New Motion Control DSPs From TI

More here

Nice devices, but no samples till September 07, so still a little way off ?

-jg

Nice devices.. just what i was looking for.. until i found out the package sizes.. jeez.. why can't they make something small thats NOT 1E6 pins ? if it was 64 pins i would buy 1E6 of them.. !! lol

yes, its seems the chip designers bonuses are directly proportional to the pin count of their designs....

If those volumes are serious, then I'd contact TI.

I'd also imagine TI have smaller members comming. If their sales pitch is to be believed, they see this as an 'Energy Saviour' device, and many of those apps, do not need large pincount.

They do have a BGA variant, and packages are quite a flexible choice - it is after all just a holder for the Die, so as long as the die fits, packages are reasonably open.

I've heard figures from semi vendors, of $2M-$5M of business, as being enough to trigger (very) serious package variant research.

-jg

On Tue, 5 Jun 2007 07:35:48 -0500, "TheDoc" wrote in comp.arch.embedded:

I don't think I'll be violating the NDA if I tell you that they didn't tell us about any upcoming smaller packages. These are all at the top end of the line, aimed at those who are using or considering the 2810,

There are the 280x members of the family in smaller packages and no external bus. Haven't used any of these -- yet -- but they do look interesting. 100 MHz and no floating point planned that they told us about.

I currently use the F2811 and it's 128 pins.. very difficult to fit into our space limitations.. and then it's an overkill on I/O.. if there was a 64 pin quadflat type package then i'd be very pleased.. BGA is not an option for us in this particular design..

thanks

This is just what I was hoping for! I started using the F2812 because it's a really fast micro with lots of peripherals. Seemed easier to work with than the higher ARMs with all their memory management. Freescale chips just didn't have enough speed. I don't need RTOS at this stage, so simple but fast does it.

Plus the plethora of counter/timers/EVMs/QEP on the C28x is essential for my applicaitons.

The addition of an FPU to a platform that I already understand makes me jump for joy.

The only thing I could possibly continue to wish for is even more speed. I have been monitoring TI's roadmap and new devices for a while hoping to see some sign that at least a 33% boost to 200MHz might occur some day. I suppose it is difficult with the flash integrated, but there's no reason flash based code can't just run with more waits, while critical code could be put in RAM to run at full speed.

Thus, the new ADI device ADSP-21369 (I was reading about it lately on Danville Signal's site) if it only had the peripherals of the C28x to make it useable as a straight micro, would be the ultimate machine that could pull me away from the C28x. It would also incline me toward wanting to play with assembly more too. I need to optimize an algorithm for the C28x right now that needs assembly, but jeez does this C28x have a lot of instructions and addressing modes!

So Jim or someone with connections in ADI, is there any hope of ADI producing something like the integrated peripheral capabilities of the C28x on a SHARC-like machine?