LM3405A: Maximum Duty Cycle

Hi Infant, 

The maximum duty cycle is indeed affected by temperature. But the major impact is from device to device process corner variation. So, I recommend to keep taking 85% as limit even at 40C. 

Vfb=0 is just a test condition to easily trigger max-duty operation by automictic test program. Max duty is the same in normal operation because their basic trigger circuit is the same. 

Could you please share more about the application you have, such as VIN, IOUT, LED number and so on. I could recommend latest devices for you which have better performance and lower cost. 

Thank you! 

Best, 

John. 

Hi John,

Thank you for your response. 

In our application the LEDs are switched through a MOSFET. During normal operation we don't have an issue. The maximum LED voltage is around 7.2V and VIN = 15V. In order to detect stuck ON/OFF failures of the MOSFET, there is a BIT.

BIT Details: VIN = 15V, IOUT = 60mA, VOUT = 12V Zener. There is a 12V Zener at the LM3405A output (in parallel to LED+FET). During BIT, current driver is enabled when the MOSFET is in open state (12V Zener biased). Output power is higher and hence input current is also higher than normal mode. This difference in input current is used to detect the MOSFET stuck failures. In our testing we found the 12V Zener is not biased when the input voltage is around 14V or less. Hence the BIT also fails at 14V or less. Accounting for the Zener diode voltage variations, the LM3405A maximum duty cycle measured is around 92%.

If we consider the worst-case duty cycle of 85%, then theoretically the BIT could fail even at 15V. This is why I am interested in the worst-case duty cycle only accounting for device-to-device variations and at 40degC.

We could measure the worst-case duty cycle during production. But I am not sure how this parameter will vary over time. Will the worst-case duty cycle become worse over time?

Thanks, Infant Jesuraj

Hi Infant, 

TI's products go through reliability simulation and test before release to market. So, I do not need to worried too much about the spec shift over time. 

Actually, LM3405A is a quite old device. I could share you some latest device with better integration and performance and also lower cost. 

For your case, I recommend TPS92200, which is a sync-buck solution integrating both HSD and LSD power FET. And for the max duty you caring about, it can reach 99%. With TPS92200, I believe your design would be much more easier. 

 Here is the datasheet. 

https://www.ti.com/lit/ds/symlink/tps92200.pdf?ts=1723793714997&ref_url=https%253A%252F%252Fwww.ti.com%252Fsitesearch%252Fen-us%252Fdocs%252Funiversalsearch.tsp%253FlangPref%253Den-US%2526nr%253D145%2526searchTerm%253DTPS92200

Could you please also share the end-equipment application please? 

Hope this could help. Thank you! 

Best, 

John. 

Hi John,

Let me re-phrase my question.

If during production test the max duty cycle is found to be 92%, Can it drift to 85% (still within the datasheet spec) over time? When I say production, I mean our PCB's production.

Regards, Infant Jesuraj

Hi Infant, 

Got your point. 

That will not shift to 85% because the major contributor of variation is process, which is set for each single chip. 

Thanks! 

Best, 

John.