How Japan Learned Semiconductors 

In the early 1980's the Japanese were so dominant in semiconductors that my father had to take Japanese language courses at his employer, VLS Technology, Inc. just so that he could collaborate with his Japanese counterparts.

I worked in the DRAM industry for many years. It is pronounced dee-ram by everyone in the industry. I like your content, keep up the good work!

Those guys worked hard. Out of school in 1984, I spent a few months in Texas Instrument's Hatogaya fab (northern Tokyo, their first Fab, started in 1970, Miho in the 1980s). First VLSI CMOS ("CSAG"), 4.5 micron on a 3" wafer. We'd come to work at 8:00am and some would leave at 6:00PM, but Jibiki-san (who I worked with closely) would work every day until 10:00pm, when we'd finally go for dinner (usually Dennys, being one of the few places still open) and in bed at midnight, to wake up at 7:00am the next day and go again. 90 hour work weeks were common. The ethos in Japan (and I heard it repeated many times), the Nation came first, the company second, and personal ambition could be priority #3. And it worked for the country, it built the companies, and indeed it did build the people. Deming was a national hero to Japan. The focus on teamwork and refusing to let the ball drop as a team meant that quality was religion. Quality is what built Japan, and Quality came from hard, hard work.

As someone who grew up in Japan, hearing the rice cake saying was just amazing. The lack of research is what makes 99% of youtube channels trash - you are in the 1% my good sir. Thanks for the amazing video essays as always, love to hear about the industry as both my parents work at tech companies in Japan 🇯🇵

As an electronics hobbyist & chip collector, Japanese manufactured parts are well sought after for their quality, there are still NUMEROUS new old stock Japanese made chips manufactured in the 1970’s through the 1990’s, the 1980’s & early 1990’s being Japan’s chip manufacturing heyday.

Anecdotal: in 1980 I was the lead engineer on a project that required an early microcontroller. In those days there was no flash so the code eas “masked” in ROM - that is, defined as part of the top layer of the semiconductor chip at time of manufacture. It was an 8049 MCU which was designed and originally sold by Intel. But in the end we chose Mitsubishi, who were able to provide the mask at 1/3 the set-up cost and deliver the parts in half the time at a much lower price. I don’t recall a single MCU failure in the two years we built the device, over which time thousands were sold.

I've never heard DRAM pronounced as "DRAM" before. Always as "D-RAM". Probably to denote the difference between DRAM and SRAM, or Static RAM,

Having worked in the US electronics/computer_chip industry for 40 years (retired 2013) , I find these walks down 'memory lane' quite enjoyable, and I definitely did learn some things. Thanks!

I very much enjoy your work as someone who, from the American side, was involved in Pacific rim technology, specifically Semiconductors, for 35 years until I retired. Minor complaint: in all my work (including working on the JEDEC memory standardization committees) I have never heard it pronounced “dram”. The general classification is random access memories (RAM) with different sub areas such dynamic-ram (pronounced in two syllables dee-ram), static - ram (pronounced s-ram into syllables) And other varieties that are less well-known and substantially less successful. Maybe this is an American - centric pronunciation, however in my discussions with representatives from Toshiba, NEC, Hitachi, Samsung, TSMC and others it was universally pronounced in two syllables Hal Hanson Senior principal semiconductor design engineer (ret)

"Japan's reign as king would not last for long". And such a statement may cause some to feel a sense of despair, or 'what's the point in even trying'. But the back room benefits are still huge: The knowledge attained, the growth in educational resources, the basic technologies and machinery to start a new adventure. That is perhaps the real reward.

At 15:10 DRAM Dominance As a hardware designer (starting in 1972) I specified only Japanese 64K DRAM for memory when it became available because no board manufacturer wanted to set up the re-work post assembly streams needed if American DRAM parts were used.

I worked for TI in a research group. I remember in 1977 we were have serious discussions as to whether semiconductor memory would ever replace core. The concern was that core memory would retain data, even during a power outage. TI's was putting it's hope in bubble memory. I think the incredible growth of winchester technology came as a surprise.

Your research is very well done. Sure - there is probably an inaccuracy here and there. We all make mistakes. But that doesn’t take away from the general high quality of both your research and your presentation. You know what you are doing. By the way: I like your fine and creative humor on your slides.

"Bro chill" that one cracked me up. Good video as always my guy

This series of mini documentaries are só good, I wish I had found this series much earlier. It’s good enough to be part of an academic course (history of the electronics industry). My father did a strategic study of Japan’s electronics industry for Philips in 1985- concluding that Philips should try to merge with Sony. Philips wasn’t ready for such profound conclusion.. 👍😊. Thanks for producing this great content!!

I still remember taking apart my first transistor radio when I was a kid back in the 70's to see what these magical transistor everyone was talking about looked like.

Excellent video. I find all the content on this channel interesting and informative.

17:31 Nice reference! I believe the original quote is "Nobunaga pounded the rice, Hideyoshi baked the cake and Tokugawa ate it" in reference to the final winner of the Sengoku Wars. Although, the industry is so fast moving who knows what the outcome will be in 10 to 20 years?

One minor detail missing. In 80s in the midst of Cold War, Japanese supplied top of the range semiconductors (military grade) to Soviets. USA found out about such arrangements and slammed the door shut for the Japanese imports and export of lithography machines to Japan. Japanese developed their own lithography machines (which are still fairly reasonable even in today’s standard). The problem was two folds, i) Japanese were largely shut out from the world’s largest market hence lacked the momentum in sales - R&D cycle. ii) Plaza agreement drove Yen more expensive making Japanese products expensive in cut throat chip market. From 256k Japanese were en passed by Koreans and never took the lead back. They focused more on the materials and machines supply where they still maintained competitive edge not because of technology but because they were relatively reliable and cheap. . In 2019 Japanese blocked export to Korea and Koreans started to start to drive for domestic manufacturing of not inly the 3 products controlled by Japanese gov but all supplies. All Korean manufacturers demanded their suppliers comply with ‘Japan free’ supply chain. Dupon and other top names quickly joined the frenzy and Korean producers quickly replaced Japanese in an once in a life time oppotunity. As many suppliers work in partnership with chip makers from the development stage, it is notoriously difficult to make into the supplier club. Some Japanese producers moved their production to Korea, outside Japanese gov export control. Others followed suit and created R&D Centres in Korea to work with the next gen development, making Korea de facto the hub for global materials, equipment and final product hub. The talent pool that will be created is only matched by the USA. Even ASLM is building its R&D hub in Korea. Japanese gov just gave Koreans the best gift in a generation to consolidate their dominance in the sector. The end result is, far less Japanese producers are working with Korean on the next gen product development and the opportunity to supply in the future. It would not be surprising to see Japanese material and equipment supply vanishing in next 5-10 years. And total annihilation of Japanese market share and talent pool for any future opportunities. It is a classical case study of bad gov decisions (supply to Soviets, Plaza agreement, export control for Korea) demolishing once flourished industries.

Anybody interested in the subject should check out a documentary series from the 90's called "A videohistory of Japan's electronic industry" If you look for the channel RC286 you can find it here in RU-vid. It includes interviews from some of the people who were directly involved in this history, and was made by Japan's Ministry of Foreign Affairs. I think it was originally on four VHS tapes, so it's a couple of hours all together. It has some objectively catchy episode titles like Circuits in Stone and The Calculator Wars. The fact that Japan could just arbitrarily reorganize what separate companies were doing, and tell them what to work on is interesting. The US economy just doesn't work that way. Back in the 80's, pretty much all Cyberpunk sci fi predicted a huge Japanese influence in 21st century because so many American companies assumed that the Japanese conglomerate alliances would be able to outcompete any individual American company. Go figure, that turned out very differently from what a lot of people expected.

Dude, your knowledge and research into these videos is astounding. I love your videos and even the sense of humor you have with some of the subjects. Keep up the good work.

1985 seemed to be peak Japan period. Can you do a video about the Plaza Accords? There seems to be some discussion in Asian media about whether Japan's problems came about as a result of the Plaza Accords, or as a result of the Japanese government's monetary policy in response to the Plaza Accords.

Something to add to the Bell patent on the transistor. The anti-trust lawsuit that Bell was involved with the federal government eventually ruled that Bell labs Western electric AT&t Bell telephone could not manufacture and sell to any entity except themselves. It also prohibited them from operating outside the telephony industry. However they could license patents and collect royalties. Which prohibited them from getting into the computer industry or other various industries. After the lawsuit until 1984 when the giant was broken up in another Anti-Trust lawsuit. All equipment that was manufactured and used by the Bell group had stamped on it the property of Bell telephone not for sale. In both law cases due to Bell telephone being a massive company and wanting to protect their business interest these were actually negotiations with the federal government in which Bell telephone and it's arms agreed on certain conditions to end the lawsuits. Negotiations that were finally agreed upon in approximately 1982 and 2 years later would lead to the breakup of the Bell system. Removed the previous negotiations on the first antitrust lawsuit which was filed in the early 1930s.

興味深い内容でした。 日本人として知っておきたい歴史ですね。

Note that Bell Labs had both the patents on the point contact transistor (which is really not possible to mass produce) but also by 1954 had the patents on the diffused-junction transistor which WAS possible to mass-produce. Sony got interested in the transistor but by the time they actually licensed the technology what they were licensing was far more advanced than what they had originally been interested in. They were the right people at exactly the right time.

A radio receiver in your pocket, imagine that! A cassette player in your pocket, imagine that! A video game console in your pocket, imagine that! A phone in your pocket, imagine that! A solid state music player in your pocket in with all of your music collection, imagine that! ...Now you're entire digital life in your pocket with a super computer, Imagine that! :D

You learn something new every day. I never heard about "Dram" before ;)

I've noticed that people who get really hung up on small, inconsequential details like VLSI tend to have severe emotional regulation issues. You're fine. Hope that guy gets some therapy. Maybe he can make a living doing research for RU-vid channels.

I recall the scare of Japan's 5th generation computer in the early 80s. It had the US fearing domination from Japan. As Japan had the technological capacity at that point, it was deemed dead serious they would be introducing systems that made existing big-time computing systems irrelevant and ultimately folding the big US firms. In reality, it was just one of the biggest forms of fear mongering as it never came to be.

I really appreciate how you clearly are passionate about this stuff and you share you knowledge in such a well put togewther manner. :)

"to troll" is not what you would use to describe the things this channel is good at, yet I notice the trolling here and there, every once in a while. I really like it and your lack of research on the subject of Dram is peak Asianometry!

0:10 I read Kyoto Station(in Japanese), wondered where in Kyoto this is... Only to have it click that it is a past photo. Having spent hours waiting for the 59 Bus to my apartment right in the middle of the photo. Wow.

Most informative and interesting as always! 😃

I studied electronics back in the 1970s in high school. My teacher, a great man who recently passed away at 102 years old, told a variation of the story you told in the beginning of the presentation. In his version Bell Labs was demonstrating novel transistor technology at a trade fair and Sony (presumably a party including Akio Morita and Masaru Ibuka) was present. Bell had lamented (or as my teacher said were frustrated) at its limited potential over traditional, non solid state components. He said that Sony immediately saw the potential, and jumped on the capability and applied it to consume electronics. The Teacher's lesson for us was to take what we have and understand how to solve problems with that understanding. As a kid, transistors and ICs were the rage, and companies like Motorola led the charge (remember HEP?). But in assembly and manufacture, the sun was setting in America and (no pun intended) rising in the East (東方日出).

No. NTT in not a spin off of ETL. Do not confuse the MITI with the Ministry of Communications. ETL(電総研, part of the AIST) and former NTT-CRL(武蔵野通研、横須賀通研、物性科学基礎研究所 etc.) have indipendent origin.

Thanks for the history lesson. I'm a US babyboomer, my 1961 8-th grade graduation present was a Motorola AM transistor radio. Later in the 1990s I worked for the NEC technology PC group. Unfortunately that business was short lived due to hyper competitive PC (IBM compatible) market.

wow, this was fantastic.

Your videos are like the "Rocky" movies but with the great exception of the analysis of why Rocky comes up on top!

I used read some of this stuffs in different books, websites(like wikipedia but I have to verify the refence) and magazines. Gladly, we can now learn this informations on your channel, research made easy ☺️

Interesting.. AT&T and the Japanese didn't stay in the semiconductor for different reasons - one because of antitrust, the other business competitiveness. If they could've foreseen how much growth the industry would have and will have, would they have remained? Currently, the USA is trying to hold on, Korean and Taiwan hold leads and China is trying to get in. What role should government play? Time and again, we've seen new technologies invented in the US, but made it big overseas.

Keep up the high quality job.

Great video! Possible to also do a video on Canon and nikon Lithography machines vs those by ASML? Tia

DRAM Pronounced DEE-RAM. SRAM Pronounced ESS-RAM. (I have a degree in Electronics Engineering.)

Japan getting ahold of the transistor was a huge leap for the whole world. While Bell Labs deserves credit for developing it, others made it a true success.

In '67 the Raytheon fabs were not always in clean rooms. I was working on TTL for the FAA inter-route computer - to keep track of flights between airports. The fab was right across the room in a glassed in enclosure.

i have to mention ex-Intel CEO Craig Barrett (also Bob Noyce co-founded Intel w/ Gordon Moore); who went to Japan to learn from the Japanese how they were able to do HVM (high-volume manufacturing). Ironically the Japanese earlier learned from the Americans i.e. Deming

Nitpick: I've never heard DRAM pronounced DRAAM, it's always DEERAM. Perhaps certain areas say DRAAM, but I've been in the industry for over 2 decades and have never heard it said that way.

You teach me so much Asianometry. Worth paying for!!

I'll connect the dots on why they gave it to Japan. It was about building a capitalist power base in Asia. South Korea was later added. It was a military decision to give Japan keys to the technological kingdom.

IC bonding was generally not to a board. It was to a lead frame (pronounced "leed" ) i.e. a chip package. 14 and 16 pin DIPs (Dual Inline Packages) were quite popular. Surface mount was starting to show up, but the parts were too small to handle. These days I'm fond of SOT23s (vs TOs) and the TSSOP (vs DIP) form factor. But I have to wear a magnifying glass to build anything.

Was the VLSI project, where the etching technology was invented that allowed the buyer capacitors, or did that come later. I remember see in some corrections of nearly vertical wells more than 10 times the depth of the diameter during my university courses. I remember mention of a Japanese project of endless etching experiments which ultimately led to the capability to etch "any" 3D shape into Si substrate.

I enjoy how you pronounce "DRAM". It's very dramatic :)

Striving after war, climbing up from zero's recipe: Horizontal growth spices. East Asian countries in general booksmart enough to do horizontal step, but bad at vertical one, meanwhile the west good at vertical but bad at horizontal

Bell Labs was the most important US business in the mid 20th century.

One thing he got wrong was the Regency TR1. Texas Instruments made the silicon transistors for I.D.E.A. which made the Regency TR1 radio. Thanks.

The way I see is that the fact that 5 different companies could work together, "helping each other" made Japan achieve that. It would never be possible with US companies.

interesting sidenote. I live next to a holiday hotel owned and exclusively for the Tokyo electron company (TEL), an IC production equipment manufacturer.

Great episode per usual. I would love to hear how you do your research for an episode.

12:26 - Good old Kimwipes. Still use them every time I deflux a PCB.

Fascinating story. Thanks for the hard work and research.

Great video. Side note: I’ve never heard anyone pronounce DRAM any way but Dee-RAM.

Hate to burst your balloon but the Germans were playing with FET's in the 1920's/ 30's.

Technology transfers led to a technology arms race but the whole reason why it benefited Western countries is due to everyone using the same core standards based on Western designs. And I think the Americans knew that having a technology arms race would help accelerate the evolution of the technology by having so many governments and industries working to push the state of the art forward. However, it is insane that while the USA was determined to break up the monopoly in AT&T they had no problem supporting monopolies and public private partnerships overseas which were a no no in American markets. Because in reality some things such as technology R&D work better when you have economies of scale where a monopoly allows for large sums of money to be spent on future innovations. And that was the benefit of AT&T being a monopoly in the telephone industry (keep in mind they developed the technology of the telephone). This is also why TSMC as a near monopoly on chip manufacturing also works as well, as economies of scale allows for absorbing the cost of manufacturing and design while also promoting better quality control and integration.

Thanks for letting us know that wasn't actually footage. Could a fooled me.

Another ShOcK & AWE learning moment. One correction however, VLSI actually stands for: Viable Link Section IV Yeah! That's right baby. How about that for a breakthrough ;) Thank's so much. Cheers!

Actually, I think there were multiple ideas of what VLSI meant. In the logic business we had SSI, MSI, LSI and VLSI Integrated circuits. I don't believe that there was a clear consensus of whether this meant "X" scale Integration or "X" scale Integrated circuit. It was common to say things like "X"SI Integrated circuit. I started designing stuff in about 1965 and felt it was a fairly loose concept. Mostly marketing BS so who really cared very much. Eventually the powers that be ran out of energy, so this cute naming convention withered away. LOL What made it even worse was that these abbreviations were also eventually used in the names of several different companies. At any rate, I enjoy your stuff very much.

Japanese quality control was legendary. It was invented by Americans but adopted wholesale by the Japanese. It was reshored by studying how Japan did it. I was reading all the books in the "QC" era.

11:54 I remember my father doing business with the Mostek company back in the 1970s.

"Japan was still recovering from its ill-advised participation in WWII..." You win Internet understatement of the day.

Great video as always

I just want to read comments and all there is is some noise about how a word is supposed to be pronounced.

My man you’re on a roll! Incredible work and super grateful 🙏

Wow, thank you so much for your great work! Excellent!

Thanks for this excellent content

In looking at semiconductor theroy, learnt that the germanium atom has an extra electron orbit when compared to the silcon atom. This may explain the lower conduction voltage and the higher reverse leakage current that germanium devices have as the outermost electron orbit is slightly less influenced by the central nucleus in attraction properties. This may be a fairly old idea as to the behavior of the 2 different atoms under thermal and field charge conditions.

I have never heard D-RAM pronounced as "dram" before. I think I need to have a dram of whiskey to calm down.

Can you make a video on NTT. It being a massive company and not having any proper competition.

Thank you for your video! I've found the subtitles you provide crucial to parse the information you present, thank you so much for those. But I've noticed that often text you put like "Winners of the 1956 Physics Nobel for the transitor effect" at 0:32 overlap with the positioning of youtube subtitles. Maybe something to keep in mind when you position stuff? Thanks!

he says DRAM... isn't it "D-RAM" ? ( not calling you out just wondering)

One observation regarding emphasis: The VLSI project outcome successes were primarily process-based, rather than innovations in DRAM and PAL designs. Another way to say this is that, as you mention in this video, the Japanese participants became extremely proficient at manufacturing with low defects, a sort of Japanese Six Sigma success, and the Japanese labor costs were lower. However, design and tool innovations in the late 70s and 80s were still primarily American, but the Americans outsourced their back end to satisfy Wall Street, and the result was lower American yields. So the thousand patents you mention were not necessarily all innovations in technology, they were more frequently patent encirclements and incremental improvements in the process for manufacturing. Another American mistake was allowing Japanese engineers to tour and photograph the facilities at Motorola, IBM, Monolithic Memories, TI, etc., since the Japanese then replicated and perfected everything they saw on the floor. When Japanese companies began to make sales to concerns in the USSR, those American companies started to tighten up security, but the horse had already left the barn.

On the other hand, SONY acronyms stands for Sound of New York?

I'm sure other people have pointed this out, but when talking about RAM, the unit of measurement is a byte, not a bit. There are 8 Bits in a Byte. Bits are used for calculation and transport. Microprocessors work in bits. Ethernet, WiFi, and fiber optics are measured in bits. DRAM is measured in Bytes, kilobytes, megabytes, gigabytes, and terabytes.

During the Japanese 2nd VLSI project, the US government tried to get American manufacturers to cooperate to counter "Japan Inc". They put an Admiral in charge. It failed miserably.

Excellent. Thank you.

Wait, so we have: VLSI the concept VLSI the product group VLSI the Japanese government project VLSI the DARPA project VLSI the American chip company And VLS1, the Brazilian space launch the vehicle. That's one hell of a flexible acronym!

We need to bring "Made in Japan" Back up to the top again, same to Made in the U.S, Made in Indonesia, Made in Taiwan, Just ANYTHING elsewhere than Made in China. If we ever want globalization then countries need to step up their game and stop relying on the Factory of the World Country for help.

“Dram” not “D Ram” is a new one Sick video, thanks

Excellent video as always!

So much D-rama in the comments

Very good. Thank you for all the work, and sharing.

Very helpful. Thankyou :)

Good research. Thanks

Great video as always, thank you for your work. The story, narration, images, jokes are all amazing.

Very interesting. I was hoping to see something about the analog side of the industry, e.g. JRC.

People talk about the transistor like it was some immediate major change in technology, but the fact of the matter is that despite their invention and entry of production in the early 50s, most consumers never saw a single one until the mid 60s and no electronics companies really took them seriously until the early 70s, after the moon landing had finally proved how capable they were. By the time solid-state electronics really started to catch on, the integrated circuit was already in full swing. The first all transistor television, the RCA XL-100 does not even show up until 1971. Also worthy of note, the way we see "Made in China" products now is the way "Made in Japan" products were viewed in America during much of this time. It was typically assumed that the Japanese product was cheaper, of inferior quality or design, and just good enough to get by. Obviously, this view has dramatically reversed in the last 40 years. Excellent video though very well researched, I just think it could have used some context on the state of consumers and consumer products for the era to really put all of this into perspective. I would love to hear the story behind the Korean start ups of the time and their meteoric rise, and the "bounce back" you briefly mentioned in the American semiconductor industry.

DERAM :) i guess the correct pronunciation landed in that 1% of research

The picture at 2:03 is a building being moved, it has nothing to do with WWII bombing damage.

bruh DRAM is pronounced DEE-RAM not dram, which is a glass of whiskey

All these subtle puns... Love your work as always

Good info but one major issue missing. Most of the CPUs and RAM were made overseas with Japan at the lead for some time. The Lithography and other Equipment and processes such as VLSI were created and most advancements were made in the USA and much of it via grants from the US Government. I had to watch over and over as the USA companies on tax payers dime developed new tech and since most CHIPs and RAM were made in Japan soon Japan was leader in Lithgraphy Hardware Software and procedures because We paid to develop and Japan would purchase one or two of the new systems and reverse engineer the newsest lithography tech and mostly they used it to make the CPUs. So i had to watch my and others Jobs get moved to Japan while we paid in many many ways. Mostly the reverse tech. Now China and Tiwan are added to the mix and Korea. So tired of the US government moving our jobs and tech to Japan and paying to develop more for them when we cannot ship much of anything to them due to thire terriff on US products.