Math and how to think about constant-current sources?

[ghawkins]

I'm a hobbyist whose background doesn't include anything much more than Kirchhoff's laws and the kind of stuff one covered in high school. I wasn't even aware that there was such a thing as constant-current sources until recently and I'm confused about how to think about them.

Constant-voltage sources have a stated voltage and a maximum current, e.g. 5V and 2A. If my particular circuit consumes less than 2A then this voltage source should be adequate.

1. Are things as simple with constant-current sources? E.g. if I have a  load of LEDs with a forward voltage of 3.3V, that I want to run in series, can I just sum up the voltages and then be confident that things will work if this value falls within the stated voltage range for my constant-current source?

Imagine I have two different types of LED both with a forward voltage of 3.3V but one with a forward current of 350mA and one with a forward current of 700mA. Will a 350mA 24-86VDC constant-current source support the same maximum number of 700mA LEDs as 350mA LEDs (with the 700mA LEDs obviously running dimmer than they could)? I.e. is it really as simple as just considering the forward voltages or does the forward current of the devices also come into play in my calculations of how many LEDs (or whatever) I can use with a given source?

2. Constant-voltage sources are rated for a maximum current, e.g. 2A. But no one states a minimum current (at least with the devices I've used), I assume they can support from ~0A to e.g. 2A.

Constant-current sources however always seem to have a minimum rated value as well as a maximum, e.g. 24-86VDC.

So what typically happens when you connect a load that requires a lower voltage, e.g. I connect just one of the 3.3V LEDs mentioned above? Does the source cut-out, i.e. refuse to provide any voltage? Or does it "work" but fail to limit the current to the specified value as I failed to limit my circuit to the specified voltage range?

I presume different loads and different constant-current sources behave differently. So for two concrete examples, let's take:

* The 45W DC-DC step-up constant current sources from MeanWell (datasheet, Digikey).
* The 700mA 3.3V LEDs from Broadcom (datasheet, Digikey).

3. Finally, one question that's specific to the datasheet linked to above. For the input voltages, it says in the notes "output voltage must step up 3V from the input voltage". So e.g. for one of the devices the input voltage range is given as 18-32VDC and the output voltage is given as 21-126VDC - is the note saying that the lower value of the output range is only possible if the input voltage is at least 3V lower, i.e. 18V? And that if my input voltage is e.g. 32V then the output range is effectively constrained to 35-126VDC?

Thanks - and sorry for such basic questions. While it seems easy to find beginner introductions to constant voltage sources, I didn't find it so easy to find similar introductions for constant current sources.

[TimFox]

For question (1):  are you connecting the different types of LEDs in series or parallel?
If in series, they will all conduct the same current, so the 700 mA diodes will not be as bright as they would be at 700 mA.
Connecting LEDs in parallel (if they have no internal resistors or current limiters) is a bad idea:  the ones that require slightly less voltage will take most of the current, due to the exponential behavior of voltage vs. current on the individual diode.
(2):  constant voltage  sources are like batteries and have a no-load voltage value.  Ideal constant-current sources (like lab supplies) work down to zero voltage across the load (short circuit).  Practical lighting sources should be operated within the manufacturer's specs.

[bdunham7]

You have to differentiate between a theoretical or lab-quality power source and the LED-drivers that you are using.

A 'proper' constant current source, like a lab PSU or the current source in an ohmmeter just to name two, will ensure that a specific constant current flows through whatever is connected to it up to a certain maximum voltage, called the compliance voltage.  Beyond that, it will be unable to maintain the constant current. 

The LED boost drivers you are referring to can only operate within certain defined parameters which is why you need to read the detailed datasheet.  The ones you linked I would call boost converters with current regulation within limits, not to be confused with a true constant current source.  There are other LED-driver specific current sources that don't have these limitations, but they probably aren't as powerful or cheap.  Boost converters often do need a minimum step-up, and according to the datasheet you linked, those converters will be outside their operating specifications if the output is not at least 3 volts above the input.  IOW, the output is constrained in the way you have stated.

[Doctorandus_P]

It is indeed pretty much that simple.
A constant current source just delivers some fixed current, regardless of what it's output voltage is, as long as it's lower then the maximum voltage that it can deliver.

Any decent lab power supply has an adjustable current limit, and you can therefore also use it as a constant current source. Just set it's current to a lower value then the connected circuit is trying to suck out of it.
So if you set a lab power supply to an output voltage of 30V (often it's maximum), but limit the current to your 350mA then it will deliver 3V to one led, or 6V to two leds in Series (9V for three, etc).

In practice there are some limitations to circuitry.

When you want to use a regular power supply as a constant current current source, you have to be aware of it's output capacitor, and control loop speed. Setting it to 30V output and 20mA, and then connecting an LED will probably instantly destroy the LED. What you can do is set the current limit to 20mA, and it's output voltage to 0V, then connect the LED and increase it's output voltage untill it goes into current limit.

For a voltage regulator such as the LM317 for example, it has a minimum output current of around 1.25V/240 Ohm = 5mA (Have not checked the datasheet) That is one of the reasons that a low ohmic resistor is often used in it's feedback circuitry. That resistor sucks up it's minimum output current.

An LM317 is also perfectly suitable as a constant current source or sink. Just use a single resistor between the adjust pin and the output pin, and then tap the current off from the adjust pin. It will start limiting the current when the reference voltage (1.2V) is over the external resistor.

You mentioned you thing there is a minimum output voltage for a constant current source. Why do you think so? Usually it's perfectly all right to short circuit a current source, and it has an output voltage of 0V under those conditions. When you do this, you may have dissipation problems with your circuit, but that is a practical limit depending on your circuit and not a limit of a constant current source.

Edit:
Those constant current drivers for LED's often have additional limits indeed. Voltage Boost circuits usually have a diode directly between input and output, and therefore  can not deliver an output voltage that is lower then a diode drop from it's input voltage. An LM317 in constant current configuration can be safely shorted on it's output, with only a limitation of it's temperature (and therefore dissipation, and size of it's heatsink).

For breadboards, I like to use 6 pin sectons of dual row headers. The strong square pis do not bend easily, and you can solder each pin of an TO220 to two header pins. This makes it stand firmly in a breadboard.

[rstofer]

Think about a 1 mA constant current source driving a 1 Mohm resistor.  It would need to have a voltage of 1000 Volts.  Ohm's Law (E=I*R) still applies.

There is always an upper limit to the load resistance or, actually, the maximum deliverable voltage at the rated current.

[TimFox]

Nomenclature:  for a true constant-current supply, the maximum output voltage is called the “compliance” limit.  Lab supplies often have a front-panel control to reduce this limit (to avoid accidental high voltage into open circuits).

[tggzzz]

With a constant voltage power supply, the output current is whatever is necessary to maintain the selected output voltage.

With a constant current power supply, the output voltage is whatever is necessary to maintain the selected output current.

Typical lab power supplies are often marked as having a constant current mode. A better name would be a current limited mode.

[TimFox]

Yes, most lab supplies have a “current-limited” mode, which is not as accurate as a true current stabilization, but suffices for protection.
Some better ones “cross over” from voltage stabilization to current stabilization (depending on the voltage and current settings and the load), often with indicator lights to signal which mode is controlling the output voltage.

[bdunham7]

Typical lab power supplies are often marked as having a constant current mode. A better name would be a current limited mode.

Why then would you not refer to the other mode as voltage limited mode?  True well-designed CV/CC mode PSUs should do both modes fairly well.

[tggzzz]

Typical lab power supplies are often marked as having a constant current mode. A better name would be a current limited mode.

Why then would you not refer to the other mode as voltage limited mode?  True well-designed CV/CC mode PSUs should do both modes fairly well.

Personally I would be very happy with that nomenclature. But I doubt my opinion is going to change the world

[TimFox]

“Current-limited” power supplies may have a less accurate current control, or even a fold-back limit, and are designed for constant voltage applications.  The cv/cc supplies that cross over between modes are typically more expensive;  obviously, they are better for applications that require a stable current.
Note that in modern terminology, a “regulator” changes the output (e.g. the pass transistor in a linear supply or a manual valve in a hydraulic system), and a “stabilizer” is a servo that uses feedback to maintain a stable value of a parameter.  The older terminology “regulated power supply” persists, as does the misnomer “source regulation”.

[Joel_Dunsmore]

Many supplies are dual modes and sometimes name the modes "Voltage Priority" and "Current Priority".  In voltage priority you set a current limit and a desired voltage output. The supply will maintain the desired output voltage unless the current exceeds the limit, in which case it drops the voltage to maintain a constant current at the limit value.  In current-priority mode you set a voltage limit, and the supply will maintain the selected current by raising or lowering the voltage until the proper current is achieved. If the resistance is so high that maintaining the desired current would require a voltage above the voltage limit, then at that limit-voltage the source turns into a constant voltage source.  There might be some minimum current needed (milliamps or microamps) for proper operation and same for minimum voltage for constant current sources. 

[tggzzz]

Many supplies are dual modes and sometimes name the modes "Voltage Priority" and "Current Priority".

That's not unreasonable terminology, but I've never seen it. Can you give an example?

[Terry Bites]

Nearly every bench supply would have the contant current limit and constant voltage modes as standard. Though the term contant current can be a bit misleading misleading. A true contant currne source will adjust its self   by raisning or lowering the voltage across the load untill the set current reached- most power supplies impose a current limit, a maximum that cannot be exceded at any given output voltage set.

A indeal current source providing I Amps into an infinte resistor R would produce I*oo Volts. Of course it cant really do that (simulation software CCS can generate teravolts, beware of the illusion). Ideally, the output impedance of the CCS is assumed to be infinite, ie current only flows in the load. Its not though, but in practice many meg ohms can be achieved. https://www.electronics-lab.com/article/current-sources/

CCS for LED drivers work on averaged pulse current for effciency and are very application specific devices. Though any swiching power supply can act as an LED driver.

A controllable source of any kind uses negative feedback to force a a voltage across aload untill the current current through a load is equal to the current in the sense reistor, Io= Vref/Rsense
This can be achieved by linear and swiching ciruits.