The lesser-known Initial Chip Set (ICS) mystery

[andreas]

(Well, this thread might be moved somewhere else, but I did not find any appropriate place...)

What is the talk about this "ICS" thing all about? Toni has hardware-exact ICS emulation in very latest WinUAE builds, and some others claim the A1000 had OCS, not ICS; even Wikipedia!

http://en.wikipedia.org/wiki/Amiga_1000

The only place in the net where I could find out about this ICS is this New Zealand site (greetings Codetapper )

http://www.titan.co.nz/amigaak/AA020310.htm

[Oli_HD]

Originally Posted by andreas

What is the talk about this "ICS" thing all about? Toni has hardware-exact ICS emulation in very latest WinUAE builds, and some others claim the A1000 had OCS, not ICS; even Wikipedia!

The only place in the net where I could find out about this ICS is this New Zealand site (greetings Codetapper )

Cool, My A2000 (Old German version) must have the ICS chipset, nice!
Thats two new chip names I have to remember though

[Calgor]

Denise without EHB would be one of the differences if it is indeed Initial Chip Set? My later PAL A1000 came with the EHB Denise.

I was not aware that any A1000s were released with the Portia and Daphne chips instead of the Paula and Denise respectively. Surely the information is incorrect unless it was an internal build?

[Jope]

Calgor, indeed. the ICS has the same name chips as OCS but it was missing EHB.

[NLS]

I am not so sure.

EHB came "much" later ("much" is in quotes because all these things are very relative now). I remember I had Amiga for maybe 1-2 years before I even heard about EHB. I think my first A500 couldn't make EHB and definitely had OCS.

I think Portia and Daphne came only with VERY few A1000.

And since we are in this... Who has ever SEEN those chips? I mean so many Amiga people here... nobody with an A1000 with ICS?

[Jope]

All PAL Amigas have halfbrite mode..

No A1000s were with Portia or Daphne. IIRC the names were changed before they had silicon.

[Oli_HD]

I will have to try ehb on my A2000 and see, cant say I use it much.
(I guess VT may have a command to use EHB instead of HAM)

[girv]

Wow, all these years on Amiga and I'd never even heard of any chipset before the OCS !

[Jope]

Originally Posted by Oli_HD

I will have to try ehb on my A2000 and see, cant say I use it much.
(I guess VT may have a command to use EHB instead of HAM)

DPaint 2 or DPaint3 should have halfbrite in the screen mode selector.

[OddbOd]

I have to side with Jope and Calgor on this because there seems to be an assumption that Portia and Daphne were actually ICs rather than just the names temporarily carried over from the Lorraine prototype. Either that or we start referring to bits of the A1000 as War, Peace, Agony etc.

The names Paula (short for Ports Audio Uart And Logic) and Denise (for Display Encoder) were probably chosen after the initial US A1000 release (the so-called ICS model) to bring them into line with Commodore's tradition of naming custom ICs after the function they perform in addition to being peoples names. Without some sort of evidence to the contrary it would be safe to assume that there is no relationship between the name changes and any functional changes.

[NLS]

I don't think anybody mentioned anything about functional changes.

If there were, they are just "pre-1.0" limitations (ie. things that were not active because bugs didn't allow them to be enabled).

I don't think any of these reached RTM machines (although some say the contrary for some VERY early A1000 NTSC).

[andreas]

Originally Posted by NLS

And since we are in this... Who has ever SEEN those chips? I mean so many Amiga people here... nobody with an A1000 with ICS?

Well, mine has both the paw print from Jay Miner's dog on the inside and a female serial connector, so it MIGHT have ICS

[Jope]

All A1000s have the signatures + the wonky serial + parallel connectors.. :-)

[joeksterz]

I can confirm that I did have an ntsc A1000 with the non-halfbright denise chip. I dont know how early it was because I got it used. All I did was swap out the denise for a newer one and halfbright worked. I also had an A1000 with the newer denise out of the box. I assume at some point they switched from the old to new chips in production.

[mark_k]

Regarding EHB support: apparently about 50% of all A1000s made had the original non-EHB 8362R5 Denise chip. Most of the later production with the 8362R6 or 8362R8 Denise (which does support EHB) was sold in Europe. So probably the majority of A1000s in the USA don't have an EHB-capable Denise.

The EHB-capable 8362R6 Denise was available -- kind of -- either at or very soon after the Amiga 1000 went on sale to the public. A batch of machines was produced with that chip in. But since Commodore had a large stock of the earlier 8362R5, later production moved back to the non-EHB chip until stocks of the R5 chip were exhausted.

[Some developer machines, sold before the A1000 went on general sale, had 8362R5 chips which were not tested for HAM functionality. So it's possible that HAM mode might not work properly on a (small?) proportion of early developer machines. But that's a chip yield issue rather than HAM not being implemented. Even very early documentation -- the March 1984 preliminary Hardware Manual -- mentions HAM.]

References:
EHB test program posted by Bob Pariseau to net.micro.amiga on 9 Nov 1985
That post contains some incorrect info which he later corrected:
Posting to net.micro.amiga on 12 Nov 1985. "Meanwhile, my statement that consumer machines handle HAM displays correctly is true. This function is also handled in the 8362 and works fine in R5. Some R5's were not, however, tested for this function back when we were making pre-release machines for developers and are not running fast enough to do HAM properly. All consumer
machines HAVE BEEN properly tested for HAM function along with the rest of the battery of tests and burn-in procedures done as a matter of course on the Amiga assembly line."
HAM mode info and test program: "BEWARE! Some of the older developer machines and store demo machines do not produce reliable HAM images. ALL of the consumer machines do HAM just fine."
Posting by Leo Schwab to comp.sys.amiga on 22 Jun 1987
Posting by George Robbins to comp.sys.amiga on 8 Jul 1987

The acronym OCS only really came into use once Commodore introduced the Enhanced Chip Set. I don't think Commodore ever used it themselves.

This talk about "ICS" is quite confusing IMO. As far as I recall no-one talked about ICS back in the day; most people didn't distinguish between the OCS chips as in e.g. the A500 and those in the A1000. The only different chip, early non-EHB Denise aside, is the Agnus (original DIP or "fat" PLCC).

ICS is probably used to refer to the original chip set, as used in the A1000 and German-designed A2000 (not the US-designed A2000 a.k.a. A2000-CR, B2000). That consists of Paula, Agnus and Denise. The original Agnus (8361 NTSC, 8367 PAL) is in a DIP package.

For the A500 and US-designed A2000-CR, the "Fat Agnus" chip (8370 NTSC, 8371 PAL) was produced. That chip is in a square PLCC package and has a couple of significant differences from the original Agnus:
- Blitter busy bug fixed. With the original Agnus, in certain cases the blitter busy bit reads as zero immediately after starting a blit. (The ROM WaitBlit routine takes that bug into account.)
- Fat Agnus integrates some circuitry that was implemented with discrete logic chips or PALs in the A1000 & A2000-A. It also controls access to "slow RAM" at $C00000, which is why access to that memory is as slow as to chip RAM on the A500 and A2000-CR. ($C00000 memory in the German A2000-A is actually on a card in the CPU slot, and is true fast RAM. Legend has it that by setting the jumpers on the A2000-A RAM card appropriately, it can function like the A1000's Kickstart RAM/WCS.)

All A500s and A2000s have an EHB-capable Denise (probably all 8362R8).

-- M

[NLS]

thanks for the info

[andreas]

from me too, perfect research, mark

[desiv]

Originally Posted by mark_k

Regarding EHB support: apparently about 50% of all A1000s made had the original non-EHB 8362R5 Denise chip. Most of the later production with the 8362R6 or 8362R8 Denise (which does support EHB) was sold in Europe. So probably the majority of A1000s in the USA don't have an EHB-capable Denise.

The EHB-capable 8362R6 Denise was available -- kind of -- either at or very soon after the Amiga 1000 went on sale to the public. A batch of machines was produced with that chip in. But since Commodore had a large stock of the earlier 8362R5, later production moved back to the non-EHB chip until stocks of the R5 chip were exhausted.

OK, to revive this 2 year old thread for very little reason.. ;-)

Looking at my new (to me of course) NTSC A1000.
I haven't tested EHB yet, but I'm looking at the chips.
My chip next to the letter D is an 8362R6, so I'm guessing that it means I have an EHB capable A1000.

The motherboard say "REV A" on it. I've no idea if it came that way or if Denise was replaced at some time.
When I put it back together, I'll test something with EHB.

desiv

[xArtx]

This might help one day (Amiga 1000) Denise/Paula.
Google points here with the chip number.
They are yours if you are first to demonstrate you have a machine that is better off with them.

[gulliver]

Originally Posted by xArtx

They are yours if you are first to demonstrate you have a machine that is better off with them.

Easy: I have a DraCo, and its better without them.

Instead of Paula (8 bit 27khz audio) I have Toccata 16 bit 48khz audio and insted of Denise (12bit gfx and 2MB chipmem limit), I have 24 bit RTG gfx thru Altais and 4MB of chipmem.

So I am better off with them