1
\$\begingroup\$

It is almost the reference design from the DS. The only differences are those quite large inductances I had to add to remove the noise.

The neg voltage is used to power RR OPamps feeding the ADC the current is never larger than about 20mA

What can cause the problem? At some point (after a day, sometimes month or mostly 5-6 months) the pump dies and stops generating negative voltage.

enter image description here

\$\endgroup\$
10
  • \$\begingroup\$ What is the temperature of the part when operating under full load of 5V-? Do you have other DC to DC's in the design? What is the voltage on Vin? can you prove it's stability? Part death usually comes from exceeding a maximum rating... \$\endgroup\$
    – Voltage Spike
    Jul 13, 2019 at 22:51
  • \$\begingroup\$ actually the full load is about 20-30mA. It is not warmer than the board. WEhen it stops generating it starts to be a bit hotter (about 3-4 degrees) \$\endgroup\$ Jul 13, 2019 at 22:55
  • \$\begingroup\$ I do not have any others DC-DC converters. The voltage is between 4.5 and 5V (USB power). \$\endgroup\$ Jul 13, 2019 at 22:56
  • \$\begingroup\$ Hook an o-scope up to the part and look at the voltages on Vin and Vout while you hot plug the USB, make sure the ground on the scope is on pin2 \$\endgroup\$
    – Voltage Spike
    Jul 13, 2019 at 22:59
  • \$\begingroup\$ I did of course and everything looks 100% fine. Also oscillations on the capacitor look exactly as they should (frequency an length of the bursts) It works 6 months and the suddenly dies. Sometimes (on some boards) it does not want to start at all. \$\endgroup\$ Jul 13, 2019 at 23:04

1 Answer 1

1
\$\begingroup\$

Since you don't know what is killing the LM2776, it's probably a good idea to bullet proof the design at this point.

The absolute maximum ratings are being exceeded in some manner. If it's not from hot plugging, then ESD is a likely contender for failures in this device, most of them are fast, and usually happen when you don't have a scope plugged in.

With devices that are hot plugged, there can be voltage spikes. Put a 5.2V zener and\or TVS on the output and input of the LM2776. The diodes across Vin and GND and the output would ensure that the voltage is kept within spec at all times.

Then do your testing and handle it and see if the LM2776 still dies, if it does, then you need to look elsewhere and eliminate all possibilities of failure.

To make sure the current is within range, put a series limiting resistor on output and input. (if you can on the output, the drop might be too much.)

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.