What's causing this current ripple when the PWM is turned low?

I've calculated everything up and came up with the following.

Vin – 12vdc
Vout – 3.1vdc
Iout– 350mA
Led driver – LT3517
Led – Cree Xlamp XM-L Color (White Die)
IC frequency – 500kHz
Pwm frequecy - 500kHz

Inductor selection:

Dbuck = Vled / Vin

3.1/ 12 = .2583

Tsw = 1/Frequency

1/500,000 = .000002

ΔI = Percentage of current ripple x Iout

.30 x .350 = .105

L(µH) = Dbuck • Tsw(µs) • (Vin – Vled) / ΔI

.2583 x .000002 x 8.9 / .105 = L(µH) = 0.000043788

Inductor = 43.7uH

There isn't a calculation for the output capacitor in the document for the IC, so I sourced a document from a google search and found a Texas Instrument document on buck converter design and it says to use the following to calculate the output capacitor. Also read in that same document that 25mV is a good set point for output voltage ripple as the following equation require you to put in a value for the
desired output voltage ripple.

Capacitor selection:

Cout(min) = ΔIL / 8 x Fsw x Δvout

.105 / 8 x 500,000 x .025 = .00000105

Capacitor Out = 1.05uF

Voltage probed after mosfet and before LED = Green(n003)
Voltage probed after led before inductor = Teal(n004)
PWM voltage = Red(n012)
Current through LED = Blue(D2)
Current through inductor = Purple(Inductor)

Schematic of the circuit

 

What is the the trr (Reverse Recovery Time) for the diode used in the simulation? To large a value will cause turnoff spikes like the one you observe. You might find the following interesting.

https://www.fairchildsemi.com/techn...covery-and-Its-Effect-on-Switching-Losses.pdf

I might use something like this in an actual design

http://www.st.com/content/ccc/resou...df/jcr:content/translations/en.CD00110329.pdf

 

What's causing this current ripple when the PWM is turned low?

I've calculated everything up and came up with the following.

Vin – 12vdc
Vout – 3.1vdc
Iout– 350mA
Led driver – LT3517
Led – Cree Xlamp XM-L Color (White Die)
IC frequency – 500kHz
Pwm frequecy - 500kHz

Inductor selection:

Dbuck = Vled / Vin

3.1/ 12 = .2583

Tsw = 1/Frequency

1/500,000 = .000002

ΔI = Percentage of current ripple x Iout

.30 x .350 = .105

L(µH) = Dbuck • Tsw(µs) • (Vin – Vled) / ΔI

.2583 x .000002 x 8.9 / .105 = L(µH) = 0.000043788

Inductor = 43.7uH

There isn't a calculation for the output capacitor in the document for the IC, so I sourced a document from a google search and found a Texas Instrument document on buck converter design and it says to use the following to calculate the output capacitor. Also read in that same document that 25mV is a good set point for output voltage ripple as the following equation require you to put in a value for the
desired output voltage ripple.

Capacitor selection:

Cout(min) = ΔIL / 8 x Fsw x Δvout

.105 / 8 x 500,000 x .025 = .00000105

Capacitor Out = 1.05uF

Voltage probed after mosfet and before LED = Green(n003)
Voltage probed after led before inductor = Teal(n004)
PWM voltage = Red(n012)
Current through LED = Blue(D2)
Current through inductor = Purple(Inductor)

Schematic of the circuit

Can you post your .asc file?

 

Thanks for sourcing the documents, I'll start digesting those, ha ha.

When I click upload a file nothing is showing up but folders.

 

Thanks for sourcing the documents, I'll start digesting those, ha ha.

When I click upload a file nothing is showing up but folders.

Okay, I put it in Spice. One thing I think is you PWM frequency is to high. Try like a 10MS period.
Lots of little switching spikes. Which one is bothering you?

 

Hey man, thanks for the response, and taking the time to punch it all into LTspice.

I sent an email the moderator that sent me the welcome message to try and figure out why I can't upload any attachments.

After you letting me know that I read in the IC document that min on/off time is 4 x Tsw of the micro controller. At 500kHz that would be 8uS, I had it set to 2uS.

Thanks man! Made the square wave square!

I am still having current spikes, but now at the beginning and end of the pwm switching.

Led current spike on/1.4a off/.788a

On Spike

Off Spike

P-P led ripple is 26.7mA

Max current inductor about 408mA

P-P inductor current ripple 112.85mA

Thanks again man, much appreciated!

Any idea what's causing the spikes?

 

Hey man, thanks for the response, and taking the time to punch it all into LTspice.

I sent an email the moderator that sent me the welcome message to try and figure out why I can't upload any attachments.

After you letting me know that I read in the IC document that min on/off time is 4 x Tsw of the micro controller. At 500kHz that would be 8uS, I had it set to 2uS.

Thanks man! Made the square wave square!

I am still having current spikes, but now at the beginning and end of the pwm switching.

Led current spike on/1.4a off/.788a

On Spike

Off Spike

P-P led ripple is 26.7mA

Max current inductor about 408mA

P-P inductor current ripple 112.85mA

Thanks again man, much appreciated!

Any idea what's causing the spikes?

The big one is from charging the cap when the PWM turns back on. I'm not clear yet on the little one.

 

Hi

Just a couple of questions...
Why reduce the switching frequency? You can use smaller value components if you keep the 1 Mhz switching frequency.
Why are using a 500 Khz PWM frequency? Why not use a much lower frequency like 500 Hz?

I think the spike is caused by the model. If you insert a 1k resistor in series with the gate, it greatly reduces the spike.

 

If I move Cfilter to the left of the mosfet or take it completely out of the circuit, the current waveform and ripple is the exact same on the LED and inductor @112mA p-p ripple. But does do away with the current ripples at turn on and off.

Am I messing up on capacitor calculations?

Current waveform diode:

Current waveform inductor:

Current ripple diode:

Current ripple inductor:

I can increase the frequency, there's really no reason other than EMI, for me keeping the frequency low. If you guys have any suggestions like that
let me know. I appreciate the help and knowledge I'm gaining from you guys.

I slowed the frequency to 500hz, plan to drive this from Arduino that I will integrate into the PCB.

 

If I move Cfilter to the left of the mosfet or take it completely out of the circuit, the current waveform and ripple is the exact same on the LED and inductor @112mA p-p ripple. But does do away with the current ripples at turn on and off.

Am I messing up on capacitor calculations?

Current waveform diode:

The cap gets rid of the high frequency ripple current thru the diode. I'm not sure how important that is since you can't see it anyway. Try a 4.7Ufd. ceramic.
If your worried about the spike, I don't think I would. It's only a few nanoseconds in length. Not enough to do any damage.