0 Members and 1 Guest are viewing this topic.

#### David

• Frequent Contributor
• Posts: 279
• Country:
« on: January 22, 2010, 01:40:14 pm »
Whilst drinking coffee and scratching my head over some ideas for my latest project, I thought it could be useful to have a thread for small questions that don't deserve there own topic or for questions that people feel too silly to ask

I shall start off with something that has been confusing me. Whilst looking at several schematics (For Altera FPGA development boards) I have noticed that 330 ohm resistor values have been used in series with LED's running from 3.3V. I don't understand how the LED's are working off such a low current (3.3V - 1.85V = 1.45V/330 = ~5mA). For a typical LED with a forward voltage of 1.85V and forward current of 20mA, surely a resistor around the 100 ohm mark is required? The only conclusion I can come to is that they are using super low current consumption LED's?

Dave
« Last Edit: January 22, 2010, 05:43:39 pm by David »
David
(United Kingdom)

#### Andrew

• Guest
« Reply #1 on: January 22, 2010, 05:32:53 pm »
Mistake 1: Forward current 20 mA. 20 mA is just what is often used by LED manufacturers to specify some characteristics. But it is not the current one needs to operate an LED at.  There should be a diagram in the LED datasheets of every self-respeciting LED manufacturer, showing relative luminous intensity over forward current. You'll notice the diagram doesn't just have a single dot at 20 mA, but shows quiet a range.

Mistake 2: Forward voltage 1,85 V. Forward voltage is not fixed. It is a function of forward current (if you drive the LED with current). Or you can also say forward current is a function of forward voltage (if you (try to) drive with a constant voltage).

So let's pick some realistic values. Lets say you want to drive the LED with 5 mA. You find in the datasheet that its forward voltage at 5 mA is approx. 1.5 V. So you need a resistor of approx. (3.3 V - 1.5 V) / 5 mA = 360 ?. Hu, pretty close to the 330 ? you found :-)

But there is a problem. With that 360 ? resistor the LED wouldn't run at exactly 5 mA. LEDs come with some manufacturing tolerance and temperature sensitivity. The forward voltage over forward current function varies widely. So the above procedure, picking the forward voltage out of the diagram for a given forward current and then using that to calculate the resistor, is a rough approximation only.

What that 360 ? (or 330 ?) resistor ensures is that the current can never exceed 9 mA (10 mA respectively). And that's how you usually dimension such resistors. Decide what current you don't want to exceed. The 330 ? engineers apparently decided 10 mA is a nice value (and it is). 10 mA = 3.3 V / 330 ?. Then you let the non-linear system (the diode makes it one), find itself an operating point within that absolute limit. Because that's good enough for the purpose.
« Last Edit: January 22, 2010, 05:39:54 pm by Andrew »

#### armandas

• Frequent Contributor
• Posts: 336
• Country:
« Reply #2 on: January 22, 2010, 05:40:28 pm »
Mistake 1: 1.45V / 330R = 4.4mA, not 0.5mA

#### David

• Frequent Contributor
• Posts: 279
• Country:
« Reply #3 on: January 22, 2010, 05:43:54 pm »
Mistake 1: 1.45V / 330R = 4.4mA, not 0.5mA

Ah yes, I meant 5mA (Rounded up).
David
(United Kingdom)

#### Machina

• Guest
« Reply #4 on: January 23, 2010, 11:37:34 pm »
Okay here's one that's bugged me.

My Rigol DG1022 function generator can generate a sine wave up to 20 MHz. If I set the Vpp (Voltage peak to peak) to 5 volts, 1 Khz and put a scope on it, I read about 5 volts Peak to peak. As expected.

However if I jump the frequency to 20 MHz, the Vpp according to the oscilloscope jumps to 6 volts Vpp. This was verified with a Rigol 1204B Scope, a Tektronix TDS220 and a Tektronix 2465B ( perhaps the finest analog scope ever put out). If I set the frequency value between these two extremes, I still can see a voltage rise between 5 volts and 6 volts.

I might say that my Rigol FG is to blame, but my venerable Tektronix 280CFG (11 MHz function generator) does the same thing, when jumping from 1 MHz to 11 Mhz. So it seems to be a common problem for function generators.

Any one have a electronics explanation?

#### EEVblog

• Posts: 38135
• Country:
« Reply #5 on: January 24, 2010, 12:34:43 am »
Mistake 1: Forward current 20 mA. 20 mA is just what is often used by LED manufacturers to specify some characteristics. But it is not the current one needs to operate an LED at.  There should be a diagram in the LED datasheets of every self-respeciting LED manufacturer, showing relative luminous intensity over forward current. You'll notice the diagram doesn't just have a single dot at 20 mA, but shows quiet a range.

In fact the LED luminosity is usually fairly linear with current. So at 5mA any LED should have roughly 1/4 the lumens as it does at 20mA. Nothing special about these LED's at all, most LED's will be the same.

The 20mA figure is a classic beginners assumption, in most apps you don't need to run anywhere near 20mA. 5mA is about the figure I normally run with. Only if I'm after the absolute brightest thing possible will I consider going over 10mA.

But I'm an electron miser, I don't like throwing away current :->

Dave.

#### qno

• Frequent Contributor
• Posts: 422
• Country:
« Reply #6 on: January 24, 2010, 09:27:43 am »
Remember from the Photo voltaic cell blog that the light sensitivity of the eye is non lineair!
Why spend money I don't have on things I don't need to impress people I don't like?

#### jahonen

• Super Contributor
• Posts: 1055
• Country:
« Reply #7 on: January 24, 2010, 12:07:44 pm »
Okay here's one that's bugged me.

My Rigol DG1022 function generator can generate a sine wave up to 20 MHz. If I set the Vpp (Voltage peak to peak) to 5 volts, 1 Khz and put a scope on it, I read about 5 volts Peak to peak. As expected.

However if I jump the frequency to 20 MHz, the Vpp according to the oscilloscope jumps to 6 volts Vpp. This was verified with a Rigol 1204B Scope, a Tektronix TDS220 and a Tektronix 2465B ( perhaps the finest analog scope ever put out). If I set the frequency value between these two extremes, I still can see a voltage rise between 5 volts and 6 volts.

I might say that my Rigol FG is to blame, but my venerable Tektronix 280CFG (11 MHz function generator) does the same thing, when jumping from 1 MHz to 11 Mhz. So it seems to be a common problem for function generators.

Any one have a electronics explanation?

Did you use decent coaxial cable with inline 50 ? terminator at the scope end? If you are using common 1:10 probe especially with the long ground clip, I'm not at all surprised. Tektronix AFG3101 function generator at my work goes quite flat upto 100 MHz (the maximum), only with very slight amplitude variation along the way using that terminated setup.

Regards,
Janne

#### Machina

• Guest
« Reply #8 on: January 24, 2010, 08:15:24 pm »
100 percent correct Janne! Adding a ZL of 50 ohms took care of the problem. I thought it might be the reflected wave, however I thought it would show up as a wave distortion. In retrospect its obvious the reflected wave is so quick, it simply attenuates the signal.

Thanks again.

#### Zero999

• Super Contributor
• Posts: 19717
• Country:
• 0999
« Reply #9 on: January 25, 2010, 02:10:37 pm »
Doesn't the signal generator have a 50? source impedance to prevent problems like this?

Still I suppose if the impedance of the load varies with frequency, the amplitude will vary. If the 'scope input is 1M then adding a 50R load should solve the problem but the voltage will drop to 2.5V because it's a voltage divider.

Regarding the LED

For 3.3V operation, I'd use anything between 47R and 330R depending on the required current and forward voltage but for a standard red LED I'd go for 150R for a current of about 10mA.

Here's a challenge: how would you drive a blue LED (typical forward voltage of 3.5V) from a 3V MCU (no other power supply other than 3V) with only using two additional components?

#### Zero999

• Super Contributor
• Posts: 19717
• Country:
• 0999
« Reply #10 on: January 27, 2010, 01:53:56 pm »
No one took on my challenge?

You could of course use an inductor but capacitors are cheaper and more compact.

It might also be a good idea to use a Schottky diode because it will work at lower voltages.

EDIT:
Corrected text on schematic.
« Last Edit: January 27, 2010, 04:21:51 pm by Hero999 »

Smf