When protection fails

“Don’t Stop Til you Burn Enough!”

IMG_0436.JPG

Merry Christmas to you all! Christmas Eve ended up with some smoke and the party was over a bit early. We played music really loud last night and got carried away with some dancing around whilst playing “Don’t stop til you get enough” from Michael Jackson. And literally, the 814 SE didn’t get enough!

Let me explain briefly what happened. The 814 output stage has a crowbar protection circuit. It is configured to trigger around 200mA. The crowbar works brilliantly well, however, what is tricky on my design is the fact that you do get proper grid current in A2 operation. This grid current adds to the cathode current and flows through the crowbar sensing resistor. Well, volume was so loud so am sure that when the drum or bass kicked in, the crowbar was triggered. Interestingly enough, the shunt resistor is a 330Ω / 50W piece which should (in theory) blow the 500mA fast fuse. Well, it didn’t.  The current peak wasn’t big enough before the output voltage of the 600V supply dropped significantly. Bear in mind that 330Ω should take serious current out of a 600V supply in theory!

The result was evident in a couple of seconds. The shunt resistor went madly hot, burned the plastic stand-off isolator quickly and fell over one of the current meters and burned the plastic cover badly as you can see on the image above. Luckily only one channel crowbar got activated.

When I rushed into the 600V supply mains switch I saw the secondary pair of dampening resistors (100Ω 7W wire-wound) melting and red hot. They were acting as a fuse and obviously preventing the supply to deliver further current.

Of course, the 814 was intact as it was the OT. At least the crowbar did it job, but far too early 🙂

Fixed it this morning after doing the Christmas present stuff. I couldn’t end up without music in Christmas!

Lesson learned here. Crowbar needs proper testing! I will buy smaller fuses – probably a pair of 200mA slow burn will work fine next time.

Happy Christmas!