TPS2660: TPS26602 window comparator

John Duquette

Part Number: TPS2660

My TPS26602 seems to be operating as a window comparator. When I have only a 1 uF cap on the OUT pins it indicates a fault. If I put a resistor from OUT to GND and pull 1 mA the fault clears. When I pull > 800 mA an overcurrent fault properly occurs (Rlim = 15K). This 'undercurrent' fault is unexpected.

My input voltage is 12V.

UVLO has a 178K (R1) and 22K (42) divider (~10V).

OVP is tied to RTN (R3 = 0 ohm).

MODE is open circuit.

DVDT is open circuit.

Is there a minimum DC load (current path) required on the OUT pins?

over 8 years ago

0 John Duquette over 8 years ago

Intellectual 710 points

Upon further inspection I do have RTN tied to GND (section 10.4) disabling the Reverse Polarity protection. Can that cause this under-current issue?

0 John Duquette over 8 years ago in reply to John Duquette

Intellectual 710 points

my RFLT is connected to an external 3.3V supply, not to OUT.

How does Vgs get established across the OUT FET when OUT (source) is an open circuit? Is there an internal bias path?

0 John Duquette over 8 years ago in reply to John Duquette

Intellectual 710 points

With a scope I can see FLT is toggling at random rates. It goes inactive for a minimum time of 11us and goes active for a minimum time of 570 us.

0 Karikalan Selvaraj over 8 years ago in reply to John Duquette

TI__Genius 16305 points

Hi John,

Could you please monitor if VIN is changing or having glitch for some reason and share the waveform of VIN, and VOUT and FLTb.

Regards,
Kari.

0 John Duquette over 8 years ago in reply to Karikalan Selvaraj

Intellectual 710 points

CH1=~FLT, CH2=IN

IN is 11.87V (measured with a multimeter) from a switching supply with about 100 mVpp of ripple.

CH1=~FLT, CH2(AC)=IN

This capture shows VIN at a FLT event. Ripple is 250 KHz from switcher, approximately 148 mV / 2us = 74 KV/sec

CH1=~FLT, CH2(AC)=OUT

No correlation between ~FLT and OUT. OUT is at 11.83V

0 John Duquette over 8 years ago in reply to John Duquette

Intellectual 710 points

Hmm, the previous post appears to have lost my images. See the attached file.

AUX_current_limit_issue.docx

0 John Duquette over 7 years ago in reply to John Duquette

Intellectual 710 points

Any thoughts on this?

0 Karikalan Selvaraj over 7 years ago in reply to John Duquette

TI__Genius 16305 points

Hi John,

TPS2660 is type of hysteritic controller (ORing control) and it can oscillate under light load conditons.

1. Reduce the VOUT capacitor OR

2. Increase the VIN capacitor OR

3. Increase the light load current

Please refer to the applicaiton notes for more details.

Regards,
Kari.

0 Karikalan Selvaraj over 7 years ago in reply to Karikalan Selvaraj

TI__Genius 16305 points

Hi John,

FLTb is active because of Reverse trip comparator of TPS2660 is falsely tripping as it sees more than -10mV reverse voltage (IN-OUT).
This is mainly because of IN having ripple and OUT is relatively not changing (due to OUT cap).

Note that FLTB has a delay of 45uS and 70uS for reverse compartor and forward turnon.
This is why there is no direct corrlation between FLTB and IN or OUT.

Please follow the above recomendations to fix the issue.
Let us know if you still have problems.

Regards,
Kari.

0 John Duquette over 7 years ago in reply to Karikalan Selvaraj

Intellectual 710 points

Thanks. This makes sense. Unfortunately my Cin is already 22 uF on the output of a PI filter on a switching power supply output. The plot above shows the switcher ripple that is enough to cause 1uF on the output to back feed.

I currently have this functioning with a hand modification of an 8.2K resistor shunting my 1.0 uF Cout. This band-aid will require a layout change that I need to avoid.

This no load condition is only when an accessory is not plugged into the system. Thus I don't care about output stability when nothing is plugged in, I just need to prevent false positive events from the ~FLT pin. When the accessory is plugged in I can easily meet a minimum current draw requirement.

The datasheet Section 7.3 recommends a minimum capacitance of 0.1 uF for Cin and Cout. Is that the minimum value that I can use for Cout? What if I just put a resistive load (8.2K for about 1.4 mA) in place of Cout? Additional capacitance will be present when the accessory is connected, but it will be after a common mode choke and then at the far end of a 10 meter (typical) cable.

TPS26602 -> choke -> 10 m cable -> additional capacitance

0 Karikalan Selvaraj over 7 years ago in reply to John Duquette

TI__Genius 16305 points

Hi John,

We recommend 0.1uF of minimum output capacitance, but in this case replacing with light load resistor is fine.

Regards,
Kari.

Power management

Power management forum

TPS2660: TPS26602 window comparator