Forums

Atmel Introduces World's smallest Microcontrollers?

Started by Bill Giovino May 14, 2010
http://microcontroller.com/news/Atmel_smallest.asp

Teeny tiny uDFN package that is only 2mm on each side.


Bill Giovino
Executive Editor
http://Microcontroller.com
Follow Me: http://twitter.com/BGiovino




On May 14, 2:00=A0pm, "Bill Giovino" <contac...@microcontroller.com>
wrote:
> http://microcontroller.com/news/Atmel_smallest.asp > > Teeny tiny uDFN package that is only 2mm on each side.
Wow, I want a Beowulf cluster of those! -a
MICROCHIP HAS BEEN & IS SUPERIOR TO ATMEL OR ANYONE!
IF WORLD'S SMALLEST PACKAGE (as you claim - Atmel) DOESN'T DO ITS JOB AS 
WELL AS MICROCHIP INC. THEN THERE'S NO USE FOR IT.

MICROCHIP HAS BEEN OFFERING uDFN packages and/or waiting for market reaction 
anyway, as per their V.P who emailed me in response to my suggestion to keep 
PDIP packaging for easier prototyping, which is the opposite end - IC's 
largest package.

He said uDFN is extremely difficult to install on circuit boards, also 
reliability is an issue.
But technologically Microchip has uDFN, it just doesn't know if market woudl 
care for IC package which in essense is BARE SILICON DIE with a thin layer 
of passivation/protection.  The next step would be simply a bare die which 
cannot be installed on circuit boards, b/c it should hermetically sealed

Baredies is the only way to create RF/Microwave ciruitcs above 3GHz or so, 
I've spent years with people doing it, upward to 65GHz and believe me 
designing a circuit is much easier than building it when package size is 
smaller than certain limit - we even have many people who CRUSH components 
with the forces of tweezers used ot pick up component & drop into circuit 
housing.

uDFN is an extremely difficult package to work with without a Microscope.
I personally prefer Microchip to Atmel, anyway.

You should see new Microchip offers on their parametric search webpage, I've 
been using PIC18F4520 on theprevious project and now they offer a PIC with a 
dozen Timers, 80MHz clocks, enormous number of A/D & PWM channels, etc.
And this beauty costs $3-5??

Try to beat that Atmel.

=========================
Stan Starinski
Phone (mobile): +1 (646) 416-2052
Web:      www.Interengineers.org  www.Nanoinfocenter.com   Currently OFF for 
maintenance

Consulting Engineer (EE+ME, ECAD+MCAD [3D/2D]), R&D + Prototype, 
Embedded/Firmware ["C" or ASM for Microcontrollers], computers).
Past clients: MTT Corp, XigoNanotools, Amplitech, Miteq, Vematech, 
KirbyLester, Universities, WaltDisney/ESPNzone 

"Stanley Starinski" <no@no.com> wrote in message 
news:hskfvj$oj7$1@news.eternal-september.org...
> MICROCHIP HAS BEEN & IS SUPERIOR TO ATMEL OR ANYONE! > IF WORLD'S SMALLEST PACKAGE (as you claim - Atmel) DOESN'T DO ITS JOB AS > WELL AS MICROCHIP INC. THEN THERE'S NO USE FOR IT. > > MICROCHIP HAS BEEN OFFERING uDFN packages and/or waiting for market > reaction anyway, as per their V.P who emailed me in response to my > suggestion to keep PDIP packaging for easier prototyping, which is the > opposite end - IC's largest package. > > He said uDFN is extremely difficult to install on circuit boards, also > reliability is an issue. > But technologically Microchip has uDFN, it just doesn't know if market > woudl care for IC package which in essense is BARE SILICON DIE with a thin > layer of passivation/protection. The next step would be simply a bare die > which cannot be installed on circuit boards, b/c it should hermetically > sealed > > Baredies is the only way to create RF/Microwave ciruitcs above 3GHz or so, > I've spent years with people doing it, upward to 65GHz and believe me > designing a circuit is much easier than building it when package size is > smaller than certain limit - we even have many people who CRUSH components > with the forces of tweezers used ot pick up component & drop into circuit > housing. > > uDFN is an extremely difficult package to work with without a Microscope. > I personally prefer Microchip to Atmel, anyway. > > You should see new Microchip offers on their parametric search webpage, > I've been using PIC18F4520 on theprevious project and now they offer a PIC > with a dozen Timers, 80MHz clocks, enormous number of A/D & PWM channels, > etc. > And this beauty costs $3-5?? > > Try to beat that Atmel. >
Aimed at high vol professional use, not small business/hobbyist. Each package has its merits...
On May 15, 9:00=A0am, "Bill Giovino" <contac...@microcontroller.com>
wrote:
> http://microcontroller.com/news/Atmel_smallest.asp > > Teeny tiny uDFN package that is only 2mm on each side.
hmm...NXP are already way ahead of that on density. They pack TWICE as many pins into a BGA just over 2mmx2mm http://www.nxp.com/news/content/file_1701.html NXP also has MUCH more MEMORY!! NXP: 32KF, 8KR vs Atmel: 1KF/32bR So, you'd need 32 Atmel parts to match in CODE, and a whopping 256 to equal the RAM.... -jg
but the cost!
"-jg" wrote...
> On May 15, 9:00 am, "Bill Giovino" wrote: > > http://microcontroller.com/news/Atmel_smallest.asp > > > > Teeny tiny uDFN package that is only 2mm on each side. > > hmm...NXP are already way ahead of that on density. > > They pack TWICE as many pins into a BGA just over 2mmx2mm >
The NXP page you cited lists the dimensions at "2.17 mm&#2013266098; x 2.32 mm&#2013266098;, thickness of 0.6mm" Can I safely assume that they actually meant to write "2.17mm x 2.32mm"? Because if they didn't, I'm not sure how to interpret or trust those dimensions. However, most Engineers would agree that the Atmel 2mm x 2mm package is still smaller. Bill Giovino Executive Editor http://Microcontroller.com Follow Me: http://twitter.com/BGiovino
On May 17, 7:39=A0am, "Bill Giovino" <contac...@microcontroller.com>
wrote:
> "-jg" wrote... > > On May 15, 9:00 am, "Bill Giovino" wrote: > > >http://microcontroller.com/news/Atmel_smallest.asp > > > > Teeny tiny uDFN package that is only 2mm on each side. > > > hmm...NXP are already way ahead of that on density. > > > They pack TWICE as many pins into a BGA just over 2mmx2mm > > The NXP page you cited lists the dimensions at "2.17 mm=B2 x 2.32 mm=B2, =
thickness of 0.6mm"
> Can I safely assume that they actually meant to write "2.17mm x 2.32mm"? =
Because if they
> didn't, I'm not sure how to interpret or trust those dimensions. > > However, most Engineers would agree that the Atmel 2mm x 2mm package is s=
till smaller. Sure, but most engineers I now, are smart enough to grasp the difference between Density (which is what I wrote) and Area ;) Most designers are more interested in Functionality per area, than absolute area, some are interested in Price/area, but vendors do try for whatever wriggle space they can find, for bragging rights. You also missed out on your banner claim "Atmel introduces the worlds smallest microcontrollers", because someone else already has an 8 bit, 2mmx2mm package Microcontroller, (with 10 pins, more code and more Ram) so Atmel is second even there.... oops. -jg
-jg wrote:
> On May 17, 7:39 am, "Bill Giovino" <contac...@microcontroller.com> > wrote: >> "-jg" wrote... >>> On May 15, 9:00 am, "Bill Giovino" wrote: >>>> http://microcontroller.com/news/Atmel_smallest.asp >>>> Teeny tiny uDFN package that is only 2mm on each side. >>> hmm...NXP are already way ahead of that on density. >>> They pack TWICE as many pins into a BGA just over 2mmx2mm >> The NXP page you cited lists the dimensions at "2.17 mm&#2013266098; x 2.32 mm&#2013266098;, thickness of 0.6mm" >> Can I safely assume that they actually meant to write "2.17mm x 2.32mm"? Because if they >> didn't, I'm not sure how to interpret or trust those dimensions. >> >> However, most Engineers would agree that the Atmel 2mm x 2mm package is still smaller. > > Sure, but most engineers I now, are smart enough to > grasp the difference between Density (which is what I wrote) and > Area ;) > > Most designers are more interested in Functionality per area, than > absolute area, some are interested in Price/area, but vendors do try > for whatever wriggle space they can find, for bragging rights. > > You also missed out on your banner claim "Atmel introduces the worlds > smallest microcontrollers", because someone else already has an 8 bit, > 2mmx2mm package Microcontroller, (with 10 pins, more code and more > Ram) so Atmel is second even there.... oops. > > -jg
I'm always a bit suspicious when I read the phase "Atmel Introduces". What this usually seems to mean is "Atmel's marketing dept aspires to...", rather that "Atmel is shipping". Is this actually a real product, or just something they might make in the future if enough customers show interest? Martin
"MartinWalton" wrote...
:
> > I'm always a bit suspicious when I read the phase "Atmel Introduces". What this > usually seems to mean is "Atmel's marketing dept aspires to...", rather that "Atmel is > shipping". > > Is this actually a real product, or just something they might make in the future if > enough customers show interest? > > Martin
If you discover that the product isn't real, please let me know. -Bill Giovino