[sdiy] TZ triangle core

[Veronica Merryfield]

On 2011-03-02, at 11:06 AM, Neil Johnson wrote:

> Hi Ian,
> 
>> Turns out it doesn't. The reason is that it only will work with one input polarity.  The opposite polarity puts the integrator/ST out of phase, so the system runs immediately to the rail.  <headslap>
> 
> Yes, 5 minutes thinking about this and I come to same conclusion: once the Schmitt has reversed the signal trajectory it won't flip until the other bound is reached.  If the signal reverses due to external control then the Schmitt won't see it as its looking in the wrong direction.
> 
> So a window comparator is needed, coupled to a T-flip flip as you say.  This should catch the signal whatever bound it reaches, irrespective of its history, and then flip its direction.  I would also add a safety window comparator set to wider limits (say +/- 7V) to catch any unfortunate startup conditions.

As long as you have one comparator set to "greater than" and the other to "less than" you will catch the normal excursions plus any out of bound startup conditions. And one could use an SR flip flop.