0
\$\begingroup\$

A TLC5940 can sink up to 120ma per channel. I am having difficulty finding the max current for the entire chip on the Datasheet. Is there one, or is there another limiting factor for this chip?

Additional Hypothetical Question: If I try to drive 10 LEDs (each with a voltage drop of 3.2) in series, with 38 Volts, will the voltage be dropped to 4 volts before the current enters the chip?

Also, I neglected to mention that my LED's are 4 prong RGB LEDs. I will be unable to wire them all in series. Them must be wired in parallel.

***Personal Tests ==============

I have been able to power over 50 LED using this chip without an issue. to do this, you must make sure that each group of series LEDs on a channel (The chip can control 16 groups via 16 channels via 16 pins) drops as much voltage as possible (between the positive connection and the chip.) For example, if you are using a 9 volt power supply, and LEDs with a 2 volt drop, each channel should have 4 LEDS in series, plus some resistance to drop most of the remaining 1 volt. A good way to find the best value for the resistor is to build your LED circuit with a potentiometer, and use it find the resistance value at which the LEDs start to dim.

Also, if you want to use a TLC594x chip, read this article about additional issued which you need to be aware of:

Test Update ==============

I am testing this chip with 8 different LED configurations:

A) One LED only

B) One LED in series with a 10ohm resistor

C) One LED in series with a 220ohm resistor

D) Three LEDs in series

E) Three LEDs + 10ohm resistor in series

F) Three LEDs + 220ohm resistor in series

G) Three LEDs in parallel, each in series with a 10ohm resistor

H) Three LEDs in parallel, each in series with a 220ohm resistor

Note: brightness is given in ranked order, higher numbers are brighter

Source set to 6.7 ma. Chip slightly warm to the touch

A) volts dropped: 2v[LED]0v
Bright: 2

B) volts dropped: 2v[10ohm]1.9v[LED]0v
Bright: 2

C) volts dropped: 3.4v[220ohm]1.9v[LED]0v
Bright: 2

D) volts dropped: 5.7v[LED]3.9v[LED]2v[LED]0v
Bright: 2

E) volts dropped: 5.7v[10ohm]5.6v[LED]3.8v[LED]1.9v[LED]0v
Bright: 2

F) volts dropped: 7v[220ohm]5.5v[LED]3.7v[LED]1.9v[LED]0v
Bright: 2

G) volts dropped: 1.7v[10ohm]1.7v[LED]0v
Bright: 1

H) volts dropped: 2.2v[220ohm]1.7v[LED]0v
Bright: 1

Source set to 11.3 ma. Chip comfortably hot to the touch

A) volts dropped: 2v[LED]0v
Bright: 2

B) volts dropped: 2.1v[10ohm]2v[LED]0v
Bright: 2

C) volts dropped: 4.5v[220ohm]2v[LED]0v
Bright: 2

D) volts dropped: 6v[LED]4v[LED]2v[LED]0v
Bright: 2

E) volts dropped: 5.9v[10ohm]5.8v[LED]3.9v[LED]2v[LED]0v
Bright: 2

F) volts dropped: 8.3v[220ohm]5.8v[LED]3.9v[LED]2v[LED]0v
Bright: 2

G) volts dropped: 1.8v[10ohm]1.7v[LED]0v
Bright: 1

H) volts dropped: 2.6v[220ohm]1.7v[LED]0v
Bright: 1

Source set to 16.9 ma. Chip almost painfully hot to the touch

A) volts dropped: 2.1v[LED]0v
Bright: 2

B) volts dropped: 2.3v[10ohm]2.1v[LED]0v
Bright: 2

C) volts dropped: 5.8v[220ohm]2.2v[LED]0v
Bright: 2

D) volts dropped: 6.2v[LED]4.2v[LED]2.1v[LED]0v
Bright: 2

E) volts dropped: 6.2v[10ohm]6.1v[LED]4.1v[LED]2.1v[LED]0v
Bright: 2

F) volts dropped: 9.1v[220ohm]6.1v[LED]4.1v[LED]2.1v[LED]0v
Bright: 2

G) volts dropped: 1.8v[10ohm]1.7v[LED]0v
Bright: 1

H) volts dropped: 3v[220ohm]1.8v[LED]0v
Bright: 1

Source set to 21.1 ma. Chip touchable but too hot for sustained contact

A) volts dropped: 2.3v[LED]0v
Bright: 2

B) volts dropped: 2.5v[10ohm]2.3v[LED]0v
Bright: 2

C) volts dropped: 6.8v[220ohm]2.3v[LED]0v
Bright: 2

D) volts dropped: 6.4v[LED]4.4v[LED]2.3v[LED]0v
Bright: 2

E) volts dropped: 6.6v[10ohm]6.3v[LED]4.2v[LED]2.3v[LED]0v
Bright: 2

F) volts dropped: 9.2v[220ohm]6.0v[LED]4.1v[LED]2.2v[LED]0v
Bright: 2

G) volts dropped: 1.9v[10ohm]1.8v[LED]0v
Bright: 1

H) volts dropped: 3.4v[220ohm]1.8v[LED]0v
Bright: 1

Additional test 1:

source set to 21ma

9.4 volt max supply

PWM set to max

8 channels used

4 LEDs in series per channel

no additional resistance

voltage drop per channel: 7.8 volts

32 LEDs total

total current 150ma

Chip is only slightly warm

Note: 80 ohms can be added in series to a channel without lowering the brightness.

\$\endgroup\$
  • \$\begingroup\$ Link to datasheet? Your circuit schematic so we know how you're connecting them? \$\endgroup\$ – Transistor Jul 7 '16 at 21:35
  • \$\begingroup\$ Your schematic is wrong. The leds are backwards. And your much better off with a higher voltage and leds in series. The IC is better for that than parallel. \$\endgroup\$ – Passerby Jul 7 '16 at 21:52
  • \$\begingroup\$ Why is a higher voltage and leds in series better? \$\endgroup\$ – Hoytman Jul 7 '16 at 22:15
  • \$\begingroup\$ Because then you don't have to worry about inevitable minor structural differences between the LEDs, even within the same batch. They're all forced to draw the same current. \$\endgroup\$ – Ignacio Vazquez-Abrams Jul 7 '16 at 22:21
  • \$\begingroup\$ The limiting factor is how much heat you can generate in the chip before the silicon starts breaking down. \$\endgroup\$ – Ignacio Vazquez-Abrams Jul 7 '16 at 22:26
3
\$\begingroup\$

Sinking

enter image description here

Figure 1. Section 9.2 shows TL5940 sinking LED current.

Sinking means that the chip will connect the negative leg or cathode of the LED to ground. This is the opposite of your schematic. Figure 1 shows the LED anodes connected to a positive supply, \$ V_{(LED)} \$.

Minimising current

Provided \$ V_{(LED)} \$ is high enough, LEDs can be series connected to have the same current power multiple LEDs. This will reduce power dissipation in the chip. The chip can work with \$ V_{(LED)} \$ up to 17 V (Section 1, Features).

Maximum voltage

Pay attention to Section 10 Power Supply Recommendations for the \$V_{LED} \$ voltage.

[OP's comment] If I try to drive 10 LEDs (each with a voltage drop of 3.2) in series, with 38 Volts, will the voltage be dropped to 4 volts before the current enters the chip?

No. While it seems it might be true, consider what happens when the output transistor turns off. The full 38 V will be fed through the LEDs to the transitor collector but it is rated at 17 V max. The transistor will break down due to electric stress.

Current control

The chip also has current control which allows you to set the maximum channel current (Section 8.3.7). By adding \$R_{(IREF)} \$ the you get

$$ I_{max} = \frac {1.24}{R_{(IREF)}} \times 31.5 $$

You don't need the current limiting resistors. If you insist on paralleling LEDs or strings of LEDs you might want to add a low value resistor to drop maybe 0.2 V or so at the individual LED current to balance out the currents a bit.

Power calculation

Run the power calculation given in section 11.3. The max power dissipation will be based on ideal heatsink mounting. You'll need to derate this based on your ability to match this. Anyway, remember that those are best case.

How many LED's can a TLC5940 16-Channel LED Driver support (without additional driver chips or transistors)?

That depends on LED current, supply voltage and series / parallel connection.

Can this chip actually supply 120ma to all 16 channels at the same time?

The datasheet says so. See Power calculation above.

Additional: My current is intended to enforce uniformity across all LEDs. I am aware that the chip can regulate current, but I doubt that each channel will have the same number of LEDs.

You will have a problem here. The chip will try to drive the same current through each channel as determined by Current Control above.

Does an LED's voltage drop vanish when current stops flowing?

enter image description here

Figure 2. LED forward and reverse voltage curve. Source: Wikipedia.

Yes. The relationship between voltage drop and current is non-linear but when you get to very low currents the voltage drop falls off rapidly. At zero current there will be zero volts. Similarly when the output transistor switches off the voltage on the collector will be \$ V_{LED} \$. If it was any lower then current would flow through the LEDs.


Constant current and PWM

It appears that this chip uses both constant current and PWM to control the LED brightness. The constant current control (discussed above) sets the maximum current the LEDs will ever see. The PWM controls the brightness. Let's look at the current control and the thermal problem it causes.

schematic

simulate this circuit – Schematic created using CircuitLab

Figure 3. \$ V_{LED} \$ set at 6 V. \$ I_{OUT} \$ set at 50 mA. (a) Driving two green LEDs. (b) Driving one green LED.

  • In Figure 3a we can calculate that at 100% PWM the current sink \$I_0\$ will adjust it's "resistance" to sink 50 mA from the \$ V_{LED} \$ supply. Let's say the green LEDs drop 2.2 V at 50 mA then we can see that OUT_0 will be at 1.6 V. We can calculate the power dissipated in the chip as \$ P = VI = 1.6 \times 50m = 80~mW\$.

  • If we examine the case in 3b we find \$ P = 3.8 \times 50m = 190~mW \$. Driving one LED makes the IC heat more than twice as much as it would for two LEDs! (for the voltage I chose).

It should now be clear that the selection of the \$ V_{LED} \$ value is critical to maximising the number of LEDs which can be run. Looking at the power dissipation calculation formula in section 11.3

$$ P_C = (V_{CC} \times I_{CC}) + ( V_{OUT} \times I_{MAX} \times \frac {DC_n}{63} \times d_{PWM} \times N) $$

The first part is the power dissipated by the logic and you can't do much about that. (It's small anyway.) The second part tells us that power will be proportional to \$ V_{OUT} \$. This confirms our need to keep this value low.

Choosing \$ V_{LED} \$

Neatly tucked away in Table 4 we find the answer.

enter image description here

Figure 4. The calculation for \$ V_{LED} \$.

  1. Decide on an LED current.
  2. Look up the IC knee voltage for that. (See below.)
  3. Figure out your worst case LED forward voltage at the desired current.
  4. Figure out how many of these you can get in series with your chosen power supply and still have a little voltage for the IC knee voltage (step 2).
  5. Re-check your calculations and then set the current limit for the chip using the resistor calculated in section 8.3.7.

The knee voltage

enter image description here

Figure 5. Section 6.7, Figure 3 shows the knee voltage at various output currents. If we were setting for 120 mA output current we would need to allow 1 V for the voltage drop across the chip as indicated at (1). If \$ V{LED} is a volt higher we would end up at (2), etc. For minimum power keep just to the right of the knee at your chosen current._

Power dissipation ability

enter image description here

Figure 6. Section 6.7, Figure 2, shows the maximum power dissipation. I haven't studied this much but if you select a maximum temperature rise of 60°C relative to ambient then you can work out the maximum heat dissipation before thermal shutdown.

Again heatsinking to the PCB will be critical here.

RGB LEDs

schematic

simulate this circuit

Figure 7. Converting an RGB LED to a white LED.

If you still want to use your RGB LEDs you can use a couple of resistors on each to compensate for the different forward voltage drops.

$$ R_{RED} = \frac {(V_{blue} - V_{red})}{I_{red}} $$

and similarly for the green.

\$\endgroup\$
  • \$\begingroup\$ This my be way off, but If I try to drive 10 LEDs (each with a voltage drop of 3.2) in series, with 38 Volts, will the voltage be dropped to 4 volts before the current enters the chip? \$\endgroup\$ – Hoytman Jul 7 '16 at 22:27
  • \$\begingroup\$ See "Maximum voltage" update. Add the LED specs into your question. Please be careful deleting sections of your post. I've referenced your incorrect schematic which you have now deleted so it makes my answer look a bit nonsensical. I put a lot of effort into making them "sensical". It's OK for a question to evolve but leave the history there. \$\endgroup\$ – Transistor Jul 7 '16 at 22:36
  • \$\begingroup\$ Sorry, I deleted it before I saw your post. \$\endgroup\$ – Hoytman Jul 7 '16 at 22:38
  • \$\begingroup\$ Does an LED's voltage drop vanish when current stops flowing? \$\endgroup\$ – Hoytman Jul 7 '16 at 22:40
  • \$\begingroup\$ The absolute maximum current spec says 120mA. I think that's the max ground current allowed. So no, 2 Amps at once will not work here. \$\endgroup\$ – Passerby Jul 7 '16 at 22:42
1
\$\begingroup\$

...sink 120ma per channel, which should be able to support 6 LEDs at full brightness.

...

Additional: each LED will have its own 500ohm resistor, and none of the LED's will be wired in series with other LEDs

That is absolutely not how you want to do it. You want to put the LEDs in series and then supply a voltage high enough to light them, and none of the LEDs should have a series resistor. The IC is capable of current limiting on its own when IREF is properly connected.

If you need to run the LEDs with a higher voltage than the IC can handle then read SLVA280, "Using TLC5940 With Higher LED Supply Voltages and Series LEDs". Heck, read it anyways, because it will help you understand the IC just a bit better.

\$\endgroup\$
  • \$\begingroup\$ Thank you for the referecne to the document. That is very helpful. Please put comments about the question in the comment section \$\endgroup\$ – Hoytman Jul 7 '16 at 21:51
  • \$\begingroup\$ Yes, you commented on my circuit design, but you did not address the question: How many LED's can a TLC5940 16-Channel LED Driver power? \$\endgroup\$ – Hoytman Jul 7 '16 at 21:54
  • \$\begingroup\$ @Hoytman: As many as you can supply the voltage for. \$\endgroup\$ – Ignacio Vazquez-Abrams Jul 7 '16 at 21:55
  • \$\begingroup\$ I don't think I will need to use MOSFETs in this design, the chip alone can manage LED Power Supply Voltage up to 17 V. \$\endgroup\$ – Hoytman Jul 7 '16 at 21:59
  • \$\begingroup\$ Also, I neglected to mention that my LED's are 4 prong RGB LEDs. I will be unable to wire them all in series. Them must be wired in parallel. \$\endgroup\$ – Hoytman Jul 8 '16 at 2:24

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.