LM3481: Startup issues

Mark Bussis

Part Number: LM3481

I am using the LM3481 in a circuit that has input voltage of 11-24VDC or 24-32VAC. Output is 12V@750mA.

When I run the circuit from a DC power supply, it functions as expected (power supply is current limited at 2A). When I run the circuit with a 28VAC input through full wave bridge rectifier, the VIN of the LM3481 immediately gets damaged. I do not believe this damage is from an over voltage condition, but rather from a high current through the IC. After the IC fails, VIN is shorted to GND. I have attemped to place a 100 ohm resistor in series with VIN with cap to ground to slow input voltage rise time, but the resistor immediately burns up.

I have placed a cap on the UVLO pin to try to slow down the function of the regulator at startup, but this did not help.

My assumption at this point is that the FET drive from the regulator is latched on such that my input voltage from VIN gets latched to the ground pin and fries the part.

I have had it function correctly a time or two on the AC input (which is a clean 28V sine wave), but most of the time it damages the IC on startup.

Any insights appreciated.WBDesign96_Steady State-2 45v max input.pdf

over 3 years ago

0 Yinsong Cai1 over 3 years ago

TI__Expert 7450 points

Hi Mark,

Thanks for your interest in LM3481.

If the input is grounded from the FET the inductor or FET should damaged. Have you ever checked the input after rectification?

Thanks,

Yinsong

0 Mark Bussis over 3 years ago in reply to Yinsong Cai1

Prodigy 50 points

The FET and inductor are fine - no issues there. The only damage is to VIN (pin 10) on the regulator.

0 Mark Bussis over 3 years ago in reply to Mark Bussis

Prodigy 50 points

I find that if I slow down the VIN to about 25mS the part survives just fine and everything works with 24VAC input. Is there a rise time requirement on VIN? Without slowing it down VIN fails.

0 Yinsong Cai1 over 3 years ago in reply to Mark Bussis

TI__Expert 7450 points

Hi Mark,

Thanks for the information.

Sorry but I'm a little confused here. By "slow down" the VIN, do you mean the AC input or do you give the DC input directly?

Thanks,

Yinsong

0 Mark Bussis over 3 years ago in reply to Yinsong Cai1

Prodigy 50 points

Hi Yinsong

I put a FET in series with the VIN (after the bridge rectifier) with an RC slowing the gate drive, such that the VIN voltage is slowed in ramp up through the FET. VIN at the IC is therefore not such an immediate ramp up but slowed down a bit. With this scenario the LM3481 does not destruct as it did with a very fast turn on time.

0 Yinsong Cai1 over 3 years ago in reply to Mark Bussis

TI__Expert 7450 points

Hi Mark,

I see. But again, it would be good to have the actual input voltage waveform (fast rise) at the VIN pin. A guess is that the input has a big overshoot somehow when the slew rate is high.

Thanks,
Yinsong

Power management

Power management forum

LM3481: Startup issues