The MCU process has continued to develop, from 90nm to 40nm, and now adapts to high -performance applications such as core cores to 28nm and even more advanced nodes.However, the process upgrade of EFLASH embedded EFLASH is more difficult than logical circuit.storageThe circuit of the instrument requires an additional light hood layer. After 40nm, the miniature difficulty is greatly improved, and it no longer has cost -effective advantages.In contrast, many types of new ENVM (embedded non -easy storage) technology are favored by the original factory because of its more advantage in performance, power consumption, and logical circuits.
Under the continuous promotion of the main original factory, several new ENVM technology has gradually developed and mature and moved towards applications, such as EPCM, Emram and other solutions. MCU adaptation of advanced process.The products based on these ENVM schemes will better serve intelligence to intelligencecar, End side AI,Internet of ThingsFields.
Deedsemiconductor(ST.
STThe current products are widely used by EFLASH as embeddedstorageDevice, such as the most widely usedSTM32 series.On the 28nm FD-SOI platform, ST applies the EPCM solution, and the specific product is Stellar P and Stellar G seriescarMCU.
EPCM (embedded phase changestorageDevice)Use 锗锑碲 (G (GST) The change between alloys between amorphous and crystal statestorageData, which has the advantages of high writing speed, low power consumption and high durability. It has a significant advantage of reading and writing speed under low voltage than EFLASH and other ENVMs, and the manufacturing cost of 28nm nodes is more than EFLASH than EFLASH Low.
Texas Instrument (TI)
TI's MCU products mainly use Eflash and EEPROM as embeddedstorageEFLASH is used to store program code and data, and EEPROM is used to store a small number of parameters that need to be blocked frequently.In addition, the TI's MSP430 series has ultra -low power products embedded in FRAM memory.
Fram (iron electricstorageDevice)Use the unique physical characteristics of iron electrical materials to achieve datastorageIt has a faster reading and writing speed than EEPROM and Flash, and has non -loss characteristics.Because the current does not need to refresh the current when reading, the power consumption is lower.
British Fei Ling(Infineon)
British Fei LingExisting MCU products mainly use EFLASHstorageplan.In advanced products, Ying Fei Ling andTSMCCooperation to apply RRAM technology to its next -generation Aurix MCU.
RRAM (RRAMresistanceFormula random accessstorageDevice)As the simplest structurestorageTechnology, by changing the electrical mediumresistanceCome to work.RRAM can have both DRAM's reading and writing speed and SSD non -easy loss, so it has the characteristics of high reading speed, high durability, and a single unit that can store multiple digits of data.And due to structural characteristics, RRAM's power consumption is much lower than flash memory.The introduction of RRAM in the 28nm node is conducive to the size reduction of MCU products, reduced power consumption, and performance improvement.
NXP (NXP)
NXP's MCU generally uses EFLASH as an embeddedstorageInstruments, including the most widely used Kinetis and LPC series.In advanced products, NXP andTSMCCooperating the development of the embedded MRAM IP of the 16nm FinFET process, used in S32Z and S32E processors, facingcarapplication.
MRAM (magnetic resistance random accessstorageDevice)Have non -prone loss, the number of reads and writing is almost unlimited, the writing speed is extremely fast, the power consumption is low, and the logicchipHigher technical characteristics of integration.Emram can be seamlessly integrated with advanced process processes such as 16nm FinFet, which is the characteristic that EFLASH cannot reach.Therefore, NXP's high -end vehicle MCU is also very suitablestorageplan.
Risa(RENESAS)
RisaThe most widely used MCU products are EFLASHstorageplan.In terms of advanced technology, I also chose the MRAM route.RisaThe MRAM opportunity is independently developed, and the category belongs toSTT-MRAM (self-spin injection MRAM)Essence
RisaTests based on 16nm and 22nm node logic circuits have shown that MRAM can bring stronger read and write speed and energy efficiency performance, thereby enhancing the overall performance of MCU.Different from other originals, the advanced ENVM scheme is used for the core of the car, Risa integratedSTThe MCU of T-MRAM technology will be mainly appliedInternet of ThingsIn the field, the possibility of the core is also used in the future.
In summary, the node after 40nm, the embedded type of the MCUstorageThe plan also needs to be updated.The technical routes of major original plants in this regard are different, but they all strive to achieve higher storage density, faster reading and writing speed and lower power consumption to cater to intelligencecar, End side AI and other market needs.
Various ENVM schemes often have both the speed of RAM and the non -easy loss of Flash, thus laying a solid foundation for the technological innovation of the MCU.After these new solutions have matured and stable, the product form of the MCU will also be renewed.