Spice model for TPN3021

[UKilc.1]

Hi,

I need the spice model for thyristor surge protector device TPN3021RL to simulate a circuit. However I could not find it, could you provide the model?

[Peter BENSCH · Accepted solution]

Unfortunately, there is no Spice model available for TPN3021.

Regards
/Peter

[UKilc.1]

Hi Peter, I see. I observe the crowbar operation at currents (~50mAs) much smaller than breakover current given in the datasheet (0.3A). Can that parameter have a wide range? What I do is applying a relatively high AC voltage (150V amplítude, 1kHz frequency) with series resistance (around 2k).

[Peter BENSCH]

Ideally, the limiting voltage for TVS should be exactly at one voltage level. However, with Zener in general and TVS in particular, a more or less steep slope of the characteristic curve can be observed. The beginning kink in the characteristic curve, at which a relatively low current still flows, is referred to as the breakdown voltage VBR; the breakover voltage VBO is then the value from which significant current is diverted.

According to the data sheet, the breakover voltage VBO is a maximum of 38V when a current of 300mA is flowing. At 150V and 2kohms, only a peak current of (150V-38V)/2k ~ 56mA can flow, so the resulting VBO should be significantly lower than 38V. Simplified, this could be extrapolated linearly if the dynamic resistance or at least VBR were specified in the data sheet - which unfortunately is not the case. We can therefore only estimate that VBR is somewhere between 28...38V and therefore the dynamic resistance is between 0...33Ohms.

For a single TVS with similar parameters, a dynamic resistance of approx. 1..2ohms can be obtained (depending on the characteristic range), so that VBO at 50mA should be approx. (1...2ohms)*(300mA-50mA) = approx. 250...500mV below (VBO for 300mA).

Hope this rough estimate helps?

Regards
/Peter

[UKilc.1]

Thank you for the explanation, Peter. I'm a little confused, though. I would assume that with this level of series resistance, the device should not enter the breakover region but instead remain in the clamping region—where it limits the voltage across itself—since the current is nowhere near the breakover threshold. In that case, the power dissipation would be high, around 2W, which could eventually cause permanent damage to the device unless proper thermal management is in place.

[Peter BENSCH]

Your signal has 1kHz, i.e. 500µs per half-wave. This is not a problem for one half-wave, but it is for several consecutive ones or even continuous operation. The small SO-8 package, which does not even have a cooling surface at the bottom, is not suitable for the power loss that then occurs. According to the data sheet, the thermal resistance is 170K/W, so that at 2W there is a temperature difference of 340°C, by which the temperature would increase. A bit much, even for better thermal management.

[UKilc.1]

I'm wondering whether that high power dissipation causes the device to go into breakover region. Because normally, it should not crowbar with that of a low current. Am I right or missing something? 

[Peter BENSCH]

It is not the power dissipation that leads to the breakdown, but essentially zener effect and avalanche effect. As mentioned, the breakdown starts at the breakdown voltage VBR (not specified in the data sheet for this device) and is specified with a maximum current of 300mA at VBO of about 38V.

As you can also see from the data sheet, the main application of the TPN3021 is the overvoltage protection of network interfaces, where spikes can occasionally be coupled in. However, you obviously want to use it for a completely different application in continuous operation, for which it is not designed.

Perhaps this would be an opportunity to describe the target function of your device as far as this limitation is concerned, so that suggestions can be made if necessary?