How Do Constant Current Sources Differ From Other Power Sources?

[Dash-IQ]

How are constant current sources different than any other source?
Also, would they really have a constant current in any situation? Even when back-emf is induced?

Can anyone explain the use of a constant current source, and how different they are from other sources?

 

[Dash-IQ]

"A current source is an electronic circuit that delivers or absorbs an electric current which is independent of the voltage across it." -- How?!

 

[Nick O]

True "constant current sources" don't exist. Imagine removing such a device from a circuit; it would arc, and continue to arc until it either destroyed itself or was replaced.

But, it can be useful to consider something a constant current source. If you have a constant voltage source V in series with resistance R, it will behave exactly the same (in terms of the circuit) as a theoretical constant current source I=V/R in parallel with resistance R.

 

[sophiecentaur]

True "constant current sources" don't exist. Imagine removing such a device from a circuit; it would arc, and continue to arc until it either destroyed itself or was replaced.

But, it can be useful to consider something a constant current source. If you have a constant voltage source V in series with resistance R, it will behave exactly the same (in terms of the circuit) as a theoretical constant current source I=V/R in parallel with resistance R.

This is precisely what can happen to a current transformer, used for measuring the current flowing in HV cables. They are always kept with their terminals shorted together until they are actually connected to the measurement circuit.

It is strange that the 'Voltage source', as a concept gives us very little intuitive trouble. But perhaps not so strange, I suppose, since the invention of the Car battery, which maintains a pretty good PD, even when you roast and melt a thick piece of wire, held across the contacts. There, you have a source of 'as much current as is needed' to give you 12V. A very familiar piece of kit.

If you still have a CRT based TV, the beam of electrons has a current which is set by the electron gun and the grid and the same current will fall on the screen for a massive range of screen / cathode voltages. There, you have 'as much PD as you need' to ensure the required current; a 'current source'.

 

[Dash-IQ]

But what is a "constant current source"? I don't understand what it is, and what differs it from the other sources?

 

[MrSparkle]

constant current source is to current like a battery is to voltage.

Think of the I-V curve. An ideal constant current source is a straight horizontal line, while an ideal voltage source (like an ideal battery) is a straight vertical line.

 

[sophiecentaur]

A Constant Current Source is a Fiction. It is not a real piece of equipment or component. It is an Ideal which is used in circuit analysis.* Its characteristic is that it delivers the required current into any load, whatever its value. Constant current supplies are less familiar, which is why they give problems for students. It is valid to assume a constant current in many real circuits, even when the source consists of real components.

There is a parallel with the way that a constant voltage source is used. For example, an ideal power supply would normally maintain its nominal volts, whatever load you give it (in practice there would be limits). A good audio amp would behave very much like a constant voltage source, providing the loudspeaker with the same voltage signal for a wide range of load impedances. We often assume a constant voltage in these cases.

A constant current supply will behave in a similar way but will deliver the required current whatever the load impedance. You can achieve 'near enough' constant current by using a high supply voltage in series with a very high resistance. The current that comes out will be (more or less) the same over a large range of (low) load resistances. A high enough supply voltage and high enough source (series) resistance will behave a a constant current source. Work out the current from a supply of 1000V in series with a source resistance of 1MΩ, when you connect it to 100Ω load or a 10Ω load. Near enough the same value for practical purposes.

*Just the same way that a Voltmeter or Ammeter is included in a circuit; we usually ignore the fact that meters actually consume Power. Engineering is full of such dodges.

 

[jim hardy]

Have you studied Limits yet in math ?

A constant current source would be an infinite voltage in series with an infinite resistance. We all know that can't happen, how would you ever pick it up without getting electrocuted?.

But how would a circuit behave if you had numerically equal voltage and ohms in series ?

Let's start with one volt and one ohm, then see how it behaves as we increase them.

One volt and one ohm would deliver one amp into a short circuit
and one half amp into a one ohm load
and one tenth amp into a nine ohm load,
and so on. That's a long way shy of constant current...

Now take 10 volts in series with 10 ohms.
That'd deliver one amp into a short circuit
and 10/11ths of an amp into a one ohm load
and one half amp into a ten ohm load...
better, but not close..



Now take 100 volts and 100 ohms
you get one amp into a short circuit
0.990099.. amp into a one ohm load
and 0.90909... amp into a ten ohm load

Try the same arithmetic exercise with 10kv and 10 k ohms...


As you see, the higher the internal voltage and internal resistance, the more nearly it resembles a constant current source.

So: A constant current source behaves mathematically like the limit as n approaches infinity of n volts in series with n ohms.
In reality no such device exists, but there do exist pretty good approximations of it.

A real word example is a "Capacitive Coupled Potential Device"
where perhaps 500 kilovolts is the driving voltage.
http://www.gedigitalenergy.com/multilin/notes/artsci/art08.pdf

Or Sophie's current transformer.
http://www.electronics-tutorials.ws/transformer/current-transformer.html

An op-amp can be arranged to behave like one within the limits of its power supply. They're used a lot for providing constant current to electronic gizmos.

So form yourself a mental image and work it in your head until it agrees with the math - it's a really useful concept, if a hard one to believe in .

 

[Baluncore]

You normally consider constant voltage sources, called batteries or regulated power supplies. The circuits are designed to draw the current they need from the constant voltage sources. Oscilloscopes and voltmeters are used to observe the signal and supply voltages. You only know the current if you also know the circuit's series resistance, and can compute reciprocals or do arithmetic division.

In the other half of the universe, where I am, we use constant current supplies. I have linear constant current generators and switching current regulators that work really well. My circuits are designed to drop the voltage they require when powered by a constant current. I can still measure the voltage across the components, but I know the current is constant, so I can compute the circuit resistance as being directly proportional to voltage. I have no problem with circuit inductance.

In both our worlds, the power distribution grid systems are based on constant voltage levels so any customer can independently draw, and pay for, the current they require from the supply. If a constant current was passed around the neighbourhood, then we would all need to drop only the voltage we required and there would be all sorts of problems with earthing and insulation breakdown, on very thick wires.

Sometimes a constant current circuit is needed by an engineer in your world, where a circuit from my world must be used. But your education system grew up in the time of “batteries” of electrochemical cells, with all those crudely fixed voltage sources, so it has forgotten to teach you about the other half of the universe.

 

[Dash-IQ]

I'm trying to grasp the idea... but Jim thanks to your relation to limits that accelerated it.

 

[Dash-IQ]

Thank you for the help everyone.

 

[jim hardy]

I'm trying to grasp the idea... but Jim thanks to your relation to limits that accelerated it.

nowadays you'll see them in lots of configurations , used for LED drivers .

http://www.ti.com/lsds/ti/power-management/led-driver-products.page

http://www.ti.com/lsds/ti/power-management/led-driver-overview.page

you can think of it as a variable voltage source that adjusts its voltage to maintain the desired current. That's how you make an op-amp do the job.
see the 'getting started' links here
http://www.ti.com/analog/docs/microsite.tsp?sectionId=560&tabId=2213&micrositeId=7

 

[Baluncore]

For a constant current source:

you can think of it as a variable voltage source that adjusts its voltage to maintain the desired current. That's how you make an op-amp do the job.

In just the same way as you think of a regulated voltage supply, as a device that adjusts it's current to maintain the desired voltage. You can also do that with an op-amp and series pass transistor.

 

[Okefenokee]

A good use for real current sources is lighting at airports. If you used a voltage source you would need to run 2 wires in parallel across every single light, one hot and one neutral. a lighting circuit at an airport can be miles long so by the end the voltage would be weak because of the resistance in the wire. Now if you use a current source you can put all the lights in series and you only need one wire going into and out of each light. Each light receives the same power because it has the same current. This also saves a fortune in copper because a good wire layout will have a single wire making a big loop around a runway or taxiway as it goes to each light.

In case anyone was wondering, each light in an airfield is connected to an isolation transformer. If the light burns out, as they do, current can still flow through the primary winding of the transformer to all the other lights. One bad light will not take out the entire circuit.

The current sources they use come in large sizes like 5kW to 25kW and more. A typical current is around 5 amps for a light circuit, depending on what kind of light intensity you want. That means that the current sources max out at 1kV to 5kV and higher because power = current X voltage.

The circuits do fail sometimes and the danger of this setup comes from electricians failing to understand how it works as they troubleshoot it. Electricians spend most of their time working on voltage sources like you have in your house or business. Like others said, if a current source sees a high resistance it will pump out a high voltage to maintain a steady current. If there is a an open circuit fault on the airfield and an electrician tries to measure voltage on that circuit then the electrician may experience a short peak voltage of 10,000 Volts or more before the source shuts itself down.

I worked in the airfield lighting industry for three years and I've seen some close calls.

 

[sophiecentaur]

That's a revelation! No wonder you have had "close calls". Your average EE would be totally flummoxed if he hadn't read the handbook!

 

[anorlunda]

A good use for real current sources is lighting at airports.

Very interesting Okefenokee, thank you.

I can't help wondering how history might have been different if Edison had chosen this scheme for the Edison Electric Illuminating company. Edison designed the whole electric utility business, (including generation, transmission, loads, and even billing models) for the express purpose of providing illumination. Contrary to public opinion, Edison was quite aware of AC and it's advantages. He may well have considered and rejected a design similar to what Okefenokee described. Instead, he made his own brilliant invention, the feeder, which is still vital today in power distribution.

Why did he choose parallel circuits? Perhaps the complexity of manufacturing an isolation transformer with each light bulb. Perhaps the designs of isolation transformers back then were not good enough. But I suspect the real reason was that Edison planned on making big money in copper. He invested in the predecessor to Anaconda Copper. If thick copper cables became the norm, he would have profited greatly.

 

[cabraham]

Any battery or generator can be designed for constant current. The utility company spins their turbines at constant speed, resulting in constant frequency and constant open circuit voltage. Once load current is drawn, voltage drops due to synchronous reactance due to stator inductance (self). Adjusting field current in rotor regulates voltage to constant value.

The power company could just as well spin their turbines at constant torque and adjust field current for constant current instead. Doing so is not done because it would mean that full current is always being distributed. Voltage varies with loading. Conductors lose more power than insulators, I²R vs. V²G. So constant voltage is used. Also, with constant torque and current, the speed varies, as does the frequency. Synchronous motor speeds would not be fixed, and paralleling generators with differing frequencies is not done.

At dc, however, constant current would make sense for LED lamps. Connecting the lamps in series, and providing constant current is the best way to run LED lamps, as they are naturally amenable to current drive. A switching power converter can be set up for constant current to drive LED lamps. Did I help?

Claude

 

[sophiecentaur]

A series system wouldn't be too convenient round the house though. Some of the appliances would have to be operating at many kV Potential above Earth. Then there would be the problem of different powered devices and switching them in and out. It could just be my innate conservatism but I can see many problems which, although they have their constant voltage feed equivalent, would seem to be harder to solve. For instance, which is more likely in an electrical installation, a short circuit or an open circuit? All the sockets in the house would need to short out when an appliance was unplugged.
I think History got it right for normal living but we may have missed out in some circs which are like the Airport lights.

 

[psparky]

So to sum up, a controlled current source basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

 

[anorlunda]

History did get it right and Edison got it wrong, but not because of the technical details. Edison's vision was that the geographical extent of a utility company would be no more than a few blocks, and no applications for the power other than light bulbs. There was no idea of a general purpose socket in houses. True, he did accommodate limited use of motors and other stuff, but that was secondary. Remote generation, intercity transmission, and diverse applications of electric power were beyond Edison's vision. He was a brilliant engineer, but a mediocre businessman. The name of his company said Illumination not Electric and that was his focus.

The fascinating alternate history would have come if Edison had found a much superior way to supply illumination. If that had happened, we might have evolved two electric grids, one for illumination, and another for everything else. It would be the realm of SF to project how that might have changed the 20th century, and what artifacts of that history might linger today. Come to think of it, that might make a fun science fiction story. It would be a challenge to make readers understand how technology/business/society interact.

 

[anorlunda]

So to sum up, a controlled current source basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

Yes, more or less. The terms are loosely defined.

 

[psparky]

Yes, more or less. The terms are loosely defined.

"Loosely defined" fits me as well. I'll take it.

 

[cabraham]

So to sum up, a controlled current source basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

I don't think so. A controlled voltage source which makes up for voltage drop or transmission distance is still by definition a voltage source. I've worked with power supplies where the output voltage sense terminals are remotely located from the main supply parts. Voltage drops due to loading and distance are corrected by the servo loop. This is still, however, a constant voltage source.

In utility power generation, transmission, and distribution, the generator always outputs full voltage, and the current varies with loading. When current is drawn, two things happen. A counter-torque exists in the stator windings whose direction is opposite to the present rotational speed. To keep constant speed/frequency, more fuel must be burned, more steam pressure must be exerted on the turbine blades.

Another issue is that the self inductance of the stator is very high, typically 75% to 250% of the load impedance. As current is drawn, the voltage drop is very great, much more than miles of transmission line. T-line impedance is typically a few percent, 5% typical, inductive reactance, maybe 1 to 2 percent resistance. So the voltage drop in the stator inductive reactance must be corrected. Adjusting rotor field current is how it's done.

Anyway, to sum it up, Mother Nature has decreed that insulators have very low losses, but conductors have very high losses. It's less lossy to generate full voltage then let current be dictated by demand. That way, losses are kept low. This approach is taken a step further with transformers. After generation, voltages are stepped up, transmitted at hundreds of kilovolts, then stepped down to distribution levels of 240/120 volts.

The idea is to minimize current, which requires that we maximize voltage. Again, constant voltage source was the right way to do it. It could have been set up for constant current but losses are greater.

A CVS (constant voltage source) is a power source whose feedback control monitors output voltage then adjusts field current and fuel consumption to maintain fixed voltage despite varying load current. A CCS (constant current source) is similar except that fuel consumption and field current are varied to maintain constant output current under varying load.

CVS by its nature makes it easier to fix the turbine speed, and get fixed frequency as a result. Fixed frequency makes multiple generators connected in parallel on the grid much easier. It also makes synchronous motors run at fixed speed, good for clocks, or anytime fixed speed is needed.

CVS is definitely the right choice. At dc though, when driving LED lamps, I would go with CCS. We may see CCS lines on the utility poles someday after street lamps are transitioned to LED.

Claude

 

[sophiecentaur]

I think this thread may be getting bogged down in specifics as the OP has told us he understands this:

a controlled current source basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

There is nothing in the description of a Constant Current Source that specifies how it is achieved. It is, as I wrote earlier, just a fiction, and a shorthand for any device that produces much the same current over a wide(ish) range of loads. How this is achieved is not really relevant to someone who looks at a textbook with Thevenin or Norton and I have a feeling that this could be the direction he's coming from. Its just another example of something you encounter when entering Engineering.

There are many examples of things that behave, for the purposes of simple circuit analysis, as CCSs. These include, a transistor collector, a thermionic Anode, a very high impedance power supply, a beam of electrons, the Solar Wind AND special circuits, designed specifically (using feedback) to deliver current which is constant to a desired degree in an overall design.

 

[cabraham]

I think this thread may be getting bogged down in specifics as the OP has told us he understands this:


There is nothing in the description of a Constant Current Source that specifies how it is achieved. It is, as I wrote earlier, just a fiction, and a shorthand for any device that produces much the same current over a wide(ish) range of loads. How this is achieved is not really relevant to someone who looks at a textbook with Thevenin or Norton and I have a feeling that this could be the direction he's coming from. Its just another example of something you encounter when entering Engineering.

There are many examples of things that behave, for the purposes of simple circuit analysis, as CCSs. These include, a transistor collector, a thermionic Anode, a very high impedance power supply, a beam of electrons, the Solar Wind AND special circuits, designed specifically (using feedback) to deliver current which is constant to a desired degree in an overall design.

I would be careful not to confuse *constant* current source with *controlled* current source. A transistor, bjt or fet, is not CCS, but controlled current source as its value of drain or collector current is contrplled by an input signal. The CCS for dring LED lamps is constant, not controlled by a signal. As far ad a CCS being fiction, I would agree that ideal ccs is fictions, but a real ccs is just as real as a cvs.
Claude

 

[psparky]

I would be careful not to confuse *constant* current source with *controlled* current source. A transistor, bjt or fet, is not CCS, but controlled current source as its value of drain or collector current is contrplled by an input signal. The CCS for dring LED lamps is constant, not controlled by a signal. As far ad a CCS being fiction, I would agree that ideal ccs is fictions, but a real ccs is just as real as a cvs.
Claude

So to sum up, a CONSTANT current source is basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

Fixed?

 

[jim hardy]

I was aware streetlamps are sometimes series wired, but that requires provision for one in the string open circuiting.
I suppose those individual airport light transformers have some sort of magnetic shunt to limit the voltage when a lamp burns put?

Lamps are pretty robust devices. Series connection allows one conductor to serve a lot of loads, but opens the door to magnetic induction.
However - sensitive electronics , big loops and changing magnetic fields sounds to me like a risky mix(think power lines and geomagnetic storms) . The phone company learned of that decades ago when they shared poles with the electric companies..

i'll stick with parallel conductors in close proximity.

 

[sophiecentaur]

So to sum up, a CONSTANT current source is basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

Fixed?

That is not a definition, though. In a textbook (whatever we conclude must agree with that, surely) a constant current source is defined as producing a certain value of current. It's an ideal component that is placed in equivalent circuits of many amplifying devices.

How it's achieved is not considered. Feedback is only a smart way of avoiding the need for 100000MV and a Gizillion Ohm series resistor. In the linear world, any source of current has an equivalent of an ideal constant current source and a resistor. A "Controlled Current Source' is a common (?) description of a real circuit / device that behaves a bit like a Constant Current Source.

Whatever the practicalities are behind this, I don't think you can get away with describing a Constant Current Source in any other way than its original (formal) form unless you want to re-write a lot of very basic stuff. Implementation is an entirely different matter.

 

[jim hardy]

Whatever the practicalities are behind this, I don't think you can get away with describing a Constant Current Source in any other way than its original (formal) form unless you want to re-write a lot of very basic stuff. Implementation is an entirely different matter.

quite agreed.

We think in terms of ideal circuit elements. Ideal current source is one of them, and it's no more achievable than an ideal voltage source.

OP, who presumably is early on in his studies, needs to have it in his repertoire.

 

[cabraham]

So to sum up, a CONSTANT current source is basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

Fixed?

NO! I've already addressed that. What you just described is a constant voltage source! If the output being measured and controlled is voltage, then it's a voltage source. Say we generate a voltage at a power plant, using feedback to hold the terminal voltage constant. Five miles down at the load end, a small voltage drop takes place. So we modify the system by observing the voltage at the load end and adjusting fuel burn and field current to keep the load voltage constant to compensate for 5 miles of T-lines and the incurred drops.

This is still a constant voltage source. The constant voltage is at the load end of the T-line whereas in the first case it was at the source end.

A constant current source is produced by a system which measures output current, then adjusts fuel burn and field current so that as the load varies, the current is held fixed. Of course the voltage varies with loading. Any battery can be built for CCS mode. It's not done because lifetime is much better in CVS mode. Generators can go either way, but due to vast difference in loss between insulator and conductor, CVS is better.

A car alternator can be set up for CCS mode, but since the battery is CVS, it is built to regulate voltage. Current sources can be built just like voltage sources. An ideal CVS has 0 series resistance and 0 series reactance. With ac sources, the 0 reactance is impossible. With ac or dc, 0 resistance can be done with superconducting windings. But inductance is always present so 0 reactance is not going to happen. As soon as current is drawn, inductive reactance drops voltage. To keep terminal voltage fixed, field current needs adjustment.

An ideal CCS has a shunt impedance that is infinite, i.e. infinite resistance, infinite reactance. In ac domain, capacitance is non-zero, so infinite shunt reactance cannot happen. A real CCS has limited compliance.

To summarize, ideal CVS and/or CCS cannot be made. But real CVS & CCS can offer very good performance. By the way, to make a CCS, we do not need a CVS plus high series resistance. That is a poor way to construct a CCS, very lossy.

Google constant current LED drives and you will find that switching converters and inductors are used to keep LED current constant. The voltage generated is only the LED forward drop plus a little to cover drops incurred in wires, traces, sensing resistor, etc. If I wish to drive an LED with 100 mA, and it's forward drop is 3.0 volts, a CCS is the way to do it. A 100 volt supply, with a large series resistor of 970 ohms is a terrible way to do it. The drop across the resistor is 97.0 volts, and at 100 mA, that is 9.7 watts of heat loss. The LED power is 3.0V*0.100A = 0.300W. That is an efficiency of 3%, terrible.

Designing a CCS is not that hard, but please avoid large CVS with large resistors, esp. if large power is involved.

Claude

 

[sophiecentaur]

So to sum up, a CONSTANT current source is basically a controlled voltage source to make up for voltage drop or voltage increase due to distance in lines and loads starting or turning off?

Fixed?

What I would like to know is why you imply that it's ok to talk in terms of a controlled voltage source, as if it's a more basic component than a controlled current source. This controlled voltage source is just as difficult to achieve as a controlled current source would be.

Would there be any fundamental difference between using the collector of a PNP transistor or the emitter of an NPN transistor as the output stage, giving you a constant ("controlled") current generator (if you use appropriate feedback)? Those two terminals can be treated, in an elementary way, as constant current and constant voltage sources, respectively. Imo, you are giving too much emphasis on the way the source is actually achieved and seem to be assuming that you'd be starting with a low impedance amplifier output stage. If you had spent your life with thermionic valves, I suggest you might be viewing it the other way round.

 

[jim hardy]

If you had spent your life with thermionic valves, I suggest you might be viewing it the other way round.

thermionic valves ? Beam Power Pentode, add to that the inductance of impedance matching transformer, ? Small wonder you mentioned audio way back there...

6AQ5's forever !

 

[Okefenokee]

I was aware streetlamps are sometimes series wired, but that requires provision for one in the string open circuiting.
I suppose those individual airport light transformers have some sort of magnetic shunt to limit the voltage when a lamp burns put?

Lamps are pretty robust devices. Series connection allows one conductor to serve a lot of loads, but opens the door to magnetic induction.
However - sensitive electronics , big loops and changing magnetic fields sounds to me like a risky mix(think power lines and geomagnetic storms) . The phone company learned of that decades ago when they shared poles with the electric companies..

i'll stick with parallel conductors in close proximity.

No they don't have shunts. The primary has a constant current so the secondary has a constant EMF at all times.

It's funny that you should mention sensitive electronics. Some of the newer traffic signs had sophisticated electronics for brightness calibration and power factor correction. An airport ordered one for a circuit powered by an ancient CCS. The sign didn't work. The computer in the sign kept shutting itself down. I called the manufacturer and they took me through their troubleshooting steps. At the end they told me that I must have a poor meter and I need to buy a really expensive one then try again. That sounded like a last resort to me. Even though the CCS had the proper power rating to handle the addition of this sign I had the thought that maybe it wasn't operating well enough to power it. Perhaps it was unstable. I rewired the airport at a junction box to cut out half of the sign circuit and the sign starting working. I also rewired the airport to power the whole sign circuit with one of their newer sources and the whole system worked. I had a phone conversation with my boss to try and convince him to sell the airport a new circuit. That wasn't going to happen. It finally occurred to me that the airport bought the absolute cheapest isolation transformer there is. The one they ordered was square but better models are toroidal. I drove to our office and got a toroidal transformer with identical ratings and wallah! the sign worked.

 

[jim hardy]

The primary has a constant current so the secondary has a constant EMF at all times.

Sounds like an interesting transformer to have in one's "bag of tricks".

old jim

 

[Babadag]

The current equivalent source –and the voltage equivalent source-it is an ideal source – as per Norton [or Thevenin] theorem. See:
http://en.wikipedia.org/wiki/Norton's_theorem
IEC 60909 Short-circuit currents in three-phase a.c. systems – employs an equivalent voltage source- c*Un-in the short-circuit location in order to simplify the short-circuit current calculation.