Emulator Design

[FC-Emulator]

Hi

I'm designing an electronic emulator for a (electrochemical) fuel cell. The design must mimic the behaviour of the fuel cell (see attachment). From the curve, the voltage is inversely proportional to current (which does not satisfy Ohm's Law: V is directly proportional to I).

I tried few approaches but can't get to agree with the characteristic curve.

May you kindly offer some ideas of how to approach this design.

Thank you

 

[ericgibbs]

Hi

I'm designing an electronic emulator for a (electrochemical) fuel cell. The design must mimic the behaviour of the fuel cell (see attachment). From the curve, the voltage is inversely proportional to current (which does not satisfy Ohm's Law: V is directly proportional to I).

I tried few approaches but can't get to agree with the characteristic curve.

May you kindly offer some ideas of how to approach this design.

Thank you

hi,
Is this a software or hardware cell emulator.?

If is software/program look at LTspice simulator from the web, its a free download.
It has a PWL parameter option for unusual curve tracking.

 

[dougy83]

Does your circuit have to emulate the cell charge (i.e. a load), discharge (i.e. a source) or both?

 

[FC-Emulator]

It is hardware emulator (that will work similar to a battery).

The circuit have to emulate the cell discharge.

 

[dougy83]

Here's an idea of how I might approach it.. The 3 regions can be represented by 3 separate thevinen sources. The sources are switched to the output depending on the load using an analogue multiplexor. See attached cct. for an example. To ensure that the regions are accurate and line up with one another, I would suggest using low-value pots (or opamp-buffered pots) for the voltage source of each region, and another pot for the series resistance of the region (allows you to adjust for the resistance of the MUX also). It's a bit of a crazy design, I know, but it's pretty late here..

 

[dougy83]

Another dumb idea. You can use a lookup table stored on an (E)EPROM, referenced to the output current, driving the output voltage. If you offset the voltage into the ADC, you can provide the response to allow the cell to act as both source and sink.

 

[FC-Emulator]

I looked at both design and I still cannot figure out how they work. May you break it down for me. Thankx

 

[dougy83]

They are just rough sketches of 'aproaches' you could take to obtain the emulator. Both measure the current through the load by amplifying the the voltage across a shunt resistor, and use the obtained voltage to select what to connect to the output.

The first sketch depicts a mux (multiplexor) switching a different thevenin equivalent source to the load (Depending on the current, the voltage to the 2 comparators on the left will connect the mux input 0,1 or 3 to the mux output). To work out the correct thev. voltage and resistance, extend the lines on your transfer drawing until they intersect the y axis: this gives the thev. voltage; the thev. resistance for each region is given by the slope of the line within each region. There is 3 regions in your graph, hence 3 thevenin sources.

The second sketch uses a analogue to digital converter (ADC) to index a lookup table stored on an eprom to provide a programmed output voltage through a digital to analogue converter (DAC). The eprom may be programmed to give any weird transfer function you could think of.

 

[FC-Emulator]

Thanks for the explanation...

 

[FC-Emulator]

Can you help me with finding the dc voltage component on LTSpice IV

 

[dougy83]

If you're talking about the thevenin voltage, simply extend the lines in the graph you posted. Where they cross the y-axis, that's your thev. voltage. e.g. region 1 is ~2.25V, region 2 is ~1.15V.

 

[FC-Emulator]

No,actually I need the (common) dc voltage component.