TPS92518HV: What are the boundaries for the PWM frequency

Hello,

I assume you are not shunt FET dimming, you are using the PWM pins?  The maximum PWM frequency has more to do with the design such as the inductor value, input voltage, output voltage, switching frequency and resolution/accuracy desired.  By using the PWM pin the MOSFET will turn on and off however the inductor current still has to ramp up or down.  This slew rate depends on the voltage applied, for turn-on it is (Vin - Vled), for turn-off it is (Vled + Vdiode) using the equation V = L * (di/dt), di is the peak current threshold.

If you need accurate dimming at 32 KHz I would look at the shunt FET dimming method though there are things you have to take into account with that as well.

Improvements can be made with enable dimming such as lowering the inductor value/raising the switching frequency and optimizing the input/output voltage.

Best Regards,

Hi Irwin,

Thanks for the reply. I am indeed using the PWM input pin of the controller rather than shunt PWM.

The input Voltage is 48V, output voltage 36.5V, current 7A. L = 47uH, Rsense = 33mR,  PKTH_DAC = 240, ToffDAC = 21, ToffMAXDAC = 255
Somehow, the input current is only about 0.6A, when I'm expecting a lot more than that when the PWM signal is at 100%.

Any ideas on where to look?

Kind regards,
Jean

Hello Jean,

According to my calculations you should see 7.17A average, inductor di/dt of 0.21A.  Toff = 269 ns.  Fsw = 890 KHz.  This is fairly high switching frequency for this value of inductor and output current.

What is the inductor you are using?

What is the output voltage at?

Have you looking at the sw pin to see what it looks like?  Also verified that the other pins are approximately correct?

Also, you will not get dimming to work accurately at 32 KHz with this value of inductor, the ramp up time is 29.7 us and the ramp down time is 9.4 us.

Best Regards,

Hi Irwin,

Sorry for not getting back sooner, it was a holiday here as well as no internet :(

I've re run my calculations with a switching frequency of 350kHz and get

toff = 443ns I avg ~ 7.87A  di/dt = 0.266A

The inductor I'm using is https://www.we-online.de/katalog/datasheet/74435584700.pdf

The actual output current at maximum duty cycle is avg ~255mA amplitude 320mA

The output current looks sinusoidal-ish with a frequency of ~145kHz

The output voltage is at 26.9V

I've attempted to get some waveforms on the switch node as well as the transistor gate (that was a mistake as the probe adds too much capacitance) just need to figure out how to get the waveforms to you.

Kind regards,

Hello,

If I use your output at 26.9V and toff at 21 (decimal) I get toff = 365.5ns, if I use 21 (decimal) and 36.5V I get toff = 269 ns.

You can just look at the switch node, if it's not running continuous it will show in the duty cycle/waveform.

Do you have a schematic?

Best Regards,

Hi Irwin,

I can provide a schematic. But not on the forum.

Is there any way to get it to you?

Jean

Hello Jean,

You should have an email I sent you.

Best Regards,