Odd Ram Expansion - Power Computing A500+ PC504...

ElectroBlaster

Active member
Joined
Aug 2, 2010
Posts
1,316
Country
Exeter, Devon
I picked this ram board up ages ago and I have never tested it. Decided it is time to find out exactly what it is. The chip count suggests this has more memory than one would assume for an A500+ expansion:

Ram chips are marked as - TMS 4256-10NL.

Has "A500+ PC504 ISS2" written on the board (both sides)

attachment.php


Anybody have a clue what this is? I will stuff this into an A500+ promptly and see what happens. Googling seems to throw hardly anything up other than the possibilities of it holding 2mb or even 4mb.

attachment.php


attachment.php


attachment.php
 

BLTCON0

Math inside
AmiBayer
Joined
May 7, 2011
Posts
2,221
Country
Hellas (Greece)
Region
Chania, Crete
@ElectroBlaster

My guess would be this is not a true 4 MB RAM expansion - that's not possible without the usual Gary board to intercept addresses and handle bank switching.
More likely it's 1 MB true RAM + 3 MB custom RAM disk, accessible via driver software (something like the well known KCS PowerPC board for the A500, which has 1 MB of RAM but of course only 512 KB of it can be seen in Amiga mode, the other 512 KB can conditionally be accessible as a custom RAM disk after installing a driver).
The simplicity of the board is remarkable. Now, pins 41-50 are dedicated to the real time clock which this board of course lacks so there would be no obvious reason to have any traces leading to any of those pins. Check that out - if there are traces leading to those pins, it probably uses some of the clock signals to selectively "block" specific banks when under driver control. Just a wild guess.
 

roy_bates

resistance is futile!
AmiBayer
Blogger
Joined
Dec 18, 2011
Posts
8,387
Country
england
Region
birmingham
tms-4256 can be used as chip ram on revision 5's,256X1 drams i think

16 of them makes 512kb

32 makes 1mb...so is a old chipram card for the 500+ to make 2mb...

sorry people,hopes that helps.:)
 
Last edited:

BLTCON0

Math inside
AmiBayer
Joined
May 7, 2011
Posts
2,221
Country
Hellas (Greece)
Region
Chania, Crete
lol roy good one (y)
Of course you're right, these are 2x8-pin chips on second look...
mystery solved :)
 

SaviorX

VCCUG
Banned
Joined
Jul 12, 2010
Posts
1,863
Country
Canada
Region
Calgary, Alberta
lol roy good one (y)
Of course you're right, these are 2x8-pin chips on second look...
mystery solved :)

yes, you are right, i never bothered to count the pins, they are shorter than proper 4256s arent they (2x10) :p
 

BLTCON0

Math inside
AmiBayer
Joined
May 7, 2011
Posts
2,221
Country
Hellas (Greece)
Region
Chania, Crete
Yep, those are 2x10 and dubbed 44256, not just 4256.
These should be dubbed 41256, so they were probably made when the 44256 ones hadn't even been conceived?
So another guess here would be Power Computing ran into a batch of dirt-cheap old chips and took advantage of it - I mean, 44256 ones were already used on rev6A's and other trapdoor expansions since early 1989, it's at least weird seeing these chips in an expansion made for the A500+ halfway through 1992.
 

Bryce

Active member
AmiBayer
Joined
Aug 21, 2012
Posts
2,586
Country
Germany
Region
Nordrhein Westfalen
Yep, those are 2x10 and dubbed 44256, not just 4256.
These should be dubbed 41256, so they were probably made when the 44256 ones hadn't even been conceived?
So another guess here would be Power Computing ran into a batch of dirt-cheap old chips and took advantage of it - I mean, 44256 ones were already used on rev6A's and other trapdoor expansions since early 1989, it's at least weird seeing these chips in an expansion made for the A500+ halfway through 1992.

The TMS4256 shouldn't be called 41256, both are valid part names. Different companies had different ways of naming there ICs. The 4256 and 41256 are compatible with each other too. However, the really strange thing about this PCB is that some ICs have a 1992 date stamp and some (the RAM) have a 1988 date stamp. This shouldn't really be possible, the 1988 parts would have been out of date in 1992, so it was obviously a case of getting rid of old stock they found in their storeroom.

Bryce.

P.s. The 41256 and 44256 were both released around the same time. The 44256 isn't an upgrade from the 41256, it is just an alternative. 1bit memory chips had many uses and reason to exist even if there was a 4bit version available.
 
Last edited:

roy_bates

resistance is futile!
AmiBayer
Blogger
Joined
Dec 18, 2011
Posts
8,387
Country
england
Region
birmingham
Yep, those are 2x10 and dubbed 44256, not just 4256.
These should be dubbed 41256, so they were probably made when the 44256 ones hadn't even been conceived?
So another guess here would be Power Computing ran into a batch of dirt-cheap old chips and took advantage of it - I mean, 44256 ones were already used on rev6A's and other trapdoor expansions since early 1989, it's at least weird seeing these chips in an expansion made for the A500+ halfway through 1992.

The TMS4256 shouldn't be called 41256, both are valid part names. Different companies had different ways of naming there ICs. The 4256 and 41256 are compatible with each other too. However, the really strange thing about this PCB is that some ICs have a 1992 date stamp and some (the RAM) have a 1988 date stamp. This shouldn't really be possible, the 1988 parts would have been out of date in 1992, so it was obviously a case of getting rid of old stock they found in their storeroom.

Bryce.


exactly, for the parts numbers.

same sort thing was very much going on with the 8 bit machines as well,different companys had different part numbers but were the same part and compatible with each other.

a good one for comparison is the drams used in the 48k spectrum and c64 and grey spectrum 128 machines(using the same memory cells),youll know what i mean when youve seen a few.
parts are all the same just different part numbers and speed rating..
same a bit later,with the c64c and spectrum 128A/B's

i have seen 20 pin 256x4 drams that have part numbers on them that are really obscure unless you recognized them by sight or know there purpose in the machine,you might not know what they were otherwise.
same was true of the old 4116's or parts found in Russian machines etc.
 

ElectroBlaster

Active member
Joined
Aug 2, 2010
Posts
1,316
Country
Exeter, Devon
So I guess now the question is! If a program is run that then enables the rest as a Ram disk.... what program was used? I have even thought of trawling the old Magazines to see if this thing was ever advertised. All I can find is the usual standard Power Computing single and double-page adverts with hardware that we all know about. Was this card part of something obscure... like an information koisk or something? Would not surprise me tbh.
 

roy_bates

resistance is futile!
AmiBayer
Blogger
Joined
Dec 18, 2011
Posts
8,387
Country
england
Region
birmingham
when you plug this in,it shows as chipram the chipram will show as 2mb in avail.

i should of made myself clear really:)
 

BLTCON0

Math inside
AmiBayer
Joined
May 7, 2011
Posts
2,221
Country
Hellas (Greece)
Region
Chania, Crete
P.s. The 41256 and 44256 were both released around the same time. The 44256 isn't an upgrade from the 41256, it is just an alternative. 1bit memory chips had many uses and reason to exist even if there was a 4bit version available.

Quite true sir :) Took a look at A1000 motherboard photos - 4bit chips seem to have been used there (4bit x 64K). Same capacity chips for rev3/rev5 A500s but 1bit x 256K. But the 1bit chips help get away with a single RAS line per 512KB :)
 
Last edited:
Top Bottom