Announcement

Collapse
No announcement yet.

CooliPi 4B Is The Best Raspberry Pi 4 Heatsink We Have Tested Yet

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

  • #51
    Originally posted by Paradigm Shifter View Post
    Yes, that's correct. It's one of the reasons I ignore MTBF numbers, they're just too long to be meaningful. If something is going to die, in my experience it is usually, a) dead on arrival, b) dead in the first two weeks, c) dead just after warranty expires or d) never dies and is still kickin' 28 years later (an old 850MB Maxtor drive...)
    From a manufacturer's point of view, it's just the opposite. The MTBF mean mean time between failures and it's a statistical figure, made out of individual measures. But it has presumably poisson (normal) distribution - some units die sooner, some later. I dare to say expected lifespan, because that's more understandable to uneducated end users. But it's statistical. So, if I want to be comfortably sure, that some equipment lasts say 10 years, I need to overengineer a part accordingly. Because after 17 years, half of the units would be dead. Perhaps all, statistically. That's about what I feel MTBF means in this statistical case - correct me if I'm wrong.

    I can be reasonably sure, that some 10 years with lower voltage or lower spinning rate is archievable with most units. And if controlled via PWM, it can be even those 17 years, if it's bearings wouldn't dry out.

    Go apply this to a Pimoroni Fan shim. How many units will be dead by then ? What is their MTBF? how long does a fan run with rattling sounds before it finally dies? But we'll see over time. How much dustup will be there, needing service, lowering efficiency? How about smoker's fumes sticking the dust together with SOCs, shorting their I/O when being moisted? I've been cleaning some PSUs and notebooks of smokers. Not possible.

    For me, MTBF is a measure how much is a manufacturer certain that a product is worth of, bar noname manufacturers which evaporate after sale.
    Last edited by CooliPi; 19 December 2019, 06:12 AM.

    Comment


    • #52
      Originally posted by stiiixy View Post
      Extremely high temps 42 degrees?

      heh.
      I feel so, literally :-)

      But given some decent airflow, some 60-ish degrees are pretty much within the reach of CooliPi without throttling. If you see the graph here
      https://www.coolipi.com/Performance.html
      you'll see that with a fan, the temperature delta between ambient and BCM temp is 8 degrees Celsius at max synthetic load. It's 1.5 degrees at idle.
      Could go even higher with this efficiency. Dare to say - 70 degrees Celsius? But the PMIC would need a thermal bridge (in the works) and I'd be scared of microSD card lifetime.
      Last edited by CooliPi; 19 December 2019, 06:09 AM.

      Comment


      • #53
        Where I live, you can be sweating it out in 30-45 deg temps for most of a 14 hr day, then by night, sitting dropping to 10. It'd sure be interesting to see the mainboard resist cracking in the extreme variances day in day out, UV indices notwithstanding.

        I personally lost a fair amount of consumer hardware in a tropical area (lots of parts were rusting out rapidly, constant high heat coupled with constant use) so its interesting to see results like this, instead of a quoted manufacturer spec.
        Hi

        Comment


        • #54
          Originally posted by stiiixy View Post
          Where I live, you can be sweating it out in 30-45 deg temps for most of a 14 hr day, then by night, sitting dropping to 10. It'd sure be interesting to see the mainboard resist cracking in the extreme variances day in day out, UV indices notwithstanding.

          I personally lost a fair amount of consumer hardware in a tropical area (lots of parts were rusting out rapidly, constant high heat coupled with constant use) so its interesting to see results like this, instead of a quoted manufacturer spec.
          CooliPi was designed with even this in mind. I can recommend you DOW CORNING's 3140 rtv coating, it's a non-corrosive silicone rubber. I've coated some electronics going into highly condensing area, say 93% humidity. I'd coat the Pi on every place except contacting areas for (BCM, ETH, USB) chips, I'd add the soon to be developed PMIC shim and coat it around it to close access to PMIC and then use CooliPi on top of that. It has a riser - GPIO breakout - that takes out the GPIO port to the side, this can alleviate some stress on the Pi itself, and can be exchanged if necessary(destroyed). Heatsinks are anodised, should not rust.

          We have expected scenarios like this too - that's part of the "no compromise" paradigm. It can remove heat from every major heat source on a Pi, essentially allowing it to be coated with some protective jelly/coating. This particular model of coating is viscose, but with higher viscosity than water. Say - like honey. It creates about 0.3mm high layer on top of SMT components. The GPIO port can be easily "sticking out" of coated area , i.e. being coated at it's plastic heels. I've tested this too. Gold contacts don't rust, that's ok.

          I don't think the Pi motherboard would crack, it's a bit loose in the casing. The heatsink is mounted via 4 screws to the pi. The case is mounted via 4 screws to this sandwich. See picture at
          https://www.coolipi.com/GPIO_Breakout.html
          and
          https://www.coolipi.com/Industrial.html

          I'm interested in performance reviews from tropical areas - would be superb to log temperature and humidity. Would you be interested in testing it in your area? Some DHT11 sensor should be enough for it. Logging temperatures and actual speed of a Pi along with it would be essential. The sensor can be connected to the riser from a side. Protective coating however - can you source it? This particular one is expensive, some lighter one would suffice as well. I'd expect some few pictures of the setup and some log and user experience. WiFi antenna performance could be another issue when coated, but if coated with thin layer, I don't see much of a problem.

          Comment


          • #55
            Originally posted by CooliPi View Post
            From a manufacturer's point of view, it's just the opposite. The MTBF mean mean time between failures and it's a statistical figure, made out of individual measures.

            ...[snip]...

            Go apply this to a Pimoroni Fan shim. How many units will be dead by then ? What is their MTBF? how long does a fan run with rattling sounds before it finally dies? But we'll see over time. How much dustup will be there, needing service, lowering efficiency? How about smoker's fumes sticking the dust together with SOCs, shorting their I/O when being moisted? I've been cleaning some PSUs and notebooks of smokers. Not possible.

            For me, MTBF is a measure how much is a manufacturer certain that a product is worth of, bar noname manufacturers which evaporate after sale.
            Haha, I know the "fun" with computers/electronics in houses with smokers...!

            You just explained why I distrust MTBF. Although since many components are more expensive to repair now than replace, it would be better called MTTF (Mean Time To Failure). How are tests for it done? It is normally "accelerated life" testing, using extreme conditions so it doesn't take 17 years to find out when something will break.

            I take warranty longevity (and willingness of a company to back it up, rather than try to wriggle out of repair/replacement) as a better sign: a company that guarantees that their kit will work for 10 years I will more likely purchase from than one that guarantees for 1, because the implication is that it will be reliable for that long.

            That said, I'm one of those people who actually read the warranty terms, and if they are so locked down that there is almost no way to make a claim in the event of a failure, I will either not purchase or return the item immediately.

            Comment


            • #56
              Originally posted by CooliPi View Post

              CooliPi was designed with even this in mind. I can recommend you DOW CORNING's 3140 rtv coating, it's a non-corrosive..........oating however - can you source it? This particular one is expensive, some lighter one would suffice as well. I'd expect some few pictures of the setup and some log and user experience. WiFi antenna performance could be another issue when coated, but if coated with thin layer, I don't see much of a problem.
              Sorry mate, I've since moved from said tropics in to more drier, albeit hotter (during summer) and colder (during winter) climes (and currently surrounded by bushfires, so my time's a bit limited =D). And I don't own either of the devices, sorry.

              The silicone treatment actually sounds very interesting and is something I've long wanted for a lot of other sensitive pieces (especially when I start installing things in to my 4WD for camping).
              Last edited by stiiixy; 20 December 2019, 08:33 AM.
              Hi

              Comment


              • #57
                Originally posted by Paradigm Shifter View Post
                Haha, I know the "fun" with computers/electronics in houses with smokers...!

                You just explained why I distrust MTBF. Although since many components are more expensive to repair now than replace, it would be better called MTTF (Mean Time To Failure). How are tests for it done? It is normally "accelerated life" testing, using extreme conditions so it doesn't take 17 years to find out when something will break.

                I take warranty longevity (and willingness of a company to back it up, rather than try to wriggle out of repair/replacement) as a better sign: a company that guarantees that their kit will work for 10 years I will more likely purchase from than one that guarantees for 1, because the implication is that it will be reliable for that long.

                That said, I'm one of those people who actually read the warranty terms, and if they are so locked down that there is almost no way to make a claim in the event of a failure, I will either not purchase or return the item immediately.
                You've inspired me to advertise CooliPi with lifetime warranty :-D

                You're right, there are MTTF, MTBF, MTTD and MTTR statistical measures. Anyone can find it online. They're based on accelerated models, but for some type of device - say a fan - the model is more reliable than for some generic case. In the case of a fan, most problematic is the bearing. I'd dare to say almost all of the fan chicanery lies in the bearings.

                I trust my own experience and it tells me that - long shaft prolongs lifetime, because vibrations caused by rotating the unbalanced mass of a fan stresses the bearings to the sides. Longer shaft exerts less force sideways (less wiggling). Low quality bearings dry out or wear out and kill the device. Higher spinning rate kills bearings, approximately quadratically with speed - because the centrifugal force is proportional to the spinning rate squared. I've repaired many fans in notebooks (a drop of silicone oil year after year prolonged lifetime of 2008 Mac to 2018 when it finally died), seen many fans from smoker's offices (it's everywhere in the fan, on it's electronics etc).

                Based on this experience, I suggest all the people to get a high quality fan that lasts 10 times longer than other noname fan, even though it's twice as expensive. This is the reasoning behind choosing that particular Noctua fan - it's a monster, long shaft, best bearings, quiet - but does very fine job and will do so for 10 years or more. But I have very hard time explaining this to some non-technical people, mostly its price.

                One last comment regarding to "it either dies right after purchase, or after the warranty has passed" - see this chart
                https://www.backblaze.com/blog/what-...-should-i-buy/
                it's about survival rate of HDDs, but I think it tells much more than MT*F numbers. For example WD drives - at the beginning, the failure rate is higher, then it steadily dies one after one - constant failure rate. Hitachis don't have this early peak. But Seagates show strong dropoff at about 18 months age. So here are very clearly three stages of hdds' life, or any other gadget you've mentioned. This is known in the industry and has interesting consequences.

                Quoting from https://www.ti.com/support-quality/r...rminology.html
                • Early life failure rate (or infant mortality): This phase is characterized by a relatively higher initial failure rate, which decreases rapidly. The failure rate during this phase is typically measured as "defective parts per million" (dppm).
                • Normal life: This phase consists of a relatively constant failure rate, which remains stable over the useful lifetime of the device. The failure rate is described in units of "FITs", or alternatively as a "Mean Time Between Failures" (MTBF) in hours.
                • Wearout phase: This represents the point at which intrinsic wear-out mechanisms begin to dominate and the failure rate begins increasing exponentially. The product lifetime is typically defined as the time from initial production until the onset of wear-out.
                What has happened in the industry is optimisation for "Normal life" period to fit tightly in the warranty period. After that, "Wearout phase" of multitude of components start to emerge and kill a product. It has become apparent after Chinese bought many IT divisions of western companies. The race to bottom has finally brought its fruit - 2004 notebooks survive today, >2010 notebooks die after 2 years.

                It's the consumers ( who decide to buy the cheapest sh*t ) who move the momentum of an industry. And regulators too - if they don't create trading categories for products with longer lifetimes, it all falls to the lowest common denominator - to barely survive a minimal legal warranty.
                It has eroded car industry too, whatever industry next.

                Comment


                • #58
                  Originally posted by CooliPi View Post

                  It has eroded car industry too, whatever industry next.
                  There's a lot of pushback in my region to buying new car's. A lot of the younger guys along with older are holding on to 'ancient' car's. The cost of an individual part, while higher, is vastly less expensive than the constant costs of maintenence/parts replacement that is standard in new car's and their ridiculous service maintenence/warranty 'contracts'. 7 years? wow. I can't smell the BS for the rubbish pile. I personally wont ever give up my TJ Wrangler. Fuel's a costly factor in the old 4.0L donk, but I can get under it and hit on that baby wherever, whenever. And I've on more than one occassion fixed a piece or two in metre-high water.

                  It's a way of life as much as a money factor and it's a good indication that not everyone's interested in 'planned obsolescence', 'she'll be right I'll just buy another $2 one tomorrow' (duh, time=money?) mantra.

                  *shrug* Each to their own, but you're right; Cheap shit is ruling the worst of the 'advances' in modern tech. Ingenuity has taken a backseat to 'replaceable parts' and the number of engineers running around in my country without a clue as to what 'engineer' even means is ridiculously scarey.
                  Hi

                  Comment


                  • #59
                    Originally posted by coder View Post
                    Only because your government effectively subsidizes shipping of packages from China with funds collected from your own taxes. China is a Group III country, in the IPU, meaning it has to pay lower terminal dues than a Group I country, like the UK and most/all EU members.

                    https://en.wikipedia.org/wiki/Universal_Postal_Union


                    Other considerations include ethical and ecological sourcing & processing of materials, as well as the energy mix used to power the process. I think EU countries all participate in the OECD, whereas China does not.

                    https://en.wikipedia.org/wiki/OECD
                    Not only that. See this
                    https://www.businessinsider.com/how-...e-state-2018-4

                    Also because the've made all the people slaves to their dictator, stripped them of all rights and used western technology to enslave them.

                    Do you still want to save $10 on a Pi from China? The difference you take away from local companies and give it to a high-tech terror state George Orwell would stare at in amazement and Hitler would be ashamed by?

                    Make no mistake, western "elite" is craving to get it done in the west too, only some people like Ed are standing in the way.

                    The end looks like this. On the road to the end, you have an option every time you shop, whether you vote for the end right now (it's cheaper) or later (buying local goods and as a byproduct, partly financing the rest of our democracy or whatever has left of it).

                    Comment

                    Working...
                    X