[Coco] OT: 63B09

[john dumas]

On 11/1/2011 5:28 PM, gene heskett wrote:
> On Tuesday, November 01, 2011 07:25:14 PM john dumas did opine:
>
>> On 10/29/2011 5:37 PM, gene heskett wrote:
>>> On Saturday, October 29, 2011 06:09:27 PM Louis Ciotti did opine:
>>>> The processor is not for the COCO. I am debugging a radio
>>>> controller that has a 63B09P in it. I found a ?NTE6809 which has
>>>> the exact same  Pinout, but the crystal does not oscillate when
>>>> I insert it into the MCU socket, but The old one does. I am
>>>> wondering if this means the NTE6809 is DOA. It was NOS in the
>>>> bag when I got it. 
>>>>
>>>> I need to know if the problem is a compatibility issue between the
>>>> NTE6809 and the HD63B09P or is it just that the NTE6809 is bad.
>>> Well, an ordinary 6809 is rated I believe a PMOS tehcnology version
>>> rated for a 1 Mhz clock, the B version, which your about 10 years
>>> newer Hitachi CMOS is, is rated at 2Mhz.  I have a C version, rated
>>> at 3Mhz in my coco, and from the waveforms I can read external to the
>>> chip, I get the impression it could run quite a bit faster than 3Mhz.
>> The MOT versions were all NMOS Depletion-Load design. The "A" and "B"
>> versions
>> were simply the same die graded out for speed.
>>
> Thanks John.  Why I had PMOS, a very old tech stuck in my memory I have no
> clue.  But you are quite correct.  :)  Wasn't it the PMOS stuff that
> suffered the purple plague&  had a lifetime of 15 years at best?
Probably. I seem to remember it at about the same time
period. I was at TI at that time and we had - created - our
own lifetime problems. The process in those days was PMOS.
To get low Vt, we would use 110 material, but then the field
breakdown voltage became too low. So someone there decided
that the solution to low Vt was the "Nitride" process (Nitride/Silicon
dioxide sandwich in the gate area).

Sounded great but somehow it was ignored that IBM had published
a white paper detailing the problems. Well TI moved to the process,
got low Vt and shipped a whole pot load of parts. In about a year, they
started coming back nonfunctional.

Baking them overnight restored function. Sure enough IBM was right;
after so many hours of operation, the electric fields would sweep the
cr at p in the oxide to the nitride/silicon dioxide interface in the gate
and RAISE the Vt - stopping operation. Baking the part caused thermal \
agitation to move the "stuff" out of the gate area- restoring operation.

A life of about a year!
Unless you designed in a heater to "bake" the chip....HA!
Needless to say, ion implantation arrived in a hurry....(grin)

Don't ask about early metal problems that got worse when you
got BETTER alignment of the photo masks..... Learning a new
technology is _such_ fun.

johnd


>
>> I can't imagine anyone redesigning the '09 for PMOS!
>> That was "previous technology" at the time of the '09 design.
>>
>> This is from the best of my memory; I was a member of the MPU
>> design group while the '09 was designed........
>>
>> cheers,
>> johnd
>
> Cheers, Gene


-- 
One useless man is a shame, two are a law firm, three is Congress.