Just checked on DigiKey.
An ATmega808 in a 28-pin SSOP package can be gotten for for $0.89 USD in single quantity. In larger quantity, its $0.75 each.
This device has 8K of flash and 1K of ram! And, it has all the great 0-series peripherals shared by the larger members of that group.
What a buy!
Jim
True, they have good value, but always keep in mind the caveat that this series doesn't have the quartz crystal driving circuitry.
The problem is just that you can get a ARM with 32K Flash and 4KRAM for the same price.
The problem is just that you can get a ARM with 32K Flash and 4KRAM for the same price.
Well, actually you can get a Cortex-M3 64K/8K for 50 cents: https://lcsc.com/product-detail/...
The advantage of 8 bit chips IMO is just that they are simpler to configure and support 5V (if you need that) which ARM usually don't at this price range.
My little "essay" about complexity in the programming forum addresses the issue of ARM vs AVR 0-series, sort of. Sometimes, cost is not the only factor.
Jim
My little "essay" about complexity in the programming forum addresses the issue of ARM vs AVR 0-series, sort of. Sometimes, cost is not the only factor.
Jim
This I would assume:
https://www.avrfreaks.net/forum/complexity-trap
Sorry, that is indeed the cross-link.
Jim
The problem is just that you can get a ARM with 32K Flash and 4KRAM for the same price.
My question now is: Why it is no possible to get simple 8Bit controller with higher frequencies, much more flash and sram?
Why do I have to accept a higher (32Bit ARM) complexity today in order to get much more from it? Many applications do not require the elaborate flexibility and rich functionality = complex & confusing configuration of these types. Only more performance and memory space.
gchapman Why it is no possible to get simple 8Bit controller with higher frequencies, much more flash and sram?
An 8-bit advantage is wide voltage.
Only more performance and memory space.
What are the AVR28DA128 AVR32DA128 AVR48DA128 AVR64DA128 ?? | AVR Freaks
clawson 
My question now is: Why it is no possible to get simple 8Bit controller with higher frequencies, much more flash and sram?
Why do I have to accept a higher (32Bit ARM) complexity today in order to get much more from it? Many applications do not require the elaborate flexibility and rich functionality = complex & confusing configuration of these types. Only more performance and memory space.
They don't want to make a simple 200MHz AVR...it might scare the ARM chips away (too many developers might go for simplicity).
The Man is holding the designers who want faster AVR's back.
Maybe the AVR's can actually run at 100MHz at 5V , but they don't want us to know it...I wonder how fast they've been successfully "overclocked"?
Maybe there is a speed cartel, along the lines of the famous light bulb conspiracy, first organized by Osram head William Meinhardt
https://spectrum.ieee.org/tech-history/dawn-of-electronics/the-great-lightbulb-conspiracy
the argument may well go that if you are going to do that anyway you might as well be learning Cortex-M.
32 bit MCUs always seem to have much more complex peripherals. It's as if hardware designers, realizing they have 32 bit SFRs instead of 8/16 bit ones, feel some kind of compulsion to keep adding features until most of those 32 bits have some role in the peripheral 
Why it is no possible to get simple 8Bit controller with ... much more flash and sram?
One reason is that you can't address much memory with only 8 bits!
So you have to start adding special, extra-wide registers to give you a decent address size.
And you need extra logic to handle those "wide" values.
And the instructions have to get bigger to fit the addresses in.
So you end up saying, "why not just make it a 32-bit processor?"
32 bit MCUs always seem to have much more complex peripherals
This is true.
Most of the "ARM is so complex" whingeing is really not about ARM at all - it's about the peripherals that the chipmakers have chosen to bolt onto the ARM.
One reason is that you can't address much memory with only 8 bits!
Many 8 bit mcus can address much more than 256 flash bytes.
That can not be the reason 
So you end up saying, "why not just make it a 32-bit processor?"
I can not and do not want to understand exactly this conclusion (that means of course including all the consequences of designing today's 32 bit controllers).
They don't want to make a simple 200MHz AVR...it might scare the ARM chips away (too many developers might go for simplicity).
That seems to me the real motivation.
And simplicity as an attractive feature is still underestimated.
I can only guess
I only know what I want and need...
But I'm pretty sure I'm not the only one with this crazy wishes
Many 8 bit mcus can address much more than 256 flash bytes.
Of course they do - and that's what the rest of the post was about!
So you have to start adding special, extra-wide registers to give you a decent address size ... etc ...
clawson Another thing with ARM (or any 32 bit) versus AVR (or any 8 bit) is that memory is generally accessed in 32 bit (4 byte) so RAM tends to be 4 times the size. A 16K flash 8bit might typically have 1K RAM but a 16K flash 32 bit will likely have 4K RAM. So you always tend to find the flash/RAM ratio is larger (x4) with the larger bit width.
