The (Super) Fat Agnus situation

eXeler0 · 15 June 2021, 23:02

Howdy,

This is an old story, but it still doesn't have a happy ending.

Many have talked about the Agnus problem. Short recap:
While old OCS Agnus is plentiful and not really desirable there a lots of them out there.. But then we have the newer variants ands 8372/8372A are, while obtainable, already hitting pretty painful prices. Similar situation for 8375..
But then we get into "unobtanium" territory.. Agnus 8372AB and 8372B.

(Please tell me when you last saw a 8372B for sale, anywhere on planet Earth. ;-) )

**First question / Problem.. If any of the several "Work in Progress" FPGA drop- in Agnus replacement projects materialize, how exactly will this be solved considering the numerous Agnus versions out there?

**Second problem.. The fabled Rejuvinator (for the A1000) reverse engineering projects seems like its going to be an exact clone of the original (at least initially) which would probably mean that it will require the unobtainable 8372B to work. So without the Agnus, all that hard work they've put into the reverse engineering will be... maybe not for nothing.. but almost..

**Third thing.. Can we write down the FPGA Agnus projects that are not yet dead. Jens S fairly recently said he hasn't abandoned the thought yet.
ReneeC has something on Github, but given the Buffee project, Agnus is likely not a priority.
StephenL seemed to have a nice thing going on but that one was AFAIK abandoned.
I know at least one more, but it might be unofficial so I wont mention it.

**Fourth thing.. Any other possibilities? Some crazy custom board + raspberry pi zero hack instead of FPGA or CPLD?

It would be nice to gather the current state of Agnus replacement projects as well as other reflections about this issue in one thread..

Thanx and Cheers

MarkJPK · 15 June 2021, 23:34

I wonder how many 8372AB/Bs are actually out there (other than in A3000s)? I have one in an A500 with the chip ram mod. I wouldn't bother trying to sell it as the chances of somebody needing it that much that they would pay a high price for such a rare item are extremely slim. It could be others are in a similar position.

AmigaHope · 15 June 2021, 23:41

It's possible to make adapter boards for various ECS Agnus to fit any system (assuming PAL/NTSC xtal is correct on the non-agnostic ones). These may become more important in the future, though it might involve e.g. turning your A500+ into an A500 since you can't find the 2M Agnus. In general it's just a matter of getting the right pins in the right place.

Also with a clean room and careful drilling work it may be possible to convert 1M 8375s into 2M ones. That would be extremely difficult though and I don't think anyone's attempted it.

The OCS Agnus are too different and there's not much you can do with them other than use them as a fallback if that's all that's available.

The best solution would be to throw together a new run of them on a modern process, maybe with some enhancements. The cost of geting a spot on rideshare wafers has gone way, way down.

A10001986 · 16 June 2021, 11:04

There are 8375 versions that are pin-combatible with the 8372B (apart from pin 41, the Pal/ntsc selector). Those used to show up on ebay from time to time. I managed to buy a couple of those a few years back, and reasonable prices.

All is not lost.

eXeler0 · 16 June 2021, 12:47

Quote:

Originally Posted by A10001986

There are 8375 versions that are pin-combatible with the 8372B (apart from pin 41, the Pal/ntsc selector). Those used to show up on ebay from time to time. I managed to buy a couple of those a few years back, and reasonable prices.

All is not lost.

Now that's interesting..
I did a similar thing on the 8372A at some point when I modded one of my A500s.. isolating tape on one of the pins:
https://www.exretro.com/galleries/co...gnusMod_01.jpg

But anyway, in this table, it would seem pin41 on the 8375 is GND, not NTSC/PAL switch
https://en.wikipedia.org/wiki/MOS_Technology_Agnus

How did you id the right version?

A10001986 · 16 June 2021, 12:59

https://www.amigawiki.org/lib/exe/fe...v09_120923.pdf

If "VBB" is printed on the chip, pin 41is vbb.

eXeler0 · 16 June 2021, 17:11

Quote:

Originally Posted by A10001986

https://www.amigawiki.org/lib/exe/fe...v09_120923.pdf

If "VBB" is printed on the chip, pin 41is vbb.

Is it confirmed that 8375VBB is a 2MB version? (And not 1MB)?

Jope · 16 June 2021, 18:13

You need to look at the long inventory number to determine anything on 8375 parts.

eXeler0 · 18 June 2021, 00:30

Quote:

Originally Posted by Jope

You need to look at the long inventory number to determine anything on 8375 parts.

You mean the number typically (but not always) printed on the row between the logo and the 8375 text?
Like "390544-01" on the 2MB Agnus (for A500+)

The VBB version i can find are "318069-16"

A4000Bear · 18 June 2021, 08:10

318069-16 is a 1MB PAL Agnus.

There is a list of Agnus variants here: https://pmsoft.nl/amiga/amiga-parts-chips.html

eXeler0 · 26 June 2021, 01:12

Quote:

Originally Posted by A4000Bear

318069-16 is a 1MB PAL Agnus.

There is a list of Agnus variants here: https://pmsoft.nl/amiga/amiga-parts-chips.html

So I ordered the 318069-10 2MB variant which is pin compatible with 8372B, so theoretically should work with the rejuvinator and it also happens to be the part that will work with the A1100 mobo.
Price with shipping + "Brexit tax" wasnt fun though ;-)

A10001986 · 26 June 2021, 01:21

According to all the lists I know, the 318069-10 is for the A500+ and not pin compatible to the 8372B.

eXeler0 · 26 June 2021, 13:47

Quote:

Originally Posted by A10001986

According to all the lists I know, the 318069-10 is for the A500+ and not pin compatible to the 8372B.

I'm on the waiting list for the A1100 and that ones uses the 318069-10, the rejuvinator is more speculative tech since its not fully reverse engineered yet so that will be a later problem.
But according to the list posted above, my interpretation was that the 8372B was compatible with all 318069- variants of the 8375..?
Hard to know the sources for these lists though. I guess some are results of trial n error.

A10001986 · 26 June 2021, 14:52

These lists are partly wrong in some parts, but they both say the same thing regarding the 318069-10, and that is that this one is for the A500+, and is therefore NOT pin compatible with the 8372B (318069-03). That's why those chips had different numbers.

If the A1100 uses that one, so be it. If one Agnus type is not listed for the A3000, it isn't pin compatible to the 8372B. The rejuvenator was designed when there was no 8375, so I am pretty sure it needs a 318069-03 (or 318069-18, 318069-19, with pin 41 connected to ground via a capacitor), unless the new design will change the pin assignment.

The "agnusliste" document, BTW, is the result of a a1k group effort from around 2004 when Georg Braun designed the GBA1000.

eXeler0 · 27 June 2021, 18:18

All A3000 Agnus variants are probably "unobtainium".
The A1100 dev was wise to use another variant. The first version of the Rejuvinator aims to be a reverse engineered copy so yea, it will likely require same Agnus as the original but unless some adapter can be used to get around this, I dont know how useful the first version of the Rejuvinator will be, which of course is a shame considering the effort they've put into the reverse engineering.
Ive seen adapters (LIV2) to downgrade an A500+ to use more obtainable Agnus versions but other that IDK whats out there.

QuikSanz · 27 June 2021, 20:04

My DKB Megachip in the A2000 has an 8375 - 318069-11 in it, it's original.

eXeler0 · 28 June 2021, 23:12

Quote:

Originally Posted by QuikSanz

My DKB Megachip in the A2000 has an 8375 - 318069-11 in it, it's original.

Right.. that would be the NTSC variant of the chip I got (318069-10), does it say VBB on this chip? (As user A10001986 mentioned earlier.. )

QuikSanz · 29 June 2021, 02:53

Yes, "4592 -20VBB" I suppose week 45 of 1992 Mfg date.

eXeler0 · 16 July 2021, 15:31

Quote:

Originally Posted by QuikSanz

Yes, "4592 -20VBB" I suppose week 45 of 1992 Mfg date.

Found (and bought) a 318069-18.. (also 20VBB) price was too ridiculous to mention but seems to be an A3000 compatible part..

alexh · 16 July 2021, 16:28

I'm fairly sure that VBB does not reference the functionality of the silicon within.

I think it has to do with whether this part requires a VBB pin or not.

I think that some different silicon process nodes they generated the VBB internally.

15 June 2021, 23:02	#1
eXeler0 Registered User Join Date: Feb 2015 Location: Sweden Age: 50 Posts: 2,946	The (Super) Fat Agnus situation Howdy, This is an old story, but it still doesn't have a happy ending. Many have talked about the Agnus problem. Short recap: While old OCS Agnus is plentiful and not really desirable there a lots of them out there.. But then we have the newer variants ands 8372/8372A are, while obtainable, already hitting pretty painful prices. Similar situation for 8375.. But then we get into "unobtanium" territory.. Agnus 8372AB and 8372B. (Please tell me when you last saw a 8372B for sale, anywhere on planet Earth. ;-) ) First question / Problem.. If any of the several "Work in Progress" FPGA drop- in Agnus replacement projects materialize, how exactly will this be solved considering the numerous Agnus versions out there? Second problem.. The fabled Rejuvinator (for the A1000) reverse engineering projects seems like its going to be an exact clone of the original (at least initially) which would probably mean that it will require the unobtainable 8372B to work. So without the Agnus, all that hard work they've put into the reverse engineering will be... maybe not for nothing.. but almost.. Third thing.. Can we write down the FPGA Agnus projects that are not yet dead. Jens S fairly recently said he hasn't abandoned the thought yet. ReneeC has something on Github, but given the Buffee project, Agnus is likely not a priority. StephenL seemed to have a nice thing going on but that one was AFAIK abandoned. I know at least one more, but it might be unofficial so I wont mention it. Fourth thing.. Any other possibilities? Some crazy custom board + raspberry pi zero hack instead of FPGA or CPLD? It would be nice to gather the current state of Agnus replacement projects as well as other reflections about this issue in one thread.. Thanx and Cheers

29 June 2021, 02:53	#18
QuikSanz Registered User Join Date: May 2021 Location: Los Angeles / USA Posts: 135	Yes, "4592 -20VBB" I suppose week 45 of 1992 Mfg date. Last edited by QuikSanz; 29 June 2021 at 02:55. Reason: Add info

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Test Fat Agnus 8372B	Franchute13	support.Hardware	2	14 December 2014 19:14
Touching Fat Agnus = Crash	tenuke	support.Hardware	7	23 October 2014 20:13
fat agnus	albino	support.Hardware	1	14 September 2012 08:39
swap fat agnus with agnus	extralife	support.Hardware	12	23 July 2008 15:35
PAL 8375 Super Fat Agnus	alexh	MarketPlace	2	24 December 2005 20:17

15 June 2021, 23:34	#2
MarkJPK Registered User Join Date: Jan 2019 Location: Preston Posts: 11	I wonder how many 8372AB/Bs are actually out there (other than in A3000s)? I have one in an A500 with the chip ram mod. I wouldn't bother trying to sell it as the chances of somebody needing it that much that they would pay a high price for such a rare item are extremely slim. It could be others are in a similar position.

15 June 2021, 23:41	#3
AmigaHope Registered User Join Date: Sep 2006 Location: New Sandusky Posts: 942	It's possible to make adapter boards for various ECS Agnus to fit any system (assuming PAL/NTSC xtal is correct on the non-agnostic ones). These may become more important in the future, though it might involve e.g. turning your A500+ into an A500 since you can't find the 2M Agnus. In general it's just a matter of getting the right pins in the right place. Also with a clean room and careful drilling work it may be possible to convert 1M 8375s into 2M ones. That would be extremely difficult though and I don't think anyone's attempted it. The OCS Agnus are too different and there's not much you can do with them other than use them as a fallback if that's all that's available. The best solution would be to throw together a new run of them on a modern process, maybe with some enhancements. The cost of geting a spot on rideshare wafers has gone way, way down.

16 June 2021, 11:04	#4
A10001986 Registered User Join Date: Jun 2017 Location: 1986 Posts: 79	There are 8375 versions that are pin-combatible with the 8372B (apart from pin 41, the Pal/ntsc selector). Those used to show up on ebay from time to time. I managed to buy a couple of those a few years back, and reasonable prices. All is not lost.

16 June 2021, 12:59	#6
A10001986 Registered User Join Date: Jun 2017 Location: 1986 Posts: 79	https://www.amigawiki.org/lib/exe/fe...v09_120923.pdf If "VBB" is printed on the chip, pin 41is vbb.

16 June 2021, 18:13	#8
Jope - Join Date: Jul 2003 Location: Helsinki / Finland Age: 43 Posts: 9,861	You need to look at the long inventory number to determine anything on 8375 parts.

18 June 2021, 08:10	#10
A4000Bear Registered User Join Date: May 2011 Location: Taradale / Australia Posts: 95	318069-16 is a 1MB PAL Agnus. There is a list of Agnus variants here: https://pmsoft.nl/amiga/amiga-parts-chips.html

26 June 2021, 01:21	#12
A10001986 Registered User Join Date: Jun 2017 Location: 1986 Posts: 79	According to all the lists I know, the 318069-10 is for the A500+ and not pin compatible to the 8372B.

26 June 2021, 14:52	#14
A10001986 Registered User Join Date: Jun 2017 Location: 1986 Posts: 79	These lists are partly wrong in some parts, but they both say the same thing regarding the 318069-10, and that is that this one is for the A500+, and is therefore NOT pin compatible with the 8372B (318069-03). That's why those chips had different numbers. If the A1100 uses that one, so be it. If one Agnus type is not listed for the A3000, it isn't pin compatible to the 8372B. The rejuvenator was designed when there was no 8375, so I am pretty sure it needs a 318069-03 (or 318069-18, 318069-19, with pin 41 connected to ground via a capacitor), unless the new design will change the pin assignment. The "agnusliste" document, BTW, is the result of a a1k group effort from around 2004 when Georg Braun designed the GBA1000.

27 June 2021, 18:18	#15
eXeler0 Registered User Join Date: Feb 2015 Location: Sweden Age: 50 Posts: 2,946	All A3000 Agnus variants are probably "unobtainium". The A1100 dev was wise to use another variant. The first version of the Rejuvinator aims to be a reverse engineered copy so yea, it will likely require same Agnus as the original but unless some adapter can be used to get around this, I dont know how useful the first version of the Rejuvinator will be, which of course is a shame considering the effort they've put into the reverse engineering. Ive seen adapters (LIV2) to downgrade an A500+ to use more obtainable Agnus versions but other that IDK whats out there.

27 June 2021, 20:04	#16
QuikSanz Registered User Join Date: May 2021 Location: Los Angeles / USA Posts: 135	My DKB Megachip in the A2000 has an 8375 - 318069-11 in it, it's original.

16 July 2021, 16:28	#20
alexh Thalion Webshrine Join Date: Jan 2004 Location: Oxford Posts: 14,337	I'm fairly sure that VBB does not reference the functionality of the silicon within. I think it has to do with whether this part requires a VBB pin or not. I think that some different silicon process nodes they generated the VBB internally.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)