Hi Brent - You are right, it has been impossible to replicate on the bench, even with = the actual loads. We do have an MOV earlier in the circuit, but this whole = triac drive is disconnected by a relay (2-pole) to isolate it. The MOV is o= n the other side of the relay. I have tried repeated relay openings (1000's= under different load conditions) but still no failure :o( I have been trying to find a mechanism for this type of fault in triacs. Ga= te current pulse, voltage spike, repetitive something... ? While I can find= reference to the failure mode, and of course all of the precautions about = dv/dt etc., I have yet to find a description of what causes this type of fa= ilure. Regards Stephen -----Original Message----- From: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu] On Behalf Of= Brent Brown Sent: Friday, 27 February 2015 2:27 PM To: Microcontroller discussion list - Public. Subject: Re: [EE] Triac "half wave" failure mode On 27 Feb 2015 at 1:43, stephen.forrest@agilent.com wrote: > Hi Brent - The triacs are permanently damaged. One half cycle is=20 > permanently on. If the output is observed with a scope (and yes, I=20 > have HV differential probes) it looks like a diode. The other half=20 > cycle works fine. The gate is driven by an opto-triac so I believe the=20 > gate phasing is 1 & 3. Ok, found your original post & circuit. To be honest I have a very limited amount of experience with TRIACs, but th= is may be of some value... I once built a circuit that used a TRIAC opto-coupler to directly power a 2= 30VAC solenoid (load of 6VA =3D only 26mA =3D within the specs of the opto-= coupler, and so perhaps risky but no external TRIAC was used). Just on/off,= no phase control.=20 There were some failures in the field, embarrasing. The part that failed wa= s the TRIAC opto-coupler, sometimes the symptom was half waving as you desc= ribe. The problem was thought to be line voltage transients exceeding the voltage= rating of the opto-coupler. The fix was to add a VDR across the optocouple= r (there was already a series fusible resistor, which would act to limit cu= rrent during an over voltage excursion), and we upgraded the optocoupler to= a 600V part (was 400V).=20 This seemed to cure it acceptably. Back to your circuit. No over-voltage protection components that we can see= , although they could be elsewhere. I'm suspicious of feeding the gate driv= e from the C in the RC snubber - question in my mind is what determines the= gate drive current, if it might be possible for example that a high voltag= e/high frequency transient drastically lowers the reactance of C210 and mas= sively increases gate current. Even so you'd think the absolute max 4A peak= rating of the gate of the TRIAC and the 2 x 180R resistors would let it su= rvive this. I'm guessing the fault is not easily reproduced on the bench. When the TRIA= C dies does the optocoupler also fail? I hope it's not something absurd lik= e the series resistors breaking down but not leaving any visible evidence? -- http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/chang= e your membership options at http://mailman.mit.edu/mailman/listinfo/piclis= t --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .