18Bit Binary weighted DAC layout placement

[vadim888]

18Bit Binary matrix DAC layout placement

Hi all!

I need help.

My issue is find best way to placement (switching) 18 bit Binary matrix (array).
Maybe someone can help me or just give a tips or algorithm.

In my opinion I can't use the Q2 random walk algorithm or I just not understand.

 

[asimlink]

hello vadim,

please read following topic:

http://www.hottconsultants.com/techtips/split-gnd-plane.html

This will help you understand mixed signal pcb component placement and signal routing issues.

hope this helps!

 

[vadim888]

Help me please, I really need an advice.

 

[erikl]

Re: 18Bit Binary matrix DAC layout placement

My issue is find best way to placement (switching) 18 bit Binary matrix (array).

In my opinion I can't use the Q2 random walk algorithm ...

I think you're right: the Q2 random walk algorithm isn't appropriate for such a symmetry requiring task (both re. placement and routing), nor is the simulated annealing algorithm (Cadence, Dragon). Maybe the new DPlace engine can do a better job, or the Kraftwerk placement algorithm, but these I don't know. Anyway I'd recommend to use manual PnR for such an accuracy needed.

 

[vadim888]

Re: 18Bit Binary matrix DAC layout placement

Dankeschön

Can you give a link for reading the manual PnR?

 

[erikl]

Re: 18Bit Binary matrix DAC layout placement

manual PnR?

It means: manual Placement and Routing. I.e. DIY! The only way to achieve sufficient symmetry for such a high-resolution converter, IMHO. You could ask such companies like Analog Devices and Co.

 

[dick_freebird]

I am not optimistic that you'll meet 18 bit accuracy without
some sort of trim. We found it necessary in the 12-bit days.
On the plus side, trim will fix what simple placement can't.

You should consider segmenting your higher order bits and
distributing them throughout the array, and making them all
out of unit cells (*2^N). This increases the wireball but does
help uniformity.