Hi guys,

I often see diagrams with a configuration to protect inputs from ESD and overvoltage using a series PPTC fuse and a TVS diode.
What is the point of protecting an input in this way? An ESD event is a high-voltage event.
This voltage would easily break the series fuse, which usually has a voltage <= 60V, right?

Source: Circuit for protecting ADC with TVS diode and PTC fuse

 

Hello.

TVS is Transient Voltage Supressor, so it supresses the voltage transients... by doing what? It must increase the current, so the energy gets dissipated. This increase in the current can also do harm to the PCB by opening a copper track, as the common 'a lot of current passing through the components in series'. To limit this to safe values, we put a fuse before the TVS, so the fuse gets hot and "opens" the circuit (in the case of PPTC is temporary).
Also the overall effect is to limit the voltage on the input and if the transient continues, it opens the circuit, leaving the input healthy.

 

I used a 45Ω 265V PTC thermistor and a 5V SMAJ TVS diode to protect Microphone and RS485 inputs. That combination will survive direct connection to 230V mains.

 

A thermistor will not be nearly fast enough to have much effect on a static spike. A TVS diode should be effective, probably.
Microphone and RS485 inputs are rather different, I can advise on protecting microphone inputs. For magnetic microphones, fast diodes, two in series, connected across between the input signal and common (local grounded) lines will clamp anything much over a volt. For balanced-line differential inputs that also include phantom power it is more complicated. A similar diode clamping scheme would be used past the capacitors separating the phantom power from the signal input. There are many published circuits of good commercial PA amplifiers showing that scheme applied.

 

The thermistor doesn't need to have any effect on the spike, the TVS deals with it. The thermistor only comes into play if the over-voltage situation is rather longer-lasting than a spike to the extent that the TVS would overheat. It then increases its resistance, lowering the current through the TVS to a safe amount.
The circuit's ability to deal with mis-connected mains is handy when it may be used in an environment of extreme stupidity.

 

The thermistor doesn't need to have any effect on the spike, the TVS deals with it. The thermistor only comes into play if the over-voltage situation is rather longer-lasting than a spike to the extent that the TVS would overheat. It then increases its resistance, lowering the current through the TVS to a safe amount.
The circuit's ability to deal with mis-connected mains is handy when it may be used in an environment of extreme stupidity.

OK, and of course the thermistor will need to have a POSITIVE temperature coefficient.
AND, there are those whom I will turn and walkaway from.

 

Sorry maybe I explained myself wrongly. My question was:

If I arrange a ptc fuse in series before a TVS diode, can this fuse break due to the high voltage ESD transient?
Is it still not subject to the ESD event?

 

Sorry maybe I explained myself wrongly. My question was:

If I arrange a ptc fuse in series before a TVS diode, can this fuse break due to the high voltage ESD transient?
Is it still not subject to the ESD event?

That is correct. A high voltage transient will not blow a fuse, because there is very little energy in it, because it only lasts a few microseconds.
The fuse is required if you need to protect from longer lasting, and generally lower voltage events, such as mis-connections or other stupidity.

 

I have used fuses that cleared in about two milliseconds, but that is quite a bi longer than many ESD events. Those fuses were protecting the five volt sensor supplies in a crash sensor system. Not at all like a microphone system. Fast diodes to clamp spikes are the best choice, they do work.

 

I have used fuses that cleared in about two milliseconds, but that is quite a bi longer than many ESD events. Those fuses were protecting the five volt sensor supplies in a crash sensor system. Not at all like a microphone system. Fast diodes to clamp spikes are the best choice, they do work.

Claimed response time for a TVS diode is <1ps. That should be quick enough.
The overcurrent protection (be it fuse, PTC or PPTC) does not have to clear quickly, because the TVS diode is clamping the spike, but the fuse must clear before the TVS diode melts.

 

Possibly, in addition to that TVS device, having slower power diodes to clamp the voltage between some limits will provide suppression of longer over voltages, until the fuses clear.
The fuse response that I mentioned is what I verified myself, on a calibrated oscilloscope.