This circuit shows the TLC7705 with a resistor divider network that should provide less than 4.55V at the SENSE input when V12 drops to about 8.75V (including the internal 910k and 290k resistor divider inside the 7705). When it's above 4.55V, the /RESET line should be able to drive Q9 which drives Q8 that gates the output voltage on or off. If I take off R22 and use a wire to test the output circuit connected directly to 5V or Ground, it works fine and gates the output correctly, but as soon at it touches the /RESET line of the TLC7705 it drops to only 1mV or so and does not drive Q9 to gate the output on (it ends up having about 2.5V on the output, so it's not even driven off to ground here either). This happens only when there is a load on the output (the load is about 0.5A, 1A max), and it doesn't happen (seems to work correctly) when there is zero load on the output. I tested the available current from the Voltage Regulator when the output circuit is working (either manually by touching the Q9 input to 5V or by having the SENSE pin voltage much higher than it should need to be as detailed below), and the output current is more than enough to drive at least two times the load. The TLC7705 seemed to be the issue here - perhaps with how much current was coming into the SENSE pin? or using a TLC7705 with a resistor divider network since all examples showed the sense voltage connected directly to the SENSE pin. Since the examples only showed the TLC7701 using a resistor/divider network, we changed out the TLC7705 for the TLC7701 and we changed the resistor dividers appropriately, but we still saw the same behavior (like the TLC7701 /RESET just wasn't working or able to drive Q9 properly even though there is minimal current required there and the drive voltage required for Q9 is low). I then read another post here: https://e2e.ti.com/support/power-ma...760/tlc7701---sense-threshold-voltage-setting that mentioned the resistor/divider network current should 100x the leakage current. So I reduced the resistor/divider network to use 2.5k and 18k for R20 and R10 respectively. This still resulted in the same behavior where /reset was not able to go to 5V when the voltage on the SENSE pin is at 1.47V (but it worked okay without the load). However, when I changed out the resistors on the 7701 to have 3.6V on the SENSE pin, the TLC7701 was able to have the /reset pin go high and it would drive the output circuit correctly. This is great. However, one of the main reasons for having this voltage monitoring circuit in there is to stop the output voltage any time the V12 drops below 9V or so (or 8.75V); this doesn't happen if I change the resistor divider to be have way more than 1.1V at the SENSE input. Also, there is still the issue where it doesn't fully drive the output off either (and we see about 2.5V there) if the SENSE voltage is only a small amount over 1.1V. Ouch. I assume that I must be doing something wrong or just don't understand why the TLC77x chips don't seem to be working correctly. This is a high-priority issue, so any help or thoughts are much appreciated. Thank you.
