Arduino Uno R4
store.arduino.cc
store.arduino.cc
[1] https://files.littlebird.com.au/Shared-Image-2023-06-28-08-0...
A middle ground sure would be nice.
I respect Arduino for what they have done for the maker and tinkerer scene, but landing pages like that will not keep me as a visitor.
Here's a screenshot, everything in red was blocked: https://i.imgur.com/pAEMBxD.png
That's 84 requests blocked (54% of the page, 14 domains blocked out of 16)
To do this, in Advanced User mode, under 3rd Party Scripts and 3rd Party Frames, click the left-most cell pair to make them red. (right-most pair is settings for the current site only, left-most pair is for every single site, ie default settings)
Arduino Uno R4 WiFi - https://news.ycombinator.com/item?id=36481263 - June 2023 (118 comments)
Earlier:
Arduino Uno R4 - https://news.ycombinator.com/item?id=35315424 - March 2023 (129 comments)
On the other hand, it is a really convenient standard connector.
On the other hand, I2C means you can only have one of each kind of device due to address clashes, the lack of an interrupt pin is quite limiting, I2C is too slow for things like displays, and 3V3 isn't really enough to power LEDs like WS2812Bs.
To be clear, it's not just a matter of power, but it really is the voltage: shorter wavelength LEDs use semiconductor materials with bigger bandgaps https://en.wikipedia.org/wiki/Light-emitting_diode_physics#M...
What I don't like is that the connector is a nightmare to hand solder and isn't as strong as some. But then, I suppose that's what breakout boards are for.
There's a few other use cases that can't be done easily in a modular way unfortunately, mostly to do with power. SPI needs more pins, and there's no way to handle 12v, and only one module can be supplying power.
Stuff is not quite fully plug and play yet although it would be super cool.
They keep making them, so there must some reason?
[0] https://docs.arduino.cc/retired/boards/arduino-serial [1] https://docs.arduino.cc/retired/other/arduino-older-boards
[1] https://web.archive.org/web/20070527130630/http://www.little...
The feather format has lots of shields (“wings”) too.
It's easier to see what you're doing and plug a rats next of wires into the bigger board without it falling over, and everyone's got to start somewhere! Uno format Arduinos are a learning tool first and foremost and they do that job really well.
From the Arduino Uno R4 FAQ…
> “the Arduino UNO R4 WiFi was specifically designed to ensure compatibility with previous shields and compatible hardware developed for the Arduino UNO R3. The UNO R4 WiFi maintains the same mechanical and electrical compatibility, allowing you to seamlessly use your existing shields and hardware with the new board. This makes it easy to upgrade to the UNO R4 WiFi without the need for significant changes or adaptations to your projects.”
In addition to physical compatibility, maintaining the GPIO pin layout means software libraries created for earlier Arduinos should continue to work. For example, an interface library for a sensor or actuator that’s reused project to project.
Now it also runs Micropython, so we don't have to bother compiling code, you just push it over UART and run it.
Another big shift is going to happen when these boards shall be able to run Node Red. Siemens PLCs now run Node Red. I don't like it, but many engineering professionals have a better experience using block programmingand get stuff done even if it's a waste of resources.
They have the maker devices if you want a smaller board, but it's rather expensive compared to the alternatives.
Edit: I see this topic was discussed yesterday
Powering the pi could be so much easier - this arduino takes up to 24v.
And the arduino female bus pins are so much more practical than all the pins sticking out of the pi.
It's cool, but very much tied to one specific pixel art style and not visible in a lot of applications. I'd rather they just make an official I2C TFT you can mount wherever.
Although I suppose it's a cool novelty that will probably lead to a lot of fun stuff that wouldn't otherwise happen.
I would have preferred a tiny OLED though, those seem to be cheap.
I'm also disappointed that there's no USB-PD. It can take 20V of power, it should be able to request 20v of power from the port for powering motors and stuff with just the one convenient connection.
There's also a VBAT for the RTC on a little connector that also includes an off switch button. A nice feature but it needs a coin cell to use. I wonder if they could have done a rechargable battery instead.
They would have needed a nanopower regulator to get to a proper VBAT, and a charge IC, so it might have added a dollar though.
(You might get a free CGNATed IPv4 as well, but the point of v6 really is not to have to.)
Even with both one might decide only to use the v6 uplink. There's not really (ordinarily) a good reason not to - you can still use v4 locally if it's more familiar/easier to firewall or whatever, and unlike a v4 that you haven't paid extra for, the v6 is almost certainly static - which is either a huge benefit, or you don't care either way.
Or perhaps I am not a target audience of this hardware change.
AVR is still reasonably easy to understand and manually bit-twiddle, but virtually all ARM cores are way too complicated for that. It is essentially required to use a SDK to handle all the gory details for you.
If you want to get a proper idea of what a modern ARM microcontroller architecture is like, I recommend you take a look at the RP2040 datasheet: https://datasheets.raspberrypi.com/rp2040/rp2040-datasheet.p...
- x86 and x64 are too power-hungry for small systems.
- Power is still up and running, but also on the high end, too (see Power 8 and 9).
- MIPS32 cores are found in lower-end MCUs. AFIACT MIPS actively offers their cores for embedding in MCUs, like ARM.
- You can still buy MCUs with Atmel AVR and 8051 8-bit architectures on the lowest end.
What's lost? SPARC and Motorola 680x0, I presume. I frankly wonder what prevented 680x0 from getting into widespread MCUs. Apparently Phillips used it their custom controllers for some time.
People should just select modern ones, and not the same old 16F84/877 or something legacy like that, that's been obsolete for like 25 years already. All of the modern ones have very uniform programming interface, nicely configurable I/O and a lot of analog features.
Not anymore that I'm aware of. They laid off most (possibly all?) of their architectural engineers around 2020 and sold off the remaining rights to a Chinese firm.
There's still a few MIPS-based MCUs out there, like PIC32, but they're a dying breed. ARM Cortex-M is much better suited to these applications.
> I frankly wonder what prevented 680x0 from getting into widespread MCUs.
Poor IPC, high interrupt latency. The weird mixed 16/32-bit nature of the 68000 made it easier to implement a microcomputer with (since you only needed 16-bit RAM/ROM), but would only get in the way in a microcontroller (where memory would mostly, if not entirely, be internal). It's a shame; I liked the ISA, but the performance just isn't there.
I almost did a prank post when Apple came out with/announced the M1. I had recently had tore down my broken PSVR which has the very nicely stamped ARM M1 in it. I wanted to post something like "Apple didn't invent it, Sony did!" Regretfully, I did not.
I don't think that's correct. The Cortex-M1 is an soft-core-only design; there's no hardware implementation. (And moreover, the ARM cores used aren't usually part of the external markings on a chip.)
What chip were you looking at? I don't see anything which looks like that in the iFixit teardown: https://www.ifixit.com/Teardown/PlayStation+VR+Teardown/6934...
It was a gen1 model, so it didn't look anything like those photos on that teardown.
*Explanation for HN pedants: this instance of "everyone" does not literally mean everyone; it means the vast majority of people.
IIRC, 8 bit continues to grow. It's just that 32-bits is growing faster.
EDIT: And 8-bit can't seem to die no matter how much people try. TI's MSPM0 (Cortex M0+) chips at 39-cents this year (1k quantities) might give 8-bitters a run for their money finally though. But time-and-time again, 8-bit survives each challenge. I dunno why, it just keeps happening.
The previous Uno was a bit of a pain because it had two MCUs, with one essentially just acting as a USB interface to the other. The new Uno has the USB pins muxed both directly the MCU and the onboard ESP32, so that should be easier to work with.
In contrast, Pico is far more complicated for beginners, since it's not designed for that purpose. You need a breadboard in order to use it with anything else. And the toolchain is _far_ more complex for beginners! Hell, I had some issues setting up the Pico toolchain on my Mac (and I'm not a beginner), whereas with the Arduino, it was just "get the Arduino IDE". Yes, you can use the Arduino IDE with a Pico now, but that wasn't that way from the beginning, and the fact that you can do so now is A) not because Raspberry Pi released any official method and B) because the Arduino IDE exists in the first place specifically to support Arduinos.
In short, as I said before, they are different products with different audiences, and therefore different design goals.