Re: Obama doesnt want you to know the real truth about Global Warming

On Tue, 29 Sep 2009 16:01:55 -0700 (PDT), Bill Sloman
<bill.sloman@xxxxxxxx> wrote:

On Sep 29, 6:37 pm, John Larkin
<jjlar...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
On Tue, 29 Sep 2009 09:11:36 -0700 (PDT),Bill Sloman

<bill.slo...@xxxxxxxx> wrote:
On Sep 29, 2:40 am, dagmargoodb...@xxxxxxxxx wrote:
On Sep 27, 6:37 pm,Bill Sloman<bill.slo...@xxxxxxxx> wrote:

That's not auto calibration, that's just replacing trim pots and screw
drivers with digital pots and an eprom programmer.

Bill you're just being tedious.  Autocalibration is soooo 20th
century--we just don't feel the need to discuss it.  That doesn't mean
we don't do it.

You may do it as a matter of routine, although your personal database
doesn't show much sign of having been up-dated recently. John Larkin
doesn't seem to be entirely at home with the concept.

I'll do whatever works. Lately we've been moving entire nonlinear
channel calibration algorithms into FPGAs. But I'm not "entirely at
home" with building primary standards into measuring instruments, or
taking them offline for self-test or recal without an explicit command
from the user to do so. Even if I could do that transparently, my
customers *really* wouldn't like their gains an offsets being changed
invisibly, mid-run.

John Larkin does waste a lot of time constructing straw men. He makes
one of his typical exaggerated claims, I point that the claim isn't
valid if you put a primary standard inside the instrument, which is
possilbe, if not usual

http://tf.nist.gov/ofm/smallclock/CSAC.html

Is that in the Mouser catalog yet?


and he then proceeds to claim that I am advocating putting primary
standards in every instrument.

You've suggested it several times. Sounds like a developing
compulsion.



That's the problem having customers, I guess. It must be be very
liberating to have no customers.

Considering the amount of time you've spent building your straw man
and berating me for adopting a position that you took the trouble to
invent for me, it would seem that you must be running short on
customers at the moment, and are having to find other ways of filling
in your empy days. The anxiety seems to be making you irritable.

A day in the life:

Two customer-support calls

Five emails ditto

One customer inquiry about a new product, having to do with muons or
something

One meeting with the Xilinx folks, who I had to go fetch when they got
lost a few blocks away and couldn't find the place. That happens a
lot, actually.

Two neat part sample requests, both LC filters, Mini-Circuits and
Coilcraft.

One engineering meeting

About six mini-meetings, mostly technical

Lunch

One extended discussion about popular culture


I'm not irritable lately, but I can't design stuff with all that going
on. May as well kill time on the web between interruptions, looking at
news and checking the new product announcements and wiki-ing various
topics. I've got to get away to get anything real done these days.

I did get to design a pc board yesterday, a 24-channel high-voltage
ADC for a spectroscopy system, but that was routine part-plopping and
only took a couple of hours.


Some customers need certified data: confirm cal before a run, take
data, confirm cal after. The idea of continuous calibration is messy
there. The only things you might get away with would be tweaking
dimensionless things, like auto-zero or some such.

Our newer VME modules have a relay per channel and a test connector
(upper D9 in this case)

http://www.highlandtechnology.com/DSS/V450DS.html

so the customer can bus all the test connectors to a
tracable-calibrated DVM or whatever. He can pull in a relay and route
the channel to the cal instrument without disconnecting the field
wiring, making it a software-only thing to verify the cal of the
entire system, every morning if he wants. Why use relays? Because
Pratt&Whitney told us to use relays.

And you couldn't come up with a better alternative.

Actually, I couldn't.

John

.

Relevant Pages