# Crowbar circuit with TL431 help

So I'm trying to create a OVP using TL431 an a triac, i have checked the datasheet but since i suck at math i don't really get how to set the ref voltage..

My example circuit:

I'm having a 19v power supply, and i want to make it so that if the supply for some reason should reach over 20v, it blows the fuse.. Any tips? Thanks!

• Why a triac and not an scr?
– jonk
Oct 13, 2016 at 21:34
• what do you mean with scr? well i just googled it, i thought both where the same thing haha.. Still new at this. Have to look in the junkbox and see if i have one.
– Xane
Oct 13, 2016 at 21:34
• silicon controlled rectifier.
– jonk
Oct 13, 2016 at 21:35
• Also, I've no idea why you wired up the TL431 that way, with 19 V across it. Might be interesting to try that ... once.
– jonk
Oct 13, 2016 at 21:39
• I was not going to use that, I'm just trying to simulate different things to see if i got to understand how it works/how to set the ref voltage.
– Xane
Oct 13, 2016 at 21:41

That might work, but try flipping the triac so MT2 is grounded and MT1 connects to the input voltage, and connect the 'cathode' of the TL431 ONLY to the gate.

Calculate the gate resistors as @Tony suggests, so for about 19.5V you could use 2.49K and 16.9K, standard E48 values.

This circuit won't work with an SCR because SCRs don't generally come in the opposite polarity.

Edit: Like the below

simulate this circuit – Schematic created using CircuitLab

When the voltage at the REF input of the TL431 exceeds ~2.5V the gate current increases until the triac triggers and shorts out the supply, blowing the fuse. The triac is operated in quadrant III (gate and MT2 both negative wrt MT1), which is always acceptable. Below the trigger voltage the TL431 conducts < 1mA which should not trigger most triacs. If it is a problem (or just to be sure), you can connect a 100 ohm resistor from MT1 to gate.

• Thank you, could you explain a little more on how you meant? i tried to do it but it just wouldn't work so i prolly did something wrong.
– Xane
Oct 14, 2016 at 21:21
• Thanks, i just tried it but the fuse just won't blow.. Not sure if it is the simulation that is wrong (using multisim) i.gyazo.com/c44b50921f5268a162a07672544b9da8.png
– Xane
Oct 14, 2016 at 22:02
• Measure the gate current in your simulation. It should be more than 100mA, so the triac should be triggering. Measure the fuse current- maybe it is not modeled and the triac is actually on in the simulation. Oct 14, 2016 at 22:04
• The current won't go through the triac only the tl431.. i.gyazo.com/4d383d4016ffff9ce7b8bf8ab67aa368.png
– Xane
Oct 14, 2016 at 23:10
• Sounds like your triac model is missing or broken. Oct 14, 2016 at 23:13

• Two transistors back-back= 1 SCR
• two SCR's back-back= 1 Triac

• Since hFE & power rating of transistors have a wide range Vgt threshold is affected by Vbe bulk resistance and Vce(sat) so the gate threshold voltage is not very accurate for low voltage OVP.

Vgt's can range from 650mV to 5V in datasheets but you can roll your own if you want. The device must not FUSE before the "fuse" according to SOA and fuse times.

• There are better ways than using Vgt , instead, use a precision comparator to drive it.

• FYI on TL431 is like using Non-Inverting Op Amp gain on 2.5V reference.

• Thank you for the detailed explination, Going to try the comparator method a bit later!
– Xane
Oct 14, 2016 at 21:22
• yer welcome, ensure comparator has a bit of LPF so that tiny glitches do not false trigger OVP, ie. make a spec for OVP duration and threshold. Oct 14, 2016 at 21:35