TRSF3243 RS-232 Transceiver Overshoot Spec

Meng Chan

Hi,

Does anyone knows what is the overshoot/undershoot spec for the CMOS side this RS-232 transceiver? Is there any problem of exceeding the Absolute Input Volatge Range (-0.3 to 6.0V) for 2ns? The spec has this statement "Exposure to absolute -maximum rated conditions for extended periods may affect device reliability." Does anyone know what is the definition of the spec's "extended period"?

Thanks,

over 11 years ago

0 Ron Michallick over 11 years ago

TI__Guru**** 151276 points

Meng,

2nS is in no way an extended period. How large are the 2nS overshoots?

0 Meng Chan over 11 years ago in reply to Ron Michallick

Prodigy 60 points

Ron,

For positive overshoot, the peak is 4.8V which did not exceed the absolute max 6.0V. The duration of exceeding VCC (3.3V) is approx 2ns.

For negative overshoot, the peak is -1.78V. The duration of exceeding the -0.3V spec is approx 2ns.

The positive overshoot may be OK, as it does not exceeding the absolute max; the negative overshoot is the worst. Since this is an UART communication channel for the processor, there are always activities (overshoot is repetitive) on the channel. Even if the overshoot duration is small, does the repetitive natural of this signal make a difference? What is the mechanism dictates the absolute max and min at the I/O? and how the chip vendor specify the absolute min and max for the I/O?

Thanks,

Meng

0 Ron Michallick over 11 years ago in reply to Meng Chan

TI__Guru**** 151276 points

Meng,

The rule of thumb is -1V for less than 10ns is OK. Design rules are a little looser 30% of Vcc which in this case is about the same. The inputs have an ESD diode to ground. -1.78V is a lot of current. The cause of the overshoots should be found and fixed.

0 Meng Chan over 11 years ago in reply to Ron Michallick

Prodigy 60 points

Ron,

On the ESD protection diode subject; If the protection diode is conducting, I should see the overshoot waveforms chopping or clamping by the diodes. But in my case, I don't see this type of behaviors, see attached waveforms. Am I missing something here? What is the typical energy absorption rating for this type of ESD protection diode?

Thanks,

Meng

8055.NPU_UART0_TXD_U5_12_fall_1.TIF

4377.NPU_UART0_TXD_U5_12_rising_2.TIF

0 Ron Michallick over 11 years ago in reply to Meng Chan

TI__Guru**** 151276 points

Meng,

At -1.8V, hundreds of mili amps can flow. However a diode takes a finite amount of time (a few nS) to reach this current level. I can ask design for ESD rating, but they will say that the cells are not meant for repeated use. The energy per pulse is quite low (using rounded numbers), 500mA * 1nS * 2V = 1nJ. Even at 1Mbps (alternating bit pattern), the power is 1mW.

Is it certain that the measurement is accurate? A high quality probe and careful attention to connection points (very short loop length) is needed to measure the actual signal.