Hi,
Am 31.08.2017 um 17:15 schrieb Michael Mrozek EvilDragon@openpandora.org:
Am Thu, 31 Aug 2017 16:08:49 +0200 hat "H. Nikolaus Schaller" hns@goldelico.com geschrieben:
Hi,
notaz mentioned the EVM has over 44 capacitors for the RAM power traces only... how many do we have and can those be changed / optimized?
We don't have that many but AFAIR there are at least 2 per DDR3 chip. On the latest (not yet populated PCB) there are 13 for 4 chips, i.e.
3 per DDR3 chip.
Ah - we optimized the capacitors for the DDR3 chips as well?
I haven't said that... Just that I only counted them on the latest PCB layout. Which has not been populated and is ahead of our prototypes.
I didn't cross-check with previous boards (it is a little more difficult to locate and open the files).
Now I have done that and capacitors have not changed since the first CPU board prototype.
Then I guess we should probably populate a few to test that!
- Basically the 4GB dual die chips may need more decoupling (local
energy buffer) for VDD because they draw more current if both chips are active
Yep.
- On the other hand 2GB RAM works well with much less decoupling than
the EVM which also has only 2GB RAM... So 48 is not needed for 2GB.
True as well. The solution is probably somewhere inbetween.
If at all. It could still be something completely different.
As it works for Tony if he increases the voltage and Juliano showed that we might have too much noise, then a LITTLE improvement might already be enough.
So those >3 per DDR3 could help!
Yes, according to this completely unproven theory.
I just want to say that I would prefer new facts before drawing conclusions from hypotheses :) I understand that you prefer to really do something than still be digging in the fog/mist.
Unfortunately it is hardware and not changing some line in source code and recompiling :(
Especially since changes cost money. >3 per DDR3 chip means: * scrapping the 500 boards and * trying to squeeze in more capacitors * have PCBs built and populated
If it helps: fine. If not: wasted money and time.
The change-compile-boot-analyse cycle for software is significantly faster...
So I'd prefer to get more facts about different parameter sets that work/don't work etc. first.
BR, Nikolaus