Can some kind of LC circuit conduct constant AC current ?

Frequency being fixed 60 Hz,
Voltage supply being fixed 120 VRMS AC,
Variable resistive load,
Constant current desired 0.015A RMS AC
Could some array of coils/capacitors circuitry in series provide such ?

ACphase--------------LCcircuit-------------Amperimeter----------variable load---------ACneutral

There is no way any LC circuit, will provide constant current to a variable load. Or any other combination of passive components for that matter.

kubeek said:

There is no way any LC circuit, will provide constant current to a variable load. Or any other combination of passive components for that matter.

Click to expand...

That is not precisely true. If the variable resistance in kept in the range such that is much much less then the reactance of a series inductor or capacitance the current will be close to constant.

This is used for flourescent lights and some LED light ballasts. You only need a capacitor or inductor, not necessarily both.

15 mA sounds like you might be wanting to run an LED. A capacitor of 0.33 uF in series with LED is about right for 120vac mains. (or a 20.7 Henry inductor, but cap would be easier). For 0.33 uF cap put about a 680 1/2 watt in series with cap to prevent switch bounce surge on LED. It is also a good idea to put a 1 meg bleeder resistor across cap to discharge cap in a reasonable time when it is unplugged.

Left out you need two LED in parallel, with polarity reversed to each other. Alternately you could put a bridge recitifer after the series cap for a single device drive. Bottom line you need to ensure the series cap has equal positive and negative AC current flow.

Externet said:

Frequency being fixed 60 Hz,
Voltage supply being fixed 120 VRMS AC,
Variable resistive load,
Constant current desired 0.015A RMS AC
Could some array of coils/capacitors circuitry in series provide such ?

ACphase--------------LCcircuit-------------Amperimeter----------variable load---------ACneutral

Click to expand...

Hello there,


In theory you could construct a swinging capacitor but it may only be practical for lower current levels than you are looking for.

But it also depends what you are calling an inductor, do you consider a three or four terminal inductor (a transformer or tapped inductor) an inductor too? Then we could probably build something that uses one or two very small transformers or tapped inductors that will regulate the current.

What tolerance is there on the 15mA?

Thanks.
Tolerance can be ±33%; say from 10mA to 20mA
Coils of any kind, tapped inductors, any core, three or four terminal transformers could be used.

A 0.47uF cap in series with the 680Ω which RCinFLA proposed will give 10mA-20mA RMS for resistive loads in the range 0-~9kΩ.

Hello,


Yes what is the required resistance range also? Using tapped inductors this should be possible, with very good regulation, better than you are asking for.

Externet said:

Frequency being fixed 60 Hz,
Voltage supply being fixed 120 VRMS AC,
Variable resistive load,
Constant current desired 0.015A RMS AC
Could some array of coils/capacitors circuitry in series provide such ?

ACphase--------------LCcircuit-------------Amperimeter----------variable load---------ACneutral

Click to expand...

Steinmetz's 1917 book:

http://books.google.co.za/books/about/Theory_and_calculation_of_electric_circu.html?id=vU5Ngxn7_NsC

Look down the page at the table of contents and you'll see several entries that begin "Constant Current...". The most interesting one, which is seldom mentioned these days, is the monocyclic square.

Here's a company:

http://www.thycon.com.au/tech_ccr2.shtml

that apparently manufactures constant current regulators, and they mention the monocyclic square.

Excellent help !!! -The monocyclic square-
Thanks for the information link with great explanation, calculations and examples

Hi,


Wow that book was published around 1917? Geeze, that's amazing. Now let's see how far we have come in 100 years

Here's a circuit that could work with low current in the load resistor. T1 can be a 'tapped' inductor if no isolation is needed. T2 also.

(Note that other resistor is usually called the 'burden' resistor for the current transformer)

It might be interesting also to note that 1917 was some 10 years before the first feedback amplifier was invented, as least from the references that i can find. Feedback amplifiers changed everything, and then op amps made that idea even easier to apply.

Hi again,


Here's another solution but it will probably be very hard to implement as you will see. But theoretically it keeps the current constant via forcing the transfer function to be constant with the proper choice of L and C, except unlike the previous circuit this circuit has to have a constant frequency to work. The circuit is a simple C in parallel with the load resistor R, and there is an inductor L in series with those two and the power source E, and the power source is an AC voltage.

Looking at the transfer function of the current through the resistor we have a not too difficult amplitude to peruse:
iR(amplitude)=E/sqrt((R-w^2*C*L*R)^2+w^2*L^2)

where we see R in the denominator and it looks like it has quite an influence on the current amplitude. But one thing stands out: there is a subtraction n the denominator, and whenever we see that there is the likelihood that we can chose certain values to get that part to go to zero leaving only what is left to do all the controlling. In this case, if we could get the subtraction to equal zero we will be left with:
iR=E/(w*L)

which for constant frequency means the load resistor current will also be constant.

Ok, so the part we have to get to go to zero is the subtraction in the denominator, and expanding that part alone we have:
w^4*C^2*L^2*R^2-2*w^2*C*L*R^2+R^2

and setting that to zero we have:
w^4*C^2*L^2*R^2-2*w^2*C*L*R^2+R^2=0

and now solving for L we get:
L=1/(w^2*C)

To check this we plug that back into the original transfer function and we get:
iR=E*w*C

so we have succeeded in getting the current to be constant, and this is completely independent of the load resistance R.

Ok now here is where we start to run into a practical problem...

We want 15ma so we use that equation to compute C:
iR=E*w*C
0.015=E*w*C
or:
C=0.015/(E*w)

and assuming that E=120vac and w=2*pi*50 because f is 50Hz, we get:
C=0.015/(120*100*pi)=3.97887e-7

which is a nice small value easily obtained. But now lets compute the value required for the inductor:
L=1/(w^2*C)=1/((2*pi*50(^2*3.97887e-7)=25.46 Henries

That's quite an inductor there!

So while this does provide theoretically perfect regulation using just one L and one C, it could be hard to implement depending on the current required.

MrAl said:

Hi again,


Here's another solution but it will probably be very hard to implement as you will see. But theoretically it keeps the current constant via forcing the transfer function to be constant with the proper choice of L and C, except unlike the previous circuit this circuit has to have a constant frequency to work. The circuit is a simple C in parallel with the load resistor R, and there is an inductor L in series with those two and the power source E, and the power source is an AC voltage.

Click to expand...

Steinmetz describes this in his book.

MrAl said:

Hi,


Wow that book was published around 1917? Geeze, that's amazing. Now let's see how far we have come in 100 years

Here's a circuit that could work with low current in the load resistor. T1 can be a 'tapped' inductor if no isolation is needed. T2 also.

(Note that other resistor is usually called the 'burden' resistor for the current transformer)

It might be interesting also to note that 1917 was some 10 years before the first feedback amplifier was invented, as least from the references that i can find. Feedback amplifiers changed everything, and then op amps made that idea even easier to apply.

Click to expand...

Could work? Does it or doesn't it?

Without some description of the circuit parameters it's difficult to analyze. Will it work with any transformer ratios, primary and secondary inductances, coupling coefficients, etc.?

The Electrician said:

Steinmetz describes this in his book.

Click to expand...

Hi,

That's interesting. Does he describe it in the same way showing how the denominator affects the overall outcome?

The Electrician said:

Could work? Does it or doesn't it?

Without some description of the circuit parameters it's difficult to analyze. Will it work with any transformer ratios, primary and secondary inductances, coupling coefficients, etc.?

Click to expand...

Hello,

Well i dont always include everything about a circuit, sometimes i just preset the basic approach and then leave the design details to the end designer. The most important part is the topology, and if the topology makes sense to the reader they should know how to proceed. But i could provide a few more details as well as an example i guess. I'll see if i can get to this tomorrow sometime.

MrAl said:

Hi,

That's interesting. Does he describe it in the same way showing how the denominator affects the overall outcome?

Click to expand...

You can look at his discussion here:

https://books.google.co.za/books/about/Theory_and_calculation_of_electric_circu.html?id=vU5Ngxn7_NsC

He starts a long discussion of various constant current techniques on page 245. The table of contents says that the discussion of resonance techniques starts at about page 255.

His methods of analysis don't look quite like the modern methods, but they're not that hard to follow.

In early 1900 the constant current circuits were used to feed large numbers of carbon arc streetlights in series.

Hi there,

So i guess constant current circuits like this were more important back then.

Thanks for pointing those pages out. It was interesting to look at them although they show up a bit too small for me But i was able to discern some things, such as the monocyclic square. What i couldnt find though was a formulation for the circuit current that was enough to do a more thorough analysis, so i did a few quick calculations and what i found (if i did it right in my haste) was that the monocyclic square is not anything too extraordinary regarding it's theoretical functionality over the single inductor single capacitor solution. Unless i did something wrong, it looks like the only added benefit from using two inductors and two capacitors is that negative effects of the components may cancel out due to the balanced nature of the circuit. So for example if the inductor changed value slightly because of temperature then the other inductor would also change (presumably) and that would keep the circuit working the same as before any change had occurred. I do however think is it interesting that he thought of that solution.

So in other words, the single inductor single cap circuit looks the same theoretically, and the dual L dual C circuit just helps to compensate for component variations. Do you agree?

I determined this because the analysis had shown that we still have the same determining factors for the dual circuit as for the singular component circuit:
I=w*C*E
and
L=1/(w^2*C)

and that's for either circuit with pure resistive load.

So the monocyclic square is a component variation corrective circuit version of the single L single C circuit, apparently. I would have liked it if he pointed this out from the start, unless he did and i missed seeing it

MrAl said:

So in other words, the single inductor single cap circuit looks the same theoretically, and the dual L dual C circuit just helps to compensate for component variations. Do you agree?

<SNIP>

So the monocyclic square is a component variation corrective circuit version of the single L single C circuit, apparently. I would have liked it if he pointed this out from the start, unless he did and i missed seeing it

Click to expand...

I don't agree. In the book, Figure 117 is described as a circuit whose advantage is that the "primary side current", as he calls it, has the same "power-factor" as the secondary current. Figure 118, III has the same property if the right side of the load is connected to the mid point of the autotransformer. And finally, figure 118, IV is the monocyclic square which also has the property that if the load on the constant current port is a pure resistance, the impedance at the constant voltage port is also a pure resistance (with ideal inductors and capacitors, of course).

The single inductor, single capacitor circuit (figure 118, I) doesn't have that property.

He explains all this rather thoroughly.

Hello again,


Yes i read some more of it now and i think that is accurate. He was concerned with power factor too because of the high power systems he was targeting. Very interesting.

But i dont see any way around the requirements for the inductor value yet because the analysis comes out the same. Unless he found that there is another way to achieve somewhat good regulation without satisfying the denominator as done with the single coil and cap circuit.