[ANN] Atmel invest millions of $ to please Jim Granville

On Thu, 10 Mar 2005 21:54:56 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

>So, we all got hardened over time. Basically, if Digikey doesn't have 
>stock or worse doesn't carry a part at all we try our utmost not to need 
>that part in a design. Exceptions are very, very rare.

This is a rule I follow, as well.

Jon

In article <5li231ptov2ul492nn8oo6r04qvbjjqfkj@4ax.com>, 
jkirwan@easystreet.com says...
> On Thu, 10 Mar 2005 21:54:56 GMT, Joerg
> <notthisjoergsch@removethispacbell.net> wrote:
> 
> >So, we all got hardened over time. Basically, if Digikey doesn't have 
> >stock or worse doesn't carry a part at all we try our utmost not to need 
> >that part in a design. Exceptions are very, very rare.
> 
> This is a rule I follow, as well.
> 

I took up this rule also---after getting singed on some Maxim and
Analog Devices parts.   Now that DigiKey carries both those lines,
one-stop shopping is easier than ever.  The DigiKey website
is the de-facto authority for "what parts can I get tomorrow--
and will they be there next year".

I wonder why any company in the semiconductor business would NOT
want to be in the DigiKey catalog---or why some lines from a
company would not appear.  (I still have to get Xilinx CoolRunner
CPLDS elsewhere)    It's been at least 10 years since they outgrew their 
hobbyist origins.   DigiKey is a classic distributor---they buy big lots 
from their suppliers, maintain inventory, and process orders for smaller 
lots.  They concentrate on providing fast and reliable service and don't
try to be salesmen for preferred lines by sending out a
corps of "applications engineers".

I sometimes wonder what percentage of all the UPS and FEDEX trucks in
Minnesota visit Thief River Falls each day.

Mark Borgerson

> I sometimes wonder what percentage of all the UPS and FEDEX trucks in
> Minnesota visit Thief River Falls each day.

Probably a very small percentage, or none at all. They either have
dedicated trucks that pick up from them, or they might deliver it
direct to the local UPS/FDX sorting center with their own vehicles.

On Fri, 11 Mar 2005 09:54:39 -0800, Mark Borgerson
<mborgerson@comcast.net> wrote:

>In article <5li231ptov2ul492nn8oo6r04qvbjjqfkj@4ax.com>, 
>jkirwan@easystreet.com says...
>> On Thu, 10 Mar 2005 21:54:56 GMT, Joerg
>> <notthisjoergsch@removethispacbell.net> wrote:
>> 
>> >So, we all got hardened over time. Basically, if Digikey doesn't have 
>> >stock or worse doesn't carry a part at all we try our utmost not to need 
>> >that part in a design. Exceptions are very, very rare.
>> 
>> This is a rule I follow, as well.
>
>I took up this rule also---after getting singed on some Maxim and
>Analog Devices parts.   Now that DigiKey carries both those lines,
>one-stop shopping is easier than ever.  The DigiKey website
>is the de-facto authority for "what parts can I get tomorrow--
>and will they be there next year".
>
>I wonder why any company in the semiconductor business would NOT
>want to be in the DigiKey catalog---or why some lines from a
>company would not appear.
><snip>

Only perhaps if DigiKey has some terribly unfavorable terms in their
contracts on new items, I suppose...

In any case, since Atmel's new single-cycle '51 CPUs aren't at
DigiKey, I won't look much further into them that I have already.
When they show up at DigiKey, I'll spend some more serious time to see
if there is something worth the time.

....

By the way.... thinking about the Atmel parts makes me think of the
Cygnal parts (I mean, Silicon Labs') and this makes me wonder about
something else....

I've been in a business area that caters to custom customer needs and
requires fairly deep involvement on their processes and isn't more of
the "industrial product" areas where customer support usually isn't
nearly as 'deep'.  So one of the features that I've wanted to consider
adding to our product lines is the ability to do in-situ diagnostics
via the JTAG port.  One of the problems I run up against, time and
again, with these CPUs which have some form of JTAG debugging built in
is that the manufacturers are loathe to disclose the JTAG debugging
details.  In the case of ARM7, the JTAG debugging is disclosed.  But
pretty much elsewhere I've looked, the answer is no.

Does anyone know of ANY microcontrollers in the 8-bit or 16-bit arena
which include internal flash and ram and JTAG and where the JTAG
debugging features are disclosed or else easily available through NDA
(without a wringing of hands and gnashing of teeth?)

Jon

Hello Jonathan,

> Does anyone know of ANY microcontrollers in the 8-bit or 16-bit arena
> which include internal flash and ram and JTAG and where the JTAG
> debugging features are disclosed or else easily available through NDA
> (without a wringing of hands and gnashing of teeth?)

All I have heard so far is that the JTAG port isn't always usable for 
what it was originally conceived, boundary scan. But on a uC you have a 
lot of options. If the circuitry around it is designed with testability 
in mind you can actuate outputs and then have the uC see whether the 
expected scenario happens.

This can require additional circuitry but if the uC has ADC features 
that isn't much. Say, you had an 8 bit port that pipes data to a buffer. 
If that buffer had a wee resistor in its VCC line and then maybe a 
couple resistors as a divider hooked up to that to get into the valid 
range you could ping one pin at a time and then check for reaction. With 
CMOS buffers you'd have to send a stream of ones and zeros alternating 
as rapidly as possible because they only consume dynamic power unless 
there are terminators somewhere. There will be a slight dip, prolonged 
by the bypass cap. Just one idea.

Regards, Joerg

http://www.analogconsultants.com

On Fri, 11 Mar 2005 23:14:39 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

>Hello Jonathan,
>
>> Does anyone know of ANY microcontrollers in the 8-bit or 16-bit arena
>> which include internal flash and ram and JTAG and where the JTAG
>> debugging features are disclosed or else easily available through NDA
>> (without a wringing of hands and gnashing of teeth?)
>
>All I have heard so far is that the JTAG port isn't always usable for 
>what it was originally conceived, boundary scan. But on a uC you have a 
>lot of options. If the circuitry around it is designed with testability 
>in mind you can actuate outputs and then have the uC see whether the 
>expected scenario happens.
>
>This can require additional circuitry but if the uC has ADC features 
>that isn't much. Say, you had an 8 bit port that pipes data to a buffer. 
>If that buffer had a wee resistor in its VCC line and then maybe a 
>couple resistors as a divider hooked up to that to get into the valid 
>range you could ping one pin at a time and then check for reaction. With 
>CMOS buffers you'd have to send a stream of ones and zeros alternating 
>as rapidly as possible because they only consume dynamic power unless 
>there are terminators somewhere. There will be a slight dip, prolonged 
>by the bypass cap. Just one idea.
>
>Regards, Joerg

I'm actually thinking here of some convenience in diagnosing problems
at the customer site.  I'd like to carry in a hand-held that I can
attach to the JTAG port and set breakpoints while the device operates
normally.  (I cannot accept modifying the official code in any way.)
If and when a certain unlikely error condition takes place I can trap
at that point and then examine all of the variables.  But I'd have to
build a custom device for this -- Windows laptops often cannot be
brought in and used with existing software.

In any case, whether or not for the above circumstance specifically,
I'm still interested in microcontrollers with documented JTAG
debugging.

Jon