Running a non-L suffix part at 3 volts

Go To Last Post
14 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have a requirement to run a mega32 chip at 3 volts in an application where it is currently clocked at 14.7MHz. The non-L suffix mega32 is specified to run with Vcc of 4.5 to 5.5 volts up to 16MHz. The L-suffix part is specified to operate from 2.7 to 5.5 volts but only up to 8MHz.

Of course, I can modify the code to run at 7.37MHz in order to avoid overclocking the L-suffix part. However, I'd like to hear from people that have experience with operating non-L parts at 3 volts or overclocking the L parts running at 3 volts in production volumes.

Don Kinzer
ZBasic Microcontrollers
http://www.zbasic.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It's not a good idea to use non L spec devices at lower voltges.

BOD will definitely need to be disabled. Also prohibits EEPROM operations. Even if the chip runs....I doubt the timing will be very accurate, the oscillator might not be stable.

The wathdog and emergency oscillator just might still run.
Could boost 3 volts to 5 volts using a switching regulator.
Or...will mega48 do the job? it's probably worth a look.

Last Edited: Wed. Jan 4, 2006 - 12:38 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yikes! production volumes, I wouldn't dare overclock something that will be a actual 'product'. I haven't had any problems with running a AT90S2343-10 (10Mhz) at 14.31818Mhz but at production volumes they will eventually fail (some).


My AVR Site

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Don't try overclocking as an atempt to price undercut... Bad idea, and the only reason you would try running non L parts at 3 volts; Using a cheaper part to replace a more expensive one. Granted like all chip manufactors these numbers are all 'nominal' and some parts are more or less capable than others, this is exactly what the L prefix is for, parts that won't run 'full on' but do very well at lower voltage. Just different ends of the manufactoring extreme. I'm sure various parts from each group grey into each other so to speak, but it's if you're doing it in production always best to use parts fit for said purpose and not introduce asumptions into things. They do split them into L and non-L version for a reason. Aside from that proper 'break in' testing may reveal what your chips limits are, what they will be over time.. who knows.

-Curiosity may have killed the cat
-But that's why they have nine lives

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sceadwian wrote:
Don't try overclocking as an atempt to price undercut...

That wasn't the idea. The problem is that there is no mega32 (or better) part that is specified to run both at 3 volts and 14.7MHz. If there were, I'd buy it. Even the new, not yet available, parts like the mega324 or mega644 can do that unless I'm misreading the datasheet. Another requirement: TQFP-44 package.

Don Kinzer
ZBasic Microcontrollers
http://www.zbasic.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yeah, can't wait till the mega324 comes outs! 2 UARTS...3.3V...20MHz....November 2006 seems so far away....

Until then.....

dkinzer wrote:

Of course, I can modify the code to run at 7.37MHz in order to avoid overclocking the L-suffix part.

Dave.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
That wasn't the idea. The problem is that there is no mega32 (or better) part that is specified to run both at 3 volts and 14.7MHz. If there were, I'd buy it. Even the new, not yet available, parts like the mega324 or mega644 can do that unless I'm misreading the datasheet. Another requirement: TQFP-44 package.

Just to let you know, when the mega324 does come out it is available in a DIP40.

Quote:
ATmega324
Description: Key Parameters:
32-Kbyte self-programming Flash Program Memory, 2-Kbyte SRAM, 1-KByte EEPROM, 8 Channel 10-bit A/D-converter. JTAG interface for on-chip-debug. Up to 20 MIPS throughput at 20 MHz. 1.8 - 5.5 Volt Operation.
64KB version: ATmega644
16KB version: ATmega164

Flash (Kbytes) 32
EEPROM (Kbytes) 1
SRAM (Bytes) 2048
Max I/O Pins 32
F.max (MHz) 20
Vcc (V) 1.8-5.5
Pb-Free Packages MLF 44 PDIP 40 TQFP 44


My AVR Site

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Retarget for 7.37MHz (or 5V). The data sheet is (usually) your friend.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mneary wrote:
Retarget for 7.37MHz (or 5V). The data sheet is (usually) your friend.

Indeed. Based on what I read, that was the conclusion that I had reached. I was fishing (hoping) for a revelation, perhaps something that I had missed - you know, "That's easy, just use a megaXXX". So far, no such luck.

Don Kinzer
ZBasic Microcontrollers
http://www.zbasic.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I did some testing with an AT90S8515 (8MHz-max) running at 14.31818Mhz. It programmed but in AVRStudio it stated the flash didn't match the file. I took the crystal out and switched back to the STK500 Oscillator (3.somthingMHz) and it programmed perfectly.


My AVR Site

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LCD,

So, why not program the UC at 3.xxMHz and then see if the thing will actually run it at 14.xxMhz? This is very doable on the STK500 and, it would be practicle using the PDIP version, which, is what I believe you have. It wouldn't be as easily done using an SMT part, though...

You can avoid reality, for a while.  But you can't avoid the consequences of reality! - C.W. Livingston

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It will run at 14.xxMHz but just doesn't complete the programming cycle.In fact it actually programs the code at 14.xxMHz but just doesn't verify.


My AVR Site

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I run a mega128 at 3.3V and 18.432 MHz daily, and have done so for years. It runs at 14.7456HHz and 2.3V fine too.
According to my distributor, there really is no difference between the L and non-L versions.
However, the speed will be limited according to the voltage (the datasheet lists this to some extent). So the regular rule is that at 3.3V you can go up to 8MHz, at 5V, you can go up to 16MHz.
Also, as an example, the mega162L and non-L was replaced with a V version which was rated down to 1.8 V (1 MHz), but it was also rated for 16MHz (at 5V).

/Jesper
http://www.yampp.com
The quick black AVR jumped over the lazy PIC.
What boots up, must come down.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

i have to agree with Jesper.

My theory still holds that there is NO difference in the silicon of the -L and non-L versions.
Think about it. They make lots of chips, and some work better (faster) than others.
Those that work well upto 16MHz go through as non-L versions, and the rest are specified as -L versions, limited to half the clockspeed.

Now the non-L version is specified for 4.5-5.5 volts. WHY?
Because.....
The L and non-L parts are the same price. If the non-L part was specified for low voltage operation, who would buy the -L part then?
They specify the non-L part only for 5v operation, so that yo have to go out and buy the -L part if you want to run at lower voltages..
Meantime, the non-L part is able to work JUST FINE, as long as you don't try and run it at max clockspeed.

If they did NOT specify the non-L part as 4.5-5.5v, they would eithyer need to sell the -L part cheaper, or they will not be able to sell it. It is amazing that by changing a simple number on the datasheet can save so much money (for Atmel, I mean)

I only buy non-L versions, and use them for both 5v and 3.3v operation.
Specifically on the Atmega8, I use them all the time (hundreds of boards) at 3.3v, running at 7.3728MHz.

my 2 cents..

regards
Carel

www.pteq.net
Home of:
- Polygon Technologies CC