LM3668: Overcurrent condition

Hi Arun,

it is a new design for customer. this device is not suggested for new design. why did customer select this IC.

Hi Jasper,

This is not a new design. This is for supporting an existing design.

Thanks

Hi Arun,

from my understanding, the device just turn off the internal MOSFET until next cycle, not totally shutdown the IC.

did customer observe the issue in several system board? is the issue found in no load or heavy load condition.

Hi Jasper,

Please see the customer's observation. It appears to be not just one cycle.

Please find below more details about that issue and what we are trying to achieve.

1. This LM3668 issue is identified in battery operated units that are returned from the field with the description “Unable to be turned on”.
2. When these units are tested, LM3668 is found to have proper PVin and VDD, as well as EN. However, the output VOUT stays 0V and therefore the units are not functional.
3. If, at that point, PVIN and VDD are turned off and then on, then Vcc turns 3.3V as expected, and the units become fully functional from then on.
4. However,  once the units become functional, the fault could not be reproduced any longer, which makes it difficult to identify experimentally the conditions inducing this fault.
5. It appears that the only way that can explain this fault is a response of the limiting current circuitry to a surge in current. It follows from the paragraph in the datasheet stating that “The internal current limit comparator will disable the power device at a typical switch peak current limit of 1.85A”. ‘Disabling the power device’ implies something more permanent than just skipping one cycle.
6. The lack of detailed information on the parameters of the current limiting protection circuit in LM3668 is currently hindering the identification attempts of this failure mechanism, and consequently the possibility of finding a suitable solution.

Thanks

Hi Arun,

could you repeat such behavior in the lab? for example, add large current to the output to trigger the current limit.

Hi Jasper,

It is a bit difficult for the customer to replicate the issue as they have other protections that kick in which are hard to disable.

What would be useful is the actual data from TI from your experiments so that the customer can design their circuit accordingly.

1. The units have a fast Reset controller, which resets the unit as soon as Vcc falls a fraction under its nominal value (due to larger current induced by the external test resistor), and when the Reset becomes active, the current from Vcc consumed by the unit drops significantly, thus reducing the total Vcc current. If the LM3668 current limit circuitry is slower than the Reset, it may essentially sense only the Vcc current due to the external test resistor. Therefore, from this point of view, the requirement on the external test resistor is to have a sufficiently low value to activate the current limit circuity on its own, with no significant contribution from the unit’s circuitry.

1. However, if the external resistor on its own draws more than the current limit threshold, the Vcc may drop significantly, and then trigger another protection circuit, namely the “Short Circuit Protection”.  The latter presumably uses a low Vcc voltage threshold to identify a short and trigger a temporary and mild response, namely limiting the Vcc current to half the typical current limit value, until the short is removed. So, the value of the external test resistor may need to have a sufficiently high value to not activate the Short Circuit Protection.

1. Considering the conflicting requirements 1. and 2. above, it seems that in order to be able to prove the assumed failure mechanism, a fine tunning of the external test resistor might be required, which has not been done so far.

1. In addition, there are timing and other practical issues that make the experimental proof of this failure mechanism quite difficult, when the relevant functions of this chip are insufficiently detailed. We would appreciate a more precise description of the chip functions mentioned above, ie Overcurrent limit and Short circuit protection.

Thanks

Hi Arun,

I can try to see if the device shut down at over current condition with EVM.

what is the Vcc in customer's description.

Hi Jasper,

Let me check. Can you explain what load value you used to generate the shutdown?

Also, what was the way out of shutdown for you?

Thanks

Hi Arun,

I can increase the output current until the device trigger current and check if the device shuts down and doesn't restart after the load is removed.

Hi Jasper,

What current value did you trigger the device at? And, did it automatically restart once the load was removed or was any action required?

Thanks

Hi Arun,

unfortunately, there is no EVM for this device as it is not suggested. any possible that customer remove the external device and just evaluate the behavior of LM3668 in the system board?