EEVblog

[@351WINCHESTER]

In 1976 Harris Semiconductor in Melbourne, Fl. gave me the .50 tour as to how they made their silicon chips. Pretty remarkable especially back then.

[@demyhr]

As a retired test engineer working at National Semi, Motorola and other semi houses here's a couple tidbits about wafer probing. First, the probe card does not move, all motion is in the chuck that holds the wafer using vacuum. The chuck is positioned in X, Y and Z by linear motors in the most common Electroglas probers. The pcb is very thick because probing high pin count die requires an extremly rigid material to withstand the significant amounts of force needed to deflect the probes such that they all make contact. This Z axis distance is in the order of 1 to 3 thousandths of an inch. Second, the probe ring is mated to the pcb after it is assembled on precision fixturing to position the tip of each probe in the center of the die bond pad and perfectly aligned in Z axis. Once the probes are positioned the assembly is encapsulated in an epoxy ring, hence the name "epoxy ring probe card". The probes are tungsten which has the necessary mechanical and electrical properties required. Probe cards are periodically "aligned and planarized" after probing 25 to 100 wafers and to repair damage. This is done by highly skilled technicians under a microscope. Hopes this is helpful. Probing technology has come a long way since the 60's when I first used them.

[@GOLEG11]

that roswell Caused it all!!!

[@AlfredoMazzinghi]

Thanks for sharing!!

[@strongholds12]

David Myhr get out of here dinosaur. You're out of date 🙄

[@analogaudiorules1724]

@@strongholds12 How about don't be an asshole. 😂

[@ruthlessadmin]

Did you ever get to work with encabulators?

[@garethbull2226]

Just looking at some of these older vids. Here's a story I heard years ago about a chip fabrication factory. Once the circuit lines have been etched onto the chips, the chips pass through a kiln to harden the circuit lines on the silicon. The kilns were similar to those pizza ovens that have a conveyor belt in them, cold pizza goes in one side and hot pizza comes out the other side. There was a chip factory (I never heard exactly which one) where the defect rate of the chips would mysteriously skyrocket once in a while, then go back to normal, only to skyrocket again a few weeks later again. The engineers at the factory were tearing their hair out trying to find the cause of the defect surges. One Friday evening one of the engineers was working back late and as he was on his way out he went past the chip kiln area and found some of the night shift workers had been putting frozen pizzas through the chip kiln as a late night snack. Fat and other substances from the pizzas were then contaminating the sets of chips that went through the kiln after the pizzas had been cooked, causing the defect rate for the chips to suddenly rise. A bunch of people in the factory suddenly became unemployed.

[@thelavian4481]

LMAO!

[@RetroArcadeGuy]

You made me hungry.

[@Epic3D-k3d]

This story I do not hear for the first time. The first time was at the university of my professor. I believe it by now. Because so much what happened to me is bordering on madness.

[@davidfrisken1617]

It is an Urban Myth. You will probably hear the story again in your life, from yet another source, and so on.

[@hardscorerockkssss]

David Frisken not myth..i have also worked at chib fabrication chip and we had similar issues.like people not properly using clothing or take not needed stuff there

[@jeanious2009]

Beautiful wafers....I remember when i was in Elementary and on "career" day this geeky guy brought a wafer to show us and try his best to explain what they are and why they were so important and how in the future they would get more complex....this was about 20 years ago and I just can't believe how beautiful they are and AMAZING...wish I could go back in time and admire the wafer. No one paid attention to that geek, everyone wanted to see the stupid cop or firefighter. LOL.

[@ArieLash01]

Hi Dave A few small corrections from a wafer sort person. If you look close on the probe card you noted that some some of the probes were shorted. Those pins are power pins Vcc or Gnd supply's . The next comment is that the probe card does not come down on the pins the wafer moves under the probe. the probe card you have is the xyz accuracy of such a probe card the height diffrence between the first and last pin needs to be >5 mil

[@stoatythesecond888]

Those wafers are insane, the level of detail and scale is ridiculous. Very interesting and demonstrated how little I know about processor architecture! Thanks for posting.

[@spayderninja]

At around the 38:00 minute mark, those are not pinholes that he is seeing through the probe. Those are the reflections of his camera's LED lights. The LEDs are reflecting from the glassy surface of the probe material.

[@BillAnt]

That's exactly what I was thinking too, you can see the round illumination pattern reflecting off that glob of poxy acting like a lens. Maybe Dave got too exited and mistook it as pinholes. At these micron and sub-micron scales, one needs a good couple of hundred dollar quality optical microscope in order to see anything meaningful.

[@fredericamelowee2348]

that's a resistor

[@Sabhail_ar_Alba]

The semi industry is divided into back-end and front-end operations. The back-end - cutting,testing etc - is usually done in the far east. The front-end is the actual fabrication process which involves lithography, etch, furnace, sputter and implant - virtually all use robotics to handle the wafers. The actual fabrication process can take up to 6 weeks and the wafer might have up to a dozen layers.

[@steelthfighter]

+Angus MacLellan so aside from the wafer, how much more goes into it? or is it mostly just the wafer then all the circuitry is already in the wafer, then attach the leads?

[@AgnostosGnostos]

Nice info. One of the best videos about processors technology I have ever seen.

[@jamesdriscoll9405]

@@steelthfighter the active circuitry is all at the surface of the wafer and is "paper thin".

[@spaceframe123]

This is so interesting. I have no idea what I'm watching.

[@gustavgnoettgen]

Exactly, you're welcome 😄

[@pattyoneill91]

so cool. by bar, chip manufacturing is one of the most, if not THE most amazing production technologies we have developed. It's crazy to. Moores law is just sloing down now.

[@normanslade4903]

0kIlLtHeMuSiC 0

[@Czeckie]

speak for yourself! I haven't developed shit

[@danheidel]

FYI, the tungsten used in those 100 nm probes is partly because of the extremely high hardness of tungsten but also because you can use electrochemistry to sharpen it to a few atoms at the tip with just a few volts and some saltwater.

[@Zenodilodon]

Oh I want to get a full wafer so badly one day. They are absolutely stunning to see.

[@abpccpba]

Vincent Himpe thank you so much to be so gracious supplying these FAB parts. Marvelous : - ))

[@tjay1430]

I kept avoiding this video thinking that the video was going to be about chip reels. Much more interesting than expected. Keep up the good work. Thank you for sharing this with us.

[@emem85]

just finds this EEVblog videos...pure gold dude.

[@tomwaters86]

Haven't got a clue what your on about half of the time but your a legend and I love watching your videos as soon as they pop up. One of the best of my subscriptions! Go Dave!!!

[@van-c630]

Great example of some unbelievably microscopic work that goes on behind the scenes to provide us with modern semiconductor components. Thanks for sharing.

[@shreyasrao125]

38:28 They're not pinholes. It's the reflection of the microscope LED.

[@leocurious9919]

Shreyas Rao Jeah. I was like... wtf is he talking :D

[@BillAnt]

That's exactly what I was thinking too, you can see the round illumination pattern reflecting off that glob of poxy acting like a lens. Maybe Dave got too exited and mistook it as pinholes. At these micron and sub-micron scales, one needs a good couple of hundred dollar quality optical microscope in order to see anything meaningful.

[@micarifamily1]

Was just about to say the same thing but saw you first....lol funny to see and hear that good catch!

[@fredericamelowee2348]

the pin is soldered on a resistor and the led reflect on the blak side of the resistor.

[@EEVblog]

Thanks. I do my best to provide a decent video quality given various constraints.

[@electronash]

"Hammer time!". LMAO Awesome mailbag. Thanks to Vincent / Dave for a great vid! I have "fond" memories of wearing the ol' Bunny Suit in some factories before (optical broadband stuff mainly). Gets damned hot in those when the room was already around 35 degrees C. That probe card looks insane - I hope the people who soldered it got paid a good wage. Looking forward to hearing what the chip designs are on those wafers. They look like some sort of MCU, as you can make out the ROM / RAM blocks.

[@pklongutoobe]

That's really cool. Sort of stuff that should be on display in a museum.

[@carabela125]

Thanks for posting, I've worked in the industry for many years but never had an overview like this before.

[@embeddedhenrik]

Thank you for showing it! And a big thanks to Vincent for making it possible :D

[@rupertrooksby]

The probe card is fixed in place during setup, the test head is lowered onto it and fixed too. The wafer is loaded onto the prober chuck, it is the chuck which aligns the wafer and performs all the x, y and z movements to make contact with the probe tips. Most fabs haven't fired ink during testing for decades. The pass/fail results are instead stored as maps and used to either ink the wafer on dedicated machines later, or the die are assembled directly from the maps in an inkless process.

[@tubical71]

Absolute best mailbag episode so far!! And thank you for your in depht explanation of all the things inside this very interesting parcel. A very special and BIIIG thanx goes out to the sender: Vincent Himpe!! GREAT STUFF thank you so much for providing and Dave for sharing :)

[@numberpirate]

I used to work at the IBM 200mm fab in Burlington Vermont USA and we used to call the cleanroom suits, graph paper ninjas. Had to share. I also used to take home the old reticles for the photolithography Nikon tools. I wish I still had some because I would send one too you. Uber cool looking.

[@napablue7502]

+Marcus Taber Very cool! We were the first outside company to use the Burlington FAB when IBM opened up the facility as an ASIC vendor. I LOVE the Essex area. Very beautiful countryside, and I also love UVM. It was so amusing working with the design engineers there since before us, it was exclusively IBM personnel -- it was like the scene in "Independence Day" when the President et al go to Area 51 where the scientists "don't get out much." They loved the outside contact and asked us tons of questions about what it was like in the outside world. haha

[@gnagyusa]

Awesome video! Thanks for posting it and thanks for Vincent for the goodies.

[@Seansmit23]

Best mail bag by far Dave! Loved it. Big thanks to Vince as well!

[@FranLab]

@23:15 - Now that's a multilayer PCB! It would be great for you to X-Ray that board without the frame -- show the insanity within.....:-)

[@Paul--W]

Damn that interesting, thank you Vincent all these years on.

[@rajavijayenthiran494]

Interesting looking at the inards of chips I used to work with. But more than that, it was your presentation and the Gangam Dance that was enjoyable.

[@tonybell1597]

Excellent, an insight into a very small world,. Thanks Dave, and a huge thanks to Vincent, nice one!

[@stevetobias4890]

Vincent you are a bloody champion mate!

[@EEVblog]

There are countless variations on the 8051 core that are still manufactured today.

[@chronically_late]

I used to work in research and development for the amorphus SiO2 nanoparticle solution used to polish those wavers. I LOVED to visit our customers! :)

[@TwistedLogic93]

To answer your question, spin coating is very common. We have a clean room at my college and when we went to coat the wafer in photoresist we used a spin coater.

[@shark123476]

the probe card had 144 wires on each side of the square just in case anyone was wondering PS i counted this at 3am so i may be a little off ^-^

[@mirkomueller3412]

Absolutely amazing stuff you hardly ever get to have a look at plus lots of information provided by you equals a great video !!!

[@TheLightningStalker]

Fraps is a good screen video capturing program. It can record audio from the microphone as well. It would have made it a bit easier to see those microscope shots.

[@andiyladdie3188]

Just W.o.W nothing to say, much thanks to Dave and Vincent for sharing such a great stuff.

[@WeAreGRID]

its videos like this that are priceless to teach the next generation, words alone cannot express how much i am in awe of the process that goes on in making electronics now lol

[@StabilizeStudios]

'Sorry my probe is just way to big for this' Giggity

[@gamccoy]

Mad Props, Vincent!! That was a pretty good show. Well done, indeed.

[@SimonLindauer]

That was most impressive and informative! Thank you Dave & Vincent!

[@martijn3151]

So awesome that someone sent this over to Dave! I tip my hat to you Vincent

[@cipmars]

Man, this absolutely boggles the mind! If it's hard to imagine manufacturing anything here, imagine inventing them and then imagine finding a way to make them! Awesome!

[@roundhouse283]

The wafers are circular for polishing. This is a nice channel. Hammer time

[@DiDe273]

plz Dave, never change, you are awesome!

[@ForViewingOnly]

Awesome video Dave... totally mind-blowing for someone like me who has never seen any of this stuff before.

[@Elminator666]

What! I used to live in Naples, FL and I've never heard of that company before... Learn something new everyday.

[@k7iq]

Thank you David ! Excellent video and very educational ! Oh yes, thanks to Vincent, too !

[@emiremiroglu]

I was incredible looking forward to see wafers closely, thanks a lot!

[@Satchmoeddie1]

Most of the fabs have a least one SEM to check the probes. There are photo etch blocks & chemicals that are applied washed off and reapplied to the wafers. Sumitomo has been making wafers in North Phoenix for about 14 years now.

[@tomgt007sugden]

excellent video. I use to work in the telecommunications business making multi layered boards. Some of the techniques I witnessed in the construction and development was mind blowing.

[@metrikal1]

This is one of the most interesting videos I've stumbled into for awhile, so awesome! Thank you for sharing!

[@MrStricklen]

thank you i have been following for a year or so and Dave knows his shit and the microwatt is a cool tool thank you good work go Dave.

[@texjarhead]

Incredible job!! I have lived from the hand-crank magneto phone to this. A good time to be in this business.

[@MomentousGaming]

You know you have watched too many EEVblog when you start getting tailored electronic adverts for scopes!

[@steelthfighter]

+MomentousGaming my best one is watching an EEVblog and getting a vacation ad to Australia

[@3Dcorps]

This is just a thing of beauty..... I love the rainbow sheen on that wafer

[@EEVblog]

Please do, thanks. I'm still after a good video solution for high magnification for stuff like die.

[@johnfoy3912]

Glorious geek pron! Amazing stuff, thanks for the video and to Vincent for sharing!

[@angusbradley797]

That is a great episode! Thank you Dave. Thank you Vincent. BTW, a popular alternative to inkspots is to to note the pass/fail status of each DUT (and serial number) in a datalog file and physically separate devices in a post-processing step. Software is less messy than ink.

[@TheOldmankk]

One of your best mailbag presentation. Just love it. We called that MC hammer Bunny Suit as "Space Suit or Jump Suit", and we did use few of the thrown away for in-house maintenance ex. Wall Painting. LOL.....

[@BOMBOVA]

One of your best show and tell videos ever. clap clap clap

[@JeffWhitlatch]

Wow, cool stuff. Thanks Vincent and Dave!

[@TheMarrt]

I was working in Test-Board-Design (mixed signals PCBs for frontend and backend) for 9 years (was like a 4 guy company). If i remember correctly, the most expensive boards i made for our costumers were about 12000€ in manufacturing. We charged about 30000€ for these (we added pcb assembly, the metal stiffener, rf-relays and sockets). They had 30 layers, needed to be about 6mm thick for mechanical stability because there are thousands of pogo pins pushing onto it from the tester side (agilent 93000). Almost every second layer was a shield layer with a specific depth distance to the signal layers, so the board was 30% copper. The signal wires had a specific width in order to facilitate 50-ohm impedance (were needed). The board was about 40x50cm and had 4mil features in the middle section. It wasn't a very hard job, you had to stick to a ruleset that was verified by 2 other guys when the design was finished, so i did not have to know anything about RF-TEchnology at the beginning.

[@stevecoatesdotnet]

I love Dave's fantastic videography :) Also thumbs up to Vincent :)

[@mortensentim511]

Please do a crowd funding for getting you a microscope. I love these and would happily throw in a small amount of cash if it meant seeing stuff like this a little clearer.

[@arongatt]

mchammer lmaothese shots make you really appreciate more the tech we have

[@Lexicon345]

This guy is so happy. I'm slightly envious.

[@Mythricia1988]

You can get wafers (with chips on them) on eBay - it might cost a little bit, depending on the size and type. But it's definitely on the list of things I want to buy at some point, to put on the wall in my little workshop. They look dazzling when you see them for yourself, and it's great to think about the massively complex engineering required to create that thing that you got hanging on the wall.

[@jodycullimore6637]

Onya Dave and a big thanks to Vincent. Man if I Had this stuff I prolly wouldn't part with it.

[@lsgettier]

Dave... generally the chips are "glued" by forming a gold-chrome eutectic alloy between the chip and the lead frame. Chemical magic!

[@yanfishtwig2356]

Amazing i cranked the resolution wheel all the way up for this vid cheers dave

[@happypoo3201]

did anyone else collect old computer parts when they were little. I had a closet full of computer parts cause I just thought they were fascinating. I grew out of it but now I feel like starting a new collection lol

[@DarkRendition]

YES! I did. I remember breaking open old 486 chips to remove what I thought was gold! I dismantled old HDDs and busted open non-functioning CRT's to explore their internals. I'm 27 now, and I must have been in late elementary/early middle school when I began to really explored the world the computer hardware. That was back in the when dual Voodoo2's in SLI where the very best you get! I even remember the day Nvidia bought 3Dfx... sad day to see the Voodoo line disappear. I'd love to see Nvidia name a GK104, GK110, or even Maxwell part in honor of the Voodoo series! That would set my nostalgia on fire!

[@deanintheg]

DarkRendition Same here, always dissembling any electronics I could get my hands on when I was young

[@Guyskiy]

I actually used to work in a university research lab, and had an electron microscope and atomic force microscope under my watch. We once had a particularly annoying piece of electronics finally crap out on us, and IT didn't pick it up right away, so we tore it all up and examined it before returning it for recycling. Removing that protective epoxy packaging was challenging but eventually we got a piece of a memory module cracked open, and we got to "see" the individual transistors and interlaced arrays. Those images served as my PowerPoint backgrounds for years.

[@killercavalry]

i definitly did! i had tons of old crap that id mess around with

[@aladarhorse]

I have an old circular part kind of like that test board and believe it or not it's for a Pentium one I've had hanging up for ages it's fascinating to look at I've got some really old see-through hard drives to

[@ArmyRadioSales]

Great Video, one of the best yet! Thanks Dave and Vincent.

[@Exciting__Electronics]

Best mail bag by far love looking at silicon wafers ;-)

[@MrGuitarRadio]

I used to work for Tempress Microelectronics, a company that manufactures (or used to) bonding machines, diffusions ovens, etc. That was in the '70's. the technology has progressed a lot since then. Always interesting to watch though.

[@gexwing]

The round shape is a byproduct of the production of the silicon, you start by growing a Cylinder of (almost) perfectly aligned pure silicon molten silicon while spinning it. then you saw the cylinder into the slices which are then prepared as raw wavers. Very intricate process in itself.

[@KallePihlajasaari]

At 20:25 the long photo mask could be for a row or column driver for matrix display or similar. They are often long skinny chips directly bonded to the flex PCBs that are used to conenct to displays.

[@KallePihlajasaari]

At 26:36 the suggestion is for testing post encapsulation but this is not always practical. Sometimes the testing requires access to pads that are not bonded out (like on the bond out version of the processor) to verify internal circuits are working that are not possible to test when in normal operation. Such as watchdog circuits and copy protection fuses that have to be temporarily bypassed or op-codes have to be injected to check fault conditions. Testing may also sometimes take place before the chips are completed and if a complete wafer were to show a fault it would be scrapped before wasting time of the following steps, a whole tray of wafers may be junked with a lot of saved time. Some chips are also factory laser serialised or trimmed and this also has to take place before the encapsulation process. Many modern serialisation procedures are done with OTP ROM that is then locked from changes but sometimes that is not possible or desirable to prevent generation of multiple copies of the same serial number after they leave the factory.

[@Kalmte55]

I really love this video upload. Cheers for your enthiousm (spelling ..) and your sharp zooming. Awesome channel! Greetz from Netherlands.

[@snaprollinpitts]

Hey Dave, I just wanted to give Vincent a big shout out and thanks for this look at the tech that we use every day. I will never be able to pick up a chip and not remember the great deal of man hours and brain power that created this technology!!! Thanks again to Vincent for the great view into this world!!! mike

[@Psychlist1972]

I've put off a synth project because I was whining about the panel wiring. Now that I've seen the hand wiring on that round probe device, I'll never complain again (well, at least not for a day or two)

[@FlakeSE]

This video deserves a part 2 with the mantis and going into a bit more detail on what is going on on the dies knowing what you know now.

[@TheStowAway594]

Very interesting stuff! thank you for sharing

[@GodzillaGoesGaga]

What a wonderful mailbag.

[@EEVblog]

Doh! yeah, didn't think of that. Looked for all the world like it was see-through.

[@SaderStel]

Really cool video, I remember finding wafers near the closed chip manufacturer in my town, that was quite big in USSR times. Much smaller in size wafers, a bit bigger than mini-disk and also dark red glass for laser printing chips on the wafers. Sadly company was closed and most of the stuff was just thrown away or just "taken" by those who worked there, manufacturer was called "Nuklonas", my eldest brother bought from them our first computer that was "ZX-Specturum" or some clone of it.

[@VikasVJois]

Nice one Dave. 40 mins of awesomeness

[@mattstelmach1982]

Dave this is one of the best videos. Big thumbs up.

[@jpbartsch3]

The 'damage' on the edge of the wafers is actually a very important feature. Its a blank space for the masks to be aligned one upon the other. They usually also are on the edge of the wafer that corresponds to the crystal plain of the silicon itself.

[@supernova86]

You're more amazed by the saw cutting those wafers than what is on the wafers and what they can do?!?

[@Turboy65]

At 33:15, you talk about "a chemical process gone horribly wrong" but I believe that this spot was deliberately etched so that the depth and quality of the various deposition layers could be analyzed.

[@spyder000069]

Wow. I could watch the mc hammer clean suit dance for another 60 seconds. That was great!

[@freeman2399]

The 100nm tip one is less than the wavelength of visible light, so you probably can't see it under any magnification.

[@An0nymousMessages]

That was one of the most comprehensive deep dives I've seen. Very well done. Especially like the narrative on the magnifications and how you were achieving them.

[@anarekist]

That was really awesome. Thank you and Vincent

[@doose911]

I worked on PCBs that thick at Behlman Electronics. A lot of prototype stuff for large power supplies for the military. Drilling and milling G-10 was a real pain. So was the clean up. : )

[@rib180]

Imagine the size and purity and cost of the grown crystal slab that the wafers were sliced from. I have some wasted wafers on my roof generating electricity. Great video and story. Thanks EEVblog for sharing.