Arduino UNO - Rev4

[Ed.Kloonk]

Is there enough power to turn that LED array into a scrolling text display?

CPU power, I mean. The hacker in me would only find the LEDs useful for spewing out error msgs.

[brucehoult]

Is there enough power to turn that LED array into a scrolling text display?

CPU power, I mean. The hacker in me would only find the LEDs useful for spewing out error msgs.

Of course. Even an AVR is powerful enough to animate thousands of LEDs at update rates above human perception.

[Picuino]

The led matrix has a rather bizarre connection.

https://docs.arduino.cc/resources/datasheets/ABX00087-datasheet.pdf

[Ed.Kloonk]

I wish they'd strived for a little LCD instead.

[liaifat85]

I hope ML projects will be easier now.

[brucehoult]

The led matrix has a rather bizarre connection.

https://docs.arduino.cc/resources/datasheets/ABX00087-datasheet.pdf

If you were following this thread you'd already know what is going on there :-)

https://www.eevblog.com/forum/beginners/how-to-connect-a-7-segment-2-5-digit-module-with-only-5-legs

[bookaboo]

The Renesas sales team must be good, the ESP32-S3 has more than enough specs to be the sole processor in this application. Not to mention USB JTAG on board.

[ArdWar]

The Renesas sales team must be good, the ESP32-S3 has more than enough specs to be the sole processor in this application. Not to mention USB JTAG on board.

How absurd that will be if there's an onboard debugger? With Renesas that'll probably a RA4M2, the same as one used in their JLink-OB which is a cortex M33 core....
The one actually "usable" by user ends up being the weakest core in the whole board. Kind of weird for a non manufacturer-subsidized devkit (supposedly)

The whole thing reminds me of this

[AndyBeez]

I had not noticed this turd, but thanks to this thread I did. I thought Arduino had OLED and touch screen shields, a decade ago? Did a team of marketing executives throw a Raspberry Pi, a BBC Micro:bit and the ESP32 derivates into a feature blender and conceive their very first single board 'love child'? Ultimately, this looks like a consequence of lockdown, chipageddon and a brand that has lost its direction.

+EDIT: Yes they have decent after market display shields ...
Adafruit NeoPixel Shield for Arduino - 40 RGB LED Pixel Matrix
https://www.adafruit.com/product/1430

Adafruit 1.8" Color TFT Shield w/microSD and Joystick
https://www.adafruit.com/product/802

[xrunner]

I had not noticed this turd, but thanks to this thread I did. I thought Arduino had OLED and touch screen shields, a decade ago?

I don't get it either. I'm doing a small project right now and I have a small OLED screen right there on my bench where the Mega board is, so I don't have to use the serial link on my computer across the room to see what readings I'm dealing with

Put a small OLED display on the UNO 4 board where the LED array is and if nothing else you have a debugging display right there where it's convenient.

[PCB.Wiz]

I had not noticed this turd, but thanks to this thread I did. I thought Arduino had OLED and touch screen shields, a decade ago?

I don't get it either. I'm doing a small project right now and I have a small OLED screen right there on my bench where the Mega board is, so I don't have to use the serial link on my computer across the room to see what readings I'm dealing with

Put a small OLED display on the UNO 4 board where the LED array is and if nothing else you have a debugging display right there where it's convenient.

It may be as simple as price and reliability/ruggedness ?
Lcsc have LED from $0.0017 ! ( 1M + in stock, for one part code)

[tszaboo]

OK, so I've been thinking about the odd choice for the microcontroller (amongst other things). And I have to conclude that there aren't that many Cortex M3 or M4 MCUs that would work from 5V. They clearly wanted to keep the compatibility with their ecosystem of boards. And after doing a search, you only get a few Cortex M4 parts with 5V, some NXP (expensive ones though), some Infineon parts (Insert vomit emoji) and some parts that come from the smaller makers.
If they would accept a cortex M0, then there is the SAM C21, probably much better choice,

[ataradov]

This is not how Arduino works. They don't just select the device as a regular project would. They have enough of a pull that the can talk to the vendors. This time Renesas wanted to play ball. You can assume that they are getting the parts at a huge discount, possibly at or below the cost. It is also possible that Renesas initiated the design in a first place.

[SiliconWizard]

Yep, when you have that kind of traction, vendors will knock on your door constantly. For them that's good marketing, even if they don't target hobbyist markets.

[tszaboo]

This is not how Arduino works. They don't just select the device as a regular project would. They have enough of a pull that the can talk to the vendors. This time Renesas wanted to play ball. You can assume that they are getting the parts at a huge discount, possibly at or below the cost. It is also possible that Renesas initiated the design in a first place.

You would think, but then again, how many suppliers would run a design of their choice on a completely different node for them? There is a reason generally a microcontroller doesn't run on 5V anymore. Getting into the Arduino ecosystem would be a big deal, but not that big that you roll an entire series with different transistors in it. Judging by the fact they sold 10M Unos in about 10 years the numbers are not that high. And while a lot of products ended up using an ATMEGA328 because it's on the Uno, this is not 2010 anymore. There are just so many compatible boards with it, and not to mention the fact the parts are going to be unobtanium from Renesas anyway. They have like 5K pieces as global stock?

[ataradov]

Getting into the Arduino ecosystem would be a big deal, but not that big that you roll an entire series with different transistors in it.

They are using an existing device, nobody is making new silicon.

Judging by the fact they sold 10M Unos in about 10 years the numbers are not that high.

For the IC vendor getting into the Arduino ecosystem, especially as a primary mainstream board is huge. Sales to Arduino do not matter, sales that it generates after that are significant.

and not to mention the fact the parts are going to be unobtanium from Renesas anyway. They have like 5K pieces as global stock?

Well, they would have to work on the inventory if the boards become popular.

[tszaboo]

Getting into the Arduino ecosystem would be a big deal, but not that big that you roll an entire series with different transistors in it.

They are using an existing device, nobody is making new silicon.

Judging by the fact they sold 10M Unos in about 10 years the numbers are not that high.

For the IC vendor getting into the Arduino ecosystem, especially as a primary mainstream board is huge. Sales to Arduino do not matter, sales that it generates after that are significant.

and not to mention the fact the parts are going to be unobtanium from Renesas anyway. They have like 5K pieces as global stock?

Well, they would have to work on the inventory if the boards become popular.

Then I don't get your point at all. I was saying there is a limited selection of 5V MCUs with Cortex M4 or similar cores, and only like 3 or 4 vendors total to select from.

[ataradov]

Yes, there is a limited selection. They selected one of them. I don't see an issue here.

All I'm saying is that SAM C21 is not a better option because nobody was selling them to Arduino at cost.

[PCB.Wiz]

This is not how Arduino works. They don't just select the device as a regular project would. They have enough of a pull that the can talk to the vendors. This time Renesas wanted to play ball. You can assume that they are getting the parts at a huge discount, possibly at or below the cost. It is also possible that Renesas initiated the design in a first place.

Yes, you are very close.
Renesas are clearly motivated here and $10M helps buy you considerable influence too. 
I see it as a smart move.


This news helps explain
https://www.renesas.com/us/en/about/press-room/renesas-announces-investment-popular-open-source-company-arduino-access-huge-developer-community
"Renesas is investing $10 million into Arduino’s Series B round of $32 million"

[PCB.Wiz]

OK, so I've been thinking about the odd choice for the microcontroller (amongst other things). And I have to conclude that there aren't that many Cortex M3 or M4 MCUs that would work from 5V. They clearly wanted to keep the compatibility with their ecosystem of boards. And after doing a search, you only get a few Cortex M4 parts with 5V, some NXP (expensive ones though), some Infineon parts (Insert vomit emoji) and some parts that come from the smaller makers.
If they would accept a cortex M0, then there is the SAM C21, probably much better choice,

You are broadly correct, but there is also Nuvoton (like their M471R1E6AE, tho that has no DAC ) and Toshiba

A recent Toshiba example (June 26, 2023)
https://edacafe.com/nbc/articles/1/2013201/Toshiba-Introduces-ARM-Cortex-M3-Microcontrollers-TXZ+TM-Family-Advanced-Class-with-1MB-Code-Flash-Memory-Supporting-Firmware-Updates-without-Interrupting-Microcontroller-Operation

[ArdWar]

You are broadly correct, but there is also Nuvoton (like their M471R1E6AE, tho that has no DAC ) and Toshiba

A recent Toshiba example (June 26, 2023)
https://edacafe.com/nbc/articles/1/2013201/Toshiba-Introduces-ARM-Cortex-M3-Microcontrollers-TXZ+TM-Family-Advanced-Class-with-1MB-Code-Flash-Memory-Supporting-Firmware-Updates-without-Interrupting-Microcontroller-Operation

There's also NXP Kinetis KE series. I don't think it is that expensive. Still fairly normal price range IMO.

Not that any alternatives matter when they're already captured in by sponsorship.

[westfw]

Quote from: ataradov on Today at 01:57:18 pm

SAM C21 is not a better option because nobody was selling them to Arduino at cost.

I suspect most of the problem is that Microchip was having pretty serious supply-chain issues.  SAMC's ordered now are being scheduled for delivery in a year. :-(   None of them have native USB, either.

Atmel used to look pretty tight with Arduino, but perhaps Microchip is not so cooperative (few of the "new" AVRs have shown up in Arduino-class products, and the vendors seem to be converting to ... chips they can actually buy.  (rp2040, ESP, etc.)  (and PIC32 never showed up in many such boards, either.)

Renesas is a bit of a surprise; I don't recall them ever being inclined to servicing the "hobbyist market."

For Arduino, these are pretty nice boards at a good price point.  Less so compared to their younger competitors, but still...

[ppTRN]

I was so disappointed by the led matrix that I rage-bought a Raspberry Pi pico W.

Jokes aside, Renesas and RA world is pretty new for me so i look forward to dive into it's architecture, and surely the new Arduino UNO will be very helpfull.
It is clear that I am refering to a pure "accademical" view of things

Beside, since ESP32 pins are not available exept for the test points, why not driving the led matrix with it?

[Rudolph Riedel]

I received an Arduino UNO R4 MINIMA today and tried it with my EVE library which uses SPI for a TFT.
And the first results are terrible.
Just using spi_transmit(uint8_t data) the same code runs almost five times faster on the R3 as it does on the R4.
The logic analyzer trace reveals that there are 11µs pauses between SPI transfers on the R4 while the R3 has 1µs...2µs pauses.

The R4 SPI class is generic code and the comments date it to 2014 - and there is no DMA support.
The actual transfers happen in a binary blob called libfsp.a for which the source code is not provided in ArduinoCore-renesas.
The driver is probably running on highly generic code from Renesas with lots of runtime sanity checks.

[ataradov]

FSP source is here https://github.com/arduino/fsp/tree/8ec537bc86b1b71203d72d0cbb28e6e75b353ebd Build instructions are here https://github.com/facchinm/ArduinoCore-renesas

I would expect them to work on that a lot in the coming months. If they decide to stick with the Renesas libraries, then this board is as good as dead for general public. And I'm afraid Renesas would insist on that.

The code seems to be a typical bloatware, probably made worse by Arduino wrapping it into C++ classes. They would have to move all of that into native code.