TLE8366EV Compensation

[OzzyRules]

Hi,

Hi,

I'm helping one of my customers and found that the TLE8366EV shows sign of instability. I'm trying to understand how to stabilize the TLE8366EV. Its a 12V to 24V input converter (automotive) and output set to 10V.   Upper feedback resistor is 18.7K and lower resistor is 1.2K. There is a 4.7uF across lower resistor. Compensation is 220nF with 22Kohms. Inductor is 22uH and output is 3X 10uF (30uF total). Load is about 200mA to 500mA.

1) Is there any equations describing the IC's modulator response or  equations in for calculating the compensation ?

2) What is the purpose of the cap across the lower feedback resistor ? All I see its called a "smoothing capacitor" and its required for adjustable output power supplies.

Any help is a appreciated. Thanks.....

[OzzyRules]

Opss Typo..... Its 4.7nF across the lower feedback resistor. Not 4.7uF.

[AnilKumar]

Hello,

Thank you for posting in Infineon community,

Unfortunately, we do not have a tool publicly available currently for the stability analysis of TLE8366EV.
Can you please share the time domain waveform of output signal of TLE8366EV and also ESR values of Output Capacitor
used for us to perform simulations at our end?

Please check the following aspects additionally if not considered.

1. Recommended Rcomp, Ccomp, when internal Oscillator frequency used, is 22kΩ and 22nF respectively.
2. Additional Filter capacitors CFilter (ceramic, with low ESR) in parallel to the output can be considered. (10nF to 220nF)
3. R-C-snubber circuit in parallel to the freewheeling diode will help to damp the oscillations (only if necessary)
4. Hope Schottky diode with fast recovery time is used to minimize reverse recovery losses and oscillations.

Cap across lower feedback resistor is used to filter out noise pick up from output node to FB if routed via long trace. But since it is
connected to FB pin, it is also recommended to choose lower value for stability reasons (0.5nF recommended as per datasheet)

Reach out to us for any further query

Thanks,

Anil

[OzzyRules]

Hi Anil,

Unfortunately, we do not have a tool publicly available currently for the
stability analysis of TLE8366EV.

Bummer. I actually create simulation models for customers. Including
Infineon space group in Arizona.

Can you please share the time domain waveform of output signal of TLE8366EV
and also ESR values of Output Capacitor
used for us to perform simulations at our end?

Waveforms attached. I attached what the output ripple voltage with load
current and switch node voltage with load current. You can see the
duty-cycle changing.

The output uses 3 TDK CGA6P1X8R1E106K250AC. 10uF,25V,1210,X8R. I attached
the characteristic sheet. Assume 3mohms ESR. There's not much else for
capacitance on the output.





Additional Filter capacitors CFilter (ceramic, with low ESR) in parallel to
the output can be considered. (10nF to 220nF)
There's 3 100nF ceramic additional caps right at the loads.



R-C-snubber circuit in parallel to the freewheeling diode will help to damp
the oscillations (only if necessary)
Hope Schottky diode with fast recovery time is used to minimize reverse
recovery losses and oscillations.

Not much ringing and it's a good Schottky. The layout is good.



Cap across lower feedback resistor is used to filter out noise pick up from
output node to FB if routed via long trace. But since it is
connected to FB pin, it is also recommended to choose lower value for
stability reasons (0.5nF recommended as per datasheet)

There's 4.7nF. I'm suspicious about the value. With 1.2Kohm resistor for the
lower feedback resistor it forms a frequency pole at the same frequency of
the oscillation I see. However I do not know the history why 4.7nF was used.
I have an older app note showing 1nF with a higher resistance value in
parallel.



Btw, the CCOMP capacitor is 220nF and not 22nF in this circuit. Do not know
history of that either.



Thanks for the quick response. My client was ready for production and saw
that oscillation. I like to minimize the changes. If changing the cap on the
feedback to 0.47nF or some other value works, then that would be great. My
concern is that more we change the more chance something goes wrong.



Thanks..Ron Wunderlich







Thanks..Ron Wunderlich

<>

Name: Ron Wunderlich, PhD

Title: President/Consultant

Office: 607-754-0347

Cell: 607-341-3008

Email: rwunderlich@i2designs.com

Website: <> www.i2designs.com

[AnilKumar]

Hello Dr. Ron,

Thank you for the input.

I don't see any attachments. Can you please check and confirm? Meanwhile, I am checking with the expert team about the values proposed.

Thanks

Anil

[OzzyRules]

Hi Anil,

I thought I attached them. See if this works..







Thanks..Ron Wunderlich

<>

Name: Ron Wunderlich, PhD

Title: President/Consultant

Office: 607-754-0347

Cell: 607-341-3008

Email: rwunderlich@i2designs.com

Website: <> www.i2designs.com

[AnilKumar · Accepted solution]

Hello Dr. Ron,

We have performed stability analysis on your application circuit and here is our proposal for a stable system.

*Configure the resistor divider as in the picture

That is,

1. Add capacitor < 1nF across lower resistor (for example 500p).

2. Add 1nF across upper resistor

3. Ccomp->33nF, Rcomp-> 3.3Kohms.

The phase margin is 79deg (@ 33 kHz) which is enough to guarantee that some non-idealities not included in the model do not impact the stability.

Thanks

Anil

[OzzyRules]

Hi,

Thanks..I'll give it a try this Friday...



Thanks..Ron Wunderlich

<>

Name: Ron Wunderlich, PhD

Title: President/Consultant

Office: 607-754-0347

Cell: 607-341-3008

Email: rwunderlich@i2designs.com

Website: <> www.i2designs.com