Hi all,

I try to use the LT3757 to step up 48V to 88V in order to driver a 88V led string. I made a simulation in LTspice and it works fine. So i decided to make a prototype based on the circuit's datasheet and formulas as following:

Design:

Vcc (chip supply): 12V

Vin(at the inductor): 48V / 8A

fswitch: 200kHz

Cout: 100µF

L: 600µH

Rsense: 0.04Ohms

Desired results:

Iout: 1000mA

Vout: 88V



When i switch on the circuit, without led string,i can see high output voltage that i can change with the feedback resistor divider. So i adjust the output voltage @ 88V. But, when i connect the led string, the output voltage drops down at 76V with only 10mA in the led string. Adjusting the Vfbx voltage has no effect on the output voltage which stays at 76V and Vfbx stays under 1.6V.

Could you help and explain me why i get this issue ?

 

Have you tried a different load than LEDs? Did your LED string has a current limiting resistor?

 

Unfortunately i have no load for so much voltage. I don't think that the leds have a limiting current resistor. Here leds datasheet https://www.cree.com/led-components/media/documents/dsXPG3.pdf

 

So, you made this LED string by yourself or did you bought this 88 LED strip?

 

I made it and it works fine with 2x48V power supply @ 1000mA.

 

Why are you powering an 88V string with a 96V supply?
How many LEDs do you have in the string?

 

No..., the 2x48V are adjustable (42V to 52V each), and i can drive the led string at 88V / 1000mA. I have already checked it. Now i want to use just one 48V power supply. That's why i need to step up this voltage with the LT3757. I don't think the problem comes from the leds. Something is going wrong with my design or components. I have checked all formulas and calculations and i cannot solve the problem.

 

Have you double checked all the wiring?
Is the input voltage staying at 48V?
Do you have an oscilloscope?

 

If you have 23 LEDs in the string, as shown in your schematic, the total forward voltage at 1A should be about 2.89V x 23 = 66.5V, according to the datasheet. Why are you aiming for 88V?

 

If you have 23 LEDs in the string, as shown in your schematic, the total forward voltage at 1A should be about 2.89V x 23 = 66.5V, according to the datasheet. Why are you aiming for 88V?

For the simulation I have chosen 23 LEDs because Vf ~ 3.81V @ 1A for each. I measured it with Ltspice.
As you can see on the slope, Vout ~ 88V @ 970mA in led string.

 

Vf ~ 3.81V @ 1A for each.

That is not the Vf value in the datasheet from the link you posted, so the sim does not represent reality in this case.

 

A basic problem is that LEDs are current devices, not voltage. You need to make a constant current supply, not a constant voltage supply.
So, decide on the current you need and make a supply the controls that current. And has a sufficient voltage capability to overcome the forward voltage drop of the string of LEDS.

It may be worth trying to remove R3, change R4 to (1.6V/LEDCurrent) and adding it in the ground end of the LEDs so the regulator tries to keep 1.6V across that resistor, driven by the LED current.
I have not tried this but it may work.

 

The problem comes from the L value (600µH) to large. Calculations give 270µH so i change the induction with a 220µH avaible in my stock. The voltage drops until 83V with 320mA in led string. I suppose that my design cannot provide 88V because of inductor saturation (coil heating).
I think a 270µH 5A DC could match...
Do you think this one could make the job: https://static6.arrow.com/aropdfcon...08dc2dbc66f6c4eca00d469dde1/pm2120_series.pdf

Yes Dendad, you are right, i have already think about this way to drive the current. This first step up test was just for the power voltage stage capability.