Russian Electronics: Semiconductor and Processors

Here is an interesting article:

Russian microprocessors. There is a" Plan B". But we continue to smile slyly

Key part here:

But while walking around the Metalworking exhibition, I met a couple of CNC machines that are built on the basis of Russian processors. One of them is based on "Elbrus-8C" and the other on the basis of "Baikal-T". Moreover, the same company is also developing a meringue processor "Skif".

In short, I somehow do not notice a decrease in the intensity of development based on our processors, despite the ban from TSMC. Moreover, new products are also appearing. For example, ICL showed a new server motherboard based on the Baikal-S processor.

I will note, by the way, that not only the processor is Russian here, but also "loose", resistors and ceramic capacitors of Russian production. By the way, this is not the first time I have heard that Russian SMD components have started to be used. Yes, they are still more expensive, but in principle their contribution to the total cost is not so significant, and with the growth of production, the price will fall.

Source: https://zen.yandex.ru/media/sdelanounas.ru/rossiiskie-mikroprocessory-plan-b-est-no-prodoljaem-hitro-ulybatsia-629c802d1b901a4f263b749b?&

So baikal electronics is still getting processors made and still promoting the Baikal S. As well, company's making CNC machines are still obtaining processors of various Russian design. They admit they can't get from TSMC but had a Plan B but won't say what it is, probably to prevent any attempts to sabotage it if it's importing specific tools and materials to make at home or a third party willing to make it.

lancelot wrote:Huawei is under US sanctions in the entity list. They cannot ask TSMC to manufacture their chip designs that used to power their smartphones anymore. In addition any Western electronics they want to order have to be explicitly vetted and approved by the US government on a case by case basis. They are basically being forced to buy chips from Qualcomm and other US companies. To the point the US didn't even let them buy 5G modems, only 4G ones. This when Huawei spent hundreds of millions designing their own chips which they can't manufacture anymore because they were made for fabrication at TSMC. So my guess is Huawei just closed the stores for consumer products but probably maintain support for telecoms equipment. Mind you those Qualcomm chips are also made at TSMC in Taiwan or Samsung in South Korea.

I know Huawei did not close their R&D facilities in Russia. Those continue to operate. This is unlike what Western companies did. Eventually the Chinese will have their own sanctions proof commercial fabrication facilities. But like I said now is not the time.

If Russia wants to get Chinese help with chips, it will have to work with their MIC and companies like CETC.

Huawei had just broke into the mainstream of Canada in around 2015. Their phones were in the top tier with all the top service providers. People were talking about them. The P10 was all the talk just like the Samsung s-7 or 8 was at the time.

Then Trump got elected and basically beat Huawei out of existence in a year or so. They were gone.

My problem with China is how passive their reaction to this was. They basically just took it like a bitch. Huawei restructured the company around the sanctions.

I think this passive reaction by China prevented the chip market from restructuring properly. They should have negotiated something. Instead they just bought a few less tons of soybeans from the US.

There are a bunch of comments here. I will try to elaborate on a few of them.

The Russian government tried to make a demand based, instead of a supply based, development plan for the chip industry. The idea was to design chips on foreign fabs like TSMC for sales to government institutions. Then once demand ramped up they would make factories in Russia to produce these chips. I personally think this was a good plan but they just started on it way too late and progress was glacial. And whatever production facilities Russia has right now, they won't be able to improve much upon them in the near future because of sanctions. However current production facilities should be enough to satisfy the military and industrial control sector for next decade.

Mikron supposedly can manufacture at 65nm. But given they still operate KrF lithography machines this likely involves using multiple exposures. This means you need multiple passes per wafer and using this process will cut down your wafer production rate. It will also increase the amount of bad chips on a wafer since overall defect rate will increase. So the process will be economically non-viable unless you really need it for strategic reasons. There will be a tradeoff between just making a larger chip that takes more space on the wafer at 90nm or making one with smaller transistors at 65nm. Unless you really need the 65nm for power draw or compactness reasons I doubt it will be cost effective. Crocus Nano has more modern ArF lithography machines. These can do up to 55nm at full production rate. But supposedly the Crocus production line was for MRAM with only a couple metallization layers. This is not enough to make a complex logic chip which might have like 8 or 9 layers.

However, like I said, the Chinese basically make all the equipment classes that are used on either of these fabs. Which would be KrF and dry ArF lithography and the other machine tools. The Chinese also make all the materials for KrF but ArF materials are still in experimental production at this time.

China is already supplying fiber for making PCBs to Russian manufacturers which was previously bought in Europe. So it is untrue that China is refusing to supply materials to Russia. Of course you will have to check with the manufacturer in question. Some of the manufacturers of materials in China might be joint ventures with Western companies (typically Japanese or European ones).

There are manufacturers of discrete ICs in Russia. For example Mikron's facilities at Voronezh aka VSP-Mikron. The Voronezh facilities use smaller wafers (100mm and 150mm ones) and likely older equipment. You do not need modern equipment to make discrete ICs.

With regards to the supply of Baikal or Elbrus semiconductors which were made at TSMC my guess is both Russian companies have enough chip stockpiles for continuing sales for quite some time. But eventually these stockpiles will dwindle. In the meantime the industry will need to come up with substitutes.

As for the Chinese reaction, as you might call it, to Huawei ban I also think it was just way too weak. The Chinese market is just so huge, they could have put a major damper on US chip sales had they wanted to, I think this was a major mistake on their part. They could have easily put tariffs on any product with US semiconductors in it. But then again Asians are never reflexive in the first place.

3Logic Group, the owner of the Graviton brand, introduced its own motherboard with the Baikal-M processor. The Mini-ITX board has a very cool design, silkscreen printing all over the textolith with a visual effect of a gravitational field under the processor, looks stylish and creative. The board from Graviton is called Angara and supports 2-channel DDR4 SO-DIMM (2400) RAM, with a total capacity of up to 64 GB, as the board also has three expansion slots, including PCIe Gen3 × 4.

From the interesting, the board has control of opening the case cover (Intrusion) and a gyroscope to determine the position of the board in space, for drives there are two SATA-III ports, USB 3.2 and USB 2.0 ports, a Gigabit Ethernet 1000Base-T controller, a VGA and HDMI 2.0 video output, as well as two SSD M connectors.2. The heart of the board is the Baikal-M (BE-M1000) processor, an 8-core ARM processor with a frequency of 1.5 GHz, manufactured using a 28-nm process technology with 8 Mali-T628 MP8 graphics cores at 700 MHz.

This board is designed for installation in mini PCs, in cases of automated workstations or in monoblocks, but this is not the only model of the board from Graviton, the company also announced the "motherboards" of the server form factor and is preparing to produce its own storage systems and servers.

Source:
https://zen.yandex.ru/media/id/5f20a49cf01f506fcb80c60b/kompaniia-3logicgroup-vladelec-brenda-graviton-predstavila-sobstvennuiu-62abac81ea5792168750f843

Designers of Russian microelectronics gather in a bunch

In Russia, however, they decided to seriously consolidate microelectronics designers, bringing them together in one place and setting them a number of very ambitious tasks. There are not so many microelectronics designers in Russia, and they are all very dispersed in different companies. Therefore, we need a kind of unifying center of attraction for organizing permanent brainstorming.

In this regard, on the basis of National Research Nuclear University "MEPhI" (Moscow Engineering Physics Institute) The full-cycle microelectronics design center "Mephius" was established. The name comes from the abbreviation "MEPhI", written in English — MEPhI (Moscow Engineering Physics Institute).

Participants of the design center, in addition to the university itself, were leading Russian companies in the IT industry, such as MCST, Baikal Electronics, Hi-Tech, Astra Linux Group, EREMEX, KEAZ, "Optimizing technologies", "Kraftway", "Aerodisk" andVedaProekt.

More at source:
https://zen.yandex.ru/media/electromozg/proektirovscikov-rossiiskoi-mikroelektroniki-sobiraiut-v-kuchku-62a60a55d6061f205e633463?&

Construction of a factory for the production of processors using 28-nm technology has begun in Russia

A factory for the production of 28-nanometer processors is being built in Zelenograd. Its construction is provided for by the microelectronics development project of 2020. And it is still difficult to assess what impact the new circumstances may have on equipping the factory with equipment. But they plan to start production in 2024. The importance of the new factory for Russian microelectronics and possible obstacles to its opening will be discussed later.

Another step towards self-sufficiency of the industry

The factory is being built in Zelenograd on the site of two demolished buildings of the Angstrom company, a Russian electronics manufacturer. The territory is located in the special economic zone "Technopolis Moscow", which is also planned to be reconstructed. The developer is the MIET International Science and Technology Center, established in 2020.


Russia is one of the four countries with processor technology. But the production capacity is not enough. The appearance of such a factory is another step towards self-sufficiency of the industry. Among the advantages of starting your own production is simplified logistics and internal document management.


28-nm processors meet all market requirements. The 5 and 4 nm processes that competitors are developing are mostly designed for domestic purposes, fragile and demanding to external factors. And in industry, for example, in the automotive industry, the space industry, technical processes from 35 to 160 nm are used.

Obstacles to opening a factory

The main problem is equipping the factory with equipment. It was supposed to purchase lithographs from the Dutch company ASML. But now this may be difficult. The issue of recruitment of personnel and procurement of raw materials for production also remains open.


But, despite the difficulties, by 2030 the factory plans to start producing processors not only for the domestic market, but also for export. Further development will determine the cost of final products, which, in turn, is highly dependent on production volumes.


It is known that a fundamentally new board manufacturing technology is being developed, which can completely turn the market around. So far, experts cannot say for sure how the current circumstances will affect the purchase of equipment for the factory.

Source:
https://zen.yandex.ru/media/protech/v-rossii-nachalos-stroitelstvo-fabriki-dlia-proizvodstva-processorov-po-tehnologii-28nm-62a37a04c3044952c246300f?&

We already talked about this 28nm factory here.
The ground was cleared for construction. But with the current sanctions there is just no way the equipment for it will be available.
You would be basically be building an empty shell.

The only non sanctioned tools available to Russia right now are Chinese ones. And the Chinese are still stuck at like 55nm lithography.
The Chinese have 28nm lithography tools in advanced development but they are not available for mass production yet.

Keep your eyes peeled on SMIC Beijing fab since that should use all Chinese equipment. Expected to come online in 2024.

kvs wrote:The 28 nm news is all over the map.   First I hear talk about Russian equipment being ready by 2030 and now I see a factory being built.   I suppose that
it is using ASML equipment.   This is consistent with the talk about a move to 28 nm over the last few years.   So the 28 nm project did not start this year.
What would be useful if these journalists did their jobs and reported on the capacity for 28 nm parts being deployed.   I think it is sufficient to cover
Russian needs for the next 8 years.   In other words, Russia must have acquired a substantial amount of manufacturing equipment and did not leave
itself in a situation where it would have to wait a decade to develop its own.

As for the 2030 target for Russian 28 nm lithography equipment, I wonder whether it is true.   I do not believe that the development work on such equipment
started in 2022.  I do not care about the contracts awarded since they could just be a way to organize existing exploratory development work already
conducted.   That is, the enterprises (research institutes, etc.) that have carried out this work will just go through a rubber stamping process to deliver
it.   They get the money and also get to organize their projects into a more coherent form.   Anyone who does development work (software, hardware) knows
that it is not a linear progression along a single path.   There are lots of branches that are not necessarily dead ends.

Another aspect that needs noting is that the ultra-fine lithography ICs are a diminishing returns sort of product.   If you have an NC milling machine,
then you do not need "7 nm" CPUs with a hundreds of cores to drive it.   Much older generation CPUs already deliver the needed functionality.   I am
sure that back in the 1980s this was not true and latest generation fabrication products had a large impact on NC machine performance.   But this has
not been true for 20+ years.   Similarly a lot of military electronics have saturated years ago and do not require the latest generation of ICs.   The only
field where there is such a need is in HPC, graphics and AI.   But AI is not deployed by the US and NATzO in any way that leaves Russia behind.  Maybe
in 20 years things will change, but Russia is not standing still.  

 

Most tasks can be completed by chips on the 90nm lithography.

Aa they say, the Americans managed to land on the moon using a chip weaker than one used on a calculator.

They will slowly work on thee 28nm plant while either waiting on the equipment to become available or they have something we do not know. Seeing as none of us are there, I will continue to watch the news and see what is happening.

Till then, I suppose 90nm processors will suffice for every day tasks for ministries while people just get black market chips from China.

Actually in the 1980s it was said that the electronics in a digital watch were more sophisticated than the electronics that went to the moon with the Americans, but then next year the next big CPU released could be said to be more sophisticated than any other CPU ever made before so it is not as profound as it sounds.

The single chips in wrist watches or calculators could not be used on their own to build up the infrastructure and develop all the technologies to go to the moon... most of it was done on slide rules by mostly women... black women if you believe Hollywood.

Chip making technology is developed like any other technology... Russia now has a need to invest in some of its own technology... and I rather doubt they will bother with all that effort and money to just get what the west has got... they will want to revolutionise it and make it affordable, make it more capable, and make it better... current stuff is reaching its limits... it is clearly time to start going in different directions using new materials and new technologies...

GarryB wrote:
Chip making technology is developed like any other technology... Russia now has a need to invest in some of its own technology... and I rather doubt they will bother with all that effort and money to just get what the west has got... they will want to revolutionise it and make it affordable, make it more capable, and make it better... current stuff is reaching its limits... it is clearly time to start going in different directions using new materials and new technologies...

Chip manufacturing, everywhere it was established, took a long time and was extremely expensive. I don't believe Russia will find a way to change that. Judging by how long China takes to develop its own tech, with much more money invested and better overall situation in the industry. Russia proved capable in the design department. Not so much in manufacture.
Claiming that they will come up with something revolutionary is delusional.

caveat emptor wrote:

GarryB wrote:
Chip making technology is developed like any other technology... Russia now has a need to invest in some of its own technology... and I rather doubt they will bother with all that effort and money to just get what the west has got... they will want to revolutionise it and make it affordable, make it more capable, and make it better... current stuff is reaching its limits... it is clearly time to start going in different directions using new materials and new technologies...

Chip manufacturing, everywhere it was established, took a long time and was extremely expensive. I don't believe Russia will find a way to change that. Judging by how long China takes to develop its own tech, with much more money invested and better overall situation in the industry. Russia proved capable in the design department. Not so much in manufacture.
Claiming that they will come up with something revolutionary is delusional.

That is why I think they will go the route of localizing production of the lithography equipment they got (so a form of reverse engineering) along with assistance from Planar who used to build such equipment during Soviet times. Then they will improve upon that design on their own regarding lower nm. It will take a while. But we shall see. Maybe they will seek to look at graphite and other methods of semiconductor production. Necessity is the mother of invention.

sepheronx wrote:

That is why I think they will go the route of localizing production of the lithography equipment they got (so a form of reverse engineering) along with assistance from Planar who used to build such equipment during Soviet times.  Then they will improve upon that design on their own regarding lower nm.  It will take a while.  But we shall see.  Maybe they will seek to look at graphite and other methods of semiconductor production.  Necessity is the mother of invention.

I agree that they should first develop parts where they previously had success and competencies. Rest of the equipment, in the beginning, will be better to import from China to ensure production and then slowly work on missing pieces.

The first sets of sanctions in 2008 were supposed to destroy the Russian oil industry and therefore also collapse the country because obviously without amazing western oil drilling technology Russia couldn't access its oil and gas deposits and will come crawling to the west.

I would say the fact that they now make their own machine tooling equipment... something we were told they lost and would never get back too, why wouldn't they look at modern chip design and work out how to make it better and easier and cheaper... it might be a matter of trying different materials...

Western producers as you have said have committed to the technologies they have developed and used, but starting again from scratch would they do it the same or would they try different approaches using more modern technologies and materials... I realise you will claim I am clutching at straws suggesting they are so far behind that this is actually an advantage because they can see what the west did and look at its strengths and advantages... but also its dead ends and limits too... could using alternative materials and methods avoid those dead ends, or just make it cheaper and easier to mass produce...

There are lots of new technologies that were not available when the direction the west took with chip making was started...

The Russians are clever and don't just blindly copy... they improve and adapt... no one has described the Su-57 as being an F-22ski AFAIK, but is there a reverse description for US copies of Soviet or Russian kit... the F-15EX seems to be an Su-35 wannabe, but then it was originally an attempted copy of what they thought the MiG-25 was at the time anyway, and the F-22 is just a stealthy F-15... the lack of imagination is depressing.

At least try to write a coherent message. As i pointed out, developing a full process is extremely expensive and time consuming. In some areas Russians will start, practically, from zero. Leveraging your previous successes and importing rest of the equipment will make whole process cheaper and speedier. They can always try to develop rest and close whole technological process, if there's enough money.
I can assure you that Western and other universities and companies are constantly investing and trying to develop new technologies in the industry and not only doubling down on existing.
Your imagination very often gets better of you and you start mumbling nonsense that doesn't have any foundation in reality, hence you sound delusional.

Importing existing shit that the west can't stop you getting your hands on limits them to following their footsteps and ultimately walking the same path to the same destination where they don't have the time or the money to catch up any time soon.

The path was expensive, but Russia does not have any competition because they have been sanctioned so there is less risk for their investments...

Thank you for the therapy advice but it is not wanted.... I appreciate I am not an expert in this sphere, but if you think I need mental help or I am some sort of idiot then I think that tells me more about your ability to accept alternative views than anything regarding this industry or field.

Russia has already invested money and technology into a range of areas and generally they pay off financially too.

This is just another area they will invest money into and I doubt they will want to just blindly follow the western developments achieved so far... if they have to make their own it makes sense to develop it in directions that are useful to them.

They can reflect on the developments and progress and issues the west has had and use that to look in other directions.

There are new technologies being developed in the background that can change what we do and how we do things... perhaps it is a time for a new approach completely... oops... there I go mumbling incoherently again...

80% of domestic demand for car electronics .... pretty ambitious ...
assume he's talking about Lada and Kamaz etc too .... all the Rus brands
what would that be ? around 100-150,000 vehicles p.a. ?

By 2024, Russia will have its own automotive electronics

https://zen.yandex.ru/media/electromozg/k-2024-godu-u-rossii-budet-svoia-avtomobilnaia-elektronika-62acc48687d89902d956e87e?&

Deputy Prime Minister of the Russian Federation Dmitry Chernyshenko, who clarified some of the government's plans in this regard.

Moskvich, Aurus, and other cars produced in our country will consistently switch to Russian electronics.
Russia already owns the necessary technologies and processes for the production of chips and other electronic components.
It so happens that we know how to build these plants, and now we are expanding this production.

By 2024, we plan to meet 80% of domestic demand in a topology of 180 nanometers.

Thus, about car chips, you can be relatively calm, they will do. Now for the component base:

Major critical electronic automotive components are already being developed and significant funding has been earmarked for these purposes.
The infrastructure is being created to develop a unified electronic component base for the automotive industry, these are factories, design centers, materials (pure gases, silicon), photolithographs, and so on.
Almost all the equipment for replacing the foreign one, from which we have now been cut off, we already do it ourselves.

in another Yandex post

this Electrobrain guy has apparently driven past the site of the new factories .... to eyeball it himself

https://zen.yandex.ru/media/electromozg/novosti-o-stroiasceisia-fabrike-chipov-tehprocessa-28-nm-v-zelenograde-62ac7e901061aa2c9f02018f?&

News about the factory of chips of 28 nm process technology under construction in Zelenograd

And now a few photos from the construction site. View from Georgievsky Prospekt, there are three cranes:
Inside, work is in full swing, concrete mixers and reinforcement trucks are running, cranes are working:
etc

Gazprom was building a satellite factory in Moscow too
they will be looking for a lot of chips by 2024 as well

Gazputin wrote:
80% of domestic demand for car electronics .... pretty ambitious ...
assume he's talking about Lada and Kamaz etc too .... all the Rus brands
what would that be ? around 100-150,000 vehicles p.a. ?

well cars' drives and control computers let me guess can be applied in drones/robots as well...this should  increase effect of scale in production of microelectronics, Just a question what with investments in memory chips production capabilities?

maybe Belarus is going to be a big player in this ?

"During the USSR times Belarus semiconductor industry was taking care of the mass manufacturing of main-stream semiconductor elements, ICs and computer chips.
While the most sophisticated chips were manufactured in Zelenograd (in suburban Moscow) Belarusian enterprises were mostly oriented on consumer electronics and mid-level ICs.

The main Belarusian semiconductor consortium - Research and Production Corporation "Integral" - is in fact the largest in Eastern Europe semiconductor company.

This main belarusian semiconductor consortium "Integral" in fact consists of 6 plants and 3 design centers providing complete cycle of production of electronic components. The headquarters of "Integral" are located in Dziarzhinski Plant. "

are they creating a copy of the Integral Dziarzhinski setup at Zelenograd ?
the easiest way to quickly expand production is to clone an existing facility ... that is what I would do

interesting summary of Belarus electronics industry ...

http://www.belarusguide.com/industry1/semiconductor.htm

GarryB wrote:The first sets of sanctions in 2008 were supposed to destroy the Russian oil industry and therefore also collapse the country because obviously without amazing western oil drilling technology Russia couldn't access its oil and gas deposits and will come crawling to the west.

I would say the fact that they now make their own machine tooling equipment... something we were told they lost and would never get back too, why wouldn't they look at modern chip design and work out how to make it better and easier and cheaper... it might be a matter of trying different materials...

...And when they lost access to the old supply chain of (USSR-developed, BTW) marine and helicopter turbines from nowadays Ukraine, that was the death knell.

Until they got their own manufacturing up and running. Took time, was expensive, but still.

Same with power turbines, much ado was made about Siemens being tricked into supplying these to thermoelectric plants in Crimea (which were under sanctions), sure. Not because the tech is alien to subhuman Russians, but because of existing supply chains and whatever carry a shitload of inertia and aren't easily replaced on the fly.

I mentioned a silly example few months ago, regarding one of the main bottlenecks for Russian light-medium UCAV producrion. Namely a quite simple aviation piston engine in the 100hp class. Nobody was manufacturing that particular type of engine domestically. It was always convenient and logical to go for existing off-the-shelf products, in that case a Japanese one IIRC.

Obviously Russia can build aviation piston engines, and are doing so, just not in that particular niche. So that became an issue.

It's funny how these things work.

Re semiconductors, they're no slouches there but it's a highly underdeveloped industry in Russia when it comes to manufacture, especially considering to how good they are at development (on paper) and so on. They won't be able to compete with the market leaders in decades, if ever. But they do have the know-how and capability to meet their most pressing needs in the near future, and the market situation will change again quite soon I'm sure.

Just recently I read an article (unfortunately, the link has already been lost), where one of the representatives of Russian microelectronics plants said that one of the ways to solve the shortage could be the development of optimized circuits for older technical processes: in some cases, well-optimized chips executed on 250+ nanometer technology work just as well as those which are executed on 65 or 90 nanometer technology. Therefore, one of the solutions is the development of optimized chips based on older technological processes that have already been mastered on fully Russian equipment. Thus, more money and industrial capacity can be allocated for more responsible systems.

Scorpius wrote:Just recently I read an article (unfortunately, the link has already been lost), where one of the representatives of Russian microelectronics plants said that one of the ways to solve the shortage could be the development of optimized circuits for older technical processes: in some cases, well-optimized chips executed on 250+ nanometer technology work just as well as those which are executed on 65 or 90 nanometer technology. Therefore, one of the solutions is the development of optimized chips based on older technological processes that have already been mastered on fully Russian equipment. Thus, more money and industrial capacity can be allocated for more responsible systems.

For most military and industrial needs you can make do with rather "obsolete" technologies.

Not gonna run UE5-based video games in 8k resolution with 60fps, but that's a different story.

If I am not mistaken, things like the F-22 Raptor run on Intel 286 tech (which is essentially from 1982) and was designed using equipment running it as well. It is still considered the premier air dominance fighter in the world, right.

But I am out of my depth here so I won't speculate further. All I know is that what people consider cutting-edge military tech (and space tech etc) is more often than not based on hardware that is proper stone age compared to your average home computer or mobile phone today.

And that has many reasons. Well-proven tech goes above all, resilience goes above all, it takes a long time to develop those things for that purpose, so what they start with is usually what they also end up with 10-15-20 years later. The specialization for each component (only has to do exactly what it has to do and nothing else, in fact it's better the more one-trick-pony it is) means you don't need to overcomplicate things etc.

You are correct.

The thing is, with let's say x86, the more transistors you can fit onto the chip can potentially increase its performance. Maybe fit more instructions. Efficiency can indeed be done on much older topology mostly due to risc design. As KVS pointed out, at that point the task is up to the compiler rather than instructions on the chip. Heck, older risc chips like the cell processor used in a PS3 is a beast of a chip for its time and would be effective today if it wasn't such a bitch to program for.