K
krw@att.bizzzzzzzzzzzz
Guest
On Wed, 23 May 2012 22:35:32 -0700, John Larkin
<jjSNIPlarkin@highTHISlandtechnology.com> wrote:
morning that they're asking about reference designators that don't exist and
parts that are clearly marked as OPEN.
avoid using them.
<jjSNIPlarkin@highTHISlandtechnology.com> wrote:
Indeed.On Wed, 23 May 2012 21:35:32 -0400, "krw@att.bizzzzzzzzzzzz"
krw@att.bizzzzzzzzzzzz> wrote:
On Tue, 22 May 2012 22:28:30 -0700, John Larkin
jjSNIPlarkin@highTHISlandtechnology.com> wrote:
On Tue, 22 May 2012 21:24:50 -0400, "krw@att.bizzzzzzzzzzzz"
krw@att.bizzzzzzzzzzzz> wrote:
On Tue, 22 May 2012 14:49:48 -0700, John Larkin
jlarkin@highlandtechnology.com> wrote:
On Tue, 22 May 2012 11:30:05 -0500, John Fields
jfields@austininstruments.com> wrote:
On Tue, 22 May 2012 08:30:21 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:
You have said that you're not an engineer, and that your electronics
is self-taught. Unless you forced yourself to learn the math, that
puts you at a disadvantage when discussing theory.
One good thing about EE school is that we were forced to learn the
theory and do the math. And we were taught to be careful and check our
work before turning it in. A lot of engineering is about avoiding
risk, and the more complex a system, the more careful you have to be
if it's ever going to work. One reason we can build a very complex
laser controller, and get it to work first-rev, is that we check the
hell out of our stuff before we etch boards. It's a lot of grunt work,
but it pays off.
---
You never miss a chance to spout platitudes and pat yourself on the
back, do you?
There's nothing glorious about grunt-level checking your work. I can
see why you don't do it.
Because he's incapable of checking anything more complicated than a 555?
...even a 240V plug?
I think he's used to doing simple stuff, like 555s and 4000-series
hairball async logic. If it's just a few chips, you can take some
risks and eventually get it to work most of the time. If you've got an
8-layer board with, say, 500 parts, and simulation isn't feasible, and
you allow a 1% risk of messing any one up, let's see, the chance of it
working is 0.99^500, which is under 1%. It might work on the third or
fourth board spin, maybe.
The three boards I've designed this year (for the new PPoE) are all over 1000
components (1000, 1200, and 1500). We'll see how well I did on the first in a
few days (our CM has made a mess of assembly).
That's always scary, the first board.
Don't know yet. We haven't gotten in back from the CM. I found out thisDo all the parts fit?
morning that they're asking about reference designators that don't exist and
parts that are clearly marked as OPEN.
No FPGAs on this one. I only have one minor CPLD on these.Do the power supplies come up?
Can the uP run code?
Does the FPGA configure?
Our only manufacturing capability is in Mexico. We're specifically trying toThat's just for starters, but those are milestones.
We always assemble the first articles in-house. That way, we can talk
to the assemblers if any problems come up.
avoid using them.