Five Years Of The Raspberry Pi Model B+ Form Factor, What Has It Taught Us?

With all the hoopla surrounding the recent launch of the new Raspberry Pi 4, it’s easy to overlook another event in the Pi calendar. July will see the fifth anniversary of the launch of the Raspberry Pi Model B+ that ushered in a revised form factor. It’s familiar to us now, but at the time it was a huge change to a 40-pin expansion connector, four mounting holes, no composite video socket, and more carefully arranged interface connectors.

As the Pi 4 with its dual mini-HDMI connectors and reversed Ethernet and USB positions marks the first significant deviation from the standard set by the B+ and its successors, it’s worth taking a look at the success of the form factor and its wider impact. Is it still something that the Raspberry Pi designers can take in a new direction, or like so many standards before it has it passed from its originator to the collective ownership of the community of manufacturers that support it?

Evolution Of A Form Factor From An Accidental Success

An early Chinese-made Pi Model B from 2012.
An early Chinese-made Pi Model B from 2012.

The first Raspberry Pi that we could buy back in 2012 was not the polished production of a slick electronics company, instead it was a shoestring creation from a group of electronics and computing professionals who only expected to sell a few tens of thousands of their board in the hope that they’d teach some kids to code. Following their production blogs from those nervous few months over the autumn and winter of 2011 was a fascinating course in the challenges of preparing a complex board for production, as we learned about the Chinese assembly company creating the first samples, and then about a truly heroic SMD rework task on that very first batch.

It’s fair to say that simply getting it to the point at which we could buy a Pi was the stellar achievement, and the layout of the board itself owed more to necessity than an eye towards a future line of successors. The credit-card size was perfect but there were no mounting holes for example, and the various connectors were positioned at all points of the compass with the composite video and USB sockets protruding over the board edge.

The first of the new form factor, a 2014 Pi Model B+
The first of the new form factor, a 2014 Pi Model B+

By a combination of a winning price point, a usable and supported software distribution, and a cleverly managed dose of positive publicity, that first Raspberry Pi surpassed those initial modest expectations many times over. Its launch was massively oversubscribed, and it continued to sell in huge numbers. Production was moved from China to the Sony plant at Pencoed, Wales, and a revised version was produced that fixed a few bugs and added a couple of mounting holes where they could be shoehorned in among the other components.

Clearly something better was called for, and the B+ in 2014 delivered what was essentially the same hardware as the original with those aforementioned mounting holes, tidied-up connectors, and a new pin header. This form factor became the basis of the new HAT specification for expansion boards, with its pair of extra pins to serve a little EEPROM containing device tree instructions for the OS.

Everybody Wants To Be A Pi

The success of the Pi had not gone unnoticed by potential competitors, and within its first year there were a host of other boards being touted as Pi killers. They would however all follow their own form factors, for example the Cubieboard from late 2012 set the established route for a Pi contender in taking a tablet SoC at its heart, but did so in a custom form factor. Even the Odroid-W, the only software-compatible Pi clone to make it to market (We’re not certain Arducam’s tiny offering ever went on sale) followed its own form factor. It would not be until late 2014 that the SolidRun CarrierOne would make an appearance as the first SBC to directly copy a Pi form factor, going on to become the commercially available HummingBoard.

One of the better Pi clones using the same form factor, the Asus Tinker Board.
One of the better Pi clones using the same form factor, the Asus Tinker Board.

The launch of the B+ changed all that, as the Pi now had a well-thought-out form factor with a guarantee of being future proof. Successors to the Model B+ would follow the same format, which gave the after-market of add-on manufacturers the confidence they needed to support it. If you wanted to compete with the Pi it would be a brave move to not have at least some sort of nod to the new form factor, so we’ve seen a continuous stream of other boards following it.

A by no means complete list culled from a few minutes with a search engine brings us models from Orange Pi, Nano Pi, Banana Pi, Odroid, RockPi, Asus, Libre Computer, and our favourite, the Zberry (a Z80-based personal project we featured in 2017). There are plenty more, but it should be pretty clear that to take a Chinese tablet or phone SoC and make it look like a B+ with a random-fruit-name-Pi moniker is a winning formula if you’re based in Shenzhen and in the SBC business.

Did They Get It Right, And Can They Hang On To It For Another Five Years?

The success of all these machines with the same form factor and the huge range of HATs to go with them would indicate that the Raspberry Pi designers got something right five years ago. But it’s dangerous to assume that there is nowhere further to go, did they really succeed? Physically the dimensions of the board and choice of the 40-pin connector at one edge avoid the pitfall taken by the classic Arduino of odd dimensions, and the relocation of the larger sockets to one edge of the board avoid the need for awkward cut-outs that were a hallmark of some add-ons for the original. You may encounter Raspberry Pi owners complaining about the micro-USB power or the slow network and USB ports, but the form factor scores low on the list. If we were challenged to find a problem with it we might look at cooling, the original B+ had a bare processor while most of the more recent boards have to find space for a heatsink and possibly even a fan underneath their HATs. Perhaps this is something the designers will have to address in future versions.

The 2019 Pi 4 makes big changes to the connectors, surprisingly only the analogue output hasn't received some sort of attention.
The 2019 Pi 4 makes big changes to the connectors, surprisingly only the analogue output hasn’t received some sort of attention.

The Raspberry Pi 4 may follow the same HAT form factor as the B+, but it is the first of the line to make significant changes to the rest of the formula. It has been normal for clones to reverse the positions of USB and Ethernet sockets and USB-C for power is an obvious progression, but the Pi 4 makes a more controversial choice in its replacement of the single full-size HDMI socket with a pair of micro-HDMI receptacles. Aside from rendering a host of case designs obsolete at a stroke, is this a canny move? We’d call it “Bold” in the sense of “A bit reckless”, because while the Pi Zero has always had a mini HDMI socket it has always brought with it the annoyance of having to find an adapter. The attraction of the Pi from day one was that it could be plugged directly into a domestic television, and with this update that ease of use has been removed. Who has a micro-HDMI adapter just lying around? The answer is almost nobody, and while the prospect of dual displays is an attractive one, the feeling can’t be escaped of having lost something along the way.

The ownership of a physical form factor standard can be surprisingly fluid, and will sometimes escape its creators when its adoption becomes wide enough to eclipse their original. Readers with very long memories will recall the PC wars of the 1980s as an example, when IBM tried to move the industry away from the ISA bus as found in the earlier PCs. The industry had other ideas and persisted in developing ISA into the mid 1990s while IBM’s MCA failed to gain a foothold. The Raspberry Pi ecosystem is definitely healthy enough to follow whichever way the designers take its form factor, but has that of the Pi clones reached maturity enough to no longer follow? We’ll count the HDMI sockets on 2021’s offerings and report back.

83 thoughts on “Five Years Of The Raspberry Pi Model B+ Form Factor, What Has It Taught Us?

  1. The micro HDMI connectors (plural) are a non-issue for the vast majority of Raspberry PI users I suspect.

    Many, many Pi’s are run headless, and a fair number of Pi’s running with heads hang them off the dedicated display connector running small touchscreen boards.

    For those insisting on hanging an HDMI monitor off the new Raspberry Pi 4 B they can simply invest in micro HDMI to HDMI cables, they are widely available for less than $10 in a 2 meter/6 foot length.

    For me, I was more upset about the switch to the USB-C power connection, but the $10 ‘official’ power supply rendered this a non-issue.

    The only lasting issue is heat, but the obvious fix – heatsinks – might rule out certain ‘hats’ for the new Pi’s, so I understand leaving the fan/heatsink/nothing choice to the buyer.

  2. Display port mini would have been a better choice – those are common enough on laptops that a lot of people have an adapter for it.

    I dunno, I’m just done with Pis. Until they get some kind of onboard nonvolatile memory they’re just too much of a pain – and too limited – to play with anymore.

    1. Any form of DisplayPort would have been a better choice than any form of HDMI.
      Regarding oboard storage, well at least you can run from USB3. As limited as it is Raspberry practically has no alternatives if you want opensource accelerated graphics and latest kernel.

    2. You can boot from usb as of the 3B+. There are even SATA/SSD cases to boot straight from a drive.

      With the higher ram options on the 4, I wouldn’t be shocked if suspend to ram doesn’t become a software improvement.

  3. I find the main limitation to the form factor is the use of hats/capes/shields that have to be stacked up if you need more than one. For that reason, I have moved to using a passive backplane for my larger micro controller projects. Today that requires the use of carrier boards to mount Pis and Arduinos etc on the backplane. There are many advantages to this approach including the ability to pull individual boards, mechanically stability, ease of packaging, ability to easily provide centralized power and voltage level translation etc. I use a 30 pin 0.1 inch serial bus (power + I2C + SPI + reset + interrupt + RS485) with 15 pins for 5v and 15 for 3.3v. The bus supports 2×30 pins for extra voltage levels, higher power supply currents or GPU pins. You can also use 1×15 or 2×15 pin backplanes for a smaller form factor. If Pis and other boards had GPIO pins at right angles to the board, preferably using a standard bus similar to the one I use, putting together multiple hats/capes/shield systems would be much easier.

      1. I have set up a Hackaday.IO project with some photos and details. This project will be on the back burner for a while as I have other projects to complete first.

  4. Model B shows the whatever which way and non-aligned connectors placement which is very amateurish. I prefer straightly front and back. The external cabling affects the overall footprint, how and where you can place the Pi.

    Mechanical issues such as mounting, thermal design should be at the initial placement and should not be an after thought.

      1. Front to back on the long axis is the proper way of I/O layout. Having both front and back helps as the external cabling can be heavy tend to tip a small board. This also help say in thermal design as you can have a piece of heasink or a fan force airflow going across the sides.

        Now let’s hear you “professional” advice advocating the contrary.

        1. No way to have as many connectors and such a small form factor in the way you descibe. You want to put it all on the back it becomes hub and dongle city behind the device or the device must get a huge amount bigger.

          With the size and cost I don’t think a better layout could be created. If you double sided the connectors making the PCB much more expensive and the device thicker you could perhaps leave the channel for airflow.
          If the space you want to put a computer is really so tight you can’t make use of a pi with standard cables its time to get custom cables or solder to the connector/test pads, Even at worst case with the cables sticking out keep the pi way smaller than anything in its price/performance range.

          If your usecase absolutely needs to run the pi hard it is trivial to run an extension cable for hats and have top down access. Otherwise like the badly throttling thin macbooks put up with having peak cpu for tiny spikes of time (which for many real world situations will be fine as long as the minimum speeds are high enough)

        2. Why do I have to waste my time to prove to a troll? What I do in my spare time is up to me. If someone pay me my standard consultant fees, I am more than happy to do an industrial layout. Bare in mind that RPi foundation have enough resources to do it right. It is not a 1 person fly by night hobby any more.

          My comment is hared by a few here: Christian Nobel, oni305

        3. In the real world, there are multiple constraints. Cost is a constraint. These are board designed for students, where accessibility, cost, and standardization are prime concerns. There is probably no budget to fund an “optimized” board that simply “looks better”.

          Part of the open-source community is “giving back”. Perhaps you can donate a few hours of your “consulting” fees and give them your optimized layout? More than likely you have neither the experience or tools needed to prove your design, which currently exists only in your head, has no FCC, CE, UL certifications , and runs at some indeterminate frequency, and maybe passes EM compliance when nobody is looking.

        4. You cannot make assumptions of what someone have done. I have posted more than enough about signal integrity, multi-giga bits high speed design, safety comments in my past here as some projects sorely needed some industrial inputs.

          I have done UL/CSA, FCC and some mil-spec work. Been there done that. No need to prove to anyone other than my employer/clients.

          There are no lack of comments here from fellow poster not liking the layout. Counting 3 so far.
          Care to attack them too?

        5. “These are board designed for students, where accessibility, cost, and standardization are prime concerns.”

          Oh the we-dot-it-for-the-kids argument, and we do it out of pure altruism – baaah bullocks, because how many Pi’s do you think would be available, if it were not for a massive sales to private and companies?

          And the computemodule, that is also for educational purposes?

          And what we request would not necessary make a more expensive product, just a much more well-thought product, and with a better power supply, also a much more rugged product, so the main audience (according to “Bill Gates” – sigh!), namely the poor students, would not corrupt so many sd cards.

          Then there is the discussion about the wapor-ware aka the zero, and the stupidity in that a company in the EU are stating prices in $, and there actually is nowhere in the EU where it is possible to buy at Pi for the price Mr. Upton is bragging about.

          But OK, the Brits are so self-sufficient that the mean they at better off, converting the country to the 51′ state of US – but we will probably see a dramatic decline in the European sales of the Pi.

        6. Looking at the layout solely from a PC replacement perspective:
          Back (long) side: Display and ethernet connectors, as these are usually plug and forget.
          Left side: Back to front: Power connection, SD card, audio jack.
          Right side: USB and camera ports, and recessed reset switch.
          Front (long) side: GPIO connector. Not populated with a header, so bottom facing socket could be used. If the Pi were a hat, instead wearing one, there would be a lot more flexibility with respect to cooling and heatsinks.

  5. If the board had no video at all, that would be fine by me. I have no notion of using it as a desktop replacement or general purpose computer, rather a headless embedded board. I suspect this is what most users do with it, but could be wrong.

        1. Of course a Pi isn’t as good as a full fledged desktop or even a Shield TV, but at a far lower price, who would expect it to? Any Pi older than the Pi 4 was limited to 1080p, which is good enough for older TVs. (I see the Pi 4 as a game changer for 4K, perhaps it would become a popular fix for buggy/outdated smart TVs.)

          1. I’ve been using a PI2 for just this. Our samsung “smart” TV periodically updates, and what this means is that we loose apps and services each time. No more Youtube for example. Also playing media through our PI removes the adverts in things like Youtube and Twitch. Never had an issue playing anything with it, although we only watch 1080p offline content (and 720p online).

        1. hard data is watching my downloads from my nas take forever and unbearable media center video stutter. nas performance was a known factor due to the supported usb standards and the limits of the ethernet port. media center might be usable with the right software configuration, but that eluded me, but i think it all boils down to a network bottleneck.

    1. I’d say your wrong, my main use case for a pi is in my car running Android auto, I’d say about 80% of the pi’s I know in the wild a(in regular use) are hooked up to a TV for smart TV/media playback/emulation duties. would be an interesting thing to be surveyd

      1. Let me know where I can find a $35 board, with the same specs, but with your “optimized” layout. Maybe you can solve the problem. Please contribute your board design.

        1. please go back to your own blog site. this is hackaday.com , not i’mfinewitheverydesignyouthowatmeaday.com. discussing things that could improve knowledge about board design happens all the time here. I stopped using raspis and switched to orang pi, friendlyarm nanopi neo air (only pin headers, even for usb!!) and never looked back. same with the arduino. as soon as i found the pro micro and other breadboard friendly designs, the original form factor arduinos left the house.

          and no, you don’t have to react on this post, unless you ARE a troll of course. :-)

    1. >You need an ‘official’ one for it to work right?

      No, you need a “dumb” one which doesn’t do much in the way of power negotiation. This was pretty trivial at low amperage, but many of the higher amperage chargers are “smart”. All the same, I had a couple of old fast chargers lying around which work with my 4B fine.

    2. by your logic the only people criticizing anything must have a direct competitor for the product they are criticizing. that’s like saying that you cant have a preference in cars unless you have designed and built a car or you can no longer make a comment about feature length film until you have made a feature length film yourself.

      or you know, you could have just posted a counter argument to the opinion provided like a normal person but i guess a troll is gonna troll

  6. “Did They Get It Right?”

    And the answer is: NO!

    It is still at horrible layout, with some connectors on the “front” and some not on the back, but at one side, very close to the corner.

    This makes it more or less impossible to make some decent casing for the Pi, and leaves cords hanging out in two directions – it is in no way right, and what makes it even worse are all those products trying to sponge on the Pi’s success, instead of producing at better product.

    Olimex has their own identity, and have clearly a better layout.

    Imo there should only be (external) connectors at one side, and if that is impossible, at least only at front and back (USB at front, rest at the back.).

    And of course, a DC barrel connector, as discussed elsewhere.

      1. On the other hand, you are so lucky that you can buy the PineA64 for approx half the price that we have to pay for it in Europa.

        The layout is much better than the RPi, and Rockchip is not as closed as Broadcom.

        Imo, if not RPi shall be lost in competition, they need to make a new version of the B+, with better layout, and lower power consumption – a lot of Pi’s are used for purposes where there is no need for 4 cores etc, but power usage and heat is a concern.

        1. If you don’t need 4 cores and rather have smaller size and lower power usage, that’s what the Pi Zero is for! (I would like to see a higher end version of the Zero with 4 cores, more RAM, and 802.11ac, would be a good start for a DIY tablet.)

  7. I’m not too bothered by the micro-hdmi but I really wish they had gone for a stacked regular-sized hdmi port, similar to the stacked usb ports. Maybe they had a good reason for going the way they did though.

  8. Possibly the board size is a constraint for them.

    If they can make it work without much cost increase, an extra 1cm on the sides would give space to move sound and power connectors to the same side as the network and usb.

    Casualties would be existing enclosures, but those are a minor inconvenience/cost. The problems involving power supplies are of much bigger significance than some cheap enclosure.

  9. I would have liked to see an end to the 40-way pin connector, in favour of a 40-way socket connector. Exposing pins that can easily be shorted together is a bad idea on a board that’s often used in educational environments . A socket also allows for wires and other components to just be pushed into the holes for the simplest of physical computing projects.

    Backwards compatibility could be achieved by selling boards with through-hole pins protruding from each side, though this would increase the height a little, and additional stand-offs might be needed for HATs that use the mounting holes as well.

    1. The pin header would allow for 40 pin IDC ribbon cables, but few seem to like using that. I agree with socket pins.
      Power pins could have been assigned more intelligently to avoid the issue – by having a ground pin between two voltages or having I/O that can tolerate both voltages. That’s how it is usually done.

  10. While my RPI isn’t my “real” computer, I use it a lot to check my e-mail and hackaday throuout the day. It uses 1/20 the power of my “real” computer and boots much faster.

  11. Spot on.

    The layout is a disaster. To me they are trying to stuff too much on limited board real-estate. It would have been better to increase the size by say 1cm on the sides and get something that doesn’t look like a octopus when you connect cables to it.

    And the lack of a DC barrel connector is just plain nuts.

    All of this should have been fixed a couple of revisions ago. And if it meant increasing the price by $10 so you have better designed product, it would be worth it.

  12. The right vision, software support and price allowed them to win. The design and chips used was irrelevant. Using the goal of education is what attracted so many supporters with the promise of recreating the heady 8 bit days where every kid was coding their own games. And they have to a large part done that.

  13. IMO, the problem with a DC barrel connector advocated in a few comments above is that there is no standard for voltage or polarity. Put a DC barrel connector on it and someone is going to plug in a reversed-polarity or too-high-voltage power supply. There’s also no standard for size, so any DC barrel plug power supply you have laying around will likely not fit.

    1. Not so much cost to use a TVS or other protection part that would for example short and protect the board in case of reverse or too high voltage.

      There are standards for the usb-c connedtions, and that caused its share of problems with the pi4 and chargers also. And when people start buying cables from china, and said cables get mixed with others, more problems will also exist, to the extent of picking a bunch of cables from the BigBoxOfCables and trying to see with one of them works.

      Have seen a bunch of phones/gps burned from connecting them to cheap cigarette-lighter adapters in automobiles already. Same thing tends to occur to the pis.

    2. But it would cost very little to have on-board regulation, and polarity protection – if the input voltage could be something like 7-15V, it would also make it much easier to supply it from a battery.

      And maybe the extra cost would be 1Euro, but that could be saved on the PSU, giving the same total price – but of course it is easier to brag about a low price (and not so low in reality) on the board it self.

      1. Having guaranteed 5V input makes the power supply design much simpler compared to providing a wide input range like you describe.
        Take a look at how tiny the inductors are that are used on the RPi4 to generate the various voltages. Now take a look at the big honking inductor I have on my LiFePO4wered/Pi+ (https://hackaday.io/project/20909-lifepo4weredpi). The only reason for it to be that big is that I support a large input voltage range up to 20V. It could be much smaller if I just stuck to 5V input. Can you imagine the RPi people putting such big inductors on their board? ;)

    3. ^this^

      Unless you protect the input, barrell jacks are a horrible fit for digital electronics (in spite of their durability)

      I had a router that got fried because I plugged the wrong power supply into it. (Another supply for some other box sitting next to it). I opened it up to see if I could just replace a fuse or burnt out diode or something, but nope. There was not a single protection component, other than a few big electrolic caps, between the barrell jack and the 6ish discrete switching regs that fed the cpu and other ICs. One of the switching ics was visibly damaged (that burnt chip look), and the others might have had invisible damage. And of course, they didn’t have usable part numbers on them, so no replacing them.

      Moral of the story: Barrell jacks are good for mechanical strength and current capaticity, but that is it. They really don’t fit on electronics.

      That said, the barrell jack on my odroid xu4 is kinda nice. But I have to be careful what I plug into it.

      Honestly, USBC is a pretty good compramise. Raspberry pi’s don’t just cater to the type of people who know what not to plug into a barrell jack on a sbc. They have to make it at least somewhat foolproof (at least not foolprone). USBC has the current capaticity, just not quite the mechanical strength. It’s also much more standardized AND commonly available than a properly specked wall wart with a barrell connector.

      So at this point, we have:

      Barrell:
      Pro: mechanical strength
      Con: might blow the board, might be more expensive, not commonly available

      Usb c:
      Pro: Standardized, common, inexpensive, familiar, foolproof (mostly)
      Con: not quite as strong

      I think USBC wins this one.

  14. Everyone crying and moaning about the connector layouts. The primary goal of the Pi project is cost. I’ll say it again, cost is the primary goal. Not connector layout, power consumption, connector form factors, performance, the bottom line is cost.

    1. Agree 100%

      My take on that is it is not a flaw but a feature.

      As for bulky cables, that’s a design feature of the cables, not the Pi. Where are the right angle ended ones for example. There’s a lot of unnecessary bulk in many connectors.

      And as for the micro USB complaints, they want the moon on a stick, and then complain that the stick is too thick/short/whatever…

  15. the Pi is cheap enough, and the RPi4 brings some needed horsepower and RAM… have been using a zero w/a 2S digital 3A sound card for streaming audio server for several years, but last week built a panadapter using the Pi3B+ for the IF port on a Xiegu X5105 transceiver (biggest hurdle is the terrible engineering of SDR dongles) – US$80 for a workaday unit that usually costs almost 10X as much… but do agree that the new Pi4 dual HDMI is useless to me

  16. The Raspberry Pi Foundation is a (UK registered) charity whose aim, if I understand it correctly, is to get a functional ‘computer’ into the hands of outright noobs and feed the curiosity, creativity and enthusiasm of said noobs so that more and more could become knowledgeable and competent in ‘computing’.
    That, I think, is the driving force behind the design, not the desires of the already well qualified readers of hackaday, or the “I would do it differently, for very clever reasons” brigade.

    1. Ohh yes, that is the official story – but sorry to disturb the self-sufficient mindset of being sooo altruistic, and we-are-doing-it-for-the-kids.

      Because maybe 1 out 100 RPi’s goes for educational purposes, the wast majority for professional purposes, and a large amount for private media players etc, etc.

      And without this large non-school sales, there would be no foundation – so is it totally unreasonable just to listen a little to ones customers?

  17. If you are primarily a software person, and want to make your project look nice, I guess the connector layout is a bit of a hassle, as you’re going to use either a ready made enclosure or something easily workable that you have laying around, and you don’t wish for it to look like some sort of an octopus box.
    If, however, you are a hardware oriented person it’s a non issue. A soldering station, some wire, and some assorted connectors means you can lay it out any way you like, with whatever connections the board permits, in whatever style enclosure seems appealing. But given your nature you probably are limiting your endgame to a media box or retro arcade or some other ready to install software.
    If you’re a little of both, good for you. If not, then stop arguing the merits and pair up with a member of the other side. You’d be amazed what can be accomplished with 2 heads. Rough it out without the enclosure. Prove it works and share it. Somebody with the skills to make it look nice will probably take on the challenge if it’s something of real interest.
    I see so many projects posted here from one camp or the other that could use a little something from the other side, whether it’s an ingenious idea that looks like it assembled itself and crawled from the parts bin, or an incredibly artistic looking creation that seems to have a disappointing feature set, given its components, I often have to wonder if the creator is such an introvert that they can’t bear to work with a partner, or they are just too much of an independent cock for brains that the thought of letting someone else touch their baby is unthinkable.
    If you fall into either of the latter two categories, just learn to digest the criticism properly.

  18. Pi 4 needed some type of real storage option (EMMC at minimum, a sata/M2/pcie connector ideally) far, far, FAR more than it needed dual display capabilities. And the price increases for the higher ram configurations seems exorbitant. Does 3 additional gigs of ram really cost $20? $55 seems to throw the value argument to the wayside, especially with the crippled storage. That’s just getting too close to other options with much better specs.

    Can’t beat the software support though, gotta give them that. Even so, this is the first Pi I have no desire to buy. I’m on the sideline until they get real storage.

  19. The thing I’d really like to see is a way to access one of the USB ports from the board. That would mean it’s possible to put the Pi in a case with an internal peripheral without having to jump through hoops extending 3 USB + ethernet to the outside world.

  20. It’s interesting reading through the comments and there is as many differing desires and “should haves” as there are comments.

    It would be impossible to build a single board computer to fulfill the desires of everyone and come in at under $100.

    Yes the pi has limitations and constraints that won’t suit everyone and OMG they chose to use a part that’s not open source.

    But its CHEAP and does a pretty good job of a LOT of things. And if you don’t like that no one is saying you have to use it.

    There are fully open source alternatives one could modify to their hearts content and make it exactly what you want.

  21. The Raspberry Pi 4 is got 1-4GB of LPDDR4 RAM, a Quad-Core A72, 2 HDMI outs, two USB 3.0 ports, Gigabit Ethernet, WiFi, Bluetooth, LCD/Camera connectors, 3.5mm Audio/Analog Video out and an Accessible GPIO/UART/I2C/SPI/I2S/PWM port. It is also manufactured in the UK. All this for $35-$55. That’s a tonne of features for the price. You are seriously going to have to try very hard to find a better deal. And that’s before mentioning that you are getting the best software support in the biz.

    Most common complaints:
    – It uses SD cards for a hard drive, so access times are slow. WORKAROUND: You can always run the root filesystem off a USB3.0 hard drive.
    – It get’s hot. WORKAROUND: For Christ’s sake use a god damn heat sink. Find me another Quad Cortex A72 that doesn’t need a heat sink.
    – Raspbian is 32 bits so 64-bit hardware isn’t being fully taken advantage. WORKAROUND: I happen to agree with this one but I also understand the RPI’s foundation to stick with a 32-bit OS. One can always try a 64-bit build. I believe Ubuntu Mate has an experimental build for the RPi 3 and will very likely release a 64-bit build for the RPi 4 in the coming weeks/months. There’s also a Gentoo 64-bit OS for the RPi3 with a RPI4 version in the works. https://github.com/sakaki-/gentoo-on-rpi3-64bit

      1. Benchmarks have shown performance benefits to using a 64-bit ARM CPU over a 32-bit ARM CPU even with less than 4GB of RAM. It is not just about the maximum addressable range of RAM.

  22. The RP is,and will always be what it is. An appliance built to a price point to serve a need, it’s designers perceived to be there.I doubt the targeted end users care about open source, or are even mature enough to fully understand the concept. Hell for many adults open source has become more of an attitude, than a concept. At this price point it’s unreasonable to expect open source components, or expect broadcom not to, favor the entity that created the demand. As I read the wish list for the RP it seems RP users want a variant of the Intel atom boards., with a larger foot print and higher cost cost I have more important issues to sweat. An long as the RP is an affordable option than can do what I need, I’ll use it even if compromises have to be made.

  23. It would have been great to have all of the consumer connectors on one side (USB, network, video and power), and the the GPIO on the the other. Cable management would be much neater!

    On top of that, a socket for GPIO instead of pins, and possibly better grouping of IO into related functions.

    Finally, for the extras expense of the camera and LCD connector, I feel like it was only a very small subset of people that used it.

  24. “Five Years Of The Raspberry Pi Model B+ Form Factor, What Has It Taught Us?”

    Among other things, it–along with the ‘PoE Hat’ fiasco–have taught us that the Raspberry Pi Group can NOT do simple, entry-level electrical engineering hardware design. They are more interested in rushing ‘whiz-bang’, headline-grabbing product out the door than in doing it right–they can’t even copy the USB 3 spec!

    *****************************

    Raspberry Pi admits to faulty USB-C design on the Pi 4 ** —

    …”I expect this will be fixed in a future board revision,” says co-creator [Eben Upton]

    …”…Benson Leung, an engineer at Google and one of the Internet’s foremost USB-C implementation experts, has chimed in on the Pi 4’s USB-C design too, with a Medium post titled “How to design a proper USB-C™ power sink (hint, not the way Raspberry Pi 4 did it).”

    **
    https [colon][slash][slash] arstechnica.com/gadgets/2019/07/raspberry-pi-4-uses-incorrect-usb-c-design-wont-work-with-some-chargers/

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.