Ulf, what of these new AVRs? :)

Chris Hills wrote:

> Fair enough. Mind you I think at the time you were also "distracted" by
> a change of employment due to the company you worked for folding.  As

Yes, though one might say it had been effectively folded for a couple
of years prior to that :) It wasn't a major factor in the book not
coming out though - I successfully finished my second book during this
time, and started on my third.

> BTW I thought you said you had a directive to design out Philips?

The directive was announced the same day I gave him the board. I think
he did actually play with it anyway.

> >design in their parts again. We are a multi-billion-dollar corporation
> >with a directive to design out Philips.
>
> That is different to one person/small company  asking for a sample.

By the way, my department is also actively dumping IAR. Not to be used
for new development.

> the more money you are looking to spend the more help you get. So for a
> person looking for free samples who is not spending any money is not
> going to get anywhere.

How much money I spend on tools is nothing to do with how much money I
am going to spend on chips. We bought our license for one particular
IAR tool many years ago. It is no longer in production. We won't buy
any more because we can't buy any more. There will never be one more
dime in tool investment for that series of parts. Yet I might design in
a million dollars worth of the part tomorrow.

Semiconductor vendors don't need to know, and in my experience
practically never ask, what tools I am using.

> The way round this, in part, for the silicon distis is to let tool
> distis have some samples. Then when some one is buying tools for the
> part in question they can throw in some samples.

Very peculiar way of doing things, utterly counter to my experience. I
want bare chip samples, I ask the chip vendor (or distributor), not the
compiler or EVB vendor. Of course in many instances, there will be a
single rep covering both items.

> is not usually a problem. Surely you have good relationships with some
> silicon or tools vendors who will give you samples for you own use?

I don't like to exploit them for a couple of reasons -

1) I live close to where I work, so my area is mostly served by the
same reps as my workplace - I don't want those reps mentioning my
personal projects at work. None of anybody's business there. Conversely
I do not want those reps divulging to other people (colleagues, etc)
where I work. It's not that I'm ashamed of working here, I just don't
want the general public to know it.

2) If I'm doing a contract project for someone, it should go through
the "right" channels as far as sampling and prototyping goes. I.e. if
Atmel (or insert any other name here) is not interested in a 5000-piece
order, I will find this out when I try to order samples through the
vanilla route.

On 19 Jan 2006 08:41:42 -0800, "larwe" <zwsdotcom@gmail.com> wrote:

><snip>
>> The way round this, in part, for the silicon distis is to let tool
>> distis have some samples. Then when some one is buying tools for the
>> part in question they can throw in some samples.
>
>Very peculiar way of doing things, utterly counter to my experience. I
>want bare chip samples, I ask the chip vendor (or distributor), not the
>compiler or EVB vendor. Of course in many instances, there will be a
>single rep covering both items.
><snip>

This is one approach that never even crossed my mind, until this
thread.  If I bought software and received parts, I'd probably drop
them in a box for possible later use.  But I don't think I'd ever
consider that approach for securing samples.  Why should I decide upon
the software tool before I ever get samples to validate for the
application, in the first place???

Oh, well.

Jon

Sorry about late answer but the NEWS provider I use can suddenly
enter a mode where it refuses to accept posting.

There has been very limited engineering samples of the new mega128x/256x
during 2005 while the parts have been debugged.
I think about 80% of the parts I got went to distribution customers.
There are still a number of people waiting for samples.

I am holding my thumbs for february when rev E is out.
If everything works out OK, there should be plenty of samples by then.
ATmega1280/1s should also go into general sampling by then but production is
later.



-- 
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

I am not aware of any official policy within Atmel that says that samples
should be reserved for large customers.
I think that delays or non-deliveries are caused by the sample order system
which requires manual intervention.

There are two sample systems, one open to the public (on the Atmel web site)
and one for internal use.
The web sample order generates an email to a sales engineer
which has to enter the sample order into the internal sample system.

I suspect that there is a risk that large or known customers could get more
attention,
since the system depends on the sales engineer receiving the sample order.
We are all overworked and prioritize, right?

If I decided to open up my own shop, I would ask a *distributor* for samples
and request to have the sample order number back as a feedback.
This is a guarantee that the sample order has been entered.
I believe that if this is done this way, the system should ensure
that the samples are delivered quickly.

-- 
Best Regards,
Ulf Samuelsson

On Fri, 20 Jan 2006 12:38:19 +0100, "Ulf Samuelsson"
<ulf@a-t-m-e-l.com> wrote:

>I am not aware of any official policy within Atmel that says that samples
>should be reserved for large customers.

I don't think it was his own motivations operating, Ulf.  He was
apologetic about having to ask the questions he did.  So I'm pretty
sure that he was told to ask.  The fact that you aren't aware now of
such a policy doesn't mean there wasn't one then, or that it couldn't
have been a regional management decision and not a global one.  I
frankly don't know.  However, there is one thing I am certain of and
that is my experience and my copies of the email exchanges between the
parties.

>I think that delays or non-deliveries are caused by the sample order system
>which requires manual intervention.

That would probably _not_ explain the FAE's questions or his
explanations or his apologies.

>There are two sample systems, one open to the public (on the Atmel web site)
>and one for internal use.

This was 2002, keep in mind.

>The web sample order generates an email to a sales engineer
>which has to enter the sample order into the internal sample system.

I didn't use that approach.  If you read back through google, you will
see that I contacted my local distributor, All American, and worked
through them.  I already had spoken with the Atmel FAE on other
occasions about this particular part and some of the technical merits,
and also to begin a dialogue with Jaques about the part (which
proceeded slowly, since at first the FAE felt he needed to be in the
middle of the dialog and the 'around the world' issue added some
delays.)  I don't think the FAE was taken by surprise by All
American's request on my behalf.

It never even crossed my mind to consider using a web page for this,
at that time.

>I suspect that there is a risk that large or known customers could get more
>attention,

Well, sure.

>since the system depends on the sales engineer receiving the sample order.
>We are all overworked and prioritize, right?

 :)

>If I decided to open up my own shop, I would ask a *distributor* for samples
>and request to have the sample order number back as a feedback.
>This is a guarantee that the sample order has been entered.
>I believe that if this is done this way, the system should ensure
>that the samples are delivered quickly.

This may be true, for now, Ulf.  But believe me, I had confirmation
that All American had done their job.  The FAE and I began speaking on
this subject a few months after the order through All American and
when I was starting to wonder.  I don't believe there is any
reasonable question at all that the sample order had been properly
processed into Atmel.  He certainly didn't mention any such issue and
there was no discussion between me and the Atmel FAE that would
suggest anything like that -- quite to the contrary, in fact.

I'm pretty sure that wasn't the problem, not to mention the fact that
our discussions were honestly about other reasons why.  The FAE was
pretty clear.

...

I should add something.  If you do a google search using my last name
as the author, comp.arch.embedded as the newsgroup, and Atmel and the
AT90S2313 as keywords, and sort by date, you will find that my
earliest comments fitting that search (in 1999) were very positive.  I
still feel that way about the technical merits on many of the parts.

I appreciate your attempts to find an explanation for my experiences.
But I lived them, Ulf.  You didn't.  I will always have more
information about what happened to me than you will and that puts me
at an advantage over someone who is only able to speak generally from
quite a distance, literally and metaphorically.

So this isn't really fair to you and you should not have to try and
explain what did happen.  The most you should consider doing is to
simply appreciate it, but not try and change my mind about it.  You'll
lose that last battle, if that is where you are going, unless you are
truly willing to do a thorough investigation on the basis of emails I
have here and provide a convincing and detailed explanation of that
situation as it played out then.  Personally, I'm willing to just
leave it in the past and not worry about it.  There are many other
excellent options and I don't want or need to worry about the past.

Jon

On 20/01/2006 the venerable Ulf Samuelsson etched in runes:

> I am not aware of any official policy within Atmel that says that samples
> should be reserved for large customers.
> I think that delays or non-deliveries are caused by the sample order system
> which requires manual intervention.
> 
> There are two sample systems, one open to the public (on the Atmel web site)
> and one for internal use.
> The web sample order generates an email to a sales engineer
> which has to enter the sample order into the internal sample system.
> 
> I suspect that there is a risk that large or known customers could get more
> attention,
> since the system depends on the sales engineer receiving the sample order.
> We are all overworked and prioritize, right?
> 
> If I decided to open up my own shop, I would ask a distributor for samples
> and request to have the sample order number back as a feedback.
> This is a guarantee that the sample order has been entered.
> I believe that if this is done this way, the system should ensure
> that the samples are delivered quickly.

Well I've just put the system to the test and applied online for two samples of ATmega2561. I am
developing test code for a C compiler and need to run the code on an STK500/501 for testing.

Let's hope you are the sales enginer who receives my request.

Thanks in advance.

-- 
John B

>> If I decided to open up my own shop, I would ask a distributor for
>> samples and request to have the sample order number back as a
>> feedback.
>> This is a guarantee that the sample order has been entered.
>> I believe that if this is done this way, the system should ensure
>> that the samples are delivered quickly.
>
> Well I've just put the system to the test and applied online for two
> samples of ATmega2561. I am developing test code for a C compiler and
> need to run the code on an STK500/501 for testing.
>
> Let's hope you are the sales enginer who receives my request.
>
> Thanks in advance.

And if it works then you should get it in February
since that is when the general sampling begins of that part.

-- 
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

Jonathan Kirwan wrote:
> On Fri, 20 Jan 2006 12:38:19 +0100, "Ulf Samuelsson"
> <ulf@a-t-m-e-l.com> wrote:
>
>> I am not aware of any official policy within Atmel that says that
>> samples should be reserved for large customers.
>
> I don't think it was his own motivations operating, Ulf.  He was
> apologetic about having to ask the questions he did.  So I'm pretty
> sure that he was told to ask.

I have also been told to ask, but only for early samples
and not for parts in general sampling.
Early samples are always a special case and should not be confused
with general sampling policy.


> This may be true, for now, Ulf.  But believe me, I had confirmation
> that All American had done their job.
>
> I appreciate your attempts to find an explanation for my experiences.
...
>  You'll lose that last battle, if that is where you are going, unless you
are
> truly willing to do a thorough investigation on the basis of emails I
> have here and provide a convincing and detailed explanation of that
> situation as it played out then.

I think that my humble goal is to find out what went wrong
to be able to suggest the best approach to solve the problem or find a
workaround

In this case, only Atmel can solve the problem, but the workaround
is to order through a distributor and get the order number.
If the part is in general sampling, (and there is stock) then
the parts should arrive soon.
If the samples are ordered over the web, then it may or may not appear


-- 
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

On Sat, 21 Jan 2006 01:48:15 +0100, "Ulf Samuelsson"
<ulf@a-t-m-e-l.com> wrote:

>Jonathan Kirwan wrote:
>> On Fri, 20 Jan 2006 12:38:19 +0100, "Ulf Samuelsson"
>> <ulf@a-t-m-e-l.com> wrote:
>>
>>> I am not aware of any official policy within Atmel that says that
>>> samples should be reserved for large customers.
>>
>> I don't think it was his own motivations operating, Ulf.  He was
>> apologetic about having to ask the questions he did.  So I'm pretty
>> sure that he was told to ask.
>
>I have also been told to ask, but only for early samples
>and not for parts in general sampling.
>Early samples are always a special case and should not be confused
>with general sampling policy.

I can't say.  I ordered them when I read they were ready to be sampled
and I can't recall ever hearing the Atmel FAE mention that what I was
asking for was special in any way.  It probably would have made his
job easier with me, had he mentioned that I was asking for special
attention.  And I certainly didn't go into this expecting any special,
advance treatment.  It was a simple request for two samples of a part
that Atmel had openly advertised as sampling and where that was
confirmed by All American through their own conversations with Atmel,
relayed back to me.  But I cannot say what All American represented to
Atmel.  That would be outside my view.

Jon

Hello Tsevtan,


>>Joerg написа:
>>

Nice. Now I know my newsreader can display cyrillic letters :-)

> 
>>Suits me, I just would never buy stock in a company that thinks they can
>>survive by catering to only the big ones.
> 
> sad truth to all small co. is that they do just fine this way
> 

When I recall some of the speeches by CEOs of the big European ones I am 
not so sure their semiconductor business is doing that well. I still 
remember Jan Timmer from Philips (a long time ago) slamming his fist 
onto the pulpit and that if sales didn't improve there would have to be 
organizational consequences and all that. I could have told him about 
some of the things that they did wrong. The mistakes were nearly all in 
marketing.

Now I am not complaining about stuff you can buy from Digikey, Maplin, 
Reichelt. Or from Olimex :-)

What I am not tolerating is that when there is no way to get a part 
anywhere else. Last case was a FET. I called the big company and told 
them I or my client need x samples and we'd pay for them and the 
shipping and all. They promised and then didn't come through. What they 
don't realize is that a small consultant can have a substantial impact 
on their bottomline when he then decides to design such a part out and 
use one from ONSemi instead. Those decisions are typically final, for 
the next ten years of mass production or so. It's big $$.

There is a reason why companies such as TI are more successful than others.

Regards, Joerg

http://www.analogconsultants.com