LM3478MM boost has 6khz sawtooth output

whitecap9

Hello,

I am using a LM3478MM in boost configuration. Input range is 4-14 volts output is 12 volts. The purpose of the regulator is to maintain a minimum voltage of 12 VDC from a 12 volt lead acid battery. The problem is when the input voltage is above ~8 vdc. The output switching frequency drops from 600 khz to about 6 khz. The output voltage remains at 12 volts when this happens. The regulator seems to otherwise function however there is an annoying audible whine and the output waveform is a sawtooth with about 1 volt ripple.

Is this problem caused by the OVP circuit or could it be related to compensation? or something else when the input voltage is decreased to <8 vdc the regulator behaves as expected.

I would like to avoid going to a SEPIC strategy to reduce cost.

over 12 years ago

0 Juan Paulo Fung over 12 years ago

TI__Expert 3870 points

Hi,

Please provide us with a copy of your schematic and load current conditions when you see this behavior (unloaded or full-load)?. If you can post a couple of the scope waveforms (SW, Vout ripple). Thank you!

Best Regards,

-Juan

0 whitecap9 over 12 years ago in reply to Juan Paulo Fung

Prodigy 50 points

0 whitecap9 over 12 years ago in reply to whitecap9

Prodigy 50 points

Schematic: 4118.VRM.pdf

Top wave is under 300 mA load. Bottom wave is under no load. Both loads have an input voltage of 12 volts. When the input voltage is decreased to 9v the output is flat and clean.

0 whitecap9 over 12 years ago in reply to whitecap9

Prodigy 50 points

Ok,

I see you moved my post, how about actually responding to it now? It has been a week with no response. This is unacceptable.

0 Tommy Jewell - WEBENCH Design Center over 12 years ago in reply to whitecap9

TI__Intellectual 2930 points

Hi,

I apologize for the delay.
Yes, this looks like OVP. This part will always have a minimum load at which it will go into OVP because when switching it has a minimum on-time. The higher the input voltage the greater the current increase across the inductor during this minimum on-time and thus the increase in minimum load before OVP. What can be changed is this threshold though; this can be done by bleeding off some load on the output (i.e. smaller feedback resistors or a dummy load) and/or an increase in inductance (will decrease the energy stored in the inductor during the min on-time).

Best regards,
Tommy

0 whitecap9 over 12 years ago in reply to Tommy Jewell - WEBENCH Design Center

Prodigy 50 points

Ok, this only explains the behavior during no load. This does not explain the behavior during the 300 mA load (top wave form). Could you please explain the behavior while under a 300 mA load?

0 Tommy Jewell - WEBENCH Design Center over 12 years ago in reply to whitecap9

TI__Intellectual 2930 points

Do you have a snapshot of the switch node waveform or the FET drive pin in the 300mA case?

Power management

Power management forum

LM3478MM boost has 6khz sawtooth output