1
\$\begingroup\$

How would you consider connecting two different types of LEDs, a white (6 V @ 150mA) and red (2 V @ 350mA) in a series-parallel combination? (W W R W W R W W) << parallel multiple strings of 8 series connected LEDs.

Similar LEDs will share more or less the same characteristics within a series / parallel circuit using a constant current and a small current balancing circuit (mirror or similar).

If I use a CC driver for the LEDs, it will divide the current to each parral string for instance 150mA per string. But now I have a couple of 350mA LEDs within each series string. This will either cause the efficacy to be next to zero or they will not even light up. Driving all the strings at 350mA will probably create a small electrical fire.

Any insights?

\$\endgroup\$
  • 1
    \$\begingroup\$ It depends on your thermal and optical design goals. Will you have more cooling avail if you choose to drive at max rated current? or would you prefer to run all at the same <=150mA \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 14:06
  • 1
    \$\begingroup\$ You're doing this the wrong way around: you want two constant current drivers, one for a string containing a series of all your white, and one for a string containing a series of all your red LEDs. Basically, it's like you read an introduction to driving LEDs and then intentionally did the opposite! \$\endgroup\$ – Marcus Müller Jun 2 at 14:07
  • \$\begingroup\$ Alternatively, get some red LEDs with lower current requirements. 350mA is a lot! \$\endgroup\$ – Hearth Jun 2 at 14:11
  • \$\begingroup\$ Thanks for the feedback I do appreciate it. First off all the heat dissipation throughout the board can be assumed to be linear as I will use an 'infinite size heatsink' for the spice. The option of using a separate CC source will absolutely do the trick, I understand that yes. Let's assume I have one 2000mA source. A simple solution might be to add the R led's in a series chain before the parallel combinations of white only strings, maybe this could work. \$\endgroup\$ – Bettie Jun 2 at 14:17
  • \$\begingroup\$ simply don't. You gain nothing, and putting LEDs in parallel in a thermally stressful situation is a recipe for disaster; see thermal runaway in LEDs. \$\endgroup\$ – Marcus Müller Jun 2 at 14:22
3
\$\begingroup\$

LEDs with different forward current requirements can not be connected in series. Doing so will cause some LEDs to be damaged.

LEDs with different forward voltages can not be connected in parallel. Doing so will cause some LEDs to be dim or completely dark.

LEDs that have the same current requirement can be connected in a series string, with an appropriate resistor to limit current or using a constant-current source.

LED strings can be connected in parallel as long as they are all designed to operate at the same voltage. That is, the current limiting resistor in each string must have been chosen for operation with the same voltage. So, connecting several strings in parallel and driving them with a constant-current driver is not likely to work well...they are not guaranteed to share current equally.

\$\endgroup\$
  • \$\begingroup\$ edit 1st line req'd for using the min. rating. \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 14:07
  • \$\begingroup\$ LEDs with different forward current requirements CAN be connected in series using the smallest current rating \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 17:23
1
\$\begingroup\$

You can always use less than max rated current and the efficacy will improve.

You can also connect different current rated LEDs in series, but you need the know what the tradeoffs are for current sharing, from the Vf tolerances and not exceed the smallest rating, as well as have adequate heatsinks.

Given a preference to buy only 1 current regulator, 1 practical solution might be to create an array with 10% string drop voltage power resistors (Rs), equivalent to those tolerances in each string to normalize the voltage drop in each string for current sharing.

It is not as perfect as 2 independent current sources but may be close enough depending on your specs.

schematic

simulate this circuit – Schematic created using CircuitLab

A much better solution is to choose LEDs rated for the same current so that a CC controlled string that combines the right combination of Red to boost CRI or to customize your spectrum. https://www.samsung.com/led/lighting/high-power-leds/3535-leds/

\$\endgroup\$
  • \$\begingroup\$ I have been looking around for a nice circuit sim, I guess I will go with circuitlab. I have been using Altium, but in the same breath have been out of the scene for years. Playing catchup atm. The circuit above Is what I mentioned in one of my comments. As you mentioned 'It is not as perfect as 2 independent current sources but may be close enough depending on your specs.' I will also then add a current balancer LM3466 which is quite a usefull little black box. \$\endgroup\$ – Bettie Jun 2 at 14:42
  • \$\begingroup\$ I use Falstad for all my quick simulations, but need to customize Rs and breakdown voltage \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 14:44
  • \$\begingroup\$ The simulators are only as good as your assumptions and models \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 15:14
  • \$\begingroup\$ Totally agreed, garbage in garbage out. \$\endgroup\$ – Bettie Jun 2 at 16:09
  • \$\begingroup\$ With skill, you create a good model then U get a better understanding. The best tools are the ones you know how to use. \$\endgroup\$ – Sunnyskyguy EE75 Jun 2 at 16:16

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.