Hi.

I plan to make a pcb that drives peltier and heating pad. Only one of those works at the same time.

Load elements are parallel to capacitors. I have very little space on pcb so it would be nice if I could use only one inductor as it is in screenshot. I know it would be easier just to put two inductors.
Frequency will be 20kHz through low pass filter to act as variable constant current source.

My worry is the current flowing through inactive side diode and capacitor.
Would that be a problem?

Any advice is appreciated

 

Have you tried simulating this, for instance in LTspice? You'd have your answer. I think I understand your schematic but I'm not sure about D2 and D4.

 

That circuit won't work, since no DC from the inductor can go through the capacitors to the loads.

The top circuit below is a basic design that uses PMOSFETs, requiring only one inductor, diode, and capacitor.

The bottom circuit uses N-MOSFETs for M1 and M3, if the loads can float.

The P-MOSFET control signals must go between 12V (off) and 0V (on).

 

That circuit won't work, since no DC from the inductor can go through the capacitors to the loads.

I believe the heater+ and peltier+ are connected to common (+12V).

 

Yes, I assumed the heater+ and peltier+ are connected to common.
But why D2 and D4? - as the loads are not inductive. And why C74 and C75? These are going to cause massive current surges when the MOSFETs switch on.

 

So still don't see how it can work.

 

It won't work - once C74 and C75 charge up, absolutely noting happens. Look at @crutschow 's post #3.

 

Ok now my curiosity is piqued, so I will have to do a sim.

 

I know it would work with two inductors (attached picture) but it would be nice if I didn't need second inductor (space, cost). This is also an opportunity to learn new principles. I don't trust simulations. In case with one inductor (my first post) inactive capacitor is charging through active inductor (one side is turned on, other is turned off) so I assume that inductor oscillations will leak through capacitor and discharge in flyback diode. Correct me if I am wrong. If this is right, would these leaks cause problems?

 

The circuit seems to work, but there does seem to be an issue with current in the load that is supposed to be off, but that could just be a problem with the sim.

I could have made a mistake in the circuit, and I used the components available, not the ones listed.

And something odd, the diode that conducts when a given MOSFET is on, is not the one you might expect.

 

I guess that only solution is to cram second inductor on the pcb somehow

 

Your revised circuit runs without the issues.

 

I don't see why it wouldn't work with a single inductor as hoped. And now that I think about it, one capacitor (and one diode, if it's doing anything) should be enough as well since only one load is turned on at a time. My LTspice simulation failed with a floating node and I can't find it. Any advice on that?

 

It won't work - once C74 and C75 charge up, absolutely noting happens. Look at @crutschow 's post #3.

The load(s) are in parallel with the capacitor(s).

 

I am tempted to blame the schematic...thank you for your correction, if that is the case.

 

I will just go with dual inductors as I should in the first place. Everyone thanks for help