Page 3 of 4

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: October 5th, 2018, 6:25 am
by karlmartin95
TriMesh wrote: January 16th, 2014, 3:22 am
AmiDog wrote:Hmm, are we talking about these RAM chips?
Sadly, they won't work - the first thing is that they are "KM48C" rather than "KM48V" - so they are designed to operate from 5V rather than 3.3V. The second is that the '0' at the end indicates that they are fast page mode devices - and the PSX memory controller is designed to work with EDO chips. For reference, you can't use the KM48V2004AJ-6 either - it has 4096 row addresses, and the PSX memory controller doesn't support that.

In fact, the only other part I found that seemed to be compatible was from NEC - and nobody seemed to ever had had it in stock - I assume it must have existed, since it was listed in an old NEC shortform catalog, but nobody that I could find ever seems to have actually used it.
Hi Trimesh, I need your help.

First, I must congratulate you for this amazing tutorial, so congratulations!

Second, I cannot find those dram memories, so looking in my house I found 8 DRAM Chips from a PC SIMM Memory. These chips are nearly the same chip that you used in your proyect, but... Mine has an operational voltage of +5V. The exact model is: KM48C2104AJ-6, I readed the KM48V2104AJ-6 (the DRAM used in this proyect) datasheet and I compared with the KM48C2104AJ-6 and I didn't find any difference, only the voltage operational ratio. So.. if I supply the memory with the specific voltage, Could it work?

Thanks and I'll wait your response. Cheers !

Edit1: I discovered that my chips are EDO. It's really a shame the voltage issue.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: October 19th, 2018, 5:26 am
by Alexander_H
Hi karlmartin95,
try to get an old EDO-DIMM from an old Notebook.
The Notebook used often the voltage of 3.3 Volts.

In the online Auction House can you find sometimes
a DIMM from Samsung KMM466F213BS1-L6. This i used for.
The Chips should be KM48V2104BS-L6.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: April 24th, 2019, 4:37 pm
by JephSnow
can anyone tell me where I can find the "km48v2104aj-6"?

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: April 24th, 2019, 5:57 pm
by danhans42
As I mentioned earlier in the thread, check alibaba.com as I managed to source some through there fairly easily.

https://www.alibaba.com/showroom/ic-km48v2104aj-6.html

I used Winsource...

https://www.alibaba.com/product-detail/ ... 1192fJ23DP

They accepted paypal, were great with communication and shipped quickly. I paid for faster shipping but was reasonable and arrived in UK within a week.

Good luck

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: April 24th, 2019, 11:12 pm
by JephSnow
I saw but they didn't work. I tried to contact them, hopefully good. (sorry for the bad English, I'm using a translator) Thanks :praise

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: April 24th, 2019, 11:23 pm
by danhans42
They are based in China, so completely different time zone to where I am in England, so replies often came overnight.

That being said, they were very helpful and often followed up with emails post sale to make sure the parts I purchased were okay.

No need to apologise.. your translated English is coming through fine :-)

Good luck

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: September 3rd, 2019, 8:18 pm
by Shuji
Hello,

A while back I ended up discovering this thread upon learning what could be done I thought it would be neat
to pull off such a modification to my PU-18. Being as the more memory available I'd have access to programs
such as e.g. PSXLinux, or larger programs, or if I'm not mistaken even debugging capabilities, of course, one
would have to inject them into their software, or for instance allowing Music 2000 to have access to the additional
RAM installed assuming a patch wouldn't be required or even if it did. Despite the potential benefits listed, I was
unable to track down the four KM48V2104AJ-6 without the shipping cost is an exorbitant amount of money. Instead
of acquiring them I opted to patiently wait for someone to come up with an alternative. Months had passed and not
long after making this decision Alexander_H found what appeared to be a working replacement. What Alexander_H learned was one could theoretically desolder the four KM48V2104BS-L6 acquired from a 16 MB EDO-DIMM 144-pin
3.3V 60 ns Samsung KMM466F213BS1-L6. As it did meet the credentials those of which were: Voltage Rating, Refresh Cycle, and Memory Capacity. Since it matches the requirements the experimenter could swap out the default RAM
chips with KM48V2104BS-L6.


I decided my console would undergo the same operation since Alexander_H had success using the RAM module.
With that said I managed to snag one off of eBay for a responsible price. When it made its arrival I read on how
I could remove the RAM chips someone suggested getting the Soldering Iron and heat up the IC's legs while slowly
prying it up with very little force to prevent any damage from occurring. So I followed this method while trying to be vigilant. However, I accidentally ripped off some of IC109's pads while the other three KM48V514BJ-6 pads thankfully remained unharmed. I'll admit this procedure was not the best and the proper way of going about it is to use a Hot Air gun along with a pair of Dentist Excavator or S/S Tweezers. For the process I of desoldering the individual
KM48V2104BS-L6 I knew I couldn't make the same mistake twice. So I invested in the X-TRONIC 6040 - 2 in 1 Digital Hot Air Rework Soldering Iron Station after complicating for awhile. I'm glad I purchased It because it has made tasks like these a breeze since it took little effort whereas compared to the other method. After I separated the KM48V2104BS-L6 from the Samsung KMM466F213BS1-L6 EDO-DIMM as shown in this photo and this photo as well. I used the High-Res photo's pieced together and extracted from the PU-18 service manual that I found from where I don't remember. This allowed me to trace the remaining IC109 pins I needed to locate underneath PU-18 RAM/CPU pin outs due to nonexistent pads.


When I thought I was officially done and reassembled the unit halfway it proved to be problematic and only displaying
no signal or picture if you will :crying. I researched what could be causing the issue and to no surprise, as Trimesh
wrote in various threads from assemblergames Fuse PS604 was likely blown. So to be sure I tested with my Multi-meter and only one side was reading 3.5V while the other gave 0V reading. I was fortunate in that Trimesh offered to
supply me with some for the cost of shipping which was extremely cheap when compared to other sources. When the package arrived from Trimesh I replaced the Fuse to only be dawned by the blank screen :(. At this point, both Wisi and Trimesh brought to my attention that I should check my wiring as they may be bridging each other. I did as they suggested checking each IC9 pin for shorted connections between the CPU and continuity for bridged wires. Sure enough, the main culprits were DQ26, DQ27, and A9 being bridged as suspected here's proof image. Afterward, I corrected this by applying a tiny amount of solder to decrease their chances of crossing over one another and double-checking my other wires. I no longer was doomed by the blank screen I was once more greeted with the startup logo. Oh, the joy I felt :dance knowing I revived my console after cutting it's life short too soon.


Now after having confirmation of its status I reassembled the console, reinserted my Xploder FX Cartridge JP flashed
with Caelta v0.34 along with my my parallel cable connected which feeds to my Startech PCI Express Parallel Port
Adapter Card PEX1, executed catflap on Win7 in CMD with entering catflap /XT:POx0378 run MAIN.EXE ran the RAM
test software by Trimesh to only come across a Bus Error :shrug. Alexander_H created a fix however I am unsure
where I should inject this line of code: #define STACKSLACK 0x12800. So I would no longer get such an error anymore. I've compiled and tested the majority of the samples included from the PsyQ SDK which I experienced a couple of
crashes. while games of course work flawlessly like one would expect.



Anyhow, I've discussed implementing the idea of piggybacking an extra set of four additional RAM chips to double the amount of RAM among Wisi and Trimesh. But have yet to initiate the steps that are needed. I'd like to commence this although I'm curious as to what's your take on it Shadow or SMF on this matter? Supposedly one of the Arcade units had 16mb of CPU RAM. However, I've never seen anymore then 4mb max installed at one time being the /RAS lines or banks only had 2mb each. Perhaps the CN51 on Namco's System 12 COH-700 board allowed more memory to be installed since you can see traces the required connections leading from CXD8661R CPU to the KM416V1204BT-L5 RAM chips.


Thank you so much :praise Alexander_H for experimenting with the possibility of resourcing other RAM modules that are compatible. Also, I want to thank Trimesh for writing the guide, for shipping the necessary parts to repair my console for cheap, and for helping me troubleshoot my console as well as Wisi of course which eventually lead me to resolve the no display issue because of the two Data In/Out pins bridging each other due to wires touching.

What I had hoped to accomplish from this is that for anyone who cannot find any KM48V2104AJ-6 for a responsible price you can do what I did the correct way of course or contact Trimesh and have him or her ship some to you for barely anything.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 5th, 2020, 2:59 am
by samspin
I plan to try out this mod with a similar tactic- buying a SODIMM stick with potentially compatible ICs and desoldering them with a hot air station. The specific stick I've ordered is labled "HYMIB20016 IBM EQUIV 92G7341 16MB 144PIN EDO SODIMM". It's double-sided with four ICs per side. Going by photos I've got the part number for the individual ICs being "MT4LC2M8E7TG-6 S". According to the datasheet these match the required pinout and specifications being 60ns refresh rate with 2048 rows and 3.3V power. The only thing is these ICs have the S designation, which apparently is an optional "self refresh mode" feature. I'm not sure if this will have any implications. But it appears the above poster's KM48V2104BS-L6 ICs also have this feature (the datasheet for that part says this is what the L designation means where it says -L6). I am therefore hopeful the ones I have found will work too. Whether they do or not, I'll report back here as it may help others trying to find compatible ICs for reasonable pricing with this method.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 10th, 2020, 5:30 pm
by TriMesh
samspin wrote: August 5th, 2020, 2:59 am I plan to try out this mod with a similar tactic- buying a SODIMM stick with potentially compatible ICs and desoldering them with a hot air station. The specific stick I've ordered is labled "HYMIB20016 IBM EQUIV 92G7341 16MB 144PIN EDO SODIMM". It's double-sided with four ICs per side. Going by photos I've got the part number for the individual ICs being "MT4LC2M8E7TG-6 S". According to the datasheet these match the required pinout and specifications being 60ns refresh rate with 2048 rows and 3.3V power. The only thing is these ICs have the S designation, which apparently is an optional "self refresh mode" feature. I'm not sure if this will have any implications. But it appears the above poster's KM48V2104BS-L6 ICs also have this feature (the datasheet for that part says this is what the L designation means where it says -L6). I am therefore hopeful the ones I have found will work too. Whether they do or not, I'll report back here as it may help others trying to find compatible ICs for reasonable pricing with this method.
I've had a quick look, and I can't see any reason those parts shouldn't work. The only thing I would recommend is testing the module before removing the chips if you can, because 90's Micron memory tended to have a rather high failure rate.

The self-refresh mode shouldn't get in the way, because it seems only to operate if the RAS/ line is held low and the PSX memory controller doesn't do that.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 11th, 2020, 9:39 am
by samspin
TriMesh wrote: August 10th, 2020, 5:30 pm I've had a quick look, and I can't see any reason those parts shouldn't work. The only thing I would recommend is testing the module before removing the chips if you can, because 90's Micron memory tended to have a rather high failure rate.
Thanks for pointing that out. Sadly I don't have a laptop that uses EDO RAM so I cannot test, but for what it's worth the previous owner said it was a working pull on their end. In this case I'll try to be more careful when extracting the chips from the stick. Then solder each leg onto the pads one-by-one on the console board to avoid blasting more heat than neccessary, using plenty of no-clean flux. This should be a lot easier as these chips are slightly smaller but the pins use the same spacings and they are exposed outwards (TSOP-II) as opposed to the originals where the pins are folded underneath (SOJ).

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 15th, 2020, 10:40 am
by samspin
Update: I have performed the upgrade and wire work on the bottom of the board. The console appears to be working perfectly normal so I believe this *could* potentially be a complete success using the above-mentioned "MT4LC2M8E7TG-6 S" ICs. Unfortunately I haven't yet managed to get PSXSERIAL to work in order to upload a test to ensure all 8MB is accessible. I can boot PSXSERIAL with either a burned CD or PSIO, but I cannot get the connection to work between the console and my laptop. I'm using a USB-Serial cable I bought awhile back from DANHANS42, which *does* work okay with PS1CardLink. I get as far as "receiving data from PC" but the command window says "Waiting for playstation 1" and hangs. I know I've set the baud-rate in Device Manager properly so I'm unsure what else to try, it might well be something MS have done in the current Windows 10 build to screw things up. I'll probably have to dig my Windows XP laptop and try with that. If I do get any further I'll come back again! At least we now know that "MT4LC2M8E7TG-6 S" ICs appear to be compatible.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 16th, 2020, 12:55 am
by TriMesh
samspin wrote: August 15th, 2020, 10:40 am Update: I have performed the upgrade and wire work on the bottom of the board. The console appears to be working perfectly normal so I believe this *could* potentially be a complete success using the above-mentioned "MT4LC2M8E7TG-6 S" ICs. Unfortunately I haven't yet managed to get PSXSERIAL to work in order to upload a test to ensure all 8MB is accessible. I can boot PSXSERIAL with either a burned CD or PSIO, but I cannot get the connection to work between the console and my laptop. I'm using a USB-Serial cable I bought awhile back from DANHANS42, which *does* work okay with PS1CardLink. I get as far as "receiving data from PC" but the command window says "Waiting for playstation 1" and hangs. I know I've set the baud-rate in Device Manager properly so I'm unsure what else to try, it might well be something MS have done in the current Windows 10 build to screw things up. I'll probably have to dig my Windows XP laptop and try with that. If I do get any further I'll come back again! At least we now know that "MT4LC2M8E7TG-6 S" ICs appear to be compatible.
I don't know about that specific cable, but the Yaroze boot disc and SIOCONS are a lot more picky than most homebrew - the most important thing is that you have to have all the handshake lines connected and they need to be inverted. I have a feeling that cable you have is designed to work with SIOLOAD, which doesn't use handshaking and hence it's probable that the lines are not connected.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 16th, 2020, 11:59 am
by Squaresoft74
Something similar to what's described here ?

Configuration 1 (Net Yaroze)
Image

Configuration 2 (Homebrew)
Image

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: August 16th, 2020, 4:02 pm
by TriMesh
Yeah, that's the correct setting - obviously, that software will only work with an FTDI chip and if you're using some other USB->serial adapter you will need to invert the lines using hardware.

I've used these things just wired directly to link cables set up like that successfully with SIOCONS.
FTDI_Board.jpg

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: September 12th, 2020, 8:26 am
by szalay_1
@samspin

I've SCPH-5502/PU-18 waiting for 8MB RAM Modification
l really look forward to the result :)

so appear to be compatible, or compatible?

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: September 27th, 2020, 1:24 pm
by Deilvert
Hi everybody,

Have somebody guide for to install 8MB in a motherboard model PU-8?

Regards and thanks

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: November 3rd, 2021, 4:08 am
by szalay_1
Got some free time and lets beginning
SCHP-5502 using KM48V2104BS-6
01.jpg
02.jpg
03.jpg
04.jpg
05.jpg
06.jpg
07.jpg

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: November 3rd, 2021, 4:13 am
by szalay_1
SCHP-5552 using KM48V2104BS-6
01.jpg
02.jpg
03.jpg
04.jpg
So just tried ram test "Main.exe"
and get error, but the PSX works well
Ram Test.jpg
What could be the problem with rame test ?
data load or store error, meaning its reading (lw) or writing (sw) a bad address ?

EDIT, Thanks to danhans42 the new 8M ramtest built Done :)

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: November 25th, 2021, 1:46 pm
by Luckarusky
Got to say, I'm glad it's possible to create a PS1 Dev unit these days. Easier than doing so with an N64 in fact. (although I'm experimenting there as well.) Just as well I bought a separate x5000 series for these modifications! I'll update here once I've tracked down the needed RAM chips.

Re: [Hardware Mod] 8MB in (PU-18) Retail PSX

Posted: February 4th, 2022, 11:18 pm
by Shadow

Code: Select all

// Written by Trimesh 2015
// Fixed by Shadow 2021

//#include <stdio.h>
#include <stdlib.h>

//u_long _ramsize  = 0x80200000; // force 2 megabytes of RAM
u_long _ramsize   = 0x80800000; // force 8 megabytes of RAM
//u_long _stacksize = 0x80004000; // force 16 kilobytes of stack
u_long _stacksize = 0x80002000; // force 8 kilobytes of stack

extern unsigned long __heapbase;
extern unsigned long __heapsize;
extern unsigned long __bss;
extern unsigned long __bsslen;
extern unsigned long __data;
extern unsigned long __datalen;
extern unsigned long __text;
extern unsigned long __textlen;
//extern unsigned long _ramsize;
//extern unsigned long _stacksize;

#define STACK (0x80000000 + _ramsize - _stacksize)
#define STARTADDR 0x80200000
#define SALT 0x7E

int main()
{
	unsigned int addr;
	unsigned char value;
	
	ResetCallback();
	AddSIO(115200);
	
	printf("\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n");
	printf("Stack = %X\n", STACK);
	printf("Test the PlayStation RAM\nOriginal code by Trimesh\n\n");
	printf("Generating fill pattern...\n");
	
	for (addr = STARTADDR; addr < STACK; addr++)
	{
		if ((addr % 0x10000) == 0x0) printf("0x%08X\n", addr);
		
		value = (addr  ^ (addr >> 8) ^ SALT) & 0xFF;
		*((unsigned char *)addr) = value;
	}
	
	printf("\nVerifying fill pattern...\n");
	
	for (addr = STARTADDR; addr < STACK; addr++)
	{
		if ((addr % 0x10000) == 0x0) printf("0x%08X\n", addr);
		
		value = (addr  ^ (addr >> 8) ^ SALT) & 0xFF;
		
		if (*((unsigned char *)addr) != value)
		{
			printf("Failed at 0x%08X! Expected 0x%02X and got 0x%02X...\n", addr, value, *((unsigned char *)addr));
			return;
		}
	}
	
	printf("\nSuccess! 8 MB RAM Test *PASSED*\n");
	
	return;
}