I'm talking about the Pickit 2.
~2006...The 2010 Digikey catalog lists a "PICKit 2 Starter kit" for $50 (and another one, same name, for $25 ??) (
https://archive.org/details/digi-key-catalog-us-2010/page/691/mode/1up ) (2006 is pretty late, compared to what we're talking about. 2010 is "post-Arduino.")
At almost exactly the same time that Arduino showed up, there was also the
"USB Bitwhacker" - it used a USB-capable PIC18 and a bootloader, so it plugged directly into a PC and allowed uploading code directly from a "modern" PC. Scarily similar to the Arduino in "hardware ease of use", but it lacked the simplified software abstract or a focused target audience. (Sample software was written in several BASICs, C, or assembler. It was sort-of aimed at existing PIC hobbyists, rather than a new audience (he says, with his marketing hat on, despite not having any marketing experience! (The whole phenomena of would-be "Arduino killers" is fascinating, marketing-wise. Someone should do their MBA thesis on it.))
I'm talking about the Pickit 2.
~2006...The 2010 Digikey catalog lists a "PICKit 2 Starter kit" for $50 (and another one, same name, for $25 ??) ( https://archive.org/details/digi-key-catalog-us-2010/page/691/mode/1up ) (2006 is pretty late, compared to what we're talking about. 2010 is "post-Arduino.")
Yes, the PICKit 2 was well a truly in the era of cost cost programmers, everyone had them by then.
The PICStart was the early 90's, and it was hundreds of dollars, that's why there were low cost clones like the WARP-13.
The PICStart Plus was 2000.
I regard PICs as mainstream, even if the 16F84 is years out of date in 2025.
Contrast this with the 16F84's contemporary, I would regard Atmel's offerings like the AT90S2313 as non-mainstream, though today's incarnation ATtiny2313 seems to have a following that mirrors the hobbyist 16F84 former heyday.
I regard PICs as mainstream
They are, we are just taking the piss because they seem to have all but vanished from the hobbyist/maker world they once dominated.
To be fair, even AVRs which have had been more common in the hobbyist world in the form of Arduinos have all but vanished lately. Its all the ESP8266s and ESP32s now with WiFi and Bluetooth which dominate the hobby scene.
Its weird, hobbyists in the past preferred AVRs because of its relatively simple architecture and simple peripherals (timers and such). In the ESP world, hobbyists enjoy much more powerful peripherals but they seldom use them
Microchip PIC was THE microcontroller of choice way back,
We had once an analog(PAL) pay-TV protected by a clipped sync pulse.
Solution was 8 pin PIC.
Used a lot of PIC chips in last 30 years, PIC16 & PIC18, but moved to Cortex-Ms about 10 years ago.
Did a lot of work with various EFM32s, TG11, PG1, PG12, GG11 but now using mostly STM32s.
My new designs going to STM32H562, STM32H723/743.
Having gone up the hill with the STM 32F4xx over the past 5 years or so, I would simply not touch anything else, unless a) I needed way more power (and then a different system architecture may be the way to go) or b) the 5 quid price was too high (and then one is looking at a very high volume product with all the related commercial implications (e.g. is the customer creditworthy for a £1M order).
But then, as I often say, this is my own business, so I need to learn new stuff like I need a hole in the head
For an employee, new stuff == bread on the table at home, and a better looking CV.
The most weird chips I've used were the Atmel 90S1200, and the Hitachi H8/323. I've still got about 10k of the latter sitting there, surplus stock.
I regard PICs as mainstream
They are, we are just taking the piss because they seem to have all but vanished from the hobbyist/maker world they once dominated.
2 years ago, I was using a PIC32MX250 for some WS2812 LED projects, must be quite non-mainstream..
I sometimes do a social media search to see if people on Reddit/YouTube/Hackaday.io/etc. are posting *anything* with Microchip parts. But apart from some people
staying on 8-bit PICs or AVRs, I see practically zero dot zero mindshare elsewhere.
Looks like everyone else has moved on to 32-bit chips, in particular STM32, Nordic, RPi Pico and ESP32.
And NXP and SiLabs are not too uncommon in the more professional space.
I'd argue the most non-mainstream controllers I use are STM32U5 and STM32L0/4. Mostly because they are relatively new (the U5 that is), niche and most people will prefer to buy raw horsepower over low power features. I like the challenge of doing battery powered projects.. it often requires a total system view and in-depth look how the MCU works to get a good result.
My favourite nearly-non-mainstream mcu is the STM8. It's an interesting little chip. Feels like a hybrid of the 'old skool' 8-bit MCU's, but with a modern peripheral set and driver library, and modern debug capabilities. It's not supported by GCC, and the vendor doesn't offer drivers that are fully compatible with an open-source compiler, so, getting a toolchain set up using SDCC is an excercise.
I like playing with uber-cheap chinese chips. You know, the kind with basically non-existent documentation, and lotsa little quirks to 'discover' along the way. Learned how to stand up a toolchain, and developed a lot of re-usable skills. The WCH series is interesting - they have 8051s, ARMs, and a few flavours of RISC-V chips.
I think the bottom line here is to choose a part which really fits the bill (in terms of functional requirements, price, availability and software NRE costs) instead of just using what anyone else seems to be using. Some people are really vocal on forums about what they use but keep in mind that people also seek confirmation of their choices. Another issue I have spotted over the decades is that many hardware designers tend to choose cheap microcontrollers (and processors) but totally forget about the time it takes to write the software. Some of this is driven by the fact that hardware can be priced very easely while software development time doesn't seem to receive the same amount of scrutiny where it comes to costs. Last but not least, some manufacturers (like NXP) have their own support forums where people who use those parts have their support community aided by the manufacturer's product specialists. As a result those parts are not discussed on regular forums that much and may seem rarely used.
the bottom line here is to choose a part which really fits the bill
Being able to "evaluate" the vast array of available parts in order to make such a choice is a luxury that is all-too-frequently unavailable to the actual engineers.
Decisions get made by people further up the food chain, who are sometimes basing their decisions on "other factors."
- We're already using this chip family, and have spent big bucks on tooling.
- We're already familiar with the recommended development environment.
- We trust this vendor. (There are "second sources!") (or "don't trust.")
- It's pin compatible with...
- It's part of a big family of chips, in case we guessed wrong about requirements.
- The hardware guys liked it (but the SW guys don't!)
- The Software guys liked it (but the HW don't!)
- We don't have time to do an "Evaluation"! Just make it work!
- Undue influence of sales reps, marketing departments, and advertising.
- Someone on the team studied this architecture when they were in Univ ("Hey! PIC32 is MIPS!")
- Our competitor is using ...
And you know, usually it's OK. because the reality is most projects aren't that highly dependent on the deep internals.
My theory is that this is why "hobbyists" and other Amateurs are so important. THEY have the time and inclination to "fiddle" with multiple chip families, and communicate their findings, and etc. Consider how much you know about the RPi "PIO" state machines vs the NXP/Freescale/Motorola "TPU", for example...
I guess the smaller companies that do development help as well. It's the BIG companies where choices seem to get minimized (at least, until they're big enough to spend resources doing "nothing but evaluation.)
Some big-company engineers have a hobbyist side as well, or will happily spend "less than their budgetary authority" or personal funds on "evaluation boards" to experiment with. But ... not all of them, or even most of them.
My favourite nearly-non-mainstream mcu is the STM8. It's an interesting little chip. Feels like a hybrid of the 'old skool' 8-bit MCU's, but with a modern peripheral set and driver library, and modern debug capabilities. It's not supported by GCC, and the vendor doesn't offer drivers that are fully compatible with an open-source compiler, so, getting a toolchain set up using SDCC is an excercise.
I like playing with uber-cheap chinese chips. You know, the kind with basically non-existent documentation, and lotsa little quirks to 'discover' along the way. Learned how to stand up a toolchain, and developed a lot of re-usable skills. The WCH series is interesting - they have 8051s, ARMs, and a few flavours of RISC-V chips.
That STM8... I gave it a try. You have to ask some French company in email a license to a free compiler that only works for one year. And then you have barely any examples. It was quite frustrating and gave up in the end. I rather pay a bit more to some much more polished development.
the bottom line here is to choose a part which really fits the bill
Being able to "evaluate" the vast array of available parts in order to make such a choice is a luxury that is all-too-frequently unavailable to the actual engineers.
Decisions get made by people further up the food chain, who are sometimes basing their decisions on "other factors."
- We're already using this chip family, and have spent big bucks on tooling.
- We're already familiar with the recommended development environment.
- We trust this vendor. (There are "second sources!") (or "don't trust.")
- It's pin compatible with...
- It's part of a big family of chips, in case we guessed wrong about requirements.
- The hardware guys liked it (but the SW guys don't!)
- The Software guys liked it (but the HW don't!)
- We don't have time to do an "Evaluation"! Just make it work!
- Undue influence of sales reps, marketing departments, and advertising.
- Someone on the team studied this architecture when they were in Univ ("Hey! PIC32 is MIPS!")
- Our competitor is using ...
And you know, usually it's OK. because the reality is most projects aren't that highly dependent on the deep internals.
In a role at Thales I was the only electronics engineer in the department, so I got the chose whatever I wanted.
For one project I needed 4 serial ports and wanted to implment an easy RTOS. I saw an ad for the Rabbit 2000 somewhere and it had both of those, so Rabbit 2000 it was.
That's pretty obscure, but still available:
https://www.digi.com/rabbit
I was the only electronics engineer in the department, so I got the chose whatever I wanted.
how long did you have to make the decision? :-)
I saw an ad ...
Oh! Deep research there!
(I rather liked Rabbit, as long as you were content with a Z80-like architecture and 1MB (?) of memory. They had a clear target audience, and seemed good at addressing their needs. And they did "modules" before it was cool.)
@Selectech - may I ask why you switched to ST? I know they are cheaper but this is a pro only if you manufacture a very lot of devices. Return on invest...
My experience with ST in general is not that good. Quality wise. Their promises are... Never dig deep into STM32 but hardware friends told me things about 'did not reach specification even on standard temperature' (beginning of STM32) and in datasheet well hidden port pin restrictions, ...
In contrary since when I work with EnergyMicro/SiLabs I always found the support, clear structured peripherals, datasheets, AppNotes, astonishingly better than all what I saw before from many years working with Microchip. Of course I'm biased after about 15 years playing with EFM32 even for my hobby projects.
I never did try the Rabbit devices even though I was a Z80 fan boy back in the day.
I've used the Intel MCS-48 series including the 8035 and the first EPROM version the 8748 before moving on to the 8031 family.
The most obscure controller I've used is possibly the Dallas 87C520. I needed the speed, 33MHz and only 4 clks/instruction instead of the 8031's 12.
Obscure but probably considered mainstream in its day. OTP versions are still available, if you want to pay £45 per device.
@Selectech - may I ask why you switched to ST? I know they are cheaper but this is a pro only if you manufacture a very lot of devices. Return on invest...
My experience with ST in general is not that good. Quality wise. Their promises are... Never dig deep into STM32 but hardware friends told me things about 'did not reach specification even on standard temperature' (beginning of STM32) and in datasheet well hidden port pin restrictions, ...
In contrary since when I work with EnergyMicro/SiLabs I always found the support, clear structured peripherals, datasheets, AppNotes, astonishingly better than all what I saw before from many years working with Microchip. Of course I'm biased after about 15 years playing with EFM32 even for my hobby projects.
I had a project with an STM32F103 (I think), it was used to have USB and CAN at the same time. CAN for communication with the rest of the system, USB for programming and debugging. So these share the same pins, but you can remap one of the ports. The PCB was made that way. But on page 76583 of the reference manual, in a footer with 8px letters it tells you that these two cannot actually be used at the same time because they share some memory space for the buffers or whatever.
I had something similar with a Silabs micro recently. Every single peripheral can be mapped to every single pin, except the second I2C.
EFM32GG11s have been and continue to be a really good part for us.
Considering at some new STM32H5 & H7 parts for more serial peripherals and higher clock speed. Downside with new STM parts tends to be no bit-banding, which we've used to very good advantage with the EFM32s, but clock speed generally makes up for that.
STM32G070 have become very handy for smaller embedded card projects and we use those a lot.
We use JLC assembly and getting EFM32s to JLC has been problematic at times. STM32 parts generally directly available in JLC parts stock.
Not to go off topic but wanted to mention that the gentleman that did the AVR instruction set and later Rabbit cores went on to clone the HP41 calculator in an FPGA. Still sells a kit to build a super-charged HP41 known as the HP41CL.
HP41CL detailsDeveloper CV
That STM8... I gave it a try. You have to ask some French company in email a license to a free compiler that only works for one year. And then you have barely any examples. It was quite frustrating and gave up in the end. I rather pay a bit more to some much more polished development.
That would be Cosmic. I've never bothered with it; SDCC all the way. It does have some warts, but it works well enough. Actually, by now, it's probably the most polished and rounded compiler available for STM8, warts notwithstanding.
I've never had problems finding examples. Lots of various things on github, or in the nucleo/discovery board SDKs from STM. Code documentation is decent. If you set up an editor with code completion, it's pretty easy to just type something like 'spi' and scroll through the suggestions to find a register, struct, enum, macro or function you need. Its also pretty straightforward to develop bare metal style straight out of the RM.
(I rather liked Rabbit, as long as you were content with a Z80-like architecture and 1MB (?) of memory. They had a clear target audience, and seemed good at addressing their needs. And they did "modules" before it was cool.)
I didn't care about the architecture, C is C, I just needed serial ports and the easy to use RTOS that came with it.
Microchip PIC was THE microcontroller of choice way back, but now you hear practically nothing about them. Meanwhile, they still go into a zillion items.
It was only popular because equivalent small-scale Arm stuff didn't exist. Luckily that's now doing a good job of displacing the PIC architecture and that other abomination, MSP430.
Incidentally, you can always tell PIC and MSP430 programmers in a crowd, just look for anyone with many layers of healed-over bloody scabs on their foreheads, and possibly razor scars on their wrists.
(I rather liked Rabbit, as long as you were content with a Z80-like architecture and 1MB (?) of memory. They had a clear target audience, and seemed good at addressing their needs. And they did "modules" before it was cool.)
I didn't care about the architecture, C is C, I just needed serial ports and the easy to use RTOS that came with it.
I want the RTOS scheduling and comms to be in silicon, and libraries to do useful things like USB/ethernet comms
without impacting my task's code in any way whatsoever, and the IDE to state the performance
before executing the code.
I'll use whatever language enables those significant benefits.