Check out my first video about faked CPUs too. Great crime story. 😉 kzfaq.info/get/bejne/rdl7nbtnuq29hYU.html

I always felt a little bad for throwing away or destroying a processor. Such amazing little triumphs of technology.

The funny thing is that after almost 30 years, chinese sellers still do make weird "custom" CPUs using mobile intel chips

Actually, this CPU was pretty reliable. Intel sold it at a 100MHz clock to make it fit a specific TDP (and power draw). The higher voltage and clocks might (definitely) only decrease the efficiency, but with desktop-type cooling, the CPU might be fine for a very long time. I'm just sorry that you did not attempt to overclock it at "stock" voltage of 3.1V~3.3V and see if it goes 200MHz. I think that it could, as those chips are downclocked a lot from the factory, for efficient mobile use. I have a desktop P54C 120MHz CPU with an integrated cooler that I ran successfully at 190MHz (2x95) on an ALI Aladdin V+ motherboard for a long time, paired it with all sorts of normal or crazy (for this setup) GPUs like Matrox G200, Rage 128 Ultra, TNT Vanta, Radeon SDR, Geforce 2 MX 400, Geforce FX 5200, etc. Some of the benchmarking I did record, so I can search for it if you are curious. In other unusual tests, I ran a Cyrix MX PR 300 (233MHz real, 66MHz x 3.5) at 250MHz (100MHz x 2.5) for a very significant speed boost and pleasant performance with all of the above cards, making Unreal playable, then Glquake and Quake2 smooth as butter. LE: I see that I already had already uploaded (unlisted) for a friend some Windows XP + Matrox G200 tests with that 120MHz Pentium at 190MHz. Windows XP runs horribly on that setup with that specific motherboard and only 128MB of RAM. In Windows 98, things are significantly better, as for the DOS version of Quake (under real DOS), I remember something like over 40fps in software mode and no screen borders. kzfaq.info/get/bejne/fp-DeM1mq7rKZWQ.html LE2: Here is some testing with that Cyrix mentioned above in Windows 98 and an S3 Virge DX video card. It shows the various software rendering choices for Quake 1, GLquake1, software Quake 2, OpenGL "accelerated" Quake 2, software Unreal, software Half Life, D3D "accelerated" Half Life. Now, in all honesty, there are quite a few 3D games that work reasonable with the S3 Virge DX, among which you find Tomb Raider 1 and 2, Descent 2, Mechwarrior 2, POD, Croc, and a few others. With the Rage 128 Ultra, the Cyrix MX was OK for quite a lot of games. kzfaq.info/get/bejne/gbRmY8uYqcjKnXk.html

"Very rare and collectible" "Let's cut it apart"

Certainly a lot of work went into repackaging the chip and heatsinking it to the custom lid.

My guess, intel had an overstock of the mobile CPU's and they were happy to volume sell to a company that would produce a product that was cheaper. The cheaper chips were sold to consumers who would have purchased a competitors product anyway, so Intel still made money. Any warranty claims are easily turned around, as it is a "Chinese Knockoff".

Amazing and cool -- thanks for tearing it down so we could find out what it really is. I guess the over-voltage allows for a "stable" overclock -- LOL!

I think, that xraying before ripping apart is quite something. I love this channel

How the hell you have 790 subscribers only??! Your channel has gold premium contents and time spent on it is very entarteining!

I feel ya. I wouldn't mind running one of these actually. I enjoy collecting and using unique vintage computer parts.

For a while you could buy a laptop CPU in a desktop package on purpose. They would sell them as that. As far as the 100mhz instead of a 133mhz, it might be possible that the 100mhz was either the rating for when used in a laptop (thermal issues) or it might have come off the assembly line as being capable of running 133mhz but was binned for 100mhz for cost. As odd as it may seem under rating a CPU was common back then. A manufacturer would use a large "circle" of silicon and dope it to have many individual CPUs on it. Before thew would saw the chips apart they would put them in a jig and test them and see how fast they could go. At first you'd get lucky to find a faster chip among the batch and you would pay more for the fast one. As the manufacturing went on and settings were tweaked to get the best yield, the number of "lucky" chips grew. At a certain point there were more fast chips than slow ones, but people didn't want to spend the money on a fast one. So the manufacturer would take a fast chip from the fast pile, stick it in the package we are all familiar with, and label it as a slower one. Since the manufacturing cost was still the same, it didn't hurt sales if they sold a fast one as a slow one.

Very interesting and a very well produced video! May I suggest if you have to do this sort of thing again, see if you can get access to a mill or even just a dremel? It wouldn't take much to shave a few mm off the bottom of the package so you could leave the lead frame in place unmolested and look at the text from the underside. Alternatively, with a dremel you could just cut a hatch in the bottom since socket 7 doesn't have pins in the centre.

A very nice breakdown of those chips. I even learnt something too. So Here take my sub.

You have to admire the ingenuity that went into creating this fake. I would be intrigued to see how long it would last during a proper burn in test.

my thoughts are. maybe they worked fine since it had a decent IHS, plenty of TIM. not to mention it was dead bug style which means the IHS was closer to the heat to dissipate it. And maybe the creator discovered it could run it at a higher clock speed with more voltage than designed for because of how it was packaged? if you look at extreme overclocking you can see the same results. you can get a much higher clock with more voltage providing you can dissipate the heat well enough, while at the same time having very little or no degradation. let me know if i may be completely wrong. Awesome videos so far! glad EEVblog linked me to you!

Very interesting. Thank you for sharing.

Thank you so much for this video 👍🏻👌🏻✌🏻

In my cpu collection, I found a counterfeit intel 80486 dx2/66 that had the top ceramic sanded down and re silk screened. It was originally a 80486 dx2/50

Quite good you actually manage to overclock it to 133, I remember back in 1996 we mange to sqeeeze our 120 out of 100mhz pentium and stable 110Mhz

Great video. Thanks for showing that CPU. It makes sense that they would do that and overclock a 100 mhz pentium to 133. Some shadiness is to be expected.

Great video! I also have such a version marked Pentium MMX Tillamook-233M S85 2V. So whoever was faking these was continuing to do so years later. The packaging is exactly the same with the sandwich structure and the golden spreader… I will not cut it open to see which mobile version it is but probably a 200MHz mobile Pentium MMX.

Makes sense, over voltage generally means you can get more clock speed and stay stable with the expense of more heat and shorter life span - which explains the different lid.

Wow, this is really interesting. Just found your channel!

Vary interesting! I would love to have one of those for my collection.

I like your way of content creation I like this video

I remember that Linus Tech Tips reviewed a more modern laptop CPU converted to desktop use. I didn't know something like that was already done many years before that! I also like your accent.

Saw one pc advert in 2004, say "Genius Intel Pentium 4". :-)

I remember a more recent effort by some Chinese sellers to repackage Intel's 7th generation i7 laptop chips into desktop packages and it was for the reason that the mobile chips were cheaper and even after the whole adaptation process the CPU was cheaper than buying the desktop version.

Thanks, now I want one

It was probably an electronics engineering student's final year project (PCB board design). Someone (him?) might have then decided to make it into a complete business since the idea of using the lead carrier pentium versions is actually rather genius! It never occurred to me that you could use those chips remapped into standard PGA sockets ... its actually a really smart idea!

May I ask where you work so that you got access to an industrial x-ray machine?

Cool video.

I would have definitely kept it. Even if it's not "genuine", it's a pretty amazing collectible item in its own right 😀

I knew a company who's side business was a computer shop. They built many systems for schools and larger companies locally. The main build was a slot 1 pii 300 processor. Over time they saw performance issues and lockups because the fans would get weak or fail. Then one day the feds had shut down the store. Turned out the batch of cpu's they bought were 233 chips that had been modded to 300. They worked fine at that speed but ran hot and someone did an investigation and had contacted intel. All the chips left in inventory were confiscated and they had to provide proof they purchased them as 300's and didn't do the modding themselves. Funny thing was if you decased them and put your own heatsink and fan on them they were actually great cpu's. They also got a hold of large batch of motherboards who's caps were 85C and not great quality. Basically a 3 month timebomb. Once changed out to 105C caps they were great. I think these two issues caused them to think the computer store was more hassle than it was worth and they shut it down.

I wouldn't necessary call it a fake but, but its not original, At least it still worked and performed just like the original.

Soooo what youre saying is we can get cheap knock off pentium 1's for budget retro builds? With a pretty gold top? And being metal would take a heat sink well?

Due to the size of the processor and the ammount of heatsink plate around it. I believe they were running the 100MHz in overclocked mode to 133MHz, which is why on the benchmark it shows 132.95MHz of clock. Since there's plenty of plate to cool the thing down, and 33MHz doesn't produce much heat, I hitnk that's the choice they made.

Those are the same dye that they used on the desktop based cpu's. These ones are just binned a little higher for the 2.9 volts and low heat. In essence that cpu is not overloaded at 3.3 volts as it was essentailly designed for it. IIRC those are the same stepping which were used on the P200S and late 166's. And you could probably have gotten that to run 200 if its not multiplier locked.

This actually made me wonder, can you technically use a very fast microcontroller to simulate old CPUs? Like if you had a microcontroller with a program that translates the CPU instructions to the microcontroller, processes them then sends them back as if it were the CPU itself, using a cable that plugs in like it’s part of the socket.

The thermal engineering seems fine.

For someone interested in just the performance for late-era DOS software or early Win95 this seems like a fine option if the price is right. Don't know how easy it'd be to get the parts for the rest of the system anymore though.

👍👍👌

In China anything is possible. Even chips without a die inside. Anything, literally anything goes.

350nm process on the desktop and 350nm LV process on the mobile? I am thinking this clever hack was not actually very hard on the CPU at all. This is sort of an early 350nm+ process that likely allowed it to get away with running lower voltage, for the purpose of saving battery life, but likely could run at either 2.9V or 3.3V without any reliability problems (only a little more heat than the desktop version). Any way to measure current load on the regulator or power supply between the two chips?

I'm assuming the mobile CPU is 3.3 vt tolerant. I was able to run a desktop 133 at 2 volt long ago when I was attempting to make a home brew mobile system in order to get better battery life.

Schönes Video. Habe auch ein Abo dagelassen. Ich schätze, Du kommst aus Süddeutschland, oder? :-) Genau solche CPUs gibt es auch heute noch aus China zu kaufen. Die bauen immer noch aktuelle Mobile CPUs in Packages von Desktop-CPUs. Der Linus von LTT hatte die auch mal getestet. Was für ein Aufwand für den von mir doch recht niedrig geschätzten Gewinn bei dieser Nummer.

That’s pretty neat, it looks like there’s tiny wires coming from every pin to the die on the true Intel desktop

In the 90s they have a ton of these flood the market - celerons esp. They sold them at computer shows or in catalogs at a little lower price, but most people have no idea they were buying fakes and would be happy

Good video nice

Had one of these briefly a long time ago, was sold locally by a slightly dodgy looking guy at a weekly market. They looked strange but were sold at about 60% of the price of a "genuine" desktop intel chip. I wish I bought more of them (as I know collectors look for them) but at the time the cpu was just about superseded. Still a nice curiosity atleast, which I wish I hadn't sold.

Very interesting, that’s what I like about KZfaq,

Interesting fact: KZfaq show me Alibaba's ad in this video 🤣🤣🤣

How is the lead frame connected to the PCB, soldering looked also wire bonded?

Over volted and over clocked, but it's an intel part. It's pretty ingenious actually to build the socket adapter and heat spreader to get the thing running!

Wo hast Du das Wurst Probiermesser her? Meines ist von AVO

Back in the day, if it worked I would not have cared.

Aww you destroyed it. Thats really a nice item to collect.

I surmise that it was produced to fill the gap whereby there are not many originals available to buy on the internet for retro computer use

wow, I had no idea about these fake Chinese CPUs, thanks

I haven't seen any of these in many years as such used to turn up on eBay in the early 2000s.

I had a Toshiba Satellite 420CDT laptop with the 120MHz version of the same CPU inside. It ran a 60MHz bus with a 2x multiplier. I learned how to modify the multiplier and bus speed. I went straight for 66MHz bus and 2.5x multiplier for 166MHz and it worked, but it got hot and battery life sucked. I ended up backing down to the same 2x multiplier and keeping the 66MHz bus for 133MHz, which actually increased battery life a little bit since everything on the system bus was running 10% faster so work got done faster. That was a weird machine. Didn't have an L2 cache.

I had absolutely no idea these semi-fake chips existed, genuinely cannot work out why someone would make this! I guess there was some small market for the higher-clocked Pentium mobile chips but was it really worth it? It's weird, I definitely never saw one of these back in the day either.

I bet that 133 "100" will over cock like a BOSS!!! Mobile CPU are clocked down desktop CPU's i bet that thing would cock to 150 easy i got P233's over clocked to 300Mhz with a 100Mhz FSB

Hey, what is the name of this CPU and where did you buy the real one? Please, i really need an answer

It's pretty cool seeing all the wires that run from the CPU die to the pins in the x-rays. Not something you see every day.

i used to get these cheep. as long as the motherboard's bios could handle the cpu it would work fine. Speed wise you get some leway putting the moble into a large packige more cooling and such. I cant say for all manufactus of these but the suplier i had did sell them as surplus or reclaimed mobile cpus in adapters. for about half what you would pay for a full desktop version at the time. me and my friends being a bit on the poor side of the tracks ended up relying on these for a while and we didnt see any degregation through the life of the systems. of course back then a system wouldn't need to preform nearly as long as our modern setups wich do to the slowdown of leaps of speed tend to hang in there for years and years still can't believe my motherboard cpu and ram is as old as it is and im just now upgrading. Gen 3 i5 and only now just seeing some preformince problems. and there more related to ram speed than the cpu. just cant seem to push the video card to 100% anymore :) BTW 1 motherboard/cpu/ram has outlived 3 sepret upgrades of video cards. skipping a gen or to of video inbetween.

When I wanted to put together a cheap machine in those years, I used a Cyrix processor. Fake processors didn't arrive here in Europe. But tons of DRAM did, driving mad whoever tried to assemble a PC together.

You must be that slingshot guys long-lost brother

Science requires to hurt poor little cpu, sad, but im now convinced. I still have laptop with this lead frame, brown thing package. But mine is p2 233 i think, but looking very similar

Might have been salvaged CPU's from devices (boards) that didn't pass manufacturing specs. Ended up in China to recycle. They've been doing that for decades now. Salvage and put in a new package. Very good profit margin.

but why was this done? Were the mobile versions cheaper? Or is this a modern fake?

Hola donde los compraste esos mmx falsos

left over voltage for overclock maybe, 100-133mhz, performed like that in benchmarks, so possibly just an even cheaper way of producing a 133 clone.

man hört ganz deutlich, dass du deutscher bist, aber... very cool video i subbed and liked

I just find it so amazing at the quality and the time necessary to produce such a fake chip. Was there such a large enough profit margin to be able to acquire a mobile chip then de-lid it and the go through all the steps to reinstall it on a new desktop die, then fake the bios? for a 10% gain. maybe I'm missing something or was the chip worth double? so your gonna make $200 to $300 more I could see the effort but in 12 months your hole prosses is obsolete...But I find what your doing absolutely the coolest thing for retro chips on KZfaq...

I guess it could be overclocked easily due to increased voltage :o

Even when it is fake, pretty neat look and after all it still an original Intel CPU. I have seen some thinkpads like the 380 use this sort of mount. Low quality brands tend to buy the obsolte production lines and materia once the fabrication or integration have moved one step foward. The plant is bought for cheap, probably a laptop assembly facility that had the machinery to moun this thing into PCBs and also large quantity of unused materials. There is an entire market for this parts once their life cicle come to an end. A well known semiconductor company that tend to have good pricing and also a good policiy of keep their legacy prodcuts alive (VERY GOOD) is also knon to buy plant from other companies once they are done with them of they are not longer usable for their needs. My guess is something like that happened here too. Because mobile version probably is cheaper but the mounting process doesnt look easy ... UNLESS ... you have the proper machine that can. And I dont see them buying an sky expensive machine to fake CPUs. More like an abandoned production line bought dirty cheap

I don't think the bumped voltage do any harm if the dies are the same as the desktop Pentiums. The lower voltage only matters in a laptop where battery life is the key. Also the core gets much better cooling in this package compared to the ceramic ones. It would be interesting to see if it can run much past 133MHz.

Mobile dies are often a better binning, working on a lower core voltage, which means they offer excellent overclocking headroom. IMO there's a high probability this was stable at 133MHz at desktop core voltage.

Companies still do this nowadays, often with GPU's. Some of them function, tho usually not up to the advertised specs, some are kinda flaky and glitch out randomly, and others are just broken.

Quake! Used to run on 75 meg and up. Course I had lots of ram . To compensate. For lagging frames

Well... higher voltage => higher clock speed. With the good heat sink of an desktop pc, i assume they didnt need to lower the voltage and therefore basically overclocked the 100MHz CPU to 133MHz. Stability as you mentioned (and mabye heat excess) would be the only issues.

Arount the age of Pentium 100 mhz in Argentina we called PINTEL (PAINTED INTEL) the letters are painted with serigraphy, so bad quality paint thet cracks with the nail, i dont remember if a Pentium 75 are painted to show as Pentium 100

I think it's fundamentally the same die anyway as desktop 133MHz CPU. For sure it is cooled very nicely in this heatspreader package, so i'm not too concerned for it. Maybe some binning differences, but perhaps not to a disadvantage. The flank margin decreases at lower voltage, which may be a reason the CPU got sold as 100MHz mobile, while same quality die can really be good for 133MHz with higher desktop voltage.

Retro 🤘Rock

QL3X and QNCT grandfather :)

Man watching you cut that really cuts at the heart

ever seen Frozen? the big guy who runs the shop sounds just like you!

that was easier than i thought opps lets try that again only with 9 fingers this time

Could these chips have been salvaged from old laptops?

Also ,the original Intel cpu's was built using highly Gold coated materials, but china made is more likely painted with Gold color

A Mobile Pentium 100 @ 133 mhz desktop speeds makes perfect sense from a clocking / OC perspective. The Mobile Pentiums were 0.35 micron / 350nm instead of 0.60/600 nm. The desktop 0.35 micron chips clocked up to 200 mhz (3.3 volts), so a "100 mhz" mobile chip @ 2.9 volts should have no problem at 133 mhz @ 3.3 volts. I'd suspect you could OC this to 166 mhz without any fuss. I can only assume Intel was selling these cheap because at the time the mobile market was very small, and they were probably trying to drive adoption / get control of the mobile market, while the desktop market still commanded a premium for expandability and so forth. Or alternatively, they were simply selling these for very cheap into non-premium markets at the time, the buyer/faker got a load of these for cheap, and then remarked and re-exported to markets that were used to paying higher prices. Either way these actually look pretty cool for fakes, and were sold at the time where binning was almost completely artificial - I'd be surprised if there were any 0.35 micron Pentiums that couldn't run at 133 mhz or higher.

Intel used the exact same dies for desktop and mobile and just binned them and gave them different voltages. Only later tillamook parts were made specifically for mobile after desktop moved on to slot 1. This looks like a P54CS part by the code on the frame. This CPU will be perfectly fine with 3.3V; thats what its desktop counterparts ran at with the same CPU dies cut from the same wafers. This CPU could have been a desktop part running at perhaps anywhere between 120 MHz to 150 MHz at 3.3V, wherever it would have passed intel testing, but since power consumption matters in a laptop they chose to clock it slower (power goes linearly as frequency) at a lower voltage (power goes as voltage squared). On these older CPUs where leakage was not a big issue the power consumption of this would have been around 3.3^2/2.8^2*133/100 = 184% of the power consumption of the mobile chip at stock. So they got 33% more performance almost straight across the board at 84% more power. On a desktop that's an obvious win; especially considering the P54CS 133 MHz from intel was an 11 W part and the rest of the system uses maybe 50 W. 33% more performance at 5 W extra power consumption is very much a win. On a laptop that consumes perhaps 15-20W in total and relies on a battery that likely was NIMH or NICAD, that's just huge. My guess is that there was some great overstock or price difference that made it worth it to turn these into 133 MHz desktop CPUs and arbitrage the difference. Intel could have done it themselves since this likely could have made it as a qualified 133 MHz part, but they where too big and too slow and did not want to remanufacture finished laptop CPUs back into ceramic desktop CPUs. The Chinese likely did not do very much work to verify that all these chips indeed ran stable at 133 MHz and they didn't care about their reputation and were flexible enough to do the remanufacturing.

Seems like a huge amount of work to create a fake chip. I would like one of these mobile versions in it lead frame for my collection, never seen one before

that package is bizarre

I wish i could have a CPU like that, just for Collection Purpose

they are still selling mobile CPUs in desktop packages for Intel Core i in China

I wouldn't say it was overclocked and overvolted. I'd say it was just restored after picking it (back in Intel's factory) for a laptop part and giving it a lower voltage and clock speed rating ;)

Gotta hand it to some of the Chinese bootleggers, they come up with some ingenious stuff. What’s not right, is when they do that with the intent to rip people off