Raspberry Pi Pico W Launches For $6

[rclark]

Well, regardless of pros and cons, I'll probably pick up a couple just for fun as I like the Raspberry Pi and Pico RP2040 eco system . Like potato chips. I just picked up a couple of RP2040 powered Servo boards for fun too. Control up to 18 servos per board. Have a few other RP2040 boards from Spark Fun.... Ready to control the world ... If I can think of a project to use them in ....

[LubosD]

One point I didn't see mentioned in the discussion yet: this module is twice as big as ESP32.

[tuxd3v]

But yeah -- I agree with the general sentiment of 'just another option to pick', at least until the C5 shows up with 5Ghz wifi.

Edit: Also, the development situation is a lot easier; espressif's SDK used to be a nice makefile setup, but now they've switched over to cmake to make the windows users' lives easier, and the pico is ridiculously easy to support with gcc-arm-none-eabi being around in most distros since like 2016.

I think that toolchains was one of the things that really pushed arm up..
They are free, and you can get them via package manager or just go to download section of ARM and download the binaries or the source code and compile the toolchain..
Its free..
While that is indeed an advantage in favour of ARM, when it comes to RISCV, arm doesn't have any more that advantage, providing that your ucontroller follows the standard at 100%.

However a case maybe can be made, in favour of ARM and that case has todo with the interrupt controller.
The vectored NIVC of arm is hardware accelerated..which means that all of the burden to make irqs, what we call the house keeping burden, is made in hardware, which makes it fast.
However I don't know how many nested leves of interrupts arm can do( hardware accelerated )..

Till now I only saw in riscv world, wch( wch32v307 ) providing a hardware accelerated FPIC, which is faster than arm NVIC, 8 or in some cases 6 cycles of burden for a interrupt..
But wch, only has 3 nested levels of accelerated hardware( the others are software based )...
So if you happen to have the case were 4( or more ) interrupts happens almost at same time...nested, the last one will be done in software( in case of 4 ), and an interrupt in software is around 41 cycles in riscv..

So it would be nice to know, how many cycles takes the cores shipped in expressif riscv esp32-{Cx,Hx}, well.. and also in the extensa cores esp32{,-Sx}..to do a interrupt request and return from it..

Long story short,
In my view, maybe rpi pico w has the edge on interrupts latency..maybe, it will depend on the interrupt controller provided by the riscv cores expressif has..I believe that arm accelerated interrupts are around 12 cycles or so..

[Kamilion]

I don't see nothing in rpi pico that can make it a competitor to esp32 modules, however its good to have a fallback solution, just in case..

The only real advantage is the extreme overclocks you can throw at the RP2040; 252Mhz lets you yeet out clean DVI, and they can hit impressive 400Mhz clocks with 1.3v 'overvolting' if you change the flash divisor.

It's also useful for existing projects that use a pico, like adding wireless to that N64 flashcart without giving up any of the PIO pins.

I haven't yet seen anyone try to push the C3's riscv to it's clock limits. But that would probably also have knock-on effects to wifi and bluetooth stability; and overclocking espressifs in the past typically end up with unusable radio layers. Pico-W's independent radio chip should avoid such foolishness.

But yeah -- I agree with the general sentiment of 'just another option to pick', at least until the C5 shows up with 5Ghz wifi.

Edit: Also, the development situation is a lot easier; espressif's SDK used to be a nice makefile setup, but now they've switched over to cmake to make the windows users' lives easier, and the pico is ridiculously easy to support with gcc-arm-none-eabi being around in most distros since like 2016.

[tuxd3v]

ESP32, ESP-Sx,ESP-Cx,ESP32-Hx are superior solutions to this rpi pico.
ESP32XX offers more functionality, lower power consumption, and Wifi + Bluetooth 5.0, and sooner with ESP32-Hx thread/zigbee..
ESP modules long ago support coprocessors, so rpi doesn't bring nothing knew..

I don't see nothing in rpi pico that can make it a competitor to esp32 modules, however its good to have a fallback solution, just in case..

[JEBjames]

Michael

I think you left out the paragraph mentioning the new Pico H and Pico WH models?

[JEBjames]

Michael

Take this with a grain of salt. But there's probably a typo:

"thi 802.11n WiFi" should be "this 802.11n WiFi"

[Quackdoc]

If I need WiFi ... then why would I use this rather than ESP32?

Even after you add an external WiFi module to Raspberry RP2040,
ESP32 still has more more peripherals (and more memory, consumes less power in deep sleep mode, etc.) and is also cheaper.

cheap and I'm able to actually get my hands on it in a reasonable timeframe, rpi picos can be much easier to get a hold of at appropriate prices. with buying other systems, you order from 20 different sites, wait a few months, then cancel which ones don't come in, or you pay a 2-8x markup to purchase it from amazon.

[willmore]

The community is probably one of the largest strengths of the RPi ecosystem. The only toxicity I see is falling from your lips.

Anyhow, I'm rather amused to see that the WLAN chip has both a Cortex-M3 and a Cortex-M4 core, next to the RPi's pair of Cortex-M0(+) cores.

Maybe you should get more experience before commenting. The Rpi community is a bunch of inexperienced fools who paid too much for an SBC and who have no idea what to do with it or how moderated by liars.

The ESP8266 gained popularity because someone realized that the 'serial to wifi' chip that they were adding onto their systems was actually a better processor than their original system. It had very few GPIO as there wasn't much of a need for them when your purpose in life was to speak serial on one side and RF on the other. The ESP32 was the first product from Espresif which took this wider perspective in mind. They hired people from the community (Sprite and others) who helped guide the development. The result is a family of chips which are way more useful than this glued together crud from the Rpi foundation ever will be.

[hikingpete]

Not if I have to deal with the toxic community, these will never be worth it.

The community is probably one of the largest strengths of the RPi ecosystem. The only toxicity I see is falling from your lips.

Anyhow, I'm rather amused to see that the WLAN chip has both a Cortex-M3 and a Cortex-M4 core, next to the RPi's pair of Cortex-M0(+) cores.