There is a new application note available on atmel's website: Migration-note from ATmega48/88/168 to ATtiny48/88 http://www.atmel.com/dyn/resources/prod_documents/doc8089.pdf.
I asked my distributor - samples should be available from Q4, mass production Q1/2008 (I hope the announced schedule has not the typical delay of 6-12 months :wink: )
ATtiny48/88
Author
Message
It reads like a part designed by the marketing department, but I would like to see the datasheet before deciding. A couple explanations from the migration note caught my eye.
Quote:
... Some instructions have not been implemented in ATtiny48/88. Where ATmega48/88/168 and ATmega48P/88P/168P/328P have 131 instructions, ATtiny48/88 only has 123.... The limited instruction set is of concern chiefly when programming in assembler. High-level language (such as as "C") compilers automatically take into account the available instruction set and the end user does not need to be aware of limitations.ATtiny48/88 has less volatile and non-volatile memory than ATmega48/88/168 and ATmega48P/88P/168P/328P.... This means some firmware may have to be re-engineered or recompiled using memory optimisation algorithms.
Quote:
...Some instructions have not been implemented...
The instructions not supported are JMP, CALL and all forms of MUL. You have to perform multiplication in software (which is trivial) und use RJMP/RCALL instead of JMP/CALL. Note that JMP/CALL is also not supported on ATmega48/88(P)(/V).
SRAM-size of the tiny is half compared to the mega, EEPROM 64 bytes.
There is no USART, no Timer2, no PWM with Timer0 and no support of external resonator/crystal. The maximum clock rate ist reduced, also the PC6/RESET's source and sink capability.
BUT:
ADC6&ADC7 can operate as digital I/O PA0&PA1, one pair of VCC/GND of the 32pin mega is digital I/O PA2&PA3 on the 32-pin tiny.
From my point of view the tiny48 fits perfect for lowest cost applications whenever the pin-count of other tiny's is too low. And I have a lot of applications in this segment.
Sure, some features have been reduced (it's called tiny instead of mega) but I never needed the maximum performance of any AVR so performing some things in software to reduce price in mass-production is one important way.
jfiresto wrote:
It reads like a part designed by the marketing department, but I would like to see the datasheet before deciding.
I'd suppose the marketing department found that there is a market for a lower cost chip and the engineers made one using less ram, eeprom and the simpler AVR core. Marketing now want to sell the chip and explain the differences without painting the new chip in a bad light.
What they say is quite true and to the point. The real story depends on your application anyway. Many apps don't use much ram or eeprom and just use the mega because of flash or ports. For those it just a recompile.
What is your point ?
Markus
Nothing more than the parts appear to have been largely created and guided by marketing. Some people will be in the market for such a variant and some will not, assuming the parts come to market.
But that is true for most products on the market anyway.
Markus
True enough, but the application note suggests they will be reduced versions of two existing, like-named parts. Atmel has pruned and presumably will discount them. I am curious what else they will do to make them ATtiny parts rather than ATmega. Perhaps nothing else, but we'll know when we see the datasheet.
Level: Raving Lunatic
Joined: Wed. Mar 12, 2003
Posts: 2684 View posts
Location: Portland, Oregon USA
I think that Atmel should go the other way with new Tiny devices. Instead of getting rid of the USART, they should put out a device with one or even two USARTs in an 8 or 14 pin package. Both PIC and Z8 Encore have devices like this in the same price range (a little more expensive) as the Tiny24. Or putting a direct USB interface and a UART into an inexpensive 14 pin device.
It will be interesting to see how much cheaper that the Tiny48 will be from the Mega48. Possibly there will be a price rise for the Mega48.
I'd like to see a small, low-cost Ethernet device as well. If possible even with built-in POE (802.3af) to supply power from the network to the device.
Markus
Price rise, or worse yet, end of life. But, I doubt that. They just expanded the line with the pico power parts up to the 328P.
I got a datasheet from my Atmel rep today.
I tried to attach it but it's over the size limit. If there is interest in this document, someone let me know how to get this done.
Any one know that winavr or IAR already support tiny48/88 or not?
WinAVR will support ATtiny48 and ATtiny88 in the next release. The support for the ATtiny48 is already in the source tree. The tiny88 is still on my todo list. ;)
Tiny48 and 88 are in the pulldown in the imagecraft compiler
Quote:
Hi Bob, The post you replied to is a few years old.
Posted: Oct 26, 2007 - 06:40 AM
I don't understand the reason for developing ATtiny48/88.
Only one IO pin more should be not worth all the effort.
E.g. adding an UART, also able to be a buffered master SPI to all ATtiny should be many many more effective.
Peter
I just saw the message and didn't even notice the date.
bobgardner wrote:
I just saw the message and didn't even notice the date.
That's Old Timers setting in Bob. :wink: Takes one to know one. :)