Wiring LEDs with two different voltages

Thread Starter

gattu_marrudu

Joined May 20, 2021
20
Hello,
I'm an electronics newbie still grappling with the basics of circuitry, so bear with me.

I'm building a circuit with some high-power LEDs. Following some instructions online, I built a very simple driver and wired 6 daisy-chained red, green and blue LEDs to it. I gave it a 18V supply with two 9V batteries.

What I did not realize is that while the green and blue LEDs have a voltage rating of 3.2-3.4V, so 6 of them would have added up to 19.2-21.6V, the red ones have a much lower rating of 1.9-2.3V. Of course, one of them burned out after a minute of testing.

I have thought about 2 options to resolve this:

1. Putting a resistor between the + supply and the low-voltage LEDs, or move them to the negative side of the chain (so the more powerful ones act as resistors) -- I'm not sure how this would work out though.

2. Building two or more parallel circuits, each with a series of LEDs of the same voltage rating. For example, I calculated that three 3.2-3.4V LEDs would need 9.6-10.2V, and five 1.9-2.3V LEDs would come up to 9.5-11.5V. So applying a 9.6-10.2V supply would be in range for both circuits (drawing twice the current). Or splitting up more parallel circuits to use 3-4 D batteries.

Are my thoughts correct? Is there a better way to do this?

Thanks in advance for the help.
 

AnalogKid

Joined Aug 1, 2013
9,248
An LED, like a small signal or rectifier diode, is stupid. Technically, it is a glutton. It will attempt to suck down all of the current in the universe until it dies. This is why all LED drive circuits incorporate some form of current limiting. For a string of LEDs with a fixed voltage source, this can be as simple as one resistor. You have seen what happens without it.

Check the datasheet for your green LEDs. 3.2-3.4 V seems high. Normal little green LEDs come in at around 1.9-2.1 V, a little bit higher than reds.

Some LEDs and LED strips have current limiting built-in. This method is not for those parts.

Start with the voltage source value. Add up the LED forward voltages (Vf) in the string, and subtract this from the source voltage. The remainder is the voltage across the current-limiting resistor. Note: you cannot rely on the LED Vf total being equal to the source voltage to protect the LEDs. ALWAYS have a current limiter.

Next, determine how much current you want running through the string from the datasheets.

Using Ohm's Law, calculate the resistor value. Pick the next larger standard resistor value.

Using Watt's Law or Joule's Law, calculate the resistor power dissipation. Use one rated for at least twice that.

ak
 
Last edited:

Thread Starter

gattu_marrudu

Joined May 20, 2021
20
Thanks for the quick replies.

This is the circuit I built to manage the power source: https://content.instructables.com/O...it=bounds&md=31a868919f482358986f2dac929fe585 from https://www.instructables.com/Circuits-for-using-High-Power-LED-s ("Constant current source #1") which seemed more energy-efficient and reliable than a simple resistor.

It turns out that my wiring in the filter was incorrect. That may well be the cause of the burnout.

While I fix that, my main question remains: is it OK to mix and match voltages in a LED series, and calculate the total as @AnalogKid suggested, or shall I keep them in parallel with different voltage inputs?

FYI This is the red LED I got: https://www.allelectronics.com/item/led-250/3w-red-led-with-star-heat-sink/1.html
And the green one (the blue one is electrically identical): https://www.allelectronics.com/item/led-252/3w-green-led-with-star-heat-sink/1.html
 

dl324

Joined Mar 30, 2015
12,871
my main question remains: is it OK to mix and match voltages in a LED series, and calculate the total as @AnalogKid suggested, or shall I keep them in parallel with different voltage inputs?
It depends.

If you want to operate all of the LEDs at the same current, and you have a sufficiently high voltage, putting them in series will be the most efficient method.

If you want to operate them at different currents to match brightness, you'd need to have a separate branch for each current.

A schematic would likely avoid much of guessing required of us.
 

SamR

Joined Mar 19, 2019
3,401
LEDs are current devices that require limited/controlled current. Yes, like any diode there is a voltage drop across them and their PDF should be consulted to find what that drop is. So, think current, not voltage, when using LEDs.
 

dl324

Joined Mar 30, 2015
12,871
Attached is the schematic of my circuit.
1621531395087.png
Your design is marginal.

If the LED voltages were at the high end of their range, you'd need about 19V for the circuit to function as intended. If they're on the low end, you'd have a little head room (about a diode drop). The transistor needs to be an NPN, and the 100k resistor seems a bit large.

You mentioned that you have 2 9V batteries. Are these 9V batteries "transistor radio" types?
 

Audioguru again

Joined Oct 21, 2019
3,173
Check the datasheet for your green LEDs. 3.2-3.4 V seems high. Normal little green LEDs come in at around 1.9-2.1 V, a little bit higher than reds.
Dim very old green LEDs used almost the same chemistry as red LEDs so their voltage was low. But modern green LEDs are much brighter because they use almost the same chemistry as blue and white LEDs and have almost the same 3.2V-3.4V.
The LED sales sheet shows it.

Gattu said that he built constant current sources but the schematic and him did not say how much current or what is the important resistor value or transistor part numbers. The article said 0.75 ohms and an ordinary 0.7V or more transistor so the current might be too high at 0.7V or more /0.75 ohms= 933mA or more instead of the 700mA max allowed current.
 

dl324

Joined Mar 30, 2015
12,871
The batteries are a rechargeable Ni-MH type, the ones used commonly in smoke detectors, etc.
Those batteries are usually rated for about 300mAH. They're not going to last long and they might not like having a 700mA load on them.

BTW, with a 0.75 ohm current sense resistor, current is going to be more like 1A.
 

Thread Starter

gattu_marrudu

Joined May 20, 2021
20
I should also point out that those LEDs need to be attached to a heat sink. Are you doing that?

Bob
Thanks for pointing that out. I suspected that the built-in heat sinks would only suffice for short tests, but they may need extra dissipation for >1h operation.
 

Audioguru again

Joined Oct 21, 2019
3,173
The new schematic wrongly says PNP transistor with no part number but instead an NPN is needed.
Yeah, the same wrong resistor value was used as in the Instructable.

The math on the schematic is also wrong. A maximum of 18.2V for the LEDs plus 0.6V or more for the NPN transistor plus an unknows voltage for the Mosfet that has no part number for a total of more than 18.8V powered by 18V.
 

Thread Starter

gattu_marrudu

Joined May 20, 2021
20
Those batteries are usually rated for about 300mAH. They're not going to last long and they might not like having a 700mA load on them.
Ha. Good point. My circuit needs to be portable. Maybe I can use a parallel design and use 4 D batteries? I understand that I would have to change the LED number & voltage and the resistor.
 

Audioguru again

Joined Oct 21, 2019
3,173
PP3 size 9V Ni-MH batteries are 8.4V when fully charged and drop dead in a few minutes with the very high current of those LEDs.
Some D size battery cells have a little AA or C cell inside.
 

AnalogKid

Joined Aug 1, 2013
9,248
which seemed more energy-efficient and reliable than a simple resistor.
Nope. 0 for 2.

1. For the same LED current, the power dissipation in the components of *any* linear current limiter is exactly the same. It might be spread around in multiple parts, but if the total power in the LEDs is the same, then the total power in the limiter/regulator/driver/whatever is the same.

2. A circuit with 4 components, and two of them active devices, versus one resistor, is way less reliable. There are significant performance benefits of a constant-current regulator over a simple resistor, but reliability is not one of them. To be clear, both approaches can run for years with no problems if well-designed.

ak
 
Top