of the TED based machines)

Edited June 6, 2011 by kool kitty89

[Thorsten Günther]

Clearly, the sound is the weakest part of the Plus/4.

When its used well it sounds alright, but when it isn't it's pretty weak.

 

For the price the plus/4 should have had SID too.

That was marketing stupidity. They were convinced they could sell it as a business machine. It should have cost about half what they were asking at it's introduction.

That was the funniest fact about the plus/4: It was sold at the same price as the Sinclair QL here in Germany. But while the Sinclair QL already came with two storage devices built in (albeit them only being Microdrives), 128K RAM, a 512 pixel horizontal resolution, a full expansion bus (which was heavily used for Shugart drive interfaces and RAM expansions) and decent business software by Psion, the plus/4 came with 64K RAM and barely useful softwrae instead.

 

Thorsten

Edited June 6, 2011 by Thorsten Günther

[oky2000]

The plan was simple as directed by Jack to the engineers, build me a low end machine to plug the gap left by the VIC-20 being removed from the shop shelves. Don't forget this was for a future time when Jack wanted the C64 down by $100 to $199 to wipe the floor with everyone (and why he left when Irving Gould refused).

 

So $199 for C64 alone and $70-80 to take up the bottom end of the market. The project spiralled out of all control and is explained very nicely by Dave Haynie and Bill Hurd on their Youtube videos.

 

Beginning of the end for Commodore 8bits. The C128 is another shiny turkey with no ability to use VIC-II in 2mhz mode, for £400, crazy!

 

I am pretty sure they couldn't modify/enhance VIC-II for the same reason SID lasted forever untouched (or crippled in the case of the 8580 replacement to SID) ie the designers of those chips had left MOS Technologies.

[Rybags]

Weird about that with VIC-2.

 

Then again, kinda weird that they made the C64 CPU operate at 1 MHz instead of just putting up with variable DMA steals.

 

But, like the shonky disk access and 1976ish Kernal ROM, probably a case of knocking that extra buck or two off unit cost and meeting the initial deadline.

[oky2000]

I think the C16 would've given the Atari 400 something to think about - but sadly it's actually a more recent machine than the C64 produced in a blind panic as a knee-jerk response to machines like the ZX spectrum. It probably could've had some great software to rival the Atari but people were already on the C64 by this point and it all seemed like a step backwards (no SID, no hardware sprites, limited memory), so it got overlooked.

If they just wanted a lower-end machine, they could/should have released a cut-down C64 (but NOT the Max). A 16k C64 in the normal breadbox form factor might have been right. (especially if some of the peripheral logic was cut out -maybe one of the CIAs- with provisions for full expansion to C64 spec)

 

By the time the C16 was actually released, 16k was just too little and the full C64 was already in the low-end price range (well in the US it was . . . in Europe the Speccy was king of the low-end -but a 16k C64 really wouldn't have solved that as 16k is just too limited for tape based games -and carts were unpopular in Europe due to cost)

Hell, the VIC itself was already better than the Speccy for games in some respects (better color, trade-offs for graphics, better sound), but it wasn't as cheap -especially if they boosted the internal RAM. (maybe a DRAM based VIC could have been used instead if it didn't cause compatibility issues with the SRAM based original)

 

The Plus/4 was a different issue . . . it wasn't such a low-end system and definitely was not a successor to the VIC or C64, but something overlapping and in direct conflict with those systems. (I don't know why they didn't just release a C64 with a software bundle comparable to the Plus/4, or have the software in expanded onboard ROM like the Plus/4)

Releasing yet another machine with similar (weaker in some areas) specs and an unattractive price point was rather pointless and wasteful. (waste of R&D, manufacturing, advertising, and watering down the C64's own market)

 

At least the C128 made some sense. (though it would have been a lot better had it actually been an efficient evolution of the C64 hardware directly -apparently hindered by lost engineering staff and management decisions) Albeit, one could argue that they could have pushed a low-end Amiga derivative instead and further cut-out unnecessary hardware lines to support. (cutting things back and focusing just on the C64 and a range of Amiga models might have been the smartest option for 1985 -and if AmigaDOS was impractical to run on the low-end model, perhaps support CP/M 68k -and allow Amiga DOS to run after RAM/ROM expansion)

 

 

 

 

 

The Plus4 is virtually unrivaled among 8-bitters for static graphic capability, but that's it's trump card. When it comes to what matters for gaming, it doesn't quite have what it takes to match the A8 and C64 - ie ability to have scrolling/sprites with little CPU usage, and proper sound/music.

If CBM had updated the VIC 2 to have an extended palette (or even some sort of hack to support 16 indexed colors -from a larger palette- to displace the default palette -but otherwise manipulate those 16 colors in a similar manner).

I wonder if that was something they could have done even without the VIC's designers.

 

Regardless, the TED's graphics capabilities were still relatively limited overall . . . the ST or Amiga would have been the ones to have for that at the time, but they were both a lot more expensive, obviously -but if you were doing low-end home graphics/art type stuff, the C64 or A8 alone weren't too bad. (the MSX2 had better graphics than the TED too, but that was '86 -as was the CoCoIII, though 16 colors indexed from 6-bit RGB is a mixed bag against the TED's atribute cells)

On the other hand, both the ST and (especially) the Amiga could have been cut down to compete closer to the price range of the Plus/4 (or a lot closer to the C128). (the Amiga obviously had a lot more room to get cut down, though even the ST probably could have been cut back a fair bit more -in the case of RAM, they both would need alternate low-memory OSs to work with -and at best, they'd still have been a good bit more expensive than the Plus/4's initial $300 pricing, but the long-term value would have been a lot better)

 

 

 

 

 

All I know is Jack Tramiel basically wanted a replacement for the VIC20 to clean up at the low end. So that was meant to be a $60-70 retail machine with the rubber keys like the Aquarius/Spectrum otherwise identical to the C16 as we know.

That's not a replacement for the VIC, it's an even lower-end machine than the VIC ever was (which at least had a full keyboard and built-in peripheral ports).

 

If they just wanted a low-end model, it would have made a hell of a lot more sense to stick with existing hardware (so as to not segment the market with incompatible machines and to consolidate production). But, again, a 16k C64 probably wouldn't have been marketable at all in '84/85. (the 600XL was seen as a complete joke in Europe due mainly to the piddly amount of memory -the Spectrum 16k had never sold very well compared to the 48k, even back in '82, but by '84 16k sales were virtually nonexistent from what I've seen -or at least very tiny compared to 48k sales)

 

Maybe, just maybe they could have gotten by with a 32k machine (the 600XL might have been OK if it was at least 32k), especially since 32k on sprite/character based hardware (A8 or C64) would stretch further than the framebuffer based Spectrum/CPC where . (aside from games using framebuffers on the A8 or pseudo framebuffers on the 64) The Speccy 48k had up to 16kB reserved for graphics memory (though for normal 256x192 double buffered games, it would only be 13.5 kB, or 6.75 kB if single buffered), but on the CPC you're definitely stuck with at least 16k for the framebuffer and 32k if you double buffer (without clipping to a window smaller than 160/320x200).

 

I think had Commodore not messed about and tried to make such a hash of the whole project it would have been a nice machine for the price.

Nice hardware, perhaps, but totally unnecessary compared to what was already on the market. (the 128 is more arguable, but there should have been better alternatives for that too)

 

Expanding a compatible range of machines makes tons of sense, but pushing out a new line that was incompatible (and in pretty direct conflict with) existing machines really doesn't make sense at all. (at least the 600XL was largely backwards compatible with the 400, and readily expandable to 800XL spec)

 

I don't think it is as powerful as the 600XL/400 models except possibly if it has a C64 style char mode for very flexible total on screen colours, but I don't know so I am hoping we can have a nice discussion between the machines.

it does have fairly C64-like color flexibilities (except 2 of the colors were fixed per screen/line rather than 1 in the C64), but no sprites.

The C16 also had a nominally slower CPU (.89 MHz in RAM), but I think the Plus/4 was able to run at full speed (or close, but with wait states) in RAM. (sort of like the BBC Micro vs Electron)

 

Compared to the A8, it had weaker sound (fewer channels, less flexibility), somewhat better color capabilities (though the A8 can be made quite nice with good optimization, especially with 5 colors per line plus careful use of sprites), lacks hardware scrolling, lacks sprites (even more rudimentary Atari-style sprites), and it obviously lacks the true bitmap/framebuffer modes as well as the low res 4bpp modes of GTIA. (and that's on the hardware end, not considering software support -or the A8's huge back library- or the lack of incentive for developers to support yet another machine when it really had little to offer over the C64, the Atari 8-bit had a hard enough time getting support in the mid/late 80s as it was, let alone a new/incompatible machine -and much of what the Plus/4 did offer could have been accomplished with a C64 compatible machine, or simply a software bundle with a stock C64)

 

You could do spectrum style rendering on it pretty well, but with much more color (and more per cell in the lower res mode) and better sound than a 48k. (worse than a 128k obviously)

But again, why bother when the C64 is already there and around $100 cheaper than the Plus/4 was at release? (and the C64 was too weak to be worth much in '84, let alone later on)

 

I have seen now it has pixel scroll so really it's the sound and the sprites that are different to A8/C64.

It has scrolling? Where did you see that? (especially horizontal scrolling, since that's the really hard one to do in software -unfortunate that the MSX2 only got the V scroll register)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Edit:

Also note, I'm not trying to bash the Plus/4 or C16 out of spite or anything, nor do I dislike it personally . . . I was just speaking from a historical observation perspective. (at least from what I know)

The Plus/4 and C16 really didn't make business sense for CBM. (more so in hindsight obviously, but even at the time it made little sense to release them or even invest in their development -even the C128 seems a bit odd in a number of areas, but it at least made a lot more sense than either of the TED based machines)

 

The C16 made perfect business sense for Jack's strategy, $199 C64 and $80 C116 style rubber keyed machine. As he predicted the sales would have been astronomical. The C16 and Plus/4 are post Jack Commodore not knowing anything about good business (kept C64 price very high and tried to make another incompatible expensive machine out of their cheap $70 machine project chipset).

 

It does have pixel scroll hardware, just view the games on Youtube, too many rubbish looking games have smooth scrolling really.

 

VIC-II and SID hardware was not cheap, to then make a separate motherboard and reduce it to 16k (3 years after original model launched with 64kb so ALL old games are still incompatible) was not good business sense, as 48k less doesn't wipe out that much of the total cost in development I guess anyway.

 

Jack had the right idea, pile em high sell em cheap in the case of the C64 in 84 and ditto with new cheap low end machine. No point mixing the two, the VIC20 sold over a million units so as a concept it was fine, the C116 style machine would have taken that low end market and run with it, and for the price you can't argue with it really. The only competition was Atari 600XL/400 and Jack was making sure the price was very much below their RRP.

[Rybags]

48K less in 1984 wouldn't have made a lot of difference.

 

I remember around that time, maybe a little later, RAM retailing at $1 per K.

 

So, it'd make sense that bulk wholesale buys would probably have been about half that price, maybe less.

 

And, as stated - no point in a 16K C64 variant. 95% + of games wouldn't work on it. And it was never renowned for it's cartridge based games in any case.

Edited June 6, 2011 by Rybags

[oky2000]

48K less in 1984 wouldn't have made a lot of difference.

 

I remember around that time, maybe a little later, RAM retailing at $1 per K.

 

So, it'd make sense that bulk wholesale buys would probably have been about half that price, maybe less.

 

And, as stated - no point in a 16K C64 variant. 95% + of games wouldn't work on it. And it was never renowned for it's cartridge based games in any case.

 

Maybe, can't remember ram prices for 1984/85 really but if so my bad. Was a massive blow to the competition in 1982 having just 64kb I remember. So for Commodore in bulk trade purchasing maybe £24 for 48kb. I know in early 1982 Sinclair sold the 16kb Spectrum for 125 and 48kb for 175 so £50 difference for them so I guessed price maybe halved by mid 84 but maybe not.

 

If it sold I guess it would have had some games modified into multi-load tape games in the EU. Depends on the final memory map, the C64 steals a lot of that 64kb even in machine code. If they kept the screen memory out of the user memory map not so bad....8kb to play with = tough call so maybe carts would be the only option.

 

I find the 12bit DMA on VIC-II quite strange (8+4 really) and it is running at 18mhz if I remember right, so making a 2mhz 6510/8510 and a 36mhz VIC-II for compatibility was probably no easy task as the original designers left C= in 82 I believe. I think for the C64 it is enough for anything other than 3D or isometric games.

 

Then again just look at the Commodore Max....C64 with 4kb memory and Atari 400 style membrane keyboard. Was half the cost of C64 and sold badly (Canned same year as launched!) And this explains why the C64 cartridges were so basic and simplistic not just the cost of 32kb cartridge roms I guess.

 

Edit what I'd love to know is the cost of SID/VIC2 production vs cost of TED production.

Edited June 6, 2011 by oky2000

[JamesD]

Edit what I'd love to know is the cost of SID/VIC2 production vs cost of TED production.

As the number of gates goes up, the cost increase is not linear but steadily becomes exponential.

TED drops sprites, audio registers, analog hardware, reduces packaging, and requires a smaller motherboard layout.

The TED may or may not have half the number of gates/logic but when all things are considered, it was probably less than 1/2 the price of the SID/VIC2.

[JamesD]

The plan was simple as directed by Jack to the engineers, build me a low end machine to plug the gap left by the VIC-20 being removed from the shop shelves. Don't forget this was for a future time when Jack wanted the C64 down by $100 to $199 to wipe the floor with everyone (and why he left when Irving Gould refused).

 

So $199 for C64 alone and $70-80 to take up the bottom end of the market. The project spiralled out of all control and is explained very nicely by Dave Haynie and Bill Hurd on their Youtube videos.

Well, I don't think the 116 would have fared well in the US anyway. Every machine in the US with a 2nd rate keyboard either had a short life for was upgraded with a real keyboard. The C16 was probably a poor choice for an upgrade as well as it might have confused some people due to the similarity to the VIC & C64 case.

If the Plus/4 had come out with 16K expandable to 64k and without the +4 applications, it probably would have only cost another $10-$15 over the original price point. They ended up selling the Plus/4 for $100 for much of it's life anyway and I'm sure it could have gone lower if Commodore had stuck close to the original plan.

[JamesD]

If CBM really wanted a business machine from the TED, they should have introduced a model expandable to 128K RAM supported by BASIC. Plus/4 Basic already supported more RAM than any other version so it's not a stretch of the imagination that they could have made use of more RAM by paging. That would have at least set it apart from the C64.

[+SpiceWare]

Depends on the final memory map, the C64 steals a lot of that 64kb even in machine code.

Easily worked around. My original BBS software stored all of it's machine language code in the 4K visible at $C000. The later versions stored all their code in the RAM under the 16K ROMs for BASIC and the KERNEL, with a small jump table in $C000 that would swap the ROM out of the way as needed. The rest of the $C000 was used to store sprite graphics, text buffers and animated character data. I did a blog entry about it a while back.

[Rybags]

I believe VIC runs at 8 MHz (parts of) but it only does memory access at 2 MHz effective.

 

TED in all likelihood runs at 7.2 MHz, but again the memory access is via /4 clocking.

[barnieg]

I seem to remember reading somewhere that in the latter stages of its life the C64 cost $30 to produce? It would be about right for Commodore to have screwed their ability to cost reduce the C64 chipset, didn't they have to reverse engineer the original Amiga Chipset to allow ECS to be produced?

 

However even without complete C64 compatibility the Plus 4 would have benefited from more RAM, Sid for sound and probably 128K - they could have ditched the horrid built in software.

 

Barnie

 

 

 

Edit what I'd love to know is the cost of SID/VIC2 production vs cost of TED production.

As the number of gates goes up, the cost increase is not linear but steadily becomes exponential.

TED drops sprites, audio registers, analog hardware, reduces packaging, and requires a smaller motherboard layout.

The TED may or may not have half the number of gates/logic but when all things are considered, it was probably less than 1/2 the price of the SID/VIC2.

[Rybags]

Sounds feasible (the $30).

 

By the late 1980s the 6502 was essentially a glorified 74LS part insofar as complexity/yields.

 

In all probability the keyboard was the most expensive component by then. By 1990 I imagine RAM would have been in the order of 20 cents/K or less in bulk orders.

[kool kitty89]

The plan was simple as directed by Jack to the engineers, build me a low end machine to plug the gap left by the VIC-20 being removed from the shop shelves. Don't forget this was for a future time when Jack wanted the C64 down by $100 to $199 to wipe the floor with everyone (and why he left when Irving Gould refused).

The C64 was already down to that price in 1983, and down to $99 after the big rebate. (there was a bit of a range of pricing, but at the bix box retailers, it was already $200 in '83, plus the $99 rebate with the return of ANY computer of game console )

That was a temporary deal though, and the C64 went back to a nominal $170-200 in mid 1984. (when the Atari 800XL was down to $99 -probably a bit excessive given they didn't have the volumes to keep up with demand at that price)

 

However, I still don't see why it makes sense to go with a totally new machine rather than consolidating the market with derivatives of existing models. (they needlessly stretched themselves too thin with such overlapping products)

And, in hindsight, the C64 was already cheap enough (at least outside of Europe), and even in Europe, they could have left Sinciar to the super low end and focused on the C64 as their bottom-end machine in general.

 

Beginning of the end for Commodore 8bits. The C128 is another shiny turkey with no ability to use VIC-II in 2mhz mode, for £400, crazy!

The 128 at least had good features for a business machine in general. (though again, if they couldn't upgrade the C64's based hardware, it would have made a lot more sense to just stick with the plain C64 and push a low-end Amiga derivative much, much sooner -a stripped-down, 128k, console form factor Amiga would have been interesting to see -especially since that's more or less the configuration it was to be as a game console with Atari Inc )

 

I am pretty sure they couldn't modify/enhance VIC-II for the same reason SID lasted forever untouched (or crippled in the case of the 8580 replacement to SID) ie the designers of those chips had left MOS Technologies.

Yes, bad move not keeping those engineers on. Albeit, if they had comprehensive documents on the chip design, new engineers should have been able to make some modifications too. (and reverse engineering is always an option too -and more practical with simpler/larger chips -the sheer size of the interconnect and use of single layer would facilitate that)

 

In the odd case of the SID, they corrected an old bug, but a bug that had been heavily exploited and should have been replicated on the redesign. (they definitely should have been revising their chips, and such mistakes are just examples of messed up R&D management)

Of course, it could also be a case where the C64 dev manuals specifically warned against use of undocumented features that could be removed. (except CBM took so long to make a revision that the removed bug was anything but "undocumented" by that point )

 

CBM doesn't seem to have put enough resource into reworking and cost reducing its existing chip designs and hardware. (they should have had an all 5V version of the C64 chipset much sooner, and should have merged some of the sound/video/IO chips by the time of the 64C -if not made an SOC by the late 80s) Atari had made a 5V chipset with the A8 from day 1, and had definitely been working on further integration and cost reduction. (in '83, they'd already prototyped the "JAN" VCS on a chip and CGIA chip for the A8 -the VCS on a chip was eventually used in later Jr models by Atari Corp, the CGIA was never used, though I wonder if Atari Corp ever exhausted their initial stockpiles of ANTIC+GTIA chips -stockpiles of VCS chips probably contributed rather significantly to the delay in introducing the single chip version)

 

Hell, even without really comprehensive re-engineering, die shrinks are usually very practical. (not as efficient as a re-engineered chip for the new process, but a cost savings nevertheless)

 

 

 

 

 

 

 

 

The C16 made perfect business sense for Jack's strategy, $199 C64 and $80 C116 style rubber keyed machine. As he predicted the sales would have been astronomical. The C16 and Plus/4 are post Jack Commodore not knowing anything about good business (kept C64 price very high and tried to make another incompatible expensive machine out of their cheap $70 machine project chipset).

Yes, but why the hell bother with a totally new design? Why not have something based on the VIC or C64? (especially if such a product was started before the original engineers left MOS/CSG)

 

Why not just a 16k C64 with a rubber keyboard and some omitted I/O and perhipheral logic? (sort of like the MAX, but not as limited, and probably chiclet rather than membrane -though a full-sized membrane board is better than an undersized chiclet one )

 

It does have pixel scroll hardware, just view the games on Youtube, too many rubbish looking games have smooth scrolling really.

Can you point to any hardware documents or comments in tech/programmer comments?

 

Vertical scrolling is not surprising, but horizontal smooth scrolling in software (commonly) would be a bit surprising. (though it's pretty easy with a character based display and relatively simple character set -just use animation in leu of smooth scrolling)

 

VIC-II and SID hardware was not cheap, to then make a separate motherboard and reduce it to 16k (3 years after original model launched with 64kb so ALL old games are still incompatible) was not good business sense, as 48k less doesn't wipe out that much of the total cost in development I guess anyway.

Yes, but it at least makes a hell of a lot more sense than a totally new machine as it would be forward compatible.

And designing a modified, cost-cut motherboard would be a lot less than a totally new machine with new tooling for everything.

 

If they WERE going to launch a low-end machine, they really should have done it from day 1 (like the 400 and 800), they DID do that with the MAX, but it was even more limited than the VIC in some respects. (weak keyboard, lack of RAM expansion, reliance on ROM software, etc) They should have made it expandable to C64 spec too (like the 600XL and Spectrum 16k).

They really should have been working on cost reducing the C64 too, especially consolidating the chipset. (and cutting things down to unified 5V power prior to that)

 

But again, there was really no market for such a machine by the time it was released. They weren't going to cut-in on the spectrum unless they had it out by 1982/83 (after which, the spectrum was entrenched as the bottom-end system), the C64 itself was already pretty damn inexpensive by 1983/84 (in the US at least).

 

Jack had the right idea, pile em high sell em cheap in the case of the C64 in 84 and ditto with new cheap low end machine. No point mixing the two, the VIC20 sold over a million units so as a concept it was fine, the C116 style machine would have taken that low end market and run with it, and for the price you can't argue with it really. The only competition was Atari 600XL/400 and Jack was making sure the price was very much below their RRP.

Yes and no, the low-cost angle was good, but he went too far at times . . . except with CBM's funding he could afford to do that and still have enough to push advertising. (something that severely limited the ST in the US -in Europe, viral marketing was far, far, more effective)

 

However, segmenting the market with totally incompatible/unrelated machines makes rather little sense. Hell, not making the C64 VIC-20 compatible was a bit of a mistake (especially since it should have been possible to efficiently evolve the VIC-I into something close to the VIC-II), heh, then you could even have a low-end system sans the SID and still OK sound via the VIC. (albeit, as it was, they could have dropped the SID and used a beeper or-better- a bare DAC -especially useful with the CIA's high-res timers)

 

If it wasn't for the VIC-20's use of SRAM and low resolution limits (not even 32 column text support), it could have been a good option to continue pushing on its own.

 

 

 

In the sense of a low-end business machine in Europe (or the US), CBM was already sort of screwed by the introduction of the CP/M compatible Amstrad CPC with 80 column display (which also had better color capabilities than the C64 -odd that so many gamed used mode 1 rather than mode 0 -which looked better, used no more RAM, and was easier to render to).

The only thing the CPC lacked was TV compatibility. (albeit the 80 column mode would have been questionable via RF, even with colorburst disabled)

 

 

So, while a replacement for the VIC made some sense, by the time they actually got around to doing it, there was basically no market for such a machine, and the C64 was basically positioned in the same range as the VIC had been back in '82. (except with much better software support)

 

In that sense (ie launching such a machine in 1984), perhaps just a cut-down 64k C64 might have made some sense. (ie just a cheaper case/keyboard and maybe omitted support for the floppy drive or something else along those lines -not sure how the CIAs are configured, but maybe 1 could be removed)

 

For that matter, they never released a "cheap" form factor version of the VIC-20 itself.

 

 

 

 

 

 

48K less in 1984 wouldn't have made a lot of difference.

 

I remember around that time, maybe a little later, RAM retailing at $1 per K.

http://phe.rockefeller.edu/LogletLab/DRAM/dram.htm

 

But don't think that component cost translated to consumer pricing directly. There's a lot of factors that tend to inflate any such costs (including cost of added board space/traces needed for larger numbers of RAM chips -vs 2 16kx4-bit chips like the 600XL- and various inflationary factors for consumer level products), and then there's sheer profit margins beyond that. (higher RAM versions are usually higher-end and almost always have inflated profit margins -hence the 800XL being much more costly than the 600XL at launch)

 

So, it'd make sense that bulk wholesale buys would probably have been about half that price, maybe less.

 

And, as stated - no point in a 16K C64 variant. 95% + of games wouldn't work on it. And it was never renowned for it's cartridge based games in any case.

Yes, but 100% of the 16k games would work on a full C64 (vs zero of the C116/16/Plus/4), used existing hardware being programmed for, and (if CBM chose too) could be expanded to full C64 spec. (though they really should have done it from day 1 or not at all)

 

By the time they DID release the 116/16/Plus4, there was pretty much no reason to do so, especially at the prices CBM was asking. (if they wanted a successor to the VIC -or, indeed, an even lower end system- they should have pushed in in parallel with the C64, and really should have aimed at cross-compatibility with the C64, especially if they weren't going to be including VIC compatibility)

 

The same thing happened with the Acorn Electron, albeit a cut-down C64 might have had more of a chance due to sheer popularity. (in Acorn's case, they probably should have never released the Electron and instead focused on using the new ASICs in low-cost versions of fully capable BBC Micro compatibles)

 

 

Edit what I'd love to know is the cost of SID/VIC2 production vs cost of TED production.

As the number of gates goes up, the cost increase is not linear but steadily becomes exponential.

TED drops sprites, audio registers, analog hardware, reduces packaging, and requires a smaller motherboard layout.

The TED may or may not have half the number of gates/logic but when all things are considered, it was probably less than 1/2 the price of the SID/VIC2.

Yes, perhaps, but then there's the considerations of R&D, advertising, tooling for production (masks, molds, etc), software development, etc, etc.

 

There's also the issue where silicon gets cheaper and cheaper, but PCB/traces and chip packages don't get cheaper nearly as fast. (so that's definitely an added cost of VIC+SID, especially with the added I/O chips and such)

Though, again, they could also have dropped the SID in favor of a speccy/PC style beeper or DAC a la coco. (except with much better timers to work with than even the coco 3 -via CIA -hmm, is that what the Electron's "PSG emulation" is -software toggled square waves)

 

 

Hell, given the rather pointless nature of the Plus/4 and co, and the less than ideal nature of the C128, maybe they could have done better than either:

If they were going to go with another new, incompatible machine, and they wanted a low-cost, business oriented machine (a heavily cut-down amiga still wouldn't be cheap enough), having the C128 without the C64 components (maybe with embedded TED-like sound -or just an AY8910/YM2149) could have made sense. (a low-cost, CP/M compatible business machine with high-res RGBI graphics/text and rudimentary blitter functionality, or have graphics closer to the TED, but with 80 column support and a Z80 rather than 6502)

That wouldn't mesh with the VIC-20 replacement/Speccy killer idea (which really wouldn't have made sense unless released by 1983, or better if side by side with the C64), but it definitely would mesh with the concept of the Plus/4.

 

 

 

 

 

Sounds feasible (the $30).

 

By the late 1980s the 6502 was essentially a glorified 74LS part insofar as complexity/yields.

 

In all probability the keyboard was the most expensive component by then. By 1990 I imagine RAM would have been in the order of 20 cents/K or less in bulk orders.

RAM and the chipset (custom and off the shelf LSI and discrete logic) were still significant, but the cost of the power supply, PCB, case, connectors, and keyboard were all significant too. (keyboard was probably one of the more persisting costs as you mentioned -the things CBM didn't produce in-house in general could be less attractive)

So any such cuts in any of those areas could be quite significant. (and if we're talking 1984 here -not 1982 with a 400/800 like release- it would make more sense to have a nearly fully compatible 64k C64, but maybe missing 1 CIA, some of the peripheral support, and with a cheap-o keyboard -if they wanted to intentionally limit expandability, making it RF and tape/cart only might make sense, but support for expandability to a full C64 via a module might be nice -both more attractive to consumers, and potentially hefty profits from that otherwise cheap add-on module )

 

In the US, the C64 was basically cheap enough already, though in Europe (prior to 1985, at least), a cheaper model would have made sense.

 

Hmm, this is rather like some previous comments I made about what Atari should have done to make the A8 more attractive in the Euro market.

Edited June 7, 2011 by kool kitty89

[JamesD]

I seem to remember reading somewhere that in the latter stages of its life the C64 cost $30 to produce? It would be about right for Commodore to have screwed their ability to cost reduce the C64 chipset, didn't they have to reverse engineer the original Amiga Chipset to allow ECS to be produced?

$30 would have been very late in the C64's life.

 

I think people are misunderstanding what happened a Commodore.

While many engineers left Commodore, the C128 still integrated the VIC II logic. And the C65 integrated the VIC II and SAM with new capabilities. It was management decisions that prevented them from releasing a cheaper or better version of the C64.

 

Commodore did not have to reverse engineer the original Amiga Chipset. Management repeatedly killed projects that advanced the Amiga. Multi-CPU boards, A3000+, etc... all killed and the Amiga along with it eventually.

If even those two projects had been released, it might have changed everything. (there's a what if topic for you)

Amiga and Motorola lacked the corporate culture that would create engineering groups capable of competing with the intel world. Instead of dozens of engineers working on parallel projects, they had small groups of engineers working on one generation or product at a time. Even after Commodore's death, Motorola's culture caused them problems. They went from dominating the cell phone market to being almost wiped out. They never have recovered either.

 

However even without complete C64 compatibility the Plus 4 would have benefited from more RAM, Sid for sound and probably 128K - they could have ditched the horrid built in software.

Apple had 128K upgrades with the IIe, Atari released the 130XE, Tandy the CoCo3, Sinclair the Spectrum 128, etc... 128K or more was where everyone else was going in the 8 bit world and if what CBM engineers said was correct, it would have only cost a few dollars. A 128K C64 or 264 (Plus/4 minus software) would have made much more sense than the Plus/4 & C128. The C128 was a bit of a hack and some manager was trying to create yet another business system (Plus/4, B128, C128... I'm seeing a trend).

 

Now, the Plus/4 can be upgraded to 256K RAM (or more) if it's hacked, so there was some potential there. If Basic can actually use that then it's way ahead other machines.

 

If I were to consider any other change to the TED it would be to add some sort of blitter. It would probably take less logic than sprites and doesn't steal clock cycles when it's not actually in use.

[kool kitty89]

I think people are misunderstanding what happened a Commodore.

While many engineers left Commodore, the C128 still integrated the VIC II logic. And the C65 integrated the VIC II and SAM with new capabilities. It was management decisions that prevented them from releasing a cheaper or better version of the C64.

That's what I'd originally thought, but every time I mentioned it, I seemed to get bombarded with "they lost the original engineers" . . . though that might have limited enhancement of the original chips more than integration. (ie actually building on the original logic with enhanced modes -like a different DMA sharing set-up, 80 column text, larger color palette, etc)

 

Not cost reducing the C64 chipset was a mistake though, and one that doesn't make good business sense at all. (perhaps they thought it would die off rather quickly and have no needs for such cost reductions -but then they DID do some half-assed cost cuts with the C64C)

 

Commodore did not have to reverse engineer the original Amiga Chipset. Management repeatedly killed projects that advanced the Amiga. Multi-CPU boards, A3000+, etc... all killed and the Amiga along with it eventually.

Many of the projects they DID have were also wrong for the market in general. (like the rather questionable Ranger chipset) They needed to efficiently expand upon the original chips in a way that could cater to both high and low-end models, and also extend their range of high/mid/low-end models. (efficient use of fast page mode DRAM access was probably the number 1 thing to work on as it would double peak bandwidth with no added cost of RAM or traces on the PCB; improving the DMA set-up to allow the CPU to "steal" the bus for non-interleaved DMA -with wait states for critical chipset operations- would have been very important with faster CPU models -that, along with FPM support probably would have been more useful than the 32-bit chipRAM bus of the A3000)

And they definitely needed chunky pixel modes by the time of AGA (they really should have had it earlier), preferably with the chipset updated to take full advantage of that. (a packed-pixel optimized blitter could be really useful, especially with buffering to take advantage of fast page mode -or with DMA to fastRAM for 3DO style separate source and destination; something really neat for thetime of AGA would have been adding affine texture rendering for scaled/rotated 2D as well as warping with added CPU overhead -like the Sega CD and the Jaguar's rudimentary texture mapping function -weak due to Flare considering it a low priority back in 1990)

 

They should have had 16 (14.3) MHz rated 68000s models with built-in fastRAM as the low-end successor to the 500 by the beginning of the 90s at least, let alone a broader selection of high-end models and more options for expansion. (they had a decent expansion interface, but they didn't support it very well -3rd parties had some good stuff too, but nothing pushed hard enough to become a defacto standard as happened with many, many things on PCs -and a fair amount on the Apple II -that's the main flaw in my argument for Atari adding an expansion bus too, if it wasn't supported well -by Atari or 3rd parties taking up the slack- it wouldn't have made a difference at all -and as it was, both CBM and Atari were lacking in releasing updated models regardless of expansion support . . . though in Atari's case, it at least may have meant better support for the BLiTTER and DMA sound)

Atari actually had a much cleaner design to work with with far more options for evolution without being burdened by complex backwards compatibility, but they failed to take advantage of that in the long run. (a shame they didn't keep up the mix of modest/cheap custom LSI and logical use of the most cost effective off-the-shelf components available -and aggressively striving to find the best priced suppliers; it's pretty clear that things went down hill fast after Jack left, but the ST line had already been stagnating while Jack was still there -had the MEGA added a fastRAM -dedicated CPU- bus and 16 MHz 68k -at least optionally, and perhaps a YM2203, I wouldn't be critical of that -then there's the TT that had no mid-range or low end derivatives and lacked dual playfield or packed pixel modes -and omission of the blitter)

 

Amiga and Motorola lacked the corporate culture that would create engineering groups capable of competing with the intel world. Instead of dozens of engineers working on parallel projects, they had small groups of engineers working on one generation or product at a time. Even after Commodore's death, Motorola's culture caused them problems. They went from dominating the cell phone market to being almost wiped out. They never have recovered either.

Had it not been for powerful competition, Intel would have had a lot more problems too. AMD, Cyrix, (to some extent NEC and others like Marris), etc opened up a competitive market with much tighter margins (and thus much lower market prices) which Intel avoided for a while by continuing to push inflated prices on newer CPUs (regardless of the actual value), but they were eventually forced to compete directly with the likes of AMD and Cyrix. (though Cyrix screwed themselves in the late 90s, or rather National Semiconductor ran them into the ground)

 

Likewise, had Hitachi, ST, Signetics, or other 68k 2nd sources taken the route NEC, AMD (and later Cyrix) did, things would have been quite different. Granted, AMD had a good bit of a boost by access to Intel documentation on the 386 (and to lesser extent, 486), so more intensive reverse engineering (or internal design work -ie building their own enhanced 68000 derivatives) on the part of those 3rd parties (the likes of Hitachi certainly had the resource). In the interim (like AMD and Harris with the 286), those 3rd parties could have been pushing faster vanilla 68000s to compete, yet it was actually Motorola who offered the only versions rated above 16.7 MHz (at 20 MHz) while Hitachi/etc should have had 20 and 25 MHz models by the end of the 80s at least. (especially since Hitachi's HD68HC000 was one of the most efficient and coolest running CMOS 68ks on the market -much more so than Mortola's . . . albeit thermal dissipation was never a limiting factor on the 68000 -let alone the CMOS ones, overheating was a non-issue -unless you supplied more than 5V- and it was mainly stability of the logic at high speeds that mattered -and it seems plenty of existing 68000s from the period were stable at speeds well above their ratings, some 3rd parties were remarking chips into the 20-30 MHz range too -some ST accelerator boards used those)

Hell, another interim option (and more long-term for low-end/embedded applications) would have been moderate upgrades to the original 68000, like extended prefetch and maybe an on-chip cache (or on-chip logic for external caching), an embedded 16-bit latch for more efficient bus accesses, extending to 32 address lines should have been very straightforward too. (extending to a 32-bit data bus with 32-bit ALU would be the big step up)

 

However even without complete C64 compatibility the Plus 4 would have benefited from more RAM, Sid for sound and probably 128K - they could have ditched the horrid built in software.

Apple had 128K upgrades with the IIe, Atari released the 130XE, Tandy the CoCo3, Sinclair the Spectrum 128, etc... 128K or more was where everyone else was going in the 8 bit world and if what CBM engineers said was correct, it would have only cost a few dollars. A 128K C64 or 264 (Plus/4 minus software) would have made much more sense than the Plus/4 & C128. The C128 was a bit of a hack and some manager was trying to create yet another business system (Plus/4, B128, C128... I'm seeing a trend).

 

Now, the Plus/4 can be upgraded to 256K RAM (or more) if it's hacked, so there was some potential there. If Basic can actually use that then it's way ahead other machines.

Doesn't the C64 have such potential for hacking too? (software support is another matter, but the Mosaic was pushing open-ended RAM expansion for the Atari 400 well before the 1200XL . . . and getting software that supported it)

 

If I were to consider any other change to the TED it would be to add some sort of blitter. It would probably take less logic than sprites and doesn't steal clock cycles when it's not actually in use.

Like the rudimentary blitter-ish logic embedded in the 128's VDC?

 

But, again, if you want a low-cost business-oriented machine, why not the 128 without the C64 parts (ie just Z80+VDC and necessary I/O and DRAM logic)?

The TED's color capabilities weren't very necessary for business applications anyway. Albeit, a TED with VDC like text/graphics resolution modes could have had the best of both worlds. (other than lack of RGB compatibiltiy . . . unless the TED was made with RGB in mind)

Edited June 7, 2011 by kool kitty89