1. Home
  2. Embedded Programming
  3. Atmel Atmega48

Atmel Atmega48

I just got some spam from Atmel about mega48. It should be a smaller sibling of mega8. I tried googling it up but ended up with very little. I found some basic info from Atmel's selection guide and the AVRfreaks web site, but not what I wanted. Even the data sheet is available only on request (yes, I have requested it but not received yet). So, maybe we have here someone with a bit more information...

We are right now using mega8L in some rather simple applications, mainly for its ADC, I2C and UART. The parts are powered with 3 volts and run at

8 MHz (one application uses the internal RC, others external xtal). We use the MLF32 package.

The big question is: Can we replace the 8L with the 48 without losing any functionality? Our code will fit in 4 kB, the data will fit in 512 B, we do not write the program memory during use (we use the SPI ISP in production), and we do not need any boot sector features.

Could we actually use the same code? Are the chips binary compatible (i.e. same control registers &c.)? Does GCC support it now or in the near future.

And -- a stupid little detail -- how much does it cost? Does this mean mega8 & 8L just started their sad journey towards obsolesence (with mega88 as the direct replacement)?

TIA,

- Ville

--
Ville Voipio, Dr.Tech., M.Sc. (EE)
Reply to
Ville Voipio
Loading thread data ...

I don't know much about it. It's too new...

[...]

You probably know this, but I thought I'd mention it anyway. Be careful when you talk about "4 kB" of program space on an AVR. The ATMega8 has 8kB program memory, but that's really 4 k _words_. Some tools report program memory in words rather than bytes. If the address of the last instruction in your application is >= 0x800, it probably won't fit.

[...]

Just a datapoint: we moved some code from an ATMega16 to an ATMega32. We were in C, so we were source code compatible. But Atmel rearranged the interrupt vectors between these chips, so we had to use a different link map. The register addresses and bit positions and functions were identical, however.

Regards,

-=Dave

--
Change is inevitable, progress is not.
Reply to
Dave Hansen

Yes, I am well aware of that. Our s/w is probably around 1 kW = 2 kB.

Rearranging the registers and/or interrupt vectors is not a big deal as I guess GCC will come up with new def files. (Or if it does not, we'll make them :) However, if there is some functionality missing, then we may end up with trouble. The 8L has been chosen because of itse periphera, so missing or differently working stuff may trigger major code revisions.

- Ville

--
Ville Voipio, Dr.Tech., M.Sc. (EE)
Reply to
Ville Voipio

Ville Voipio wrote >

Functionality is the same, only memories are smaller (4KB flash, 512B SRAM, 256B EEPROM), and lower power consumption (and more PWM channels). Code is not 100% binary compatible, because some I/O registers have move to the extended I/O space.

There is no chance for the Mega8 being obsoleted: the Mega88 will be a lower-power (and probably slightly more expensive) alternative to the Mega8.

Mark

Reply to
Mark

Thank you for the information! So, it seems that we'll move to '48 if its availability and price tags allow.

From the manufacturing point of view it seems a bit odd to have two almost identical chips. My guess is that in the long run '8 will disappear. The case is similar to 2313 tiny2313, where the latter is made with a better process (faster, smaller power consumption, SPM) and has a few extra features (e.g. USI). This will take some time, though.

- Ville

--
Ville Voipio, Dr.Tech., M.Sc. (EE)
Reply to
Ville Voipio

The ATmega48 is avaialble in samples.

No it ain't, If one of your major customers wantsit, he getsit. Since this is the case, there is another 4-5 years lifetime minimum left in the ATmega8. It makes a lot of sense for other customers to move to the ATmega88. The ATmega88 DebugWire(tm) which will be supported early next year with JTAGICE mk II, will bring a significant benefit.

--
Best Regards
Ulf at atmel dot com
These comments are intended to be my own opinion and they
may, or may not be shared by my employer, Atmel Sweden.
Reply to
Ulf Samuelsson

Ok. So, what are the reasons for _not_ migrating from '8L to '88? I do understand that some industries require certified code which must not be ported or changed any way, and this will quite effectively dictate the use of the older part. But apart from this? Price? Pin-incompatibility? Something missing?

TIA,

- Ville

--
Ville Voipio, Dr.Tech., M.Sc. (EE)
Reply to
Ville Voipio

Since there is no process shrink, there is no cost advantage to switch. Why spend time on something if you dont have to. There are significant advantages in power consumption and debug functionality and there is a pin compatible upgrade 4k/8k/16kB which some people will find nice. Some small modifications in the peripherals may also be of interest.

--
Best Regards
Ulf at atmel dot com
These comments are intended to be my own opinion and they
may, or may not be shared by my employer, Atmel Sweden.
Reply to
Ulf Samuelsson

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.