Original question:
Hi team,
How does propagation delay affect dead time max/min value and dead time matching?
Datasheet says dead time range when RDT=20k is 160ns~240ns. Also dead time matching is 20ns max.
Do they include propagation delay, whose max value is 40ns?
Also do they include propagation delay matching, whose max value is 5ns?
In addition, how should we design dead time with including propagation delay?
Regards,
Ochi
Ochi-san,
The dead-time is defined as the time when neither ChA nor ChB are on/conducting.
Propagation delay itself doesn't necessarily affect deadtime if everything is well-matched.
If the propagation delay varies between turn-on and turn-off or channel to channel, and prop delay matching is not totally perfect from channel to channel, then dead-time is needed to avoid short-circuit.
Regarding your question about whether dead-time figure includes the prop delay / matching: The dead time is enforced between the rising and falling edge of output as in the Figure 30 below.
The operation of dead-time waveform is shown in this figure of the datasheet
To design dead-time, the basic principle is to find the minimum off time you need to ensure that both channels will never be switched on, plus some margin/headroom.
This can be done by: sum pulse-width distortion, channel to channel delay matching, and difference of rise and fall time for a channel. For accuracy the rise and fall time should be based on the load that will be driven.
I hope this helps. Let me know if you have more questions.
Best
Dimitri
