Talk:Amiga 4000

Separate article for the Amiga 4000T?

Should we have a separate article for the A4000T? It is a separate model with, I believe, significant differences. At the least this article should probably be cleaned up so it doesn't confuse the two so much (e.g., the way it says there are two models - the 030 and 040 versions - but the 4000T is yet another model). Mdwh 23:08, 27 June 2006 (UTC)

No objections, so I've created Amiga 4000T. Mdwh 00:16, 6 July 2006 (UTC)

This could perhaps be written more clearly

They also made use of a Lithium Ion backup battery rather than a NiCd. This backup battery is also one of the most common causes of problems in the aging A4000s: it has a tendency to eventually leak. The released fluids are somewhat corrosive and can eventually damage the motherboard.

Which battery had the leak problem? The NiCd or the LiIon? the barrel batterry that it came with- the backup as u call the fix. Its simply a standard battery used in todays computer.

Well, since the first sentence clearly states that it does not have a NiCd battery, I don't think this is very unclear? Pipatron (talk) 18:00, 17 April 2008 (UTC)

CPU cart exists in Amiga 3000

"Unlike previous Amiga models, early A4000 machines have the CPU mounted in an expansion board;" It is not true. Amiga 3000 has CPU cart compatible with A4000. Many custom CPU cards (turbo cards) compatible with A3000 and A4000 (simultaneously) were available in the market. — Preceding unsigned comment added by 83.238.150.131 (talk) 10:37, 29 February 2012 (UTC)

The early A4000s have no CPU on the motherboard. Zac67 (talk) 16:37, 4 April 2013 (UTC)

Resolution

One resolution is listed as 1024x768i. I never had that resolution on my A4000, so where does that come from? 84.49.75.226 (talk) 11:31, 4 April 2013 (UTC)

Good point - I've just sourced it. Zac67 (talk) 16:37, 4 April 2013 (UTC)

Also, I've ended up here because I was having trouble working out how 640x480 @ 72hz progressive implies "only" 60hz interlaced at 800x600, which doesn't make sense either in pixel clock or line count terms ... to find that allegedly it can do upto 72hz vertical but only 31440hz horizontal. Which works out to a maximum of about 436 scanlines including the blanking and sync interval, and is in fact not *quite* full VGA rate (which is 31468) so true 525/59.94 isn't possible. Either it needs to be 524/60, or 525/59.89 ... a very small difference, but one that puts it below rather than above standard PC display. Even so, that would allow upto about 50.4hz progressive or 100.6hz interlace for 800x600 (at 624 or 625 total lines), and with typical overscan and a fixed pixel clock that comes out to 40.3p / 80.5i ...

Is it maybe that it can do 640x400 at 72hz (finally matching what the ST could do back in 1985, albeit with a slower Hsync clock thanks to better monitor hardware, but with 256 colours so making it competitive with the Mac Color and Falcon), and 1024x768 at 60i? If we use the same number of blanking lines for 640x400 as for VGA, that means 445 total lines minimum, but SVGA standard instead implies about 424, so something in-between should be fine. Yet, doubling that takes us up to 72i (36p) at 800 active lines (and about 873 total) just in Hscan limit terms, so presumably it's a pixel clock limit... VGA line length is typically 800 pixel clocks with 640-active-pixel mode, so at 524 lines that makes 419,200 pxclk x 60 times per second = 25.152mpx/sec (which sounds at least vaguely familiar), or more simply halving the vertical scan rate allows double the clocks per frame ie 838,400, which is going to be pretty tight for 1024x768 but probably doable...

If we assume 800 total lines (seen it used for XGA before, though technically interlace should have an odd number of lines because the usual convention is to transmit each field as "X plus a half" lines, with the position of the halfline determining whether it's a "top" or "bottom" field), that leaves us with 1048 clocks per line. That's a very, very minimal amount of H sync / overscan time and probably won't work.

Paring the Vsync (which can actually get away with much more threadbare blank/sync line counts vs Hsync pixel counts, and is usually only fairly "fat" on TVs because of habitual overscan and on VGA because of the intent to maintain TV-out compatibility) back to the 24 lines of SVGA makes 792 total, with 1058.6 clocks/line. Better but still risky. Halving it from the original to 16 (allowing something like 2 lines for front porch, 3 for back, and 11 for actual sync), or maybe 17 per 2-field frame (2 fp, 2 bp, 4.5 sync per field) comes to 784/785 for 1068~1069 clocks/line, or 1024 active pixels plus 44 for blanking and sync (say 8 front, 8 back, both in black or maybe a fixed border colour, and 28 for actual sync, which, if I plumb back into the depths of memory to a time when I used to use Powerstrip to trim as much fat as possible from the frame structure of SVGA and XGA (and intermediate) resolutions in order to crank up the refresh rate on cheap and flickery second-hand monitors, sounds close to what the reliable minimum tended to come out as at high 50s to low 70s progressive Hz...). Which should just about do it, and may actually be a little more stable (technically speaking even if not perceptually) than the PC equivalent thanks to the much lower actual scan rate meaning each line and each pixel represents rather more scan-time, with the listed blank/sync periods ending up a lot closer to what the monitor would normally expect anyway. Could in fact probably add an extra couple of always-black pixels between the "border" and the sync area, shrinking the latter down a bit more, and still get away with it.

Anything higher would be basically impossible unless you really want to see where the ultimate sync limits of your monitor lie (already, at 59.9ish hz x 392.5 lines, the Hscan rate is down to 23.5khz, which not all non-multisync (S)VGA/XGAs can handle - and indeed even 80hz x 312.5 lines only rates 24.96khz), whether you have hidden photosensitive epilepsy, and if it's possible to trigger physical nosebleeds or eye disorders simply through certain frequencies of flickering light. Lower or in-between resolutions of course should be able to scan higher, within the Hscan and pixel clock limits...

So, if the Hscan rate is correct, actually the definition should be "640x400 at 72hz flicker free, 640x480 at 60hz progressive, 800x600 at 50hz progressive or 100hz interlace (...or 'upto 80hz interlaced'?), and 1024x768 at 60hz interlace", rather than "640x480 at 72hz, 800x600 at 60hz lace, oh and 1024x768 as well, somehow" ??

And this is without getting into all the gibbering insanity represented by things like 1024x1024 mode (aka sequential partial updates to the monitor's own internal framebuffer, one 60hz, 512x512 quarter-screen block at a time...).

Though I will take a moment to investigate what may be available at 56 and 50hz progressive (or 112 and 100hz interlace, if you find a monitor that allows such a thing), and the common "48"/47.5/"47", 45 and "43" (43.5) frame hz interlaced (= 95/90/87 fields) scan rates, with aspects close to 4:3 (8:5, 5:4...) which would probably have offered a pretty good boost over VGA.

56hz (equivalent to regular SVGA) - 449,142 clocks/frame, 561 lines max (= 800 clocks/line, at that combination... this might become a theme?!). With some rearrangement and shaving of blanking/sync borders, 704x528 is probably quite practical. Or 640x512/528/544, 672/688x512, 720x480/496...

50hz (equivalent to twice PAL, same as VGA being twice NTSC) - 503040 clocks, 629 lines (equals... yep). 640x576 is a natural outcome, 672/688x544, 704x528/720x512, 768x480/496 with no shaving, or slightly higher still with thinner borders...

47.5hz (95i) = 529515 clocks/662 lines, 688x576, 704x560 or 720x540/544 easy, 768x512 without much trouble.

45hz (90i) = 558933 clocks/698 lines... 720x576 is a pretty exact fit. Alternatively 768x544 isn't much bother, or 800x512/528, 704x600...

43.5hz (87i) = 578207 clocks/722 lines... 720x600 works, as does 768x560 (or 576 with a shave, for exact 4:3), 800x540, 704x624...

In fact, wouldn't be entirely surprised to find that 800x600 comes in at 43.5hz instead of 40, with blanking border widths somewhere between VGA and 1024x768. 623 lines per frame (only 11.5 vblank/sync lines per field... 3+3+5.5?) leaves 928 clocks per line, which is a fairly generous 128 spare on top of the 800, or an active ratio of 86.2% (vs 80% in VGA and the more idealized modes above, but a full 95.9% in the bleeding-edge XGA). With 625 and 627 lines it's 925 and 922 clocks, or 86.4~86.8%. So, yeah, that could work...

And just to see things through to the bitter, soul-crushing end: 56i and 50i (28 and 25hz frame rate). 56i = 898284 clocks, 1122 lines; 1056x800 might be just doable (1056x792 = 4:3), or 1024x824, 1096x768...

50i = 1006080 clocks, 1258 lines. Seems like it would be just good enough for 1088x864, 1104x852, 1120x840 (4:3), 1136x832, 1152x816, 1184x800 at a push... and Hsync down to roughly 21.4khz by this point, unless an extra nearly-50% speedup can be granted to the pixel clock somehow (...which would also take things back to about 72i as well as 31.4khz).

Thus giving us a lineup of 640x400 @ 72p, 640x480 @ 60p, 680x512 @ 56p, 720x512 @ 50p, 768x540 @ 95i, 800x600 @ 43i, 1024x768 @ 60i, 1056x800 @ 56i, 1120x840 @ 50i? ;) 193.63.174.254 (talk) 13:30, 15 March 2017 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for deletion

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:

• Amiiga4000DP.jpg

Participate in the deletion discussion at the nomination page. —Community Tech bot (talk) 03:22, 17 January 2023 (UTC)