LM5088-2 current limiter kicking in at low Vin-Vout (Low Drop Out Mode) sooner than expected

What I am finding is that If I lower the input voltage to about 26.5 Vin for a 24Vout, and the load is approximately 1 Amp, the LM5088-2 will go into hiccup mode.

This is happening even though my current sense resistor is set for 0.051 ohms (The above value of 0.51 was a typo.)

Why does the LM5088-2 enter into Hiccup mode while it is also in Low Drop Out mode>?

For example, If I continue to lower Vin, the current limit threshold seems to keep getting lower and lower (ie, a 0.5Amp load will trigger the current limit hiccup mode.)

Is this expected behavior? Is the hiccup mode just not functional while in Low Drop Out mode?

Also, is there a block diagram or explanation of the "DFC" circuitry that is identified in the datasheet?

Hi David

Have you check RES pin? I'd like to double check whether it is hiccup current limit or a behavior similar with hiccup current limit.
Switching frequency is decreased when you see the hiccup operation? I'd like to check whether it happens in DFC mode or not.
In DFC mode, current limit can be triggered early since the inductor current ripple increases and LM5088 employs peak current limit.

Regards,
EL

Hello Eric,

I shorted out the RES pin to defeat the hiccup mode feature for testing. The LM5088 then continues to work, although there is a little ripple on the ouput.

I believe the current limiter is being triggered at a lower level in DFC mode than in normal operating mode. Unfortunately TI does not give a block diagram of what the DFC does or is and the current limiter equations provided do not seem to apply to DFC mode.

So what would be the equation for the current limiter in DFC mode?

Switching frequency is indeed decreased in DFC mode.

Thanks

Hi David

- Would you please let me know Rramp and Cramp values?
- Would you please check if the switching frequency is decreased when you see the early current limit operation? I'd like to check whether it happens in DFC mode or not

Regards,
EL

Hello Eric,

The Rramp = 73.2k, 1%
Cramp = 1600pF

Ccomp = 1200pF
Chf = 47pF
Rcomp = 95.3K

Lout = 180uH
Rsense = 0.056 ohms.

Rrt = 33.34k

When Vin approaches 26.8Vdc or less, (where the vout is 24Vdc) the current limiter begins to trip. If I keep a capacitor on the RES pin, the LM5088 will go into hiccup mode. If I short out the RES pin, the LM5088 will keep operating, but the output will begin to develop a ripple.

I can observe that the switching frequency begins to decrease and keeps getting lower if I reduce the Vin further below about 26.8Vdc.

Why would the current limiter be tripping while in LDO mode?

Hi David 

Based on my calculation, I guess your current limit level is close to your full load current and I think if any noise, current limit can be triggered early. 

At 1.2A load when Vin=26.8V and Vout=24V, sensed current is around 1.087V and it is close to 1.12V minimum current limit threshold. 

Can you check it ? 

Regards,

EL 

Hello Eric,

The current limiter begins to enable at 0.5Amp to 0.6Amp load. This is what is confusing me.

Why does LDO mode cause the hiccup mode to trigger at loads much less than 0.12V / Rs?

Is this expected behavior?

If you can try it out, should be easy to spot. LM5088-2 in low drop out mode, under load. The hiccup mode triggers at levels much lower than expected.

Hi David

I think you mentioned that when Vin was 26.8V, the current limiter began to trip and you could observe that the switching frequency decrease if you further reduce the Vin below 26.8Vdc.

When the current limit begins to trip, no frequency change. Right?

Regards,
EL

Hello Eric,

I am going to get some scope captures and post them to show the behavior at light load (0.05amp) and a heavier load (0.85amp.) while in LDO mode (Vin = 26.8Vdc or less.)

I'll capture a few switching cycles to show when the switching frequency begins to reduce.

Hello Eric,

See the following scope captures.  The orange trace is the gate voltage to the MOSFET.  The blue trace is the voltage across the current sense resistor.

The two setups are: 50mA load and 850mA load, both should be well below the current limit.

I'll get the same traces with the Ramp pin monitored later on.

Hi David

Richard contacted me by email. Further discussion will go through Richard.

Regards,
EL

Eric,

The customer is still asking the following questions:

I provided the oscilloscope captures, so I'm not sure what else to send.  The main question is: Does the Low Dropout Mode current limiter activate at a lower level than when not in Low Dropout Mode?  That is what I am seeing.  I want to know if this is expected (meaning there's nothing further I can do,) or if this is unexpected (meaning I need to look at the PCB layout or components further.)

Could you please reply here in the forum to these questions?

Thanks for your help with this!

Richard Elmquist

Hi Richard 

The customer has to make it sure that the current limit was triggered in normal switching mode or in DFC mode and how much current limit level difference between the normal switching mode and the DFC mode. I observed many cases that the early current limiting happens by a noise injection at CS-CSG, not by the mode change. 

Please let them figure out the current limit level when Vin is a little bit higher than DFC mode threshold and the current limit level when Vin is a little bit lower than the DFC mode threshold. If there is small difference, I can say that the issue is caused by a noise, not by the DFC mode itself. 

By the way, as I mentioned, in DFC mode, the current limit can be triggered early since the inductor current ripple increases and LM5088 employs peak current limit. 

Regards,

EL