2
\$\begingroup\$

For example: I have two led driver chips: bcr321u and RT8474AGSP

RT8474AGSP

bcr321u

And think they both support to same output voltage and same output current etc. My question is about topological differences of them. BCR321U is definetely looks simpler and it is cheaper.

So, why do I need to use RT8474AGSP? When you consider, they both need much voltage supply and they both need an adjusted resistor for proper current flow, but RT8474AGSP configuration takes more area and so more and more expensive. In addition to that, this configuration is also an option:

OPAMP led driver

If we were constrained by supply voltage, we would use boost led driver chips however, they all are buck structured so why we should chose RT8474AGSP over bcr321u and bcr321u over this simple opamp - npn structure? Where does the value of RT8474AGSP lies?

\$\endgroup\$
1
\$\begingroup\$

Another good reason is the driver's efficiency: -

enter image description here

It's a current limited switching regulator there fore there is hardly any power lost in current limiting resistors or transistors unlike the other three circutis shown in the question. Here's ahow the data sheet describes the chip: -

With the internal 500kHz operating frequency, the size of the external PWM inductor and input/output capacitors can be minimized. High efficiency is achieved by a 100mV current sensing control.

\$\endgroup\$
  • \$\begingroup\$ Yes, looks like efficienct is the main thing that LED Drvr chips provide. But if you know the specifications of leds and the quantity, you may elegate the efficiency problem.If you have 5 leds and you know that those leds(in series) need 9.9V for maximum 100mA current, you can supply 10V and drive drive leds. But this makes it highly unflexible. Thanks. \$\endgroup\$ – Alper91 Dec 27 '15 at 11:39
1
\$\begingroup\$

Just to add to this.

The main advantage of the RT8474AGSP, as already mentioned, is the efficiency gained due to a PWM controlled current source.

The BCR321U & the OPAMP solution drop the excess voltage in a pass transistor (inefficient) but as the original poster pointed out "it uses less parts" ... that may be the case but that power that the BJT's are dissipating needs to go somewhere either into an additional heatsink or into over sized copper planes. This also limits the operating ambient such linear solution can operate at.

There isn't really a right and wrong solution (just bad implementations). The two methods both provide means to vary the intensity & are naturally tolerant of supply perturbations (to a certain degree) & a range of supply voltages.

As the RT8474AGSP is a switching topology you may find the EMC concerns on the supply side, if poorly managed, are a real concern. Likewise the additional area for heatsinking and/or operating ambient is a concern which would rule out the linear method. Or the input voltage may be quite wide that rules out the linear solution.

\$\endgroup\$
0
\$\begingroup\$

One good reason is listed on the schematic.

enter image description here

The dimming is achieved by pulse width modulation internally and results in low power losses.

\$\endgroup\$
  • \$\begingroup\$ I also do dimming on BCR321U. They all have the dimming option. Maybe analog is an extra for the RT8474AGSP, but I think in most of the applications pwm dimming is used over analog. I used all of them, all of in 120 Hz dimming(well actually dimming was not the idea, I was trying to strobe backlight in specific time with specific duty cycle but brightnes can be controlled). \$\endgroup\$ – Alper91 Dec 27 '15 at 11:04
  • \$\begingroup\$ This one appears to be dimmable with just a pot. The others need an additional PWM generator. I imagine the switcher will also give consistent brightness over a wide range of input voltages. \$\endgroup\$ – Transistor Dec 27 '15 at 11:11
  • \$\begingroup\$ No, all of them do dimming with PWM. Just, RT8474AGSP also has analog option. Meaning from scale to 0.8V and 2.8V(I am throwing out for example) you may control the current via applying to CTL pin. Other than that, all of them can be adjustable via pot and all of them can do dimming with applying PWM(with adjusted duty Cycle) on initial pins. Main reason must be about efficiency as Andy mentioned. \$\endgroup\$ – Alper91 Dec 27 '15 at 11:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.