Burning up analog switch!

Hi, Joseph. Let's go backwards on this. Your problem really sounds like one of the input voltages (A or B, depending on how you've got it wired, but not C -- that's the control signal) is outside the power supply boundaries (data sheet shows GND - 0.5V as the lower end, and Vcc + 0.5V as the upper end abs. max.) If you exceed either, the IC latches up and you're done.

This can be easy to do with thermocouples, especially if they're the grounded bead type. Try this: put your DVM on DCV 200V range, and with the T/C disconnected from the circuit, measure the DC voltage between your thermocouple - and circuit GND. Then measure the AC voltage. You might be surprised.

By the way, have you considered a very weak pullup on the + end of the thermocouple? In addition to placing your T/C at a known voltage, it might give you a free open thermocouple notification.

Good luck Chris

Chris, The problem happens prior to me attaching the thermocouple, that is why A is open (floating). B is connected to GND. You are saying A and B are definitely the "inputs", not C (although the spec sheets does call C the "control input").? This device only has the 5 pins so there are no other inputs.

The spec sheets points to one of their guides regarding the implications of leaving an input floating:

In it they just talk about keeping the inputs at either a logic low or logic high...nothing to do with analog "signals". That is why I'm asking if A and B are really considered an "input" according to TI. They are analog, not logic level. I'd like to know if despite that, I still cannot leave them floating.

Confused, joseph

I do see what you're saying, Joseph. I can see that you have C open during the uC reset/initialization period, and from the app note you provided, I can see where that might be a problem. Possibly you might want to use weak pullup/pulldown resistors on the board to help.

But it doesn't seem to make sense that the A or B sides of the analog switch couldn't be open. These are frequently used in series-parallel combinations, and one side is effectively isolated from all voltage in these configurations.

But apart from what you've said, this definitely sounds like a latchup (some voltage exceeds Vcc or is less than GND at some time (possibly during powerup) or power supply problem here. I worked with analog switches interfacing with outside world inputs before, and my problems with other logic families' analog switches (and I had quite a few) were all along those lines. Apart from the pullup/downs on the C controls, I'd look closely at what's happening with the voltage at the inputs that *aren't" isolated while voltage is coming up. I'd also look closely at the possibility of initial overshoot on the power supply on power-up. And if you've got more than one power supply here, the other one may be coming up before Vcc and feeding current into your chip. Something may be happening during those few ms. of powerup that you aren't seeing here.

If all else fails, a quick email to the apps people at TI might also provide some help.

Good luck Chris