ESD protection and transient voltage suppression devices

Are ESD protection devices for signal lines and transient voltage suppression diodes for power lines same? For ex, if I have a data line that is 5V logic level and a power input of 5V to the board can I use the same type of devices for both?

I have a regulated 5V DC coming in to the board via 2-3 mtr cable. Can I use the ESD protection diodes specified for 5V data lines here instead of transient voltage suppression devices specified especially as for data lines?

I understand low capacitance required for high speed data lines. Is low capacitance of esd devices specified for data lines an issue for power lines?

Do I need unidirectional or bi-directional for 5V DC power line?

Thanks
m.jnk

 

There're the same, but different. Both are Zener-like devices but on
different scales and tuned for different purposes.

Possible, but probably not. There are many strategies for dealing
with ESD. Some use a series impedance with a capacitor, or clamp. The
series impedance limits the energy from an ESD strike so a smaller
device can be used. There is a tradeoff that has to be made to select
the appropriate device. The chances are that your diodes specified
for 5V data lines won't be robust enough to protect a power input,
where you may or may not be able to add sufficient impedance to limit
the energy.

They make hundreds of different devices for a reason.

Low capacitance generally means low energy capability. Power lines
generally require higher energy capability.

Usually, unidirectional devices are used for DC inputs. You only want
the Zener action in the direction of the applied voltage. In the
reverse direction the TVS diode is a diode, so will protect against
excursions below ground, which will also destroy devices. There isn't
any reason to limit the negative excursion to the positive supply
voltage. The reverse diode will also help protect against supply
reversal (assuming you have a protection device to limit current).

Bidirectional TVS diodes are most often used for differential data
lines, where there is no reference to ground. The TVS diodes are
selected using the maximum common mode voltage.

 

Thanks a lot. That makes it pretty clear.

 

Diodes are more rugged when forward biased than reverse biased, so I
wouldn't go bidirectional if you don't need that feature.

The machine model only has a 200pF cap, but nearly no resistance. The
human body model is 100pF with a 1500 ohm resistor. For power supply
lines, the capacitance on the line itself should be enough to limit the
voltage. Just think of it as a charge transfer design problem, i.e.
conserve charge and you can predict the final voltage.

Generally the machine model is the one that gets you.

 

There is of course the other option of using a double diode such as
a BAV99 to clamp the signal just above the power supply or just below
ground. Then you only need to ensure the supply can't be pulled too
high. Usually the decoupling capacitors achieve this, but you if you
want add a high-power clamping device to the power supply just in
case.

The nice thing about double diodes on signal lines is that they have
low capacitance but can pass considerable current for short periods.
It is a good idea to have a little series resistance between the diode
and the outside world and some more between the diode and the
sensitive
device input to limit the current that flows during an esd event.

John

 

Today's on-chip ESD protection networks are only 100 fF or so for high
speed paths but we're doing good to survive 2kV from a HBM
(human-body-model) discharge. Fast I/O has diodes to supply and gnd and
then supply to gnd clamp circuits where the higher capacitance is
tolerated. 8-12 Gbps I/O seldom use series resistance in my (perhaps
limited) experience.