Reply by Not Really Me January 17, 20082008-01-17
"kalyanamsaritha" <kalyanamsaritha@gmail.com> wrote in message 
news:5O-dnYre6-YCSRDanZ2dnUVZ_hKdnZ2d@giganews.com...
> > LPC3180 is a ARM926EJS core which uses a 1.2V for the core, sdram, etc, > and 1.8V/3.0V for IO. > > What power supply/regulators would be recommended? Would a single > regulator be recommened to achieve these multiple voltages? And how? > > Thanks.
As a much simpler alternative, buy an eval board for the processor and interface your parts to that. You stand at least a reasonable chance of success if you get big important pieces pre-done and running out of the box. Scott
Reply by Joe G (Home) January 17, 20082008-01-17
"Clifford Heath" <no@spam.please.net> wrote in message 
news:478eba0d$0$20842$afc38c87@news.optusnet.com.au...
> kalyanamsaritha wrote: >> BTW, a budding telephony hobbyist could already benefit from work done > > Being able to read "Recipes of the Great Chefs" won't make you a great > chef. As a hobbyist in embedded myself, I have first-hand experience of > this fact :-). The more you learn, the more you learn to respect > experience. > > Hang around, listen to Joerg, JL, Phil Hobbs, and others like them, and > you'll learn more than you could imagine. > > And for g%* sake, try a simpler project first. Something with only one > or two chips, running at under 10MHz, in a pitch no finer than 0.05", > and fewer than a hundred components in all. Design and build both the > circuit and the board, then program the software and test the unit in > real-world conditions, including resilience against EMI, temperature, > being dropped, repeated power cycling, running with the batteries half > flat or reversed, etc. When you can build a device that stands up under > that kind of scrutiny, think about doubling the clock speed and/or > number of components. Rinse and repeat. In ten years, you'll know if > you're up to attempting the kind of project you've been talking about.
I 100%+ agree..... You will need a lots of dollars ($500+ USD probaly) to get a multilayer board done (8 layers probably) . Why dont you purchase a development kit first and get experience with known working hardware and profesional software. I know million dollar companies that would steer away from such complex designs, as in their experince..... they can see trouble coming! Joe
Reply by Clifford Heath January 17, 20082008-01-17
kalyanamsaritha wrote:
> I am aware that anything in a BGA package is not hacker friendly. > Unfortunately, uCs I am interested in are largely available in this > package (AVR32 AP7000 and LPC3180). These uCs appear closest to my > requirements.
Start by buying a development board with the device you want - or a more powerful version if possible - and get familiar with what it can do. Then start connecting other things to that to make your device. Expect that the first prototype will look like an elephant vomited. As long as it works, you'll be able to get help with the next step.
Reply by linnix January 17, 20082008-01-17
On Jan 16, 8:00 pm, "kalyanamsaritha" <kalyanamsari...@gmail.com>
wrote:
> >I hope you realize that making a board for such "Low Pin Count / High > >Ball Count" microcontroller is not trivial. > > I am aware that anything in a BGA package is not hacker friendly. > Unfortunately, uCs I am interested in are largely available in this > package (AVR32 AP7000 and LPC3180). These uCs appear closest to my > requirements. > > >You often need 4 to 6 layers PCB and > > Apart from the PCB routing issues, I guess I will need at least 4 layers > to conform to EMI anyways. > > >more than 1/3 of it just for the power regulators. > > Any recommendations on power regulators (parts?) with multiple outputs > (1.2V, 1.8V and 3.0V)?
I would use adjustable switching regulators. Parts (size/value) depends on your target current consumptions. This is the major part in designing your circuits/layouts.
> > I understand that I will need professional help, which invariably begins > with the parts/components (by itself quite time-consuming) and I am hoping > the community can help out. Will definitely be hanging around to continue > my education. > > Thanks.
Reply by kalyanamsaritha January 17, 20082008-01-17
> >I hope you realize that making a board for such "Low Pin Count / High >Ball Count" microcontroller is not trivial.
I am aware that anything in a BGA package is not hacker friendly. Unfortunately, uCs I am interested in are largely available in this package (AVR32 AP7000 and LPC3180). These uCs appear closest to my requirements.
>You often need 4 to 6 layers PCB and
Apart from the PCB routing issues, I guess I will need at least 4 layers to conform to EMI anyways.
>more than 1/3 of it just for the power regulators.
Any recommendations on power regulators (parts?) with multiple outputs (1.2V, 1.8V and 3.0V)? I understand that I will need professional help, which invariably begins with the parts/components (by itself quite time-consuming) and I am hoping the community can help out. Will definitely be hanging around to continue my education. Thanks.
Reply by Clifford Heath January 16, 20082008-01-16
kalyanamsaritha wrote:
> BTW, a budding telephony hobbyist could already benefit from work done
Being able to read "Recipes of the Great Chefs" won't make you a great chef. As a hobbyist in embedded myself, I have first-hand experience of this fact :-). The more you learn, the more you learn to respect experience. Hang around, listen to Joerg, JL, Phil Hobbs, and others like them, and you'll learn more than you could imagine. And for g%* sake, try a simpler project first. Something with only one or two chips, running at under 10MHz, in a pitch no finer than 0.05", and fewer than a hundred components in all. Design and build both the circuit and the board, then program the software and test the unit in real-world conditions, including resilience against EMI, temperature, being dropped, repeated power cycling, running with the batteries half flat or reversed, etc. When you can build a device that stands up under that kind of scrutiny, think about doubling the clock speed and/or number of components. Rinse and repeat. In ten years, you'll know if you're up to attempting the kind of project you've been talking about.
Reply by linnix January 16, 20082008-01-16
On Jan 16, 5:20 pm, "kalyanamsaritha" <kalyanamsari...@gmail.com>
wrote:
> >An Schwob in the USA wrote: > >> If > >> you are working with such a device as the LPC3180 I am surprised about > >> your question because routing a board for this chip is a true > >> challenge for professionals. > > >This is the same guy who's a self-professed "hobbyist" and "new to > embedded" > >who wants to bit-bang USB host, and use DSP to interface Ethernet to the > analog > >phone system. > > >Good luck to him - he certainly doesn't lack ambition. > > Yes, I am a hobbyist and all I am trying to do is understand the scope of > the project in mind, and whether a ultra low-cost solution is even > possible. Although, the project appears to require professional help, it > is still good to know.
I hope you realize that making a board for such "Low Pin Count / High Ball Count" microcontroller is not trivial. You often need 4 to 6 layers PCB and more than 1/3 of it just for the power regulators.
Reply by kalyanamsaritha January 16, 20082008-01-16
>An Schwob in the USA wrote: >> If >> you are working with such a device as the LPC3180 I am surprised about >> your question because routing a board for this chip is a true >> challenge for professionals. > >This is the same guy who's a self-professed "hobbyist" and "new to
embedded"
>who wants to bit-bang USB host, and use DSP to interface Ethernet to the
analog
>phone system. > >Good luck to him - he certainly doesn't lack ambition. >
Yes, I am a hobbyist and all I am trying to do is understand the scope of the project in mind, and whether a ultra low-cost solution is even possible. Although, the project appears to require professional help, it is still good to know. In general, folks on this forum have been most patient and helpful. I appreciate that greatly. BTW, a budding telephony hobbyist could already benefit from work done at: http://www.rowetel.com/ucasterisk/ http://www.astfin.org/ (PCB designs and all ...) Thanks.
Reply by Clifford Heath January 16, 20082008-01-16
An Schwob in the USA wrote:
> If > you are working with such a device as the LPC3180 I am surprised about > your question because routing a board for this chip is a true > challenge for professionals.
This is the same guy who's a self-professed "hobbyist" and "new to embedded" who wants to bit-bang USB host, and use DSP to interface Ethernet to the analog phone system. Good luck to him - he certainly doesn't lack ambition.
Reply by linnix January 16, 20082008-01-16
On Jan 16, 6:01 am, An Schwob in the USA <schwo...@aol.com> wrote:
> On Jan 16, 10:55 am, "kalyanamsaritha" <kalyanamsari...@gmail.com> > wrote: > > > LPC3180 is a ARM926EJS core which uses a 1.2V for the core, sdram, etc, > > and 1.8V/3.0V for IO. > > > What power supply/regulators would be recommended? Would a single > > regulator be recommened to achieve these multiple voltages? And how? > > > Thanks. > > It can be a single regulator but it better has two output voltages. If > you are working with such a device as the LPC3180 I am surprised about > your question because routing a board for this chip is a true > challenge for professionals. > > An Schwob
Some SDRAM are 2.5V. You should have on-board switching regulators for 1.2V/1.8V/2.5V/3.3V and 5.0V (if you run a hard disk).