PIC versus Atmel/ESP/STM/etc.

[0xFFF0]

Does a project with a PIC still make sense? I have the impression that the PICs are no longer properly supported. In comparison, there are 100x more code examples and projects with the new MCUs. Does a PIC still have an advantage?

[NorthGuy]

There are over 2000 different PICs. All different. Which one are you asking about?

There are also thousands of ways to use MCUs. You chose MCU based on your needs. What is your particular goal?

[AVI-crak]

There is little.
Pallets with PIC chips surprisingly have the same dimensions - as pallets with WCH chips. You can use the chip storage space without buying new racks.
And yes, PIC can be thrown away.

[jpanhalt]

And yes, PIC can be thrown away.

So can any other chip brand.  Can you suggest a brand that is indispensable?  If it works and is priced right (i.e., less) should be the determinants.  I do not include in that consideration fraudulent chips.

[mikeselectricstuff]

Every MCU has good and bad points, and also factors not directly related to the physical part.
e.g.
You can buy pre-programmed parts from Microchip Direct with minimal setup and programming costs.
Need a 32 bit MCU in a DIP or SOIC? AFAIK Microchip is the only vendor offering the former at least.
Most PICs are available in a choice of 3 or 4 packages
Want to use the same IDE and programming hardware ( and very similar peripherals) across a range  of 8, 16 and 32 bit parts, from a SOT23-6 to a 288 pin 200MHz parts - again Microchip is the only game in town

[0xFFF0]

The Arduino environment is growing and growing with tons of example whereas the MPLAB is a very sad party. Microchip simply has very little to offer.

[fchk]

Does a project with a PIC still make sense? I have the impression that the PICs are no longer properly supported. In comparison, there are 100x more code examples and projects with the new MCUs. Does a PIC still have an advantage?

Most people only know the old 8 bit PICs and forget that there are also very neat 16 and 32 bit PICs (complete different architectures). Think of 100 MHz dual core motor controllers etc.

And even today's PIC16 are much different from an old PIC16C50. They have an enhanced (more C friendly) CPU core, some have small programmable logic, numeric controlled oscillators, and other core independent peripherials. There is still a market for this.

[brucehoult]

The Arduino environment is growing and growing with tons of example whereas the MPLAB is a very sad party. Microchip simply has very little to offer.

Microchip also have AVR, Arm, and RISC-V parts. And of course some of those PICs are MIPS.

[brucehoult]

Need a 32 bit MCU in a DIP or SOIC? AFAIK Microchip is the only vendor offering the former at least.

SOIC is easy. 32 bit MCU in SOIC 8 pin (10c each) and 20 pin (14.6c each) right here:

https://www.aliexpress.us/item/3256804850399956.html

32 bit in DIP ... not easy. But here's one. 3324 in stock. Not even Microchip.

https://www.digikey.co.nz/en/products/detail/rochester-electronics-llc/LPC810M021FN8129/12102361

[PCB.Wiz]

Does a project with a PIC still make sense? I have the impression that the PICs are no longer properly supported. In comparison, there are 100x more code examples and projects with the new MCUs. Does a PIC still have an advantage?

PIC is a prefix, not a specific microcontroller.
There are 12/14/16 bit opcode PICs and PIC18 and PIC32 plus dsPIC ....

Of course, older MCUs fade as their FAB support tails off, and the price bathtub curve discourages new design ins, but they are still sold to mature products where the inertia of things like approvals may make a re-spin too costly.

Few would select a PIC12 for a new design in 2023, but the very cheap asian clones in lowest pin counts often use close to brain-dead cores.

If you only need a smart timer, any MCU will do ....

[SiliconWizard]

The Arduino environment is growing and growing with tons of example whereas the MPLAB is a very sad party. Microchip simply has very little to offer.

Fixed: "has very little to offer non-Atmel MCUs that aren't supported by Arduino."

Microchip has been owning Atmel products for a while now. Most Atmel products (even the 32-bit ARM) are supported by Arduino.

As to historical Microchip MCU lines and models (which are many as others have said), they usually aren't, although I'm not 100% sure none of them aren't (for instance the 32-bit ones.) For various reasons.
But they can still be pretty useful, there's a whole world outside of Arduino. You don't even need to wear a mask to discover it!

[brucehoult]

Few would select a PIC12 for a new design in 2023, but the very cheap asian clones in lowest pin counts often use close to brain-dead cores.

The 3c Padauk chip with 1k instructions of OTP and 64 bytes of RAM certainly has uses. 8 bit PIC-alike instruction set, supported only by a custom not-quite-C language and compiler. Can't debug it.

If you splash out 10c you can get a 32 bit RISC-V CH32V003 with 16k of flash (5000-6000 more powerful instructions), 2k RAM. You can use standard C/C++ (or Rust if you want) via upstream gcc or LLVM. You can use gdb.

Reckon it's worth the extra 7c, unless you're doing incredible volumes.

[SiliconWizard]

8-bit PIC MCUs are pretty good, are very well documented, have good support, product ranges are very wide with tons of configurations and some are ultra low power, some can withstand 150°C.
Try using an obscure chinese MCU in your next product - it may work fine but good luck with documentation, support, or any guarantee from the manufacturer. So different use cases, different requirements.

And simple 8-bit MCUs can still have their place - sometimes a very simple architecture with simple peripherals is just better to lower the overall complexity of a system.
You can pretty much fully master how an 8-bit PIC works, down to the clock cycle. Good luck with even a Cortex-M0.
Sometimes it doesn't matter, sometimes it does.

[dobsonr741]

I think we should be talking open source tools and libraries support for hobbyist purposes, maybe for light commercial use. This is where the PIC goes under. Not Microchip, happily grabbing a nice share with the Atmel lines.

[JPortici]

Does a project with a PIC still make sense? I have the impression that the PICs are no longer properly supported. In comparison, there are 100x more code examples and projects with the new MCUs. Does a PIC still have an advantage?

Who cares!
Besides, how many of those examples you mention are close to the metal so you really know how to program them? How many threads do you see in the various forums in which the person is stuck due to their lack of understanding of the library or wants to do something the library doesn't let him or the lack of information in the documentation? There could be ten thousand projects for new MCUs but if the only support and programming tools i get is platformIO or arduino, thanks but no thanks i'll grab any for which i have real documentation and a C compiler

Also: If we look at sales number in the west microchip is huge, there seems to be an inverse relationship between popularity and actual number of devices sold

Also: The newest PIC18F Q20 will be one of the first publicly available devices with I3C support, PIC18F keep getting newer and more advanced devices, even though we really are getting at the end of the line (there is no more space available for SFRs, they are resorting to banking the banks. yikes!) why invest in the architecture if it's dying due to lack of popularity? (hint: all the parts we use are firstly designed for clients that need them in huge numbers and we get the leftovers)

Also: As already said, most people think about "PIC" or "AVR" with the 16F84A/876A/18F2550 and the 328P respectively. Ancient parts with decades on their shoulder, the world have moved on and we have excellent 8/16bit parts, pretty good (ATSAM,PIC32MX) and meh (PIC32MZ/MK -- because of the erratas, and because MIPS is more or less dead, and experts that talk online are much harder to find that ARM) 32bit parts

[westfw]

32 bit in DIP ... not easy.

A bunch of the PIC32MX series is available in SPDIP-28.  Up to 256k flash and 64k RAM.
---


I don't think I'd start a new project on an 8bit PIC (PIC12, PIC16, PIC18) unless I specifically needed some unique peripheral feature that wasn't available on AVR, ARM, or RISC-V chips at similar cost (There's a PIC18 with Ethernet MAC and Phy, for instance.  Most hobbyist projects don't end up using "special" peripheral features, but they were included on the chip for SOME application!)  That doesn't mean that 8bit PICs have lost their usefulness, they're just ... less pleasant to work with.  (that bank-switched RAM.  Compilers don't like it either.)  (that said, many applications won't run into the unpleasantness.)

[mikeselectricstuff]

I think we should be talking open source tools and libraries support for hobbyist purposes,

Hobbyist volumes are insignificant. I don't care if tools are open source as long as they work.

[brucehoult]

32 bit in DIP ... not easy.

A bunch of the PIC32MX series is available in SPDIP-28.  Up to 256k flash and 64k RAM.

Microchip offering some was assumed in the post, as mentioned in the part you snipped. The challenge was NOT in finding a Microchip 32 bit MCU in DIP -- we already knew they have them -- but a NON Microchip.

[woofy]

The Arduino environment is growing and growing with tons of example whereas the MPLAB is a very sad party. Microchip simply has very little to offer.

I use a lot of PIC's from PIC16/18's through to PIC32MZ. Microchip are continually bringing out new chips with new features/peripherals even in the PIC16 flavors. They would not do that if there was no demand.

[ppTRN]

I personally uses old pics like 16Fxxx and 18Fxxx for educational purpose. I must say tho that i find still helpfull pics from the dsPIC30F and dsPIC33 series. They are fast and full of hw resources. I never worked with newer pics but i am sure that microchip is keeping up with the market even tho STM32 seems more diffuse.

PS: have you noticed that in many dyson products, in the ads where you see the internal PCB, there is a PIC uC!!

[mikeselectricstuff]

The Arduino environment is growing and growing with tons of example whereas the MPLAB is a very sad party. Microchip simply has very little to offer.

Sad in what way ?
MPLAB is a tool for professionals.
Arduino is for teaching and hobbyists.
Microchip care about commercial design-ins, not whether  hobbyists chooses PIC or ESP. They've been doing pretty well on that.
Having said that, Microchip have been more hobbyist-friendly than most other MCU manufacturers over the decades.

[WatchfulEye]

Also worth remembering that a number of the atmel MCU lines have been experienced some cross pollination, and the progeny are now called PIC.

For example the PIC32CMLx series with cortex M23 core is so similar to the atmel SAML2x series that you'd better check the docs carefully to spot the changes (and which errata have been fixed and also which ones haven't).

[westfw]

Microchip have been more hobbyist-friendly than most other MCU manufacturers over the decades.

That was their big advantage for a long time.  PICs (of the 8bit variety), and their development tools, were available to hobbyists at reasonable prices for probably a decade or more before other vendors caught on to that whole "catch them when they're young" concept.  First small (and relatively cheap) UV-eraseable chips, first OTPs (cheaper), first serial programmable EEPROM-based chip, first to appear at hobbyist vendors (yeah Parallax!)
Atmel had the "first" flash-based microcontrollers, but they were nearly impossible to obtain (on the hobbyist market) for a long time, while the PIC16C84 was gaining mind-share.


Now-a-days, we are all spoiled.  Digikey grew up, other formerly-professionals-only vendors automated to the point where they were willing to handle small orders, and thanks to Internet, many manufacturers will sell direct to anyone.
It's hard to imagine the days when the entry cost to working with a microcontroller was several hundred dollars worth of board-level products and EPROM programmers, plus several thousand for a computer (or similar equipment) to connect it to (rare!) (pre-inflation $$, too) instead of the cost of a fast food meal for something that plugs into the computer you already have (to watch cat videos and argue with people, you know.)

[brucehoult]

It's hard to imagine the days when the entry cost to working with a microcontroller was several hundred dollars worth of board-level products and EPROM programmers

I remember them, and it was daunting and kept me out.

I designed and built some things using 74xx, 555, 741 and transistors when I was in high school in the late 70s. There was a company in Auckland, David Reid Electronics, that sent a huge bound catalogue (with a Moire pattern on the cover) to anyone who asked for one, including kids. You could post an order to them and usually get your stuff in a week.

In my 3rd year at university (1983) I worked on an extra-curricula project with two others, designing and making a wire-wrapped M6809 with DRAM (with software refresh), a UART, and an 8" floppy interface. One of my contributions was the clock logic, which could independently make each of the HI and LO phases either 0.25µs or 0.5µs based on the hi bits of the address bus: 2 MHz for RAM, 1 MHz for peripherals, and 1.33 MHz (asymmetrical) for the Motorola monitor ROM based on a careful study of its timing diagram :p We also wrote a BCPL code generator for it (running on VAX), and a mini multi-tasking OS.

After that I didn't touch hardware for a LONG time.

I became aware of BASIC Stamp in the late 90s I think, but that didn't appeal. PICKit seemed too expensive and fiddly. A friend got the first Arm MBED in 2009 I think, and I was thinking about that when someone locally in NZ started selling Arduino Uno and BOOM I was in.

People criticise Arduino a lot here, but the on-ramp is SO GENTLE, and yet it doesn't constrain you in any way. You don't have to use their editor. You don't have to use their libraries (or only use parts of them). You can see exactly what commands it is running, so you can easily copy&paste the right commands to run gcc or avrdude directly if you want. But it is all there to be used up to the point that you don't want it. I actually often use the Arduino IDE terminal emulator in preference to minicom or something even when I'm working with boards that aren't supported in the IDE.

[Kim Christensen]

It's hard to imagine the days when the entry cost to working with a microcontroller was several hundred dollars worth of board-level products and EPROM programmers

I remember them, and it was daunting and kept me out.

It was a huge challenge for the hobbyist doing it on the cheap. My first foray into that world was with the Z80 in the mid 1980's. Built a system with battery backed up RAM for the program memory. Then built a TTL/CMOS keypad interface so I could enter HEX codes into the RAM and run it. (I didn't own a computer) Wrote all the code out on paper for the monitor/debugger and hand assembled it. When it was done, studied the 2716 eprom datasheet and wrote some code so it could copy the monitor/debugger code to the eprom. This was my final year project for tech school and I probably wouldn't have done it otherwise.

My first microcontroller was the 68HC11 in the 90's. I chose it because of the built in bootloader that let you program it via the serial port which was way easier than the Z80. But being cheap once again, I ended up making my own assembler in BASIC which ran on an old Tandy COCO2 and later on my first PC.

So when I discovered PICs and MPLAB  I was pretty stoked. Things suddenly got a lot easier. I still have a quite a few of them in the bin and will occasionally grab one for a project because a micro in the hand is worth 100 on backorder.
But if I was starting from scratch, I'd choose whatever is currently popular and fits my needs. It's always a tradeoff about ease of use, software support, capabilities, tools you already have, etc. No shame in using a PIC or dated tech if it works for you.