LM3429-Q1

Hello Kura,

Yes, tPULSE is a minimum PWM dimming on time. I will have to check with the IC designer to see if we could get the delay time for UVLO and VCCUVLO, but they should be very fast. I will let you know if I can get any information for you.

Thanks,

Clint

Clint-san
Thank you for your reply.

They have additional question about LM3429-Q1.
They need 0%-100% dimming quality with 0.1% resolution.
Can they achive above dimming performance using with only analog dimming?

Please give us your comment.

Regards,

Hello Kura-san,

0.1% resolution is difficult simply due to component tolerances, but it might be ok over a certain range. But the error amplifier has up to +/-7mV of offset so the accuracy will not be great over a significant range. This is not an issue only with our devices, it is an issue with any LED driver. Generally if you need a very wide range and accuracy it requires a combination of analog dimming and PWM dimming. It might even require a second control loop if you need resolution down to 0.1%. Analog dimming alone usually only gives you a range down to maybe 5% with any accuracy.

Best regards,

Clint

Hello Kura-san,

I did get a reply from the IC designer regarding the delays. His response is that the delays have not been characterized, but due to internal filter delay they are likely in the 1-2us range.

Best regards,

Clint

Clint-san
Thank you for your reply.

They consider to use LM3429-Q1 as SEPIC controller for LED drive with following requirement.(PWM dimming)

VIN:4.85 to 5.15 V
Vout:about 6.0 to 16.0 V
Iout max:350.0mA
PWM frequency:75.0 to 85.0Hz
Duty:0.0-100.0 %
Duty Resolution: <1%
LED String:TBD

Can they achive above requirement with LM3429-Q1?
In this case,please let us know MIN ON DUTY.

And about protection function of LM3429-Q1,Is below our understanding correct?
OVP:Latch (to clear this condition,only VIN ON/OFF)
UVLO:Latch(to clear this condition,only VIN ON/OFF)
Current limit:auto recovery
Thermal shut down:auto recovery

Regards,

Hello Kura-san,

I am not sure exactly how to calculate the minimum pulse width for a SEPIC, but using the boost equation with Vo being Vout+Vin (this generally works with a buck-boost) it is likely around 12.5us. SEPICs are a little slower so maybe say 25us for margin of error. That calculation assumes 22uH inductors since that information was not provided. In either case at 85Hz a pulse width of 1% duty cycle is around 117us, so you should be able to achieve less than that. One other note if using a SEPIC with PWM dimming is to use coupled inductors. Separate inductors will generally ring a whole lot and you will need to add a large snubber across the SEPIC capacitor.

Speaking of which you could use a boost with those specifications. But to answer your questions on protection:

OVP: It will not latch. It will stop switching until the OVP falls below the threshold set by the hysteresis. Then it will resume switching (and repeat over and over if the conditions persists).

UVLO: It will not latch. If the input voltage rises back above the UVLO threshold the device will resume switching.

Current limit: It will continue switching but will not let the switch current exceed this limit. Make sure the inductor saturation current rating is higher than the current limit or damage can occur.

Thermal shutdown: It will auto recover much like OVP. If the threshold is reached the device will stop switching until the temp cools some amount and then resume. It will cycle in/out until the die does not exceed the threshold any more.

Thanks,

Clint

By the way, the LM3429 can do buck-boost without being a SEPIC which only requires 1 inductor as long as you do not need one side of the LED string connected to ground (it connects to Vin in that configuration).