![]() A lower frequency will typically be more efficient, but will also need larger components. Lower switching frequencies will allow the driver to stay much cooler as it doesn’t need to provide as much power to the MOSFET if it’s not rapidly switching the gate on and off. ![]() There are a few considerations for switching frequency in this design, primarily component size, and heat from driving the gate on the MOSFETs. I’m setting the divider for 96V which will ensure that the driver can’t exceed it’s 100V rating while still allowing the LEDs to be driven at full power. For this one, I am aiming for 1.25V at the pin. The second divider is the output voltage setting. The soft-start feature of the LT3756 should help reduce the voltage sag when switching on the panel, making sure voltage below my minimum set value is a definite sign of something not performing well on the regulator. I’m setting this to 40V for this driver, as I want to allow some voltage sag when the driver switches on before the AC-DC regulator can catch up. The IC also has two voltage dividers of interest to us, the first is a typical undervoltage lockout divider to turn the driver off when the voltage on the pin falls below 1.22V. They can more than handle the current and the large package will help dissipate heat, keeping my board cooler. I’m going to use 2512 sized resistors for the design. This sense resistor is used to set the maximum switch current. When I substitute my values, this gives me a 20milliohm resistor. The second sense resistor has a formula in the datasheet, and for boost resistors it is The driver expects a 100mV drop across this resistor, and as I want to drive 1.8A in each string, a little application of Ohm’s Law tells me I need a 56 milliohm resistor. The LED controller detects the current drop across this resistor, which is placed on the high side of the LED string. ![]() The first resistor whose value I need to calculate is in series with the LED for programming the current through the LED. The LT3756EMSE-2 has two current sense resistors to monitor the load. This should give me a 1.8A load to drive. The driver is rated for 100V output, therefore I’m aiming to run 16 LEDs in series, to reach 96V per string. Higher current rated components are more expensive than higher voltage rated components, so I am going to use the boost mode which I’d originally intended. In buck mode, the driver needed 32 parallel strings of LEDs to reach the desired voltage, which meant I needed components with high current rating. I looked at running the driver, which is capable of both buck and boost modes, in whichever mode offered the better performance. Not only does it look like it has great performance, it’s also a very good starting point for the design mentioned in one of the design references. In a previous project, I tried to see how far a monolithic driver IC could go, and 65 watts was really pushing it, so for my final driver solution, I know I need a controller with external MOSFETs.Īfter considering many drivers, I settled on the Analog Devices LT3756EMSE-2. When looking for an AC-DC power supply in the 300-400 watt range, a 48V supply is the cheapest at the suppliers I use. I want to use the highest voltage I can to reduce losses and keep the current as low as possible, thus reducing the heating on the LED panel. It’s a white, high-power LED with a CRI of 95 and forward voltage of 6 volts. I want to use the Luminus Devices MP-3030-210H-40-95 for the light panels. If you’re looking for the components used in the project, you can find them in my open-source Altium Designer Library. This project is available on GitHub for you to freely use as you wish. I’m still planning to run everything from a 350W 48V power supply to have both panels powered from the same 48V source. This should also provide me some additional versatility, as I can now separate the panels to offer more creative lighting options without drawing more power. Because of this, I’ve decided to split the panel into two 150-160W panels instead. To be able to reflow the board, I’m building a vapour phase reflow oven which will have a maximum board size of 230mm by 180mm as another project. A typical low cost DIY reflow setup with a converted toaster/pizza oven or skillet just can’t take a board that large. However, I’ve come to realize that the limitation for a panel with that many LEDs is not driving them, but instead reflowing the board. A good quality video light panel is expensive, and for good reason, but as a maker, I tend to like buying things instead of just buying them. ![]() Most of my recent projects have been working towards building a high CRI (Colour Rendering Index) light panel for cinematography. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |