Yahoo Groups archive

68300

Index last updated: 2026-04-29 00:01 UTC

Message

Re: TPU Function for Variable Frequency & Duty Cycle

2003-02-02 by thisobj <fvolker@yhti.net>

Jeff,

For the variable duty cycle (PWM type) function using QOM:
It is probably best to simply write the new "on" interval parameters 
into the parameter RAM timing segments with the QOM in continuous 
mode. Worst case delay in update would be one "pwm" period + normal 
microcode latency.  I like this better than using the full host 
service with immediate update, since this could indeed cause 
considerable jitter and very narrow spikes with hi-freq content. If 
your "on" time is defined by 2 entries in the parameter RAM, then 
update with a 32-bit (atomic) write.  If more than 2 entries, then 
you might want to consider using a different T1/T2 clock rate or 
updating the value in steps.

For the variable frequency output channel:

If you're merely changing period while "on" time remains fixed, then 
use the above technique applied to the "off" time.

One thing to remember when using QOM is that the delta times per each 
entry in the parameter RAM is only 15-bits, since the low bit is used 
to determine the hi/lo pin state at the end of the interval. 

Frank  

--- In 68300@yahoogroups.com, "jeffreyslcu <jeffrey.s.sato@l...>" 
<jeffrey.s.sato@l...> wrote:
> I am using a 68332 with the 'G' TPU stamp in two applications.  One 
> application needs a variable frequency output, and the other needs 
a 
> variable duty cycle adjustment (but set frequency) and each only 
has 
> one TPU channel. They both require on-the-fly adjustments 
(operating 
> in a closed-loop control application).  I have identified the 
Queued 
> Output Match (QOM) function as the best possible candidate for both 
> applications. But, from the application note, I cannot tell if when 
> the TPU registers are updated, will I get a non-desired output at 
the 
> transition point? For example, when the update is entered will it 
> take affect mid-signal and possibly produce a transient with a much 
> higher frequency component, or does it wait for the in-process 
output 
> signal to complete before initiating the next update?
> 
> Thanks for any inputs, Jeff

Attachments

Move to quarantaine

This moves the raw source file on disk only. The archive index is not changed automatically, so you still need to run a manual refresh afterward.