Phase lead capacitor - What is it?

[tadam]

Hi,

The datasheet of a DC/DC stepup IC (LTC3429) mentions for the output capacitor: "An additional phase lead capacitor may be required with output capacitors larger than 10 uF to maintain acceptable phase margin."

I want to use 3300 uF... so I guess this applies to me

What is it, where does this cap go and how can I calculate a value for it?
Thanks!

 

[mneary]

Since this is an LTC (Linear Tech) part, you probably can use LTSpice to test your configuration.

 

[crutschow]

Why do you want such a large output capacitor? Are you trying to get very low output ripple? Such a large capacitor is usually overkill with a switching regulator since the capacitor's ESR is likely much higher than its capacitive reactance at the 500kHz switching frequency. If you really need to go to a larger capacitance than the recommended values, then you should use several capacitors in parallel to minimize the ESR.

The cap is Cpl (optional) in the Block Diagram on page 5 of the data sheet. It adds phase lead to the closed loop response to improve the phase margin.

It's value likely has to be empirically determined after you build it and test the phase margin.

 

[tadam]

Hi thanks for the hint. I played in LTSpice and added a 330 pF cap, it actually improves my transient behaviour. I don't see how it would be absolutely necessary though, just a bit less overshoot.

I want the large output capacitor because my load draws pulses the IC cannot handle continuously. I use a low-ESR electrolytic (panasonic FM series).

 

[crutschow]

Hi thanks for the hint. I played in LTSpice and added a 330 pF cap, it actually improves my transient behaviour. I don't see how it would be absolutely necessary though, just a bit less overshoot.

I want the large output capacitor because my load draws pulses the IC cannot handle continuously. I use a low-ESR electrolytic (panasonic FM series).

Be sure and check the transient response of the actual circuit to make sure it is still reasonable. It shouldn't ring more than about 1 cycle.

If your load draws large pulses, you should also decouple at the load with a large capacitor, if you can.

 

[tadam]

I made some plots of Vout and Iin with and without the 330 pF compensating capacitor.
The load is first 20 mA, then 200 mA, then 3 peaks of 2 A.

Vout without
**broken link removed**

Vout with 300 pF
**broken link removed**

Iin without
**broken link removed**

Iin with 300 pF
**broken link removed**

This removes the Vout ripple for the 200 mA load.

What do you think?

 

[crutschow]

It looks slightly more stable with the 330pF.

Have you tried different values of the lead capacitance to see if you can find an optimum value? It sounds like you just arbitrarily picked 330pF.

 

[tadam]

Yes, but it's not so simple as the battery depletes. For example, this is with 680 pF and looks quite good at 3.4 V:

**broken link removed**

But at 2.5 V input the Vout becomes quite noisy....:

**broken link removed**

 

[Speakerguy]

That is a little 600ma boost regulator. If you want larger currents why not use one of Linear's more beefy parts?