LM2734

Jon Dutra

Other Parts Discussed in Thread: LM2734

I have a customer that is failing Class B FCC , Radiated EMI by 5 dB, and it appears to be coming from the LM2734. Vin = 12V, Vout =2.5V, Iout 800mA. Boost voltage is 5V from another source. The layout is good, Inductor is 4.7 uH, and claims to be shielded. Running at 1.5 MHz. Waveforms look right. Similar LM2734 circuits on same board with less load don't radiate as much. What do you suggest ? Thank You !!

over 8 years ago

0 John Bowie over 8 years ago

TI__Intellectual 890 points

Hi Jon,

Can you send me your schematic, layout and test conditions please, have you narrowed the high EMI down to this particular device using sniffer probe or some other method, if you vary the load for this device do you pass EMI.

regards

John

0 Jon Dutra over 8 years ago in reply to John Bowie

Prodigy 60 points

0 Jon Dutra over 8 years ago in reply to Jon Dutra

Prodigy 60 points

OK I attached Schematic/ Top layer ... You can contact me directly at JODUTRA(AT)microsoft(DOT)com. Yes narrowed down with near field probe. Cannot vary this load. We are adding resistor in series with Boost pin as EMI tuner (20 ohms costs about 1% efficiency I believe) . I worry about the 1N4148 diode with 3.3V boost with 800 mA load, I think I will look for a drop in schottky ... Thank You !!

0 JohnTucker over 8 years ago in reply to Jon Dutra

TI__Guru**** 191445 points

I don't directly support LM2734, but i can give you a little basic advice. It would help if you can identify the frequency or frequencies that are failing your EMI screen. Then look at the switching node to see if there are any ringing at that frequency. Once you locate the source, it becomes much easier to tune it out.

0 Jon Dutra over 8 years ago in reply to JohnTucker

Prodigy 60 points

I should have said it is failing radiated emissions by about 5 dB with a broad peak at about 154 MHz

0 JohnTucker over 8 years ago in reply to Jon Dutra

TI__Guru**** 191445 points

See any ringing at that frequency on the switching node? That is the usual source for these type problems.

0 Jon Dutra over 8 years ago in reply to JohnTucker

Prodigy 60 points

0 Jon Dutra over 8 years ago in reply to Jon Dutra

Prodigy 60 points

Above is what I have told the PC designer to do on the next rev. ... OK Lets assume I do see a ring with a 6.49 nSec period on the switch node (1/154 MHz) ... what should I do about it ? What fixes do You recommend ? I am not a big fan of RC snubbers on the switch node but if that is what you recommend what values would you use, and would you put that RC closer to the IC than the catch diode or on the other side of the diode. What do you think about the slice in the ground plane routing all the high frequency energy back past the input cap as shown above? Thanks !!

Power management

Power management forum

LM2734