256mb edo ram kit

[delshay]

If you want the absolute smallest form factor 128MB, then you need the following PCB.

http://www.ebay.co.uk/itm/128MB-Low-profile-LP-EDO-SIMM-non-Parity-72p-Memory-RAM-72pin-32x32-60ns-TSOP-/121743177271?hash=item1c58757637:g:hnEAAOxyhTFSTeYG

The above PCB is what I used to build the worlds fastest EDO memory 28ns, using TSOP chips. I bought 64Mb version (cheaper), then filled in the other side for 128Mb.

Test has shown some listing of 50ns EDO memory on Ebay, do not operate at 50ns, even if the marking on the chip say 50ns.

You will get better success 50ns EDO, if chips were NEC, next is Micron, but I find to many faults here to be class as 50ns with Micron chips.

It seem what they are doing is off-loading as many EDO memory chips/Simms, without checking their speed.

Micron chips is what I started with, allowing 66Mhz bus low latency on Blizzard PPC, then moved to NEC with better result.

There is nothing wrong with Micron chips, it can give the same result as NEC, if it is properly checked, but NEC chips are more likely to operate at 50ns regardless if it is checked, this is what I found, but NEC chips are also not guaranteed to work at correct speed, as faults was also found with this make of chip.

I no longer use Micron or NEC chips for the ultimate 128Mb Simm, and I could have built 25ns EDO Simm, and test was showing 24ns (even quicker), but it was too expensive to complete this project. This is why no one will ever get near this Blizzard card http://www.amiga.org/gallery/index.php?n=3692

DDR1 project is coming to an end, expecting to see PC-3500+ CL1.5 4GB Quad kit/Dual Channel Ultra Low Latency DDR1 with "Enhanced Bandwidth" built into the SPD, next project SDRAM.

[delshay]

As this is a for sale thread, I am not going to list what I found with other manufactures chip, but what I would say its not the chip makers fault, it's the other manufactures who put them on SIMM card.

Their are fooling the end customer with poor quality/faulty chip that don't work at their correct speed.

The result is random crash/lock-ups, but most users don't know what causing these faults and seem to blame something else that's causing problem.

The smallest form factor SIMM PCB helps a lot, it help's release heat from an accelerator card as it is much smaller, as heat is less trapped.   ...As already pointed out, works best with TSOP chips.

Remember some chips can't handle high temperatures and begin to slow down (error) this is where computer will crash. It's not always the memory module(s) fault.

Test here has shown it's the other components/environment that can overheat them and cause error(s), this is where the smallest SIMM has a advantage, along with shorter PCB tracks.

[delshay]

I did point this out 3.3v running at 5v in another thread, but I just left it there, but some 3.3v dram do have a max 5.1v.

I did conduct test on 3.3v dram running at 5v on some dram, and it was only Micron dram that failed the Extreme Test, where most would slow down.

Micron dram got so hot, you could not keep your finger on there, even for a second, it was that hot. The surprising thing here, Micron dram did not slow down at these extreme hot temperature, it just kept on going. Only when the dram cooled down & test restarted, the Micron dram was dead. This test was repeated a number of times & only one in about 6 will still work.

Micron dram AFAIK is fine on Amiga & repeated test where dram will fail can't be done on Amiga computers.

Some 3.3v ram did not seem to show any increase in heat at 5v, even after hours of testing, but if overheated by other components will slowdown.

3.3v SIMM is fine, as long it has a regulator on-board.

Overvolting memory module is nothing new, & even manufactures are still doing it. This is the reason for DDR1 project, get the voltage requirements lower, but increasing performance. I can confirm this is already up and running, setting a new level of performance at a set lower voltage.