Announcement

Collapse
No announcement yet.

Raspberry Pi Announces RP2040 Chips For $1

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • L_A_G
    replied
    Originally posted by PlanetVaster View Post
    $1 is already really cheap, but if they're cheaper than that once they hit "reel-scale", they better make sure they can scale their manufacturing because these things will have trouble staying on the shelf.
    I wouldn't be too sure about that when companies who make embedded systems that use chips like this tend to insist on at least a decade of availability. In the embedded product world you just don't have the yearly new product cycle that you see in consumer products and instead it's not uncommon for the exact same products to be manufactured with very minimal changes for well over a decade. This is after all going to be replacing solutions based on parts with Zilog Z80, MOS 6502 and Motorola 68k-cores and there's solutions using those cores that have been in production for well over a decade.

    Leave a comment:


  • ddriver
    replied
    Originally posted by PlanetVaster View Post
    On the Raspberry Pi blog post they say

    $1 is already really cheap, but if they're cheaper than that once they hit "reel-scale", they better make sure they can scale their manufacturing because these things will have trouble staying on the shelf.
    It will be a good thing if they make some profit on that in the end, as it will at least make it economically viable enough to justify continuous production.

    Leave a comment:


  • PlanetVaster
    replied
    On the Raspberry Pi blog post they say
    The single-unit price of RP2040 is $1, giving you a lot of bang for your (literal) buck. We’re still figuring out what reel-scale pricing will look like in the autumn, but we expect it to be significantly lower than that.
    $1 is already really cheap, but if they're cheaper than that once they hit "reel-scale", they better make sure they can scale their manufacturing because these things will have trouble staying on the shelf.

    Leave a comment:


  • ddriver
    replied
    BTW, turns out this chip is a overclocking beast. From the advertised 133, it seems to be able to clock up to 300 and plus mhz in very reasonable voltage ranges.

    Leave a comment:


  • wizard69
    replied
    Originally posted by ddriver View Post

    Because that's roughly the price for a single core at 48mhz, limited io and a tiny fraction of the memory. I'd put it at a 4-5x better value than whats currently available in that price range.

    That may lead to quite strong demand, I mean yeah, it is overkill for many mcu applications, but at that price, and with that much headroom, the convenience of using it as a "one size fits all" - single codebase, single toolchain, I can easily see this as fitting a good 99% of my use cases.
    It certainly could easily fit into a lot of applications.
    I suspect the cost is subsidized, in which case, they may not be interested into selling billions and billions of it, but are more in for the publicity.
    I highly doubt that. A subsidized processor would open them up to all sorts of challenges due to unfair business practices.

    As for the cost, that depends upon the die size and the foundry used, plus any IP contracts they may have. I really doubt it cost much at all to produce.
    I wonder how much does it actually cost to make the chip.
    Probably between 5 and 25 cents a piece. Lets face it out the door at the factory isn't the end of the charges per chip. There is transportation, storage, marketing and distribution costs which get tacked on. it doesn't take long to get to $1 a chip. There is likely some profit for the PI foundation and the distributors will have to make money also.
    Also, for this type of application, having crypto hardware just doesn't make sense. It has plenty of horsepower and memory to run it in software, since it is not really a high data throughput device by any measure. It has nothing to do with openness, which is not orthogonal to security in the slightest.
    I think the desire for crypto is to protect IP in software. If a person needs this, then the processor simply isn't for them. I personally have no desire for crypto so not bothered. I'd much rather see other features added to a mark 2 model.



    Leave a comment:


  • ddriver
    replied
    Originally posted by wizard69 View Post
    products with life spans in the tens of years
    Decades is a bit unrealistic expectation. You wouldn't design such a product against any specific type of component, you'd have the logic board as a stand alone module, so it can be drop in replaceable with another conformant implementation featuring what's currently available.

    As for the charity part - foundations are free to have commercial activities, the catch is the income can only be used to further the foundation goals, unlike for profit companies, which pay dividends and whatnot.
    Last edited by ddriver; 01 June 2021, 01:46 PM.

    Leave a comment:


  • wizard69
    replied
    This is very interesting but I have to wonder how does the Raspberry Pi foundation maintain its status as a charity?

    On the other hand this is very good news for anybody building open embedded systems. That is if they have a plan to assure long term availability, and good support. The embedded world is a bit different than the system processor world and I can't see a huge buy in from people that produce products with life spans in the tens of years without some comfort level that the processor will not disappear in a year.

    Leave a comment:


  • ddriver
    replied
    Originally posted by Markopolo View Post
    I sure wish they had arm trustzone module or at least an embedded crypto engine and a programable key to use for symmetric encryption/decryption of the external flash. of course that goes against their commitments to education/openness/etc.. so I won’t hold my breath


    I don’t see why it wouldn’t be available at that price. I’m more curious what MOQ is required to hit it and the availability.
    Because that's roughly the price for a single core at 48mhz, limited io and a tiny fraction of the memory. I'd put it at a 4-5x better value than whats currently available in that price range.

    That may lead to quite strong demand, I mean yeah, it is overkill for many mcu applications, but at that price, and with that much headroom, the convenience of using it as a "one size fits all" - single codebase, single toolchain, I can easily see this as fitting a good 99% of my use cases.

    I suspect the cost is subsidized, in which case, they may not be interested into selling billions and billions of it, but are more in for the publicity.

    I wonder how much does it actually cost to make the chip.

    Also, for this type of application, having crypto hardware just doesn't make sense. It has plenty of horsepower and memory to run it in software, since it is not really a high data throughput device by any measure. It has nothing to do with openness, which is not orthogonal to security in the slightest.

    Leave a comment:


  • Markopolo
    replied
    I sure wish they had arm trustzone module or at least an embedded crypto engine and a programable key to use for symmetric encryption/decryption of the external flash. of course that goes against their commitments to education/openness/etc.. so I won’t hold my breath

    Originally posted by ddriver View Post
    LOL, just today I was trying to find the actual chip cost. That's a pretty good price, wonder how realistic it is tho.
    I don’t see why it wouldn’t be available at that price. I’m more curious what MOQ is required to hit it and the availability.

    Leave a comment:


  • ddriver
    replied
    LOL, just today I was trying to find the actual chip cost. That's a pretty good price, wonder how realistic it is tho.

    Leave a comment:

Working...
X