LPC1102, World's smallest 32-bit ARM MCU in a 5mm2 pckg

Yes, it's small, in two dimensions... However, deployed it will need a minimum layer count, so that Y axis comes at some cost. If it bumps the layer count on your board, it's suddenly very expensive.

Can they fit the die into SO16N, or QSOP16 - and so offer a single-sided solution ?

-jg

There are only 4 pads in the center. Double layers should work, but you need plugged via.

I think you could route the inner pads out between the outer, with 4-5 mil design rules

-Lasse

n I

bad

I usually limit myself to 7 mil rules. I wonder what the center pads are. If unnecessary I/Os, you can ignore them. If ground pads, you can join them with an outer pad. Anyone got inside info?

han I

t bad

Hi,

I was looking for more info too but could not find anything more specific on the NXP website. Expect to see more next week because NXP announced to show something at the ESC in San Jose.

An Schwob

te:

iar

but

than I

not bad

but

Saw it today on the NXP booth at the ESC West. Many people probably need classes to see the device :) Documentation is still a few weeks out I've been told. Well, with production scheduled for later in the year that seems not a big deal. That much information I could gather: Will be able to run @ 50 MHz if external clock is provided, bu

rote:

iliar

s but

ry than I

y not bad

, but

ds

you

must have hit the send button

... needs an external clock circuit, not just a crystal. Internal oscillator can be used for may applications.

Can't imagine were this chip will find a new home, after all there are so many applications that aim for miniaturization and this is the only ARM based chip that small.

Congratulations NXP!

te:

iar

but

than I

not bad

but

This part is not for low cost boards. It is for small devices. Sometimes, but not always the same thing.

Rick

rote:

iliar

s but

ry than I

y not bad

, but

ds

you

Rick,

you are absolutely right. You can not use an ultra low cost facility because the device need at the very least a double sided PCB, may be more layers and a more sophisticated machinery than a soldering iron and low cost labor. The lowest cost devices are the LPC1111, selling for 65 cent @ 10k according to the press release.

An Schwob

amiliar

ges but

mory than I

lly not bad

rk, but

h 4-5

pads

, you

Since when is hand soldering cheaper than pick and place??? Maybe if you are building just one, even then I have been asked for CNC information since they just don't want to do hand work. The cost in using these parts comes from the small via sizes required... at least I am assuming. Is there a breakout that does not require drills smaller than 12 mil? When I try to get boards made with 10 mil vias the board makers have "rounded" that up to 13 citing a +- 3 mil tolerance spec. of course I don't do business with that company anymore, but the point is smaller drills cost extra and that is a recurring charge. I can't say anything about the number of layers since I haven't tried to route this part yet. But I'd bet it is hard on a two layer board.

It may be that the assembly house charges extra for mounting these parts too. I've seen extra charges for pin pitch below 20 mil (0.5 mm) on leaded parts. I'm not sure where they start charging extra for fine pitch BGA type parts.

But then if this size part makes a design possible that otherwise wouldn't be, it doesn't matter if the unit costs a few cents extra.

Rick

---snip----

Hi Rick,

that was a misunderstanding. I was talking about low cost equipment for the LPC1111. Never said or wanted to imply that hand-soldering is the lowest cost. Just to make sure there is not another misunderstanding, I totally agree with you!

An Schwob