TLV3602-Q1: Dead time generator solution designed with TLV3602

Jass

Thanks for your post.  Unfortunately, I am not totally understanding what you are seeking.  I understand the concept of deadtime but your suggestion of using hysteresis to control dead time is a bit confusing.  Is the waveform we are monitoring a square wave or more of a sawtooth/triangle wave.  Would be best if you could make some hand-drawn timing diagram of what you are seeking.

Chuck

Hi Sins,

The input waveform is one PWM, but the output waveform need to be two complementary PWMs which have dead time that can be adjusted by RC network.

Jass

What would the min pulse width and amplitude of the PWM be?  Likewise, will the delay be the same for rising and falling edges?  I am thinking we need soften the edge of the PWM main signal (slew rate limit) and use a dual DAC to change the threshold of the comparators.

Chuck

The min pulse width is really large than dead time, the PWM duty is around 10-90% and the amplitude is 3.3V

The delay need to be same for rising and falling edges.

And the DAC is quite complex, it's better with external RCD network.

By the way, the complementary PWM look like this, the input is only one PWM and the output are 2 pairs which has overlap for low-level.

Jass

using an RC network to slow the edge is necessary but you still need to set the break-before-make.  Or are you saying this is a one-time thing that occurs in the factory?  For example, someone will assemble with the appropriate resistor value to accomplish a specific delay.  I was thinking the RC provides the slope and the DAC would be used to set the actual delay.  Sounds like you are thinking otherwise.  Lastly, I need to know signal amplitude.  PWM inputs are 3.3V logic levels?

Chuck

The PWM is 3.3V logic level. Customer need TLV3602 and peripheral passive device to convert one single PWM into this complementary PWM pairs. How can we design it?

Thanks for your post. We are on holiday break and will review your post when we return on Tuesday. Sorry for the delay. 
Chuck

Hi Sins, do you have any updates?

Sorry for the delay Jass.

Here is my initial thought of how we could create the dead times.

Chuck

TLV3602_dead_time_gen2.TSC

Thanks, Sins!

For the simulation results, will we have some oscillation if input signal has small noise? I don't see hysteresis area for TLV3602.

For dead time accuracy, if we are using 1kohm instead of 100ohm input RC network, can you give some formula to calculate the errors? TLV3602 have large input bias current.

Hi SIU & Sins,

Can you add some resistive network for simulation? I don't have this software.