LM5112: Fast mos driver (LM5112) + fast mos (suggestion?) for a pulsed led driver

Hi Fulvio,

thanks for your question,
whats your application? GAN FETs can allow a better efficiency due to lower Rdson for the same size FET. GAN can be easier to drive with the help of our effective GAN drivers. what MOSFET are you considering?

LMG1020 can drive a MOSFET since it has a 7A peak drive current.
Another option is UCC27611 which can also get the job done with a smaller MOSFET or GAN.

thanks,

Dear Jeff,

thanks for the reply. 

I'm still check about the mos (maybe i should ask help also for this). This should be a current driven UV-LED with a current between 5 and 100 mA with a switching frequency between 100khz and 1 Mhz.

did you have some mos or gan to suggest?

thanks

Hi Fulvio,

what end equipment will this be in? if the size is constrained we recommend the following FETs:
For GaN, you dont need a high voltage GaN so theres a limited selection since GaN has a intrinsically higher breakdown voltage.
EPC2040 and EPC2023 are single configuration. You might be on with a larger gate capacitance since your switching is only up to 1 MHz so EPC2023 might work for best efficiency.

For Silicon, we have TI CSD17381F4 which is small enough to fit in cell phones but fast enough for ns pulses and 10A peak current. Let me reach out to the TI FET team to see if theres a better FET for your application. I will update you soon.

thanks,

Dear Jeff,

i don't have any "space" problem,  all this component will be place in a custom pcb board.

I think that maybe it could be usefull if i attach my schematics.

The V-I conversion is made by a TI circuit example (slau502). The uv_led pspice model was give to me from led producer. If i use continuos mode, everithing goes well (From a fixed dac voltage i obtain the right led I/V), in this case starting from 3.3V i obtain 47mA/6.63 to the led, in according to the led datasheet.

But when i try to "pulse drive" it, i got a lot of problem, first of all the current rise up to 222mA (and the V to 10V) so i hope that using a mos driver with a fast switching mos to fix this problem. I face this problem also if i put the mos as high side or low side.

PS i don't mind that i have bot the LM5112 and LMG1020 i just attach the mos to only one of them and run the transient simulation. Btw i got the same result ( a pulsed burst of current of 222mA) with both.

Dear Jeff 

I attach the lm5112 and lmg1020 as you see in the pic (just as explicative way, i attach the lmg with a low side mos and the lm with high side mos, when i decide which is better for my application, i will remove(of course) one of them. 

About the project: due a desing choice we decide to pilot this uv-led in current and not with a pwm. So i use the TIA to convert the V to I ( 3.3V ==> 50 mA over the LED). This work fine when i need to pilot it in continuos mode, but all the problem is when i need to pilot it in pulsed mode.

Btw i i will try to add the resistor and use the csd17381f4 

Hi Fulvio,
thanks for the update and sch!

so why have you chosen 5112 for highside? 5112 is a lowside driver.
what are you trying to achieve when running in pulsed mode? why not just use continuous mode?
so V_pulse is a constant dutycycle and frequency? seperate from the TIA?

thanks,

Dear Jeff,

no i just connect the 5112 and the 1020 "just" to give an idea where i neet to place the switching mos (or high side or low side), but good to know that the 5112 is a low side.

Unfortunatlly this project demand to pilot the led in both way, continuos and pulsed. So i need to plan a circuit that could work in both mode with the same led.

2 cent on the project: here at the univerisy of Rome, we are developing a discharge system for space, based on the photoelectric effect of the uv photons over gold plated surface. this demand to have a continuos mode for fast discharging and a "pulsed" mode for "continuos" discharging.
Atcually, we develop the FPGA code in order to generate a continuos mode or a pulsed mode starting form an external trigger of 100kHz (i could decimate it up to 1/65000, and stretch the pulse from 25ns to 10 us, even if the project require a minimum signal of 500ns)
I also plan to control the current value using a V_DAC in order to set the sistem current from 5mA up to 50mA

So in my schematic V_DAC is the value of V generated from the DAC to be converted in current and V_pulse has a frequency between 100kHz and 1 MHz. (with a duty cicle of 50%, at moment i was working with 100kHz signal, so 5us up to 3.3V and 5us down to 0).

I hope that this "quick" description could help you to undestand what kind of electronics i need to develop.