LEDs in parallel don't consume the sum of individual currents?

[neoply]

Hello all. Thanks for taking up your time to read my post.

I checked the currents that passed through several LEDs for MIC2287CBD5 as the datasheet of http://www.kynix.com/uploadfiles/pdf2286/MIC2287CBD5.pdf at a particular voltage. They varied from 12.0 to 15.1 mA at 3.0 V. I then connected them all in parallel and the current running through them was about 22 mA. How can this be? I thought it would be the sum of the currents that passed through the LEDs connected individually. All the LEDs should have the same voltage across them so they should all be passing the same current they passed when tested individually ??

Can anyone help me ? I am very puzzled about this question. I do need your help.

Thanks a lot.

 

[willem2]

The pdf won't load from kynix. Is it the same as this one? http://www.mouser.com/ds/2/268/mic2287c-777801.pdf
This is a constant current source, not a 3V voltage source. You're supposed to attach the LEDs in series.

 

[Tom.G]

Both files loaded here and a File Compare reports "both files identical."
(In fact, Mouser redirects to to the Micron (kynix.com) site.)

 

[neoply]

thanks for your reply. Do you have better explanation to my question ? thanks a lot.

 

[rootone]

How many parallel LED's?. 2, or 2000?

 

[Tom.G]

Do you have better explanation to my question ?

You're supposed to attach the LEDs in series.

See the schematic on the first page of the data sheet.

It is normal for some variation in the voltage/current characteristics between different LEDs. Look at a datasheet for the LEDs you are doing and find the entry for Forward Voltage. A voltage spread of 25% as measured at the operating current is not uncommon.

 

[neoply]

thanks a lot.

 

[sophiecentaur]

thanks for your reply. Do you have better explanation to my question ? thanks a lot.

LEDs are not Ohmic so there is no simple way to predict current for a given supply voltage. If you want a string of series LEDs to receive. Particular current then you basically need a Current Source. That will do the job by producing the right supply volts.
There are many circuits that can do what you need.

 

[Tom.G]

i can find where to buy MIC2287CBD5 ，thank you for your pdf

Mouser has the MIC2287CYD5. The 'CY' version does not have Lead on the pins. The 'CB' probably has Tin/Lead solder on the pins. They would work the same.

https://www.mouser.com/Search/Refine.aspx?Keyword=mic2287c

 

[rbelli1]

They varied from 12.0 to 15.1 mA at 3.0 V.

Were these measurements taken at the input to the driver or at the LEDs?

BoB

 

[CWatters]

Old thread alert.

 

[LURCH]

Just off the top of my head, doesn’t hooking the led’s up in parallel reduce the resistance slightly, by giving the current multiple paths?

 

[CWatters]

Two identical LED n parallel should have about half the resistance of one LED so "slightly" isn't the word I would use.

If they are connected to a constant current source the resistance doesn't determine the current anyway so it's irrelevant.

 

[sophiecentaur]

about half the resistance

You have to be careful when using the term Resistance to describe a (very) non-ohmic component like a diode. The Current is certainly not proportional to the PD across an LED so you cannot use the simple Resistances in Parallel and Resistances in Series formulae. For a particular PD, there will be a value of Current and, if you insist, you can equal that to a 'Resistance' but the Ohms really don't mean much at all in that context; increase the PD ever so slightly and the current can roar away and give you a massively different 'Resistance' value. (See https://www.electronics-tutorials.ws/blog/i-v-characteristic-curves.html and many others.)

 

[rbelli1]

A typical blue diode in parallel with a typical red diode will have the "resistance" of approximately the blue diode. The red one will light little or not at all. Look at the knee voltage referenced in sophiecentaur's link. That will vary quite a bit depending on the LED. Some LEDs are actually multiple LEDs molded into one case.

BoB

 

[sophiecentaur]

A typical blue diode in parallel with a typical red diode will have the "resistance" of approximately the blue diode. The red one will light little or not at all. Look at the knee voltage referenced in sophiecentaur's link. That will vary quite a bit depending on the LED. Some LEDs are actually multiple LEDs molded into one case.

BoB

Which nicely makes my point that "Resistance is useless", as the Vogon Captain was heard to remark in the Hitchiker's Guide.

 

[essenmein]

Probably the mistake here is conflating a diode with resistance, I mean, at some voltage it will flow some current, but if your 3V Vf diode is flowing 20mA, its nonsensical to say the diode therefore is 150Ohm.

Diodes in parallel don't share well, even if they are the same there will be process variations making them not quite the same, and most have a negative temp co which basically acts like positive feedback for the diodes already taking the load resulting in them taking more of the current...

 

[CWatters]

It's not "nonsensical", it's actually in the UK GCSE Physics sylibus under "non-ohmic resistors". Take a typical diode curve, at the point (V, I) the equivalent resistance is V/I (eg not the slope of the curve at that point).

 

[sophiecentaur]

It's not "nonsensical", it's actually in the UK GCSE Physics sylibus under "non-ohmic resistors". Take a typical diode curve, at the point (V, I) the equivalent resistance is V/I (eg not the slope of the curve at that point).

I would never quote GSCE Specifications as an authority about Physics. Right or wrong, they are aimed at a very first path through Science and are full of fudges. That sort of thing is used to promote the criticism that 'Elementary Science is Wrong and you need to re-learn everything you were taught at School'. A very bad advert for Science, I think.

What is the point of having a quantity that's called Resistance? Is it just to describe the ratio of V and I or is it supposed be of use under changing circumstances? If it is just to express a ratio then it shouldn't ever be written down on a circuit diagram because any change to the circuit or its input can change that value. It would be like writing the length of a spring down on the diagram of a machine and then using that value when the load changes.
On the other hand, the value of Resistance of a 'Resistor' or of the winding in a coil (that won't change temperature) is something to work with.
This goes with what I feel is a misuse of the term Ohm's Law, which I insist is something that only applies under constant temperature. R=V/I on its own is just the definition of Resistance and does not imply Ohms Law. The filament in a lamp would follow Ohm's Law if it were given a chance and if it was full of a non (electrical) conducting coolant but it can be assigned a resistance with fingers crossed and along with a caveat. [Edit: especially in the context of an electrical supply with a controlled Voltage.]

 

[CWatters]

Humm. I did say "equivalent resistance". What should we call the property of an NTC (Negative Temperature Coefficient Thermistor) at temperature T if we can't call it resistance?

Likewise the ON resistance of a FET?

 

[sophiecentaur]

Humm. I did say "equivalent resistance". What should we call the property of an NTC (Negative Temperature Coefficient Thermistor) at temperature T if we can't call it resistance?

Likewise the ON resistance of a FET?

Oh yes. I do acknowledge there is a problem. The problem is actually far to big for GCSE students to be exposed to it.
There are borderline cases. An NTC or any Temperature sensitive resistor (including a lamp filament) can be regarded as a variable resistance for small signals and the current through these will affect resistivity. If pushed, I could suggest a symbol R_E or something similar to make it quite clear that we are not dealing with Ohm's Law.
If any component can be characterised with a fairly stable Resistance value over a range of conditions then Resistance will do. But I could ask you why a transistor or a valve is never described as a variable resistor when used in Common Emitter (or Cathode) mode? Because that's what they effectively are.
Overall, the whole thing is a bit of a mess (thanks to some too simplistic early education) and sticking to some formality has got to be a good thing in the circumstances.

 

[essenmein]

Humm. I did say "equivalent resistance". What should we call the property of an NTC (Negative Temperature Coefficient Thermistor) at temperature T if we can't call it resistance?

Likewise the ON resistance of a FET?

Those are still resistances, they do change with temp, however at a given temp, and assuming power dissipated does not change its temperature, then both Rdson and NTC follow Ohms law, ie the current is voltage/resistance.

Compared to a diode, that dramatically changes its current based on applied voltage, even if its temperature does not change.

Note there is actually no such thing as a pure resistance, I have never encountered a part that does not change its value with temperature, you might get very low temp co, but never zero.

 

[sophiecentaur]

Note there is actually no such thing as a pure resistance,

Agreed. But the conventions used in schematic circuit diagrams are there for a reason and give us a clue then Ohm's law behaviour can be expected (near enough for the particular purpose).
I can never fathom why the simple linear V/I ratio should be referred to as "Ohm's Law' when the other 'triangle formulae' that we use are just formulae and only get called a Law when they are applied in certain extra constraints.
It's amazing when you think that EE is such a monster of a discipline that there is such a lot of loose terminology used, even at high levels by some people.

 

[CWatters]

It's amazing when you think that EE is such a monster of a discipline that there is such a lot of loose terminology used, even at high levels by some people.

Indeed. You only have to look at the problems beginners have with terms and concepts like "Earth", "Ground" etc.

It's a bit like a language. Most languages (particularly English) have numerous exceptions to the rules for spelling and grammar that we could fix if we really wanted.

 

[essenmein]

I can never fathom why the simple linear V/I ratio should be referred to as "Ohm's Law' when the other 'triangle formulae' that we use are just formulae and only get called a Law when they are applied in certain extra constraints.

I think that is more historical than anything, that Ohm guy did take us from not knowing about V=IR to knowing that V=IR.

 

[anorlunda]

It's amazing when you think that EE is such a monster of a discipline that there is such a lot of loose terminology used, even at high levels by some people.

The alternative is to have language police censoring stuff. Physicists do it. Consider what happens when someone uses the phrase "relativistic mass" here on PF in a physics. If they keep policing that for another century or two, the depricated uses will disappear.

But we still have Newton's Laws but only Einstein's Theories. My favorite are principles. A principle is true because it is a principle.