Influence of size and voltage rating when replacing ceramic decoupling capacitors

Johannes Maier

Part Number: LM27762

Tool/software:

Hello

tl;dr: I am trying to figure out, if its ok to replace ceramic decoupling capacitors with capacitors of lesser size and voltage rating in a DC power system

I have developed my first power system with two linear regulators(LM27762 and TLV67), which have ceramic decoupling capacitors on Input and Output:

I used several references as guidelines for the schematic and the parts(BOM)

ADS1261 EVM for LM27762

TLV67EVM for TLV67

We can manufacture our pcbs inhouse and for the first prototype I want to replace some capacitors from the references with capacitors we already have in stock.

But I am unsure, what effect(if any) this will have on the prototype, when diverging from the references.

I know that several parameters influence the actual capacity of a ceramic capacitor:

rated voltage
size
nominal capacity
tolerance
acutal supply voltage
temperatur coefficient (X5R,X7R,...)
frequency of ripple?

Example:

Replacing

4.7 uF +-10% 25V 0805

with

4.7 uF +- 10% 16V 0603 capacitors

Questions I have:

What will happen when I replace the input and output Decoulping capacitors
Will the behaviour of my system change?
Can I ignore the change? It is DC power and not AC.
How big is the influence of each parameter?

So if anyone has some adivce and can help me, this would be much appreciated. If you have any questions feel free to ask.

Thank you in advance for your time

Regards

Johannes

8 months ago

0 Andreas Mueller 8 months ago

TI__Mastermind 24990 points

Hello Johannes,

thanks for reaching out in E2E.

I'm supporting the LM27762 but I can forward this thread later to the TLV67 experts if needed.

1. The biggest concern would be the DC bias derating, which leads to a lower effective capacitance at higher voltages. Smaller package size for ceramic capacitors usually results in stronger DC bias behavior, means the effective capacitance will be lower.

2. A lower capacitance will result in higher voltage ripple.

3. It depends how much voltage ripple are acceptable in your application. I would suggest to try it out and check the impact.

Although it is DC voltage, there will go AC current in and out of the buffer capacitors to maintain the voltage. Especially at the input of the LM27762 there are high dI/dt current spikes expected due to the switching of the charge pump.

4. Assuming all other parameters are the same and only the package size and rated voltage is smaller, then I would assume the DC bias derating is the biggest impact, unless the rated voltage is lower then the actual operating voltage, which can damage the capacitor.

I guess the capacitor manufacturer support center could give more details on the individual parameter changes.

For the LM27762 mainly the effective capacitance is relevant.

I would assume also for the TLV67 the effective capacitance is relevant for the filter performance. Let me know if I should loop in the TLV67 experts as well.

Best regards,

Andreas.

0 Johannes Maier 8 months ago in reply to Andreas Mueller

Prodigy 70 points

Hello Andreas,

thank you for your answers. They do confirm my first assumption, that changing size or rated voltage are factors I have to consider in my design.

The LM27762 will power an ADC and the current drawn will be around 1-5 mA. With a low power device like this, is it possible to make an assumption, at what nominal capacitance it is safe to assume that it will probably "work" in terms of ripple.

Kind of like a rule of thumb?

At the moment I don't think, it is necessary to add addional experts to the conversation.

I will follow your advice will check the impact with two different designs.