[TriEmbed] Suggestions Needed to Improve Quality / Reliability Re: TriEmbed Digest, Vol 44, Issue 27

[Pete Soper]

Also, with a 20% tolerance part, good  practice requires answering the 
question "how will this operate if (each, some, all) caps are at -20%, 
0%, +20%?). So the 22uF cap might be a 18uF cap before the DC bias 
correction is applied. I also agree with the forum woman's suggestion to 
consider higher voltage parts. And a single 1206 or 1210 would beat a 
pair of 0805s for both convenience and simplicity, wouldn't it?

This is all absolutely fabulous, and potentially useful to interested 
folks on the list who want to see these kinds of chips actually deliver 
their spec'd performance reliably. But can it explain the terrible state 
your system gets into such that it does not respond to reset? Are these 
changes to stray capacitance, ripple, thermal performance, etc going to 
prevent the state in which the Atmega is not restarting until the entire 
system is power cycled? Does this TI chip have any unstable states that 
could account for the symptoms you've reported?

-Pete
PS Had to trim this down to stay under the per-message limit. It's also 
very hard on folks reading this stuff on their phones to have to wade 
across fractional megabytes of redundancy. Everthing is in the archive 
(link at the bottom of every message).


On 01/25/2017 09:04 AM, Chip McClelland via TriEmbed wrote:
Jesse,

You were right on with the thermals comment. Here 
<https://e2e.ti.com/support/power_management/non-isolated_dcdc/f/196/p/569106/2089407#2089407> is 
that discussion thread with the progression of the design.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20170125/ea338e65/attachment.htm>