I've got a design utilizing a LT6100 to measure current being provided to a connected load. The LT6100 is configured for a total gain of 25. Under normal circumstances (when there is actively current moving across the sense resistor) there is no problem. The readings come back accurate to what I am measuring. Where things get weird is when I turn off the current to the load, Vout from the LT6100 shoots up to Vcc (3.3V).
Before I go farther, here's a quick schematic of how the LT6100 is connected in relation to the load being driven (I can't post the whole page unfortunately, but this should show everything relevant):
The issue is when I turn the FETs off (ENABLE and CTRL PWM both low), the Vout of the LT6100 goes low, but after about 15 seconds, will quickly begin ramping and then shoot up to full VCC (3.3V) and lock there until the ENABLE and CTRL PWM signals go active and the Vout signal drops back down to an accurate representation of the current.
I have a scope capture of the behavior. You can see Vout (Yellow) decrease as I fade the PWM back, and when the controller shuts off ENABLE (Green). Vout then sits for around 15 seconds then shoots up:
I'm confused as to why this would be happening. There's no current actually moving through the sense resistor, and even if there was, it would have to be over 2A to get all the way up to Vcc. Even stranger is the 15 second delay before the problem occurs. The only thing Vout is connected to is the high impedance analog input of a PIC microcontroller. I am wondering if it's something to do with the internals of the LT6100 IC? Something I'm missing?
Any thoughts greatly appreciated!