VOGONS


Fake ISSI cache chips?

Topic actions

First post, by cosam

User metadata
Rank Newbie
Rank
Newbie

So I took a punt on some cheap eBay cache chips, fully expecting they might not be up to spec. Of the ten, nine checked out OK in a simple USB programmer/memory tester, which agrees with the commonly quoted 10% DOA rate but also confirms they are at least RAM of some description. They however don't want to work in a 486 motherboard so I suspected they might not quite have the 15ns access time they're labeled with. They don't look glaringly fake but they do have a bit of a texture to them, possibly suggesting the top of the package has been skimmed off. If 1901 is a date code that would seem pretty late for this chip in a DIP package, or not?

The attachment issi.jpg is no longer available
The attachment issi2.jpg is no longer available

In an attempt to measure the access time I hooked up a test rig and scope. Top trace is /OE and the bottom one is a high bit on the data bus. To be honest I'm not sure what to make of the results although they are hard to distinguish from those of a 55ns SRAM I had lying around. The input isn't very square but you do see the output starting to react pretty immediately when /OE drops, although it does take a while to get to something you might deem as stable.

The attachment scope.jpg is no longer available

Unfortunately I don't have any known good 15ns SRAM to compare this with and I'm not brimming with confidence that this test is even accurate enough to mean anything useful. Anyone know if this either confirms or refutes my hypothesis that these chips are on the slow side?

Edit: added photo of underside of chip.

Last edited by cosam on 2020-03-30, 11:01. Edited 1 time in total.

Reply 1 of 12, by derSammler

User metadata
Rank l33t
Rank
l33t

Certainly faked. Date code is silly and they weren't labled like that. Most likely 20ns or even 25ns ones.

Reply 2 of 12, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
cosam wrote on 2020-03-30, 10:29:
So I took a punt on some cheap eBay cache chips, fully expecting they might not be up to spec. Of the ten, nine checked out OK i […]
Show full quote

So I took a punt on some cheap eBay cache chips, fully expecting they might not be up to spec. Of the ten, nine checked out OK in a simple USB programmer/memory tester, which agrees with the commonly quoted 10% DOA rate but also confirms they are at least RAM of some description. They however don't want to work in a 486 motherboard so I suspected they might not quite have the 15ns access time they're labeled with. They don't look glaringly fake but they do have a bit of a texture to them, possibly suggesting the top of the package has been skimmed off. If 1901 is a date code that would seem pretty late for this chip in a DIP package, or not?

issi.jpg issi2.jpg

In an attempt to measure the access time I hooked up a test rig and scope. Top trace is /OE and the bottom one is a high bit on the data bus. To be honest I'm not sure what to make of the results although they are hard to distinguish from those of a 55ns SRAM I had lying around. The input isn't very square but you do see the output starting to react pretty immediately when /OE drops, although it does take a while to get to something you might deem as stable.

scope.jpg

Unfortunately I don't have any known good 15ns SRAM to compare this with and I'm not brimming with confidence that this test is even accurate enough to mean anything useful. Anyone know if this either confirms or refutes my hypothesis that these chips are on the slow side?

Edit: added photo of underside of chip.

Nice. I searched on Ebay for IS61C1024 and gots lots of hits, the labelling is quite different from one offer to the next.
I also found one guy selling some similar to yours with datecode 1901, too. But they were marked with -10, so fake. Obviously they label now with 15ns, so it's not so easy anymore to tell about fakes, the 10ns types were all fakes as far as I know.

Reply 3 of 12, by mpe

User metadata
Rank Oldbie
Rank
Oldbie

I think you can ignore any markings or date codes shown on pictures in listings.

What you get is a lottery. I order several chips and they different types of markings even when they came in the single shipment from the same seller.

Some have bright top text others dark text that is barely legible. Some have markings on bottom side others don't.

Blog|NexGen 586|S4

Reply 4 of 12, by nzoomed

User metadata
Rank Member
Rank
Member

I just bought a bunch off ebay thinking I had got the real deal and then learned about these fake chips!
Do you think mine are legit?
They look new and shiny plating on the pins, dont see any sign of the tops being sanded down and the markings are printed white and not laser engraved like some appear to be.
2 chips look a little suspect as they are a little dirty but pins are still shiny, they are the ones on the bottom right of the photos, codes underneath are a different font and one is marked 12ns.
Did they do them in 12ns?
The one in the bottom does not show the writing underneath, but it does have printing, it just does not pick up in the camera with the lighting, looks just like the one above.
https://imgur.com/vRDRQzU
vRDRQzU.jpg
https://imgur.com/EHv0Oj0
EHv0Oj0.jpg

Reply 5 of 12, by Anonymous Coward

User metadata
Rank l33t
Rank
l33t

I've had the fake 10ns chips for quite a few years now. Indeed, a few pieces were dead. But, the ones that worked have not yet failed on me.
I agree they are definitely not 10ns, but they seem to work about as well as my legitmate 15ns parts. I was using them in a 486 at 50MHz.
There is no way these chips are more than 20ns.

If you have a memory tester, they are still worth buying due to their ultra low price.

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 6 of 12, by noshutdown

User metadata
Rank Oldbie
Rank
Oldbie
derSammler wrote on 2020-03-30, 10:34:

Certainly faked. Date code is silly and they weren't labled like that. Most likely 20ns or even 25ns ones.

i agree that they are faked for the silly date, and i also suspect that all laser engraved ones are all faked.
but, how should the real ones be labeled like? i am just not sure if the ones i bought are real...

Reply 7 of 12, by nzoomed

User metadata
Rank Member
Rank
Member
noshutdown wrote on 2020-07-17, 11:03:
derSammler wrote on 2020-03-30, 10:34:

Certainly faked. Date code is silly and they weren't labled like that. Most likely 20ns or even 25ns ones.

i agree that they are faked for the silly date, and i also suspect that all laser engraved ones are all faked.
but, how should the real ones be labeled like? i am just not sure if the ones i bought are real...

Im in the same boat, my chips as i posted above still have the textured surface, dont appear to have a polished, grinded or sanded surface, have white printing and the date codes appear legit.
Does anyone know if they have found fake chips with white printing or are they all laser? I think it must be possible to tell if the surface of the chips has been skimmed too.
Any ideas? Should I worry if they work?

I cant remember what I paid, but they were not overly cheap around $3-4US each from memory, possibly more.

Reply 8 of 12, by amadeus777999

User metadata
Rank Oldbie
Rank
Oldbie

Most chips in these packages, especially faster than 15ns, seem to be fake.

This wouldn't be a problem if the quality was good but it seems that most parts are not held to high standards. If you want genuine 12ns 32K rams you can search on ebay for the seller "huchtec"(if I remember correctly). He had/has authentic parts due to buying remnants of a HiFi manufacturer. If this isn't an option then buying ones in more recent packages and adding a adapter board seems like the only venue.

Feipoa appears to have been lucky as he acquired fake 10ns 128K(64-, 32-k) chips which seem to hold up to their spec'd latency.

Last edited by amadeus777999 on 2020-07-18, 14:55. Edited 1 time in total.

Reply 9 of 12, by mkarcher

User metadata
Rank l33t
Rank
l33t
nzoomed wrote on 2020-07-17, 06:10:

I just bought a bunch off ebay thinking I had got the real deal and then learned about these fake chips!
Do you think mine are legit?

Your chips look convincing to me. Most lots of fakes I encountered had the same datecode over the whole lot. The date code (97, 98 and 99 for the -12 chip) makes a lot more sense than 2019. Somewhere here on VOGONs, we got a post with ISSI inside knowledge that claims they did DIP SRAMs up to 12ns, but none with 10ns, so the speed grades make sense too. In my oppinion, your chips are either real or really good made fakes.

Reply 10 of 12, by nzoomed

User metadata
Rank Member
Rank
Member
mkarcher wrote on 2020-07-18, 14:32:
nzoomed wrote on 2020-07-17, 06:10:

I just bought a bunch off ebay thinking I had got the real deal and then learned about these fake chips!
Do you think mine are legit?

Your chips look convincing to me. Most lots of fakes I encountered had the same datecode over the whole lot. The date code (97, 98 and 99 for the -12 chip) makes a lot more sense than 2019. Somewhere here on VOGONs, we got a post with ISSI inside knowledge that claims they did DIP SRAMs up to 12ns, but none with 10ns, so the speed grades make sense too. In my oppinion, your chips are either real or really good made fakes.

Thats good to know. At least I know they are genuine, which is what I though when i inspected them too.
I was told by someone that even if they are genuine, expect at least 10% or so to be faulty.
If i run Memtest86, i expect I should be able to pick up any faults easily enough?

Reply 11 of 12, by Horun

User metadata
Rank l33t++
Rank
l33t++

Yes or Cachechk. You will have to do some swapping to find the bad one if there is one. Hope you bought more than needed. Hint: mark each one with a Sharpy pen 1 thru how many you have (can always remove with ISO Alc.) so you do not get confused if you get a failure...

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 12 of 12, by nzoomed

User metadata
Rank Member
Rank
Member
Horun wrote on 2020-07-19, 02:36:

Yes or Cachechk. You will have to do some swapping to find the bad one if there is one. Hope you bought more than needed. Hint: mark each one with a Sharpy pen 1 thru how many you have (can always remove with ISO Alc.) so you do not get confused if you get a failure...

yup definitely.
I might buy a bunch more anyway, that way I likely will have more with closer date codes and can use them in my 486 too