feipoa wrote:

Tetrium wrote:

I got a bunch of weird looking SIMM's. Perhaps will take a pic tomorrow.

Several 'tomorrows' have now passed. Any luck with the photos?

Sorry, have been really busy and I got like 18 small boxes of SIMMs to browse through 😵
I'll keep this one in mind for this evening

Heres one example of what server memory looks like (pic taken from my other photobucket which I lost my login of -_-)

These are 64MB FPM SIMMs. They won't work correctly in any board I have and are meant for some kind of HP server. These have chips on both sides.

I'll take some new pics later this evening (just got home and am dead tired).

Heres another example, I removed these from a dual Socket 7 rig (it was too heavy to carry home. It had a CPU card with 2 Socket 7 sockets).
These are 16MB FPM modules

I think I'll revive this thread a little bit as a cautious reminder - There are a bunch of "5 V 72-pin SIMMs" on eBay right now which are in fact 3.3 V. It is important to look at the datasheet on the individual chips. For a set of FPM memory (4x 64 MB) I bought recently, the datasheet was kinda hard to find, but alas, 3.3 V was noted on the datasheet, with 3.6 V as the max. While the chips operated at 5 V on my HOT-433 for a good hour in Memtest (without errors), they got rather hot. Running this long term will certainly cause damage. I also just realised that a set of EDO 4x 64 MB chips I bought years ago is also 3.3 V, but were marked as 5 V.

Before buying anything on eBay, try your best to get the part number of the individualised memory chips!

For reference the following FPM chips are 3.3 V:
Samsung SEC KM44V16100AK-6 (16Mx4)

The following EDO chips are 3.3 V:
Samsung SEC KM44V16104BK-6 (16Mx4)
Elpida 5165165FLTT6 (4Mx16)

There are obviously more 3.3V chips, but these are the ones I've run into so far.

It seems like some HP and IBM boards used 3.3 V 72-pin SIMMs. Were there any consumer-grade 486 or Pentium boards which used 3.3 V SIMMs?

Did anything actually use 3.3v 72-pin SIMMs? I don't recall anyone going 3.3v until the SDRAM era. Even EDO and FPM DIMMs were pretty much 5v before that. This may be a side effect of vendors trying to manufacture new SIMMs from parts that are still being made. Is the seller a vendor, of just someone trying to dump memory?

It is usually the 64 MB modules which are 3.3 V, but I've seen some 32 MB - 4 chip modules as 3.3 V. Here are some 3.3 V modules on eBay right now as an example,

http://cgi.ebay.com/ws/390613525506
http://cgi.ebay.com/ws/350478814991

The auction which does not directly specifiy 5 V (first link), I asked the seller, and he said 5 V, but both are 3.3 V as based on the memory chip part numbers.

For the second link, the photo is a standard photo used in most of this guys listings. While the part number in that image is 5 V (KM44C1000BJ-6), I received a 3.3 V item (KM44V16100AK-6).

I wonder if it is possible to cut the three Vcc traces on the SIMM module and add two 0.7 V diodes in series to drop the voltage down to an acceptable level? Two standard diodes in series would drop 5 V to about 3.6 V, which is within operating range of these memory chips. Or maybe I can get away with a single one of these diodes, depending on the operating current, http://media.digikey.com/pdf/Data%20Sheets/Ro … PDFs/rls245.pdf

According to the SIMM wiki, pins 10, 30, and 59 are Vcc.

I guess my double-diode SIMM idea is not so original. Attached is a module with four 4Mx16 chips, designated for 3.3 V. There is a double diode going to the three Vcc pins to drop the motherboard's 5 V to around 3.3 V at the memory chips. While that saves this particular SIMM from destruction at 5 V, my other set of four 64 MB EDO and four 64 MB FPM do not have diodes to drop the incoming voltage.

For those reading this who are not aware, a standard forward-biased diode, or rectifier, can often be used as a simple voltage regulator, with a standard drop voltage of about 0.7 V. The drop voltage varies slightly (non-linearly) depending on how much current is going through the diode and the ambient temperature. Forward-biased Schottky diodes drop about 0.3 V and Zener diodes, when reverse-based, usually drop between 1.8 and 36 V. For the Zener diode, you order the diode based on the reverse-bias voltage drop desired.

While I can still return my 3.3 V FPM sticks, I'll probably try the double-diode mod on my 3.3 V EDO sticks I ordered years ago.

I am eager to know if it works.
I also werent aware that i own two 32mb edo sticks that are3.3v according to spec sheet afterwards.
The ali finali 1487,9 should be supporting these 3.3v, but there was no post at all.

JaNoZ wrote:

I am eager to know if it works.

I checked my 3.3V, IBM 64 MB EDO SIMMs, however the three Vcc pins are all inner layer traces, so I won't be able to attempt a modification on them. Another possibility is to add a 3.3 V option to the motherboard via a jumper using a header and 2 diodes. Unfortunately, this would require modifying all 72-pin SIMM motherboards and isn't so practical. I will count my losses on the 3.3V EDO SIMMs and look for a proper 5 V set.

JaNoZ wrote:

I also werent aware that i own two 32mb edo sticks that are3.3v according to spec sheet afterwards.

Does the stick have a double diode configuration or other means of down regulation?

How many chips are on them? 4 or 8?
You can lift the mem IC VCC pins and insulate the pad, then connect 3.3v to the ic's so it is physically disconnected from the mobo's 5v memory lines.

But the memory will output data back to the chipset in 3.3v signals, wouldnt they be too low for the chipset to receive strong enough signals?
Do you have any umc chipset datasheet?

My 3.3v memory sticks has some sort of mosfet installed, some sort of power regulator.
Will make a picture soon when i have time.

feipoa wrote:

Were there any consumer-grade 486 or Pentium boards which used 3.3 V SIMMs?

I have a Tyan S1563D where you can jumper the SIMM voltage.

I will be surprised if you can find the UMC 8881/8886 manual. I've looked for this off and on for years.

My EDO SIMMs have 8 chips. The partial IC pull-up method you describe requires too much effort and uncertainty. I'd probably rather mod 1 particular motherboard these chips would go into.

I do not think 3.3 V will be too low for a logic high. Depending on the actual current needed for operation, it might even be receiving 3.6 V max. If the chipset is expecting 5V data, 2.2-2.6 V is normally considered a high voltage, and around 0.4-0.8 V is a low.

elianda, looks like Tyan had some forsight in the memory voltage department. I guess my SIMM jumper voltage idea is not so original either.

This is my ram, there is 3.3v as for looking in the reference list for 71v65804 it would be an 3.3v 8bit 64megabit chip, as there are 4 on the stick, so it would be a 32MB simm with an BA033 attached in the middle.
The BA033 is a 3.3v voltage regulator, so i am a little puzzled why they would use a vrm on a simm stick, is this an attempt to use 3.3v chips to a 5v circuit?

I can try maybe on a different mobo, but i am a little afraid, tell me what should i do?

I used 3.3 V EDO SIMMs which had the double-diode down regulation on an M919 for 2 years without trouble. I measured the voltage across the actual chips while it was turned on and it was 3.3 V.

I have had nothing but problems trying to get my money back from eBay member 1-800-4-memory for their 3.3V memory advertised as 5V. I'm not sure if this forum has a thread for eBay members to watch out for, but this seller might go on such a list. They don't check their part numbers before shipping, don't own up to their mistakes, and according to another member, do not test their memory before shipment.