LM5141-Q1: Test questions

Hi Wenyi,

See answers included below.

Regards,

Tim

Wenyi Zhao1 said:

Hi Tim,

Glad to talk with you again.

 

1.You say the controller enters a burst mode at 2.5Hz, so what's the difference between standby mode and burst mode? Or they are the same mode by two names because the explanation of standby mode is not very clear in the 5141-Q1 datasheet. And, the frequency when testing in a burst mode is not stable, is it a right condition or it has made some mistakes?

[TH] Yes, similar. Burst mode occurs at light load when diode emulation mode is enabled (effectively a frequency reduction to increase efficiency). The controller will operate in standby mode during the interval of pulse skipping with active mode during actual switching.

2.Sorry, i mean in light or very low load conditions, Pin decreases with the increase of Vin; while the load is heavy, Pin increases  with the increase of Vin. e.g,when load is 3A,  Pin increases  with the increase of Vin. In this condition, does the inductor core losses have an impact?

[TH] As Vin increases, the inductor ripple current amplitude increases -- this means higher core losses. Switching losses of the MOSFETs also increases with Vin.

3.I find the dutymax of 5141 is 78%, but the Vin is 5.5V while the output is 4.79V, in which condition the duty is higher than 78%(operating in CCM), am i make a mistake in measurement?

[TH] Max duty cycle is set by the chosen Fsw and the toff-min of 100ns, for example at 1MHz, Dmax = (1us - 100ns)/1us = 90%.

4.I get it. Futher more, what cause the perturbation? the power of cap or the inductor?

[TH] The load transient causes the duty cycle to change to bring the output back to its regulated setpoint. The input current has a change/perturbation, with a simultaneous change in input ripple, etc..

Hi Wenyi,

The max duty cycle for a given controller depends on the chosen switching frequency and the min off-time. If tsw is the switching period, then Dmax = (tsw - toff-min)/tsw.

The modes of operation of the regulator are continuous conduction mode (CCM) and diode emulation mode / discontinuous conduction mode (DCM) - these depend on if the inductor valley current goes negative or not. The controller itself can be in active or standby mode, depending whether it's switching or not.

Regards,

Tim 

Wenyi,

What you're describing is typical and expected behavior for light-load or no-load operation. The applicable frequency will depend on the loading condition and the chosen component values. Please send waveforms if you need further support.  

Regards,

Tim

Hi Wenyi,

See answers below.

Regards,

Tim

Wenyi Zhao1 said:

Hi Tim,

I hope these further information can make you understand my questions and give the answer.

1. for light-load operation, the controller operated when Vin high than 3.6V, and the switching frequency is 310.8k when testing and the frequency changes to 803.6k when Vin increases to 5.278V. So, during Vin from 3.6V to 5.278V, which mode does the controller operates?

[TH] This is frequency foldback when Vin is close to the Vout setpoint and the min off time is reached (effectively a max duty cycle). See the datasheet for more detail.

2.for light-load operation, when Vin is higher than 7.5V, the Iin decreases to 0A and the switching frequency is around 34.2Hz.which mode does the controller operates?

[TH] This is frequency foldback as diode emulation mode is enabled. The inductor current is not allowed to go negative. The Fsw will change to control the energy delivered to the output to match the load requirement.

3.when the load is 0.01A, when Vin is higher than 7.5V, the Iin decreases to 0A and the switching frequency is around 2kHz.

when the load is 0.1A, when Vin is higher than 12V, the Iin decreases to 0A and the switching frequency is around 15kHz.

when the load is 1A, when Vin is higher than 8V, the switching frequency is around 1.1MHz.

why does it different to light-load operation?

[TH] This is frequency foldback as diode emulation mode is enabled. The inductor current is not allowed to go negative. The Fsw will change to control the energy delivered to the output to match the load requirement. This depends on Vin, Vout, Iout, L, Cout, etc.

4.In light or low load, why does Vout*Iout>Vin*Iin when testing which means Pout>Pin? the test points are on the EVM except long wire.

for example, when load is  0.1A, when Vin>8V, Iin decreases to be smaller than 0.1A, which may be 35uA as datasheet says. So, Pout is around 0.5W while Pin<1mW, does it a normal condition and where the power?

[TH] Pin must always greater than Pout on an average basis. Check your measurements and make sure the DMM has a long aperture window setting.

5.when the load changes, from 5A to 0A, the Vout changes to see a different ripple like the photo 00093 below, why does it happen?

[TH] This is expected behavior as diode emulation mode is enabled, the inductor can't go negative so there is no active discharge of the output to bring the overshoot voltage back down to the regulated setpoint (it's effectively the load current discharging the output cap that causes the ramp down in Vout).

6. When the load is 1A, and Vin is 12V, i find the subharmonic oscillation from duty as shown below, why does it happen with the compensation inside?

Regards,

Wenyi

This is frequency foldback as diode emulation mode is enabled. The inductor current is not allow to go negative.

Hi Wenyi,

See answers below. I will follow up with a direct email.

Regards,

Tim

Wenyi Zhao1 said:

Hi Tim,

1. i have known when Vin is 3.6V to 5.2V the controller doesn't work at normal frequency due to min toff. But, why is lower Vin corresponded to lower frequency? Or, does it refer to a expression?

[TH] If max duty cycle (minimum off-time condition) is reached, the Fsw remains constant and Vout tracks Vin with a drop depending on high-side FET Rdson and inductor DCR. Also, Fsw reduces at light load if diode emulation mode is enabled.

 

2.What's more, when filtering step-down regulators from ti.com, the dmax of 5141-Q1 is 78% and does it mislead customers to ignore the controller?

[TH] The max duty cycle varies with Fsw. We will draw attention to the higher Dmax that's possible at lower Fsw in the next datasheet update cycle.

3.If you have some expression about the frequency during diode emulation mode with regard to Vin, Io or other parameters, please refer it to me.

[TH] Not available.

4.I say in light-load or low-load condition, Pin<Pout. I know it's incredible. But in these conditions, during diode emulation, Iin is very low just as iq=35uA, although Vin may be 24V or higher, Pin is still very low. So how to explain the problem?

[TH] Try setting the voltmeters and ammeters with 1 second aperture window to measure Vin, Iin, Vout and Iout.

5. So in the question 5 photo, why cause the ripple when light load? I don't see the same ripple in the test guide.

[TH]  I'll email you a ppt file to describe this.

6.You say the subharmonic oscillation is normal? Will it lead to an unstable condition?

[TH]  I'll email you a ppt file to describe this.

Hi Wenyi,

Given the high Fsw, the regulator will operate in dropout in that condition. I edited the previous post as the Fsw will remain constant in dropout.

Not sure if we have an app note on diode emulation specifically - however, it is very similar to conventional DCM operation (the low-side FET is turned off when the current reaches zero, so the behavior is similar to a diode).

Regards,

Tim

WEnyi,

I think we are discussing this in another post. Note that the duty cycle is a function of Vin, Vout as well as voltage drops of the FETs and inductor. In any event, a min off time pulse is forced during every switching cycle to charge the boot cap.

Regards,

Tim