> datasheets from Samsung.  Badly mistranslated and full of errors and
> typos.

You are right with Samsung too. Last year I programmed the S6B1713 
LCD/STN grafic controller. I understood this sheet much faster but
probably LCD grafic controllers are not as complex than DSP Timers.
Nevertheless English is not my language and typos doesnt hit
me so much but the Samsung translation was obviosly Korean English.
Thats why I was lucky that it was not required to learn Korean ...

In article <eNSdnRIBS-PB5QHdRVn-sw@comcast.com>, Clark G. Smith wrote:

> Ok Janvi, I don't have time to read all that shit, but I understand you
> complaint, I think.  My most recent project was on a  DM642 DSP, which is
> where I understood, TI has the, without any doubt, worst god-damned
> technical documentation of anybody in the world.

You may not likee th way they've organized things, but I've
always found TI datasheets to be fairly accurate and in
understandable English.  Try datasheets from Samsung.  Badly
mistranslated and full of errors and typos.

-- 
Grant Edwards                   grante             Yow!  I need to discuss
                                  at               BUY-BACK PROVISIONS
                               visi.com            with at least six studio
                                                   SLEAZEBALLS!!

Ok Janvi, I don't have time to read all that shit, but I understand you
complaint, I think.  My most recent project was on a  DM642 DSP, which is
where I understood, TI has the, without any doubt, worst god-damned
technical documentation of anybody in the world.  Its like they have a
theory, lets take the information, spread it over as many documents as
possible (all *.pdf, none printed), make that practically unuseable, and
give it to our customers..........  Maybe a note to TI...........  that
DM642 is probably the baddest-ass processing chip on the planet (ethernet
and everything) but you have to support the son-of-a-bitch..........  and
part of "support" means comprehensive documentation.........



"Janvi" <janvi@despammed.com> wrote in message
news:pan.2004.05.07.14.04.39.809898.2127@despammed.com...
> Already walked through a dozend of diffrent uC Architectures Zilog,
> Motorola, Infineon, STM, AMD and more in the recent years.
> But for the moment the Texas Instruments TMS320C24x Peripheral
> description drives me crazy.
>
> The Texas datasheet consists of a mix of feature lists and examples some
> diagramms and some descriptions. Everything shaked and then printed.
>
> There are 3 Timers. Only Timer 2&3 can be cascaded. Each timer has
> its own compare function but there are also 3 full compare units and
> 3 simple compare units. But later you read, that full compare is
> only possible for timer1 and simple compare is only possible for timer
> 1 and 2. Full compare outputs are inverted and non inverted with
> additional dead band support for push pull amplifiers while simple
> compare units have only one output pin. Becouse the simple compare
> output pin is diffrent from the full compare outputs, there is a total
> of 9 output pins for 3 timers. Maybe I start being to old
> to learn new things but for those 3 timers I had to read 100 pages !!!
> (chapter2) and when I am through I have to start again becouse I
> did not understand anything and have no overview above the possible
> functions.
>
> Maybe things are complex but the more complex something is, the more
> care should be taken on how to divide the chapters. Good diagrams can
> say more than 1000 words. The writers for datasheets may be good chip
> designers but sometimes I doubt they already heard anything from
> didactic skills.
>
> But maybe there are some Texas Counter specialists here:
>
> My Timer 2 is connected to an encoder. This works already well
> counting up and down with the axis. Only forward direction important.
> At a specific angle, a magnet valve should be switched on
> and at another angle it shoud be switched off again. The
> rising edge current for the magnet valve should be an initial pulse
> with variable length. Therefore timer 3 can possibly provide a
> time reference. After that, a hold current limited by PWM should
> be applied to the valve. Therefore timer 1 provides a PWM base
> freqency (approx 5-10 khz) and variable width. Cannot use the
> simple compare outputs becouse this would require a HW change in
> the design. Timer 3 currently makes a countdown interrupt for system
> time tic interrupt.
>
> How to setup the Action Controll Registers and what is the best
> output pin? This task requires 2 timer outputs one for initial pulse
> and one for PWM connected by wired or? Can I achieve HW action accuracy
> for the initial pulse, pulse length and PWM stop position or time or do
> I have to use Interrupt Service to enable disable the PWM outputs?

Already walked through a dozend of diffrent uC Architectures Zilog,
Motorola, Infineon, STM, AMD and more in the recent years.
But for the moment the Texas Instruments TMS320C24x Peripheral 
description drives me crazy.

The Texas datasheet consists of a mix of feature lists and examples some
diagramms and some descriptions. Everything shaked and then printed.

There are 3 Timers. Only Timer 2&3 can be cascaded. Each timer has 
its own compare function but there are also 3 full compare units and
3 simple compare units. But later you read, that full compare is 
only possible for timer1 and simple compare is only possible for timer
1 and 2. Full compare outputs are inverted and non inverted with 
additional dead band support for push pull amplifiers while simple
compare units have only one output pin. Becouse the simple compare 
output pin is diffrent from the full compare outputs, there is a total
of 9 output pins for 3 timers. Maybe I start being to old 
to learn new things but for those 3 timers I had to read 100 pages !!!
(chapter2) and when I am through I have to start again becouse I 
did not understand anything and have no overview above the possible
functions.
  
Maybe things are complex but the more complex something is, the more
care should be taken on how to divide the chapters. Good diagrams can
say more than 1000 words. The writers for datasheets may be good chip
designers but sometimes I doubt they already heard anything from 
didactic skills.

But maybe there are some Texas Counter specialists here:

My Timer 2 is connected to an encoder. This works already well 
counting up and down with the axis. Only forward direction important.
At a specific angle, a magnet valve should be switched on
and at another angle it shoud be switched off again. The 
rising edge current for the magnet valve should be an initial pulse
with variable length. Therefore timer 3 can possibly provide a 
time reference. After that, a hold current limited by PWM should
be applied to the valve. Therefore timer 1 provides a PWM base
freqency (approx 5-10 khz) and variable width. Cannot use the 
simple compare outputs becouse this would require a HW change in 
the design. Timer 3 currently makes a countdown interrupt for system
time tic interrupt. 

How to setup the Action Controll Registers and what is the best
output pin? This task requires 2 timer outputs one for initial pulse
and one for PWM connected by wired or? Can I achieve HW action accuracy
for the initial pulse, pulse length and PWM stop position or time or do
I have to use Interrupt Service to enable disable the PWM outputs?