LM5008: minimum output voltage ripple

user3193866

Part Number: LM5008

Datasheet: "A minimum of 25 mV to 50 mV of ripple voltage at the feedback pin (FB) is required for the LM5008."

What happened is that the design is modified. The 1 Ohm resistor in series with the output capacitor (tantalum) is replaced by a 0 ohm resistor. The ESR is now (to?) low, causing the ripple on the FB pin to be about 10mV instead the required 25mV min. But... The converter seems to work fine, there is no instability and also with a step response from 50mA to 300mA there is no stability issue.

I have two questions:

1. what is the real minimum ripple voltage at the FB pin?

2. what happens when the ripple is to low with a converter using the hysteretic control scheme? Now, it is not unstable, the max output current current is possible and the converter is not heating up more than expected. Also the switch node looks fine during constant load and step load, there are no extra pulses as mentioned in AN-1481 Figure 3.

Why is the converter working?

11 months ago

0 Colin Chen 11 months ago

TI__Genius 12011 points

Hi Sir,

What's your working condition? Could you attach your schematic?

B R

Colin

0 user3193866 11 months ago in reply to Colin Chen

Prodigy 170 points

I know how to add ripple just need ans answer on the questions.

In the previous version R209 was 1 Ohm, so the ripple was big enough, but in this version R209 is replaced for a 0 Ohm, and this lowers the ripple to just 10mV. I 'm surprised that is still seems to work fine, I expected an unstable output...

0 user3193866 11 months ago in reply to user3193866

Prodigy 170 points

I replaced the tantalum for ceramic (2x100u/6V3), so the ripple is now very low. But still the output voltage is stable and a step response is also no problem.

0 user3193866 11 months ago in reply to user3193866

Prodigy 170 points

Correction. The output is stable, but now the converter is not operating normally. Whats happens now is mentioned in AN-1481 Figure 3. So 10mV ripple seems fine but below that there is a point that it doesn't. But what is the minimum? I'm ignoring the noise immunity at this point.

0 Colin Chen 11 months ago in reply to user3193866

TI__Genius 12011 points

Hi Sir,

Did you test the ripple fed into FB pin? what it looks like?

A lower ripple doesn't mean converter cannot be stable but it is more sensitive to noise and easy be unstable.

B R

Colin

0 user3193866 11 months ago in reply to Colin Chen

Prodigy 170 points

"A lower ripple doesn't mean converter cannot be stable but it is more sensitive to noise and easy be unstable."

Yes I know, but there is a point that the ripple in not enough for stable operation, what is the min ripple required for stable operation? According the datasheet this is 25mV min, but I have 10mV now and that seems fine after testing in the lab.

Like I write before. If I connect ceramic capacitors, then the ripple is really low and the converter is unstable.

COT requires a small ramp injected into its feedback node to work. This ramp signal has the same shape and phase as the inductor current ripple.

0 Colin Chen 11 months ago in reply to user3193866

TI__Genius 12011 points

Hi Sir,

10mV cannot cover all condition you may meet and if you have poor layout or more noise, 10mV cannot be stable as well.

25mV in datasheet is the value considering small Vout bias and higher noise insensitive, so it is a suggested value.

B R

Colin

0 user3193866 11 months ago in reply to Colin Chen

Prodigy 170 points

So if the layout is perfect and there is no noise, the ripple on the FB can be very small. I'm not saying this a good idea :)

If talking about unstable, is it slower on step response, or is the output voltage unstable? What happens?

0 Colin Chen 11 months ago in reply to user3193866

TI__Genius 12011 points

Yes, you better keep ripple higher than 25mV due to noise resisting as datasheet suggested.

Now talking about unstable, both unstable SW and Vout will happen.

Colin

+1 user3193866 11 months ago in reply to Colin Chen

Prodigy 170 points

Thanks.

0 Colin Chen 11 months ago in reply to user3193866

TI__Genius 12011 points

You are welcomed!

Power management

Power management forum

LM5008: minimum output voltage ripple