How to smoothly switch active between constant I and constant V?

Hi - I'm working on a active load that will be used for battery testing (among other things). A standard means of discharging a battery is the constant current, constant voltage discharge. This means, say with a li-ion, begin by discharging a full cell at a constant current. Once the voltage of the cell hits a minimum voltage you switch to constant voltage and let the current taper off till a set amount.

My plan is for my load element to be a FET on a large heatsink. A microcontroller controlling a DAC that also has various IOs available will be controlling everything. I could have two DACs if needs be.

I can easily design a DAC controlled constant current or constant voltage load. But I don't know how to make a system that can transition between the two. Any suggestions?

Imagine a PI controller and two error signals (voltage and current) that are combined by a max() respectively min() function.

FvM said:

Imagine a PI controller and two error signals (voltage and current) that are combined by a max() respectively min() function.

Click to expand...

Are you talking about implementing a PI controller in software or in hardware?

In hardware I am having a lot of difficulty imagining how I can create my error signal. I mean at times I want my error signal to be dependent on I, other times dependent on V. I don't think I ever really want it dependent on both.

Any suggestions? I just can't quite figure it out...

Are you talking about implementing a PI controller in software or in hardware?

Click to expand...

I understood your question related to analog hardware, but the solution applies to both.

In hardware I am having a lot of difficulty imagining how I can create my error signal.

Click to expand...

Why? Lack of imagination? Error signal is the difference between setpoint and process value, in soft- and hardware controller as well. A max() or min() function can be basically implemented by two diodes. Perhaps you should look out for existing constant V/constant I examples. Take e.g. a look at the TL494 datasheet, how the device combines two error signals.

FvM said:

Why? Lack of imagination? Error signal is the difference between setpoint and process value, in soft- and hardware controller as well. A max() or min() function can be basically implemented by two diodes. Perhaps you should look out for existing constant V/constant I examples. Take e.g. a look at the TL494 datasheet, how the device combines two error signals.

Click to expand...

FvM - sorry for the slow response. How would I achieve a PI controller in hardware? I'm not familiar with a circuit that does this.

Thanks!

uoficowboy said:

FvM - sorry for the slow response. How would I achieve a PI controller in hardware? I'm not familiar with a circuit that does this.

Thanks!

Click to expand...

Anybody? Sorry to bump...

An inverting OP amplifier with RC series circuit in the feedback path has a PI characteristic.

One way to do it in hardware would be to build a constant current generator using your FET power transistor. So far so good. Now at your voltage setpoint, you have to divert the CC into a shunt circuit so its diverted from your battery. rather like a super zener diode, but adjustable.
Frank

FvM said:

An inverting OP amplifier with RC series circuit in the feedback path has a PI characteristic.

Click to expand...

Sorry, but I'm afraid I don't understand. Could you draw a simple circuit that achieves this? Wouldn't the inverting OA have zero DC response then?