How to design a circuit which has exponentially decreasing graph of (Vo)Vs(Io)

[AshwinNambiar]

I want to design a circuit where say,if i give Vin=130v DC,the circuit designed should be such that the output current Io increases with a decrease in the output voltage Vo. My requirement is to obtain an exponentially decreasing graph of (Vo) Vs( Io).
Example:
Vin=130v, Vo=125v, Io=10A
Vin=130v, Vo=122v, Io=11.5A
Vin=130v, Vo=120v, Io=14A
Vin=130v; Vo=114v, Io=20A
I'm a beginner,looking forward for help to design a circuit which serves the purpose as mentioned above

 

[(*steve*)]

What you ask for is not achievable.

Perhaps you can describe what you're trying to do -- what is the purpose of this?

 

[AshwinNambiar]

I'm trying to simulate an electric arc using non-linear active load on NI MULTISIM or PSPICE .The electric arc is such that the current increases as there is a decrease in the voltage;when plotted on a graph we have a negative resistance slope,an exponentially decreasing voltage with increasing current.
I want to design an electronic circuit,simulate it, which gives me the same plot of voltage Vs current as for an electric arc.
The purpose is to design a circuit which shows the simulation of electric arc using non linear active load.

 

[Harald Kapp]

In a SPICE based simulator you sjould be able to use a behavioural source (B) to simulate this effect.

 

[AshwinNambiar]

If you could elucidate on the behavioural source used.I want to design a purely electronic circuit which when simulated gives me a negative resistance slope for the V-I characteristic plotted,which is a property of an electric arc(current increases and voltage decreases).
The aim is to design an electronic circuit with the above mentioned property of an electric arc.

 

[AshwinNambiar]

I have attached an image of the V-I characteristic of an electric arc.Now,I'm looking to design an electronic circuit which gives me a similar plot,thereby calling it an electric arc simulator circuit.Looking forward for help and suggestions to design a circuit for the same.

 

[(*steve*)]

You realise you'll be designing a load, not a circuit that "if i give Vin=130v DC,the circuit designed should be such that the output current Io increases with a decrease in the output voltage Vo"?

Have a read of this. There are limitations to any practical negative resistance.

 

[(*steve*)]

You might consider a buck DC-DC converter with a constant load. For the voltages and currents you;re talking about, you may be able to use a PC power supply!

 

[Harald Kapp]

Within a simulator you can use this technique:

I=31020*exp(-0.064*V(va)) approximates the curve that fits the value pairs from your first post.
You willl be hard pressed to realize such a circuit in reality.

 

[AshwinNambiar]

Yes,precisely said,I'll be designing a "load". I'm looking forward to design a non-linear active load to serve the purpose.

 

[Harald Kapp]

Building such a circuit from analog components is possible, but no easy feat. If you look at TV circuit diagrams from pre-integrated-circuit times, you can't stop wondering how engineers in those days deisgned clever circuits to achieve every possible (and impossible ) waveform / shape.

However, today one is tempted to tackle this differently, namely in the digital doman. How fast does your voltage and/or current change? If the change is reasonably slow, say less than an equivalent frequency of e.g. 1kHz (don't nail me on that number), a microcontroller can be used to give you almost any V/I relationship you want:

• Build an active load (here is an example) that allows you to control the load current e.g. by applying a control voltage.
• Measure the voltage across the active load using e.g. a voltage divider and an analog input (ADC) of a microcontroller.
• Have the microcontroller compute the desired load curent using a suitable equation or looking it up in a table,
• The microcontroler then outputs a control signal to the active load to set the desired current.

Any relationship can be programmed. The circuit can even be improved by adding a feedback loop where the control signal from the microcontroller is dynamically adjusted to compensate errors in the active load. Depending on the sped of the microcontroller, the AD-conversion and the setling time of the active load you may achieve a satisfactory behaviour up to a few kHz.

 

[AshwinNambiar]

In the above shown simulator model,the one in which you have used a behavioural source,how did you plot the Voltage Vs Current characteristic.I 'am having a difficulty,the plot for ant circuit i design gives me a Voltage Vs Time and Current Vs time,not wanting it that way,I would like to how how did you get to plot the Voltage Vs Current characteristic.

 

[Harald Kapp]

Plot I and V of the component you want to analyze (this will at first result in V vs. time and I vs. time).
klick on the x-axis to get thsi wiondow:

Replace "time" by e.g. V(n001) or whatever voltage you want on the x-axis. Or by I(D1) or whatever current you want on the x-axis.

 

[AshwinNambiar]

Well ,are you suggesting the pwm technique when you mentioned the use of micro controllers??
The purpose for me asking this is i've taken baby steps to working on an arcjet thruster load.What should be done is:
1.Designing an electronic load such that it gives me an output characteristic similar to that of the arcjet thruster load i.e an equivalent electronic design
2.There is this specific characteristic of arcjet thruster that the arc when created has static VI characteristic(negative resistance slope),the target is to achieve the same for an electronic load design.
I could download a paper on the same but the problems faced:
1. The circuit designed has many of it s components obsolete.
2.The paper is as old as published in 1991
3.I'm unable to get an equivalent of the components that are obsolete.
I went about with the literature and trying to come up with a new design of the electronic load,the aim being to achieve the same V-I characteristic.
The link to the paper, http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19920008291.pdf

Page no.3,shows the characteristic and page no.11 has the circuit design with the then existing components.I'm not looking to do everything that is in the paper,just the electronic load design with the above mentioned characteristic.

Looking forward for help,guidance and suggestions.
Thanks

 

[Harald Kapp]

Well ,are you suggesting the pwm technique when you mentioned the use of micro controllers??

No, that is not what I wrote.

1. The circuit designed has many of it s components obsolete.
2.The paper is as old as published in 1991

Anyway, that paper and design can still give you the idea how to realize such a circuit using modern components. the importnat thing here is not a 1-by-1 replacement of obsolete components by new ones. It's about understandingthe principle behind the circuit and designing anew one based on these principles. Teh design uses opamps, transistors, a linear 4-quadrant multiplier (could be replaced by an AD633) plus a few passive components - nothing that's not available today.

 

[AshwinNambiar]

Been on that,trying to understand the principle behind the circuit.
From what I've understood ,the multiplier and the integrator circuit i.e the PI compensator ,comes into picture once the final stage circuit has been designed.
Now, all along it has been mentioned 16 paralleled transistor circuit(using IGBTs) .The thing is the "input to this final stage ( the circuit with 16 paralleled IGBTs) or its equivalent is 100v ,10A and across the output there is a decrease in the voltage and increase in the current.
That is where i'm stuck up and have used this platform to know how can i design a load where i give an input of 100v DC,10 A and obtain an output where the current increases with the decrease in voltage.
The other difficulty is the IGBT used in this paper is also obsolete,so when i design a similar circuit(on MULTISIM or SPICE) ,the values of Rd,Re and other passive components also change. I've failed to find papers or a link that will help me learn how to bias an IGBT circuit. I've not been able to find out how to calculate the values of resistors to be used for a particular IGBT.
The main question is how do i come up with a design such as 100V,10A at the input and the output i have a current that increases exponentially as the voltage decreases.
I believe there is no flaw in my understanding and if so will be glad to get it corrected.