12V G4 LED disk -- only warm at half duty.

Just to confirm that disk feeling so hot the leds are likely to fail sooner at 12VDC (disks were made for 12VAC, have a rectifier) continuous. And I bought 10 of the things on sale!

12V 3W works out to 250mA draw, I have 200mA continuous rated 2N7000 MOSFETs with a view to turn the led disk on and off and on at 500Hz. It’s not continuous, the FETs take it fine, No Smoke (maybe next time!) just light.

I got that working and added a strobe task and then added a second disk so that when one is off, the other is on with 1 line of code changing both at once. That makes 250mA (likely less) 12V per pair of disks.

Wiring:
12V ground to ground rail.

Arduino ground to ground rail.
Arduino pins wire directly to 2N7000 gates. The gate is guaranteed full open at 3V.

Both 2N7000 source pins connect to the ground rail.
Each 2N7000 drain pin connects to one of the led disks.

The led disks have rectifiers, which pin gets power or ground makes no difference as long as they’re different.
The remaining led disk pins get 12V that will only flow when the FET opens the drain.

#include <Arduino.h>

const byte statusLedPin = 13; // pin B5
const byte statusLedToggleMask = 0x20;
byte *statusLedPinsReg = (byte *) &PINB;
unsigned long startStatusMs; // measuring millis
const unsigned long statusTimeMs = 1000UL;

const byte shareLedPin[ 2 ] =
{
  2, 3  // pins D2 & D3
};
const byte shareLedToggleMask = 0x0C;
byte *shareLedPinsReg = (byte *) &PIND;
unsigned long startShareUs; // measuring micros
const unsigned long shareTimeUs = 1000UL;

byte strobeState = 1;
unsigned long startStrobeMs; // measuring micros
const unsigned long strobeTimeMs[ 2 ] =
{
  100UL, 25UL
};



void setup()
{
  Serial.begin( 115200 );
  Serial.println( strobeTimeMs[ 0 ] );
  Serial.println( strobeTimeMs[ 1 ] );
  pinMode( statusLedPin, OUTPUT );
  pinMode( shareLedPin[0], OUTPUT );
  pinMode( shareLedPin[1], OUTPUT );
  digitalWrite( shareLedPin[1], HIGH );
}

void loop()
{
  if ( strobeState > 0 )
  {
    if ( micros() - startShareUs >= shareTimeUs )
    {
      *shareLedPinsReg = shareLedToggleMask;
      startShareUs += shareTimeUs;
    }
  }

  if ( millis() - startStrobeMs >= strobeTimeMs[ strobeState ] )
  {
    startStrobeMs += strobeTimeMs[ strobeState ];

    if ( strobeState > 0 )
    {
      digitalWrite( shareLedPin[0], LOW );
      digitalWrite( shareLedPin[1], LOW );
      //      *shareLedPinsReg = shareLedToggleMask;
      strobeState = 0;
    }
    else
    {
      digitalWrite( shareLedPin[0], HIGH );
      strobeState = 1;
      startShareUs = micros();
    }
  }

  if ( millis() - startStatusMs >= statusTimeMs )
  {
    *statusLedPinsReg = statusLedToggleMask;
    startStatusMs += statusTimeMs;
  }
}

Hi,

Usually those disks have an aluminium back to dissipate the heat, sometimes to a heatsink, depends on the physical size as to how much heat 3Watts will produce.

(Have you connected gnds?)
( Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png? )
Not a Fritzy.. lol

Thanks.. Tom.. :slight_smile: :slight_smile: :slight_smile:

Not set up for scanning pics and not conversant with the Ubuntu drawing apps.

1 Arduino
1 12V 1A PS
2 led disks
2 2N7000 MOSFETs
1 breadboard & assorted jumpers

Start with 12V PS and Arduino grounds connected on the ground rail of the breadboard.

12V goes to one leg each of the two led disks. The disk has a rectifier and 4 resistors on the back, which pin does not matter.

The other pin of each led disk goes to the DRAIN pin of one of the 2N7000's
The SOURCE pin of each of those connects to the ground rail.
The GATE pin of each goes to pin 2 or 3 of the Arduino. No resistor needed. Control w/o current, Uno max total is 200mA.

When the gate is at 0V there is no flow from Drain to Source. 12V has no flow, there is no light.
When the gate is at 3+V there is full flow through the led disk and down the drain to ground.
The Arduino does more work than turning leds on/off, those use Arduino current!

Does that clear doubts? Can you tell me which text is ambiguous?

This does work. Last time I gave a disk 12VDC it was finger-hot and likely shorter life. Now they're only warm.
I wrote something similar for a project here last year or two ago.

You can test the software simply by breadboarding two leds, one to pin 2 and the other to pin 3. They will dim and strobe.

Hi,
Is this your 3W G4 12V disk?


For its size, 3W would in my opinion get quite warm, Ohms Law can't be wrong.
Tom... :slight_smile:

Led discs? rectifier on it?

You might have COB lights with a PT4115 driver chip.
No need to use a mosfet then. Post a picture or link.
Leo..

Fandyfire at dealextreme

FandyFire G4 3W Lamp Bluish White 7000K 600lm 12-5050 LED (DC 12V)

Each of those 12 squares has 3 junctions.

250mA gets split 4 ways into 3 squares in serial, all get 60mA that spilts to 3 parallel leds 20mA each.

These will run off 12VAC or DC. I think that 12VAC is used for outdoor wiring in some places.

Four strings of three LEDs, a 51ohm resistor, and a bridge rectifier.
I calculate a total of ~0.62watt on a 12volt DC supply.

33.3volt drop across the LEDs = 9.9volt.
2
0.65volt = 1.3volt drop across the bridge rectifier.
12-9.9-1.3 = 0.8volt drop across the 51ohm CL resistor = 0.8/51 = ~15.7mA per string,
= ~63mA for the four strings = ~9.9volt (LEDs) * ~63mA = ~0.62 watt total.

Absolute max per 5050 LED (three LEDs in one) = ~3.3volt * 3*20mA = ~0.2watt
12 LEDs, so 12 * 0.2watt = 2.4watt. 3watt is a 25% lie.
That amount of power (2.4watt) can only be achieved with active current limiting and a decent heatsink.
Leo..

If you are in the UK Sainsburys sell excellent 12v GU5.3 lamps for £4. They claim to produce 350 lumens and they are certainly much brighter that other LED lamps that I paid more for. Recently I have seen similarly priced lamps in Tesco which look similar but I have not tried them.

It is easy to hacksaw the lens off the lamp and either use them in the same style but with a wider spread of light or take them apart and rearrange them like the lamp in the picture in Reply #3.

The great thing about them is that the lamp unit is physically separate from the driver electronics so the electronics is not constantly overheated.

...R

Wawa:
Four strings of three LEDs, a 51ohm resistor, and a bridge rectifier.
I calculate a total of ~0.62watt on a 12volt DC supply.

33.3volt drop across the LEDs = 9.9volt.
2
0.65volt = 1.3volt drop across the bridge rectifier.
12-9.9-1.3 = 0.8volt drop across the 51ohm CL resistor = 0.8/51 = ~15.7mA per string,
= ~63mA for the four strings = ~9.9volt (LEDs) * ~63mA = ~0.62 watt total.

Absolute max per 5050 LED (three LEDs in one) = ~3.3volt * 3*20mA = ~0.2watt
12 LEDs, so 12 * 0.2watt = 2.4watt. 3watt is a 25% lie.
That amount of power (2.4watt) can only be achieved with active current limiting and a decent heatsink.
Leo…

Is it 0.62 watt or 2.4 watt?

Holy cow, manufacturers exaggerate product claims! Who knew? Call The Post!

Whatever, the disk got hotter than I like so I blinked it.

GoForSmoke:
Is it 0.62 watt or 2.4 watt?

Whatever, the disk got hotter than I like so I blinked it.

It is 0.62watt on a 12volt DC supply, and 2.4 watt absolute max if you drive it with the right voltage/current.
These LEDs are clearly designed for 12volt AC (from an old halogen transformer).

COB (power) LEDs can get rather hot, which is ok.
Leo..

I really doubt that you have all your numbers right.

Watts is volts * current.
.62W / 12V = 51mA that gets split (4 * 3) ways to go through 36 junctions? 4mA per junction?
2.4W / 12V = 200mA, 16mA per junction.
Why would they use COB leds with these currents?

12V - rectifier .7V = 11.3V to go through 4 lines of a resistor and 3 5050's in serial with 3 parallel junctions each.

Not all 5050 leds are the same color, color makes forward voltage, are you sure these are 3.3V?
Even a small difference counts for more result. Could the rectifier be Shottkey, .3V drop?

If Fv is 3.3 then the resistors see (11.3V - 9.9V) / 51R = 27mA divided into 3 = 9mA ea.
If fV is 3V then 45mA divided into 3 = 15mA each. That is more in line with what I see.

And here's the kick, forward voltage drops with heat.
You can watch these disks brighten as they warm up.

It is possible to solder to the marked + and - spots above the rectifier and get no drop.
But I dunno if that's wise given how bright they are as it is.

These LEDs are clearly designed for 12volt AC (from an old halogen transformer).

They are replacements for halogen bulbs in those G4 units.

I'm sure I got the numbers right, and yes, the slightest voltage increase is more power.
To know the exact current, you could measure the volt drop across the 51ohm resistors.

Old halogen setups used plain 12volt mains transformers (~17volt peak).
Harder to calculate average LED current during the >11volt part of the sine wave.
But I expect they will be brighter with that 12volt AC.

The bridge is there because the lights are made for AC.
You will have a lot more brightness/current if you remove the bridge, and connect direcly to the LEDs.
Leo..

I'm running them on 50% duty cycle because of the heat, and they are still bright, throw a lot of not-soft light.

You can buy just one for $1.98. Shipping is free there.