Hi, my name is Kelly Wallace. I'm your product marketing engineer, working with TI's LED driver portfolio. This is one of two videos in a series on LED dimming methods. In this video, I will cover the most basic dimming method-- analog dimming. 

The brightness of the light output of an LED is determined by the average forward LED current. Therefore, we use a constant current regulator to drive LEDs instead of a constant voltage regulator. Typically, an approximately linear relationship exists over a range of the forward LED current, although non-linearity will appear as the forward LED current increases, resulting in a less efficient solution. 

Not all LEDs behave the same, so there will be variations in the linear range based off of the LED selection. Linearity can also vary with temperature. For applications that require adjusting the LED brightness, there are many ways to dynamically control the brightness of an LED, including analog dimming or a pulse width modulation dimming. In this video, we will cover basic implementation methods, as well as the advantages and limitations of analog dimming. 

Analog dimming is the adjustment of the average continuous LED current. This method of dimming typically has the lowest dimming ratio among standard dimming techniques. The dimming ratio can vary from 10 to 1 to up to 250 to one. The performance of the dimming ratio varies based off of the control topology of the LED driver, as well as how you implement analog dimming. For dimming ratios in the thousands to one range, typically you need to use a pulse width modulation dimming method or a combination of analog and pulse width modulation. 

Some customers prefer to stay in the linear dimming range, primarily because of the simplicity of having a linear relationship between the output current and light, and it can also be more efficient. However, it is very common to keep using analog dimming outside of the linear range. The key concern during deep analog dimming is making sure the transfer function is monotonic even in the nonlinear range. A non-monotonic light output means there are unwanted fluctuations, like I've just drawn on this image, where as you dim the LED driver lower, the light output actually increases. 

Analog dimming is typically the simplest dimming method to implement. To implement it, you simply adjust the analog voltage on the current adjust pin on an LED driver, which is used as a reference to adjust the output LED current. Depending upon the specific LED driver, you can adjust the analog voltage through tying the IADJ pin to VCC, using a resistor divider or even using a signal from a microcontroller that is converted into an analog voltage through a filter. 

As mentioned earlier, the voltage on the current adjust pin is used to adjust the inductor current. For non-synchronous devices, the linear dimming range is limited by the inductor current peak to peak ripple. When the device enters discontinuous conduction mode, or DCM operation, it becomes non-linear. The example on the screen shows the behavior of the TPS92515, which is a hysteretic device with peak current control. 

On this part, VIADJ is used to adjust the peak inductor current. The average output current equals the peak conductor current minus half the peak to peak inductor ripple. This is true as long as the device is in continuous conduction mode. However, a non-synchronous part cannot have negative current, and as you decrease the peak inductor current and the valley current meets the x-axis the inductor current becomes zero and the average output current is no longer the peak current minus half the ripple, and therefore enters nonlinear dimming. This does not apply to synchronously rectified parts, because they can have negative current and do not enter DCM. 

Some customers don't mind the efficiency hit of non-linearity and use the dimming functionality outside of the linear dimming range. This image shows an ideal waveform in DCM mode. However, in some LED drivers, this condition is where monotonic behavior can creep in and cause unwanted variations in the LED current that show up in the light output. 

Some customers don't mind the efficiency hit of non-linearity and use dimming functionality outside of the linear dimming range. This image shows an ideal waveform in DCM mode. However, in some LED drivers, this condition is where monotonic behavior can creep in and cause unwanted variations in the LED current that show up in the light output. 

If you need to stay in the linear range, there are methods to extend the linear range. One strategy is to decrease the inductor current ripple in order to delay entering DCM mode by decreasing the peak inductor current. If the system is being digitally controlled, the applied IADJ pin voltage can be adjusted when DCM occurs. However, there is a practical lower limit to the IADJ pin voltage due to circuit non-idealities. For example, in the TPS92515, using a VIADJ equals 0.5 volts, results in a sense voltage of 50 microvolts, which does not allow accurate operation. 

For devices like the TPS92692 or the LM3424 that use a closed loop topology, such as peak current mode control, the implementation of analog dimming is very similar. As with hysteretic devices, non-linearity occurs when the device enters discontinuous conduction mode. Similarly, reducing the inductor ripple through increasing the inductor value or increasing the switching frequency will still help to increase the linear dimming range by delaying the moment when the LED driver enters DCM mode. 

The main difference is typically a closed loop LED driver will have a slower response time to changes in the LED current than a hysteretic topology because of the feedback loop. Closed loop topologies also typically have a slower transient response, so it's also important to keep a closer eye on the LED current over- or undershoot when you make changes. This can be mitigated through the optimization of the compensation network of the LED driver. More information on optimizing compensation networks can be found in the LED driver data sheet. 

One of the main reasons why a designer may not choose an analog dimming method is the potential for color shifting when trying to achieve a high dimming ratio. Color temperature is the metric describing the color of the LED and is quantified in LED data sheets in kelvins. As seen in the table, byte LEDs, for example, can output a range of warm to cool light. The color temperature can shift with variances in forward current junction temperature and age. Analog dimming can result in shifting the color temperature due to the manipulation of the forward LED current. In applications where LED color is critical, the dimming range could be limited to prevent color shift. 

To summarize today's video, analog dimming adjusts the average LED current through an analog voltage input. It is often the simplest and most cost-effective method to implement, and can typically provide a linear dimming ratio in the range of 10 to 1 to 250 to 1, which is sufficient for many applications. There are methods to extend the linear dimming range by design. However, if much higher dimming ratios are needed, PWM dimming methods are an alternative. 

That's all for today. Please check out the rest of the dimming method series for our similar training on PWM dimming. Thank you for watching.