This is their first shot at it.
That's how we learned to walk.
(In this case, the chemicals are different for each node that uses a different process to make the chip, in that they got to relearn how to walk every single time).
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
This is their first shot at it.
That's how we learned to walk.
(In this case, the chemicals are different for each node that uses a different process to make the chip, in that they got to relearn how to walk every single time).
Sometimes customer requirements don't make sense. Usually a company is not so stupid as to purposely introduce a product that doesn't work, they won't even submit such a product for evaluation. The internal design team has a goal that meets known market demands such as SEMI standards. Customers may then have additional requirements to fit their specific process.
Maybe they are referring to a third chip manufacturing client because in their own website the claim that ArF resist passed the validation of two chip manufacturers one for storage and another for logic.
That does bring up an interesting point. It is entirely possible it could be validated and certified by some users and not others since they may be using it for different processes (logic vs memory) and within those processes amongst various nodes. So it could be another situation of where the full commercial scale push depends on where the goalposts are set. They could be good to go, but they may be waiting to extend their reach further before pulling the trigger.
Customer requirements : Make it free
/s
Looks like Moore Thread's MTT S80 is getting more reviews
Looks like they got a lot of work ahead of them to support more games and more modern versions of direct X amongst other things. It's power consumption is still at 110W after getting the 200.2 version upgrade. I mean they are moving fast, but a lot of work/challenges ahead still
Looks like they got a lot of work ahead of them to support more games and more modern versions of direct X amongst other things. It's power consumption is still at 110W after getting the 200.2 version upgrade. I mean they are moving fast, but a lot of work/challenges ahead still
Yes, laser interferometers can actually meet the manufacturing requirements of 14nm, but the performance of planar gratings is better
It seems they used an interferometer in a previous version of the machine, but now they are moving (or have already moved) to a planar grating, that BTW seems very complex to build and requires a clean-room (see the above picture by @tonyget).
I'd suspect they moved to a planar grating, also because advanced interferometer is foreign technology. I read somewhere that China at the moment lacks a local producer with required specs.
The important thing is that they were able to overcome this technical hurdle....if grating is expensive is not a problem. Localize the full semiconductor supply chain is infinitely more expensive, but it has to be done anyhow.
Chinese SME will run at full capacity for many years to come. It's their production capacity that will determine their market share in China, not the foreign competitors. It is already like this.
And this is even more true for SMEE: if they can produce 2 machines (as long as they work), they will deliver 2 machines, if they can produce 20 they will deliver 20.
Previously, he made it sound like the current DUVi that just got delivered cannot do 14nm with multi exposures. I wonder if he is also just finding out new things and updating info as that happens. Either way, i suppose the bigger concern at the moment is for SMEE to just work with partners to get the machines validated and improved.
It's also amusing to me that he went from grating interferometer not ready and may not be doable to planar gratings is almost there in like a month. What changed?
- Status