I am trying to make a regulator that will regulate its input voltage rather than its output voltage.
I am using a LDO to work out the kinks in the control logic and keep the diagram simple. I want to get this working on a buck regulator eventually.

1. Use input voltage and link it to the adjust pin so the regulator can control it.
2. Take input voltage and first invert it so when Vin increases a falling voltage is seen on ADJ pin and the regulator passes more current thus bringing Vin down (and vise versa)
3. Compensate for a negative inverted Vin signal coming from the op-amp by adding a voltage to it giving ADJ an positive and correctly correlated response to Vin. I attempt to do this using a differential (subtracting) circuit.

So far this isn't working. Some help on how to make a functional switch mode shunt regulator would be appreciated.

 

What you are trying to do is not likely to be successful, since a voltage source will fight you at every turn. But hey, good luck with your quixotic quest. Tilting at windmills seems to be an affliction among a seemingly inexhaustible supply of tinkerers.

I'm not familiar with the LT3062, so I'll take a look at the datasheet and get back to you.

 

OK, so here is the deal:

1. You CANNOT control the ADJ pin independently of the output. The device is designed to set the voltage on the OUT pin such that the voltage on the ADJ pin is at 0.6V. This is subject to the condition that the voltage on the IN pin is above the voltage on the OUT pin by the dropout level for the current being drawn.
2. Feeding a voltage regulator with a current source is a sketchy thing to do. where did you get that quaint notion.
3. Are you an engineer or just a hobbyist, messing around with stuff you don't understand?

Se figure 1. in the datasheet for how to set the output voltage.

 

Here is a quick simulation that illustrates the problem. The design is for a 3.3V Regulator. the input voltage needs to be about 200 mv above that 3.3V output level for the device to regulate. The green trace shows an insufficient input voltage, and the blue trace shows an apparently sufficient input voltage of 3.5V to achieve a 3.3V output. The current into the adj pin is sufficiently small so as to have a negligible effect on the output.


This simulation should make it clear that you can never make the voltage on the OUT pin come closer than about 200 mV to the IN pin, no matter what you do to the adj pin. An opamp output could overwhelm the input buffer and drive the chip bat sh*t crazy. You would not be happy with that result.

 

If you connect the output to GND then the input will always be higher due to voltage drop across the power transistor (in the case for a LDO). If the input is high enough the LDO should be functional in this case 1.6V or above. If the output voltage is always zero then why couldn't you develop the inverse feedback system (described previously) that forces it to regulate its input voltage. I don't see any fundamental rules I am breaking here. When this works it will modify the base voltage on the base of its power transistor to dump enough current through the output to maintain its input voltage.

 

I really do not understand what you are trying to do. Can you please explain in a bit more detail...
What is the application?
What is the input power source, and its voltage range?
Why do you need to change it?
In my mind, the input could be from a car battery. Good luck trying to change that battery's voltage without a lot of smoke.

Are you looking to add a pre-regulator between the power source and the output regulator to control the voltage on the input of the main regulator?
That is sometimes done using a switch mode regulator for efficiency, driving a linear regulator for low noise.

 

Trying to control the input voltage sounds like you have higher impedance power supply like Solar panel. Then you need to think the regulation as an alternating load to be able to keep the input voltage steady. This is often used in solar charge controllers or load dump regulators.

 

A voltage regulator is a circuit that creates and maintains a fixed output voltage, irrespective of changes to the input voltage or load conditions. Voltage regulators (VRs) keep the voltages from a power supply within a range that is compatible with the other electrical components My Insurance Info

 

Just put a zener diode across the input. That will work, provided you get a zener diode that will dissipate enough power. What is the source impedance?

 

If you connect the output to GND then the input will always be higher due to voltage drop across the power transistor (in the case for a LDO). If the input is high enough the LDO should be functional in this case 1.6V or above. If the output voltage is always zero then why couldn't you develop the inverse feedback system (described previously) that forces it to regulate its input voltage. I don't see any fundamental rules I am breaking here. When this works it will modify the base voltage on the base of its power transistor to dump enough current through the output to maintain its input voltage.

Most regulators will have some connection to GND through a load. How do you imagine that this will result in the output voltage always being zero. Inverse feedback will NOT cause a regulator to regulate the input voltage. An LDO is NOT bidirectional. There are some types of SCC (Switched Capacitor Converters) that are but that is not the case here. It is erroneous to assume that there is a BJT inside the LDO since a BJT is not easy to fabricate with a CMOS process. Lastly dumping current through the output does nothing to maintain the input. You seem to be entertaining a number of erroneous ideas and concepts and I am unsure of how to help get you to a better place, so I intend to bow out of this thread since it appears I have no useful help to offer. It is fine with me if you want to continue your investigation and maybe somebody smarter than me will understand what you are trying to do. If that happens, I will be the first to admit that I was wrong.

 

If you want to regulate a constant-current to give a constant voltage, then you can use a simple shunt regulator.
Can you explain the reason for the complex circuit you proposed?

 

Simple really.

Just connect a regulator to a power source and rename the regulator output from "output" to "input" (to next stage).

Now you have a regulated input.

 

You need to explain what this input source is, and what you are going to do with the "regulated" output.

Some posters have assumed you are using a high impedance voltage source or a current source, but I have not seen you confirm either of these hypotheses. If you are using a low impedance voltage source, then regulating it down to a lower voltage means overloading the power supply and either damaging it or having it shut down.

So what is the power source you are trying to regulate and why?

 

Sorry for the confusion all. The source is the rectified output of a permanent magnet generator. This will create a proportional amount of current related to the RPM of the generator. The idea is to essentially create a Zener diode (functionally at least) but with the efficiency of a sync buck regulator. The way I want to go abut this it to force the buck regulator to look at its input voltage (yes assuming it is regulating above the minimum input voltage required to run the regulator itself) by connecting its output (Vout) to GND. It should pass as much current necessary through it to maintain its set input voltage. Identical to a Zener.

 

What is supposed to happen to the energy from the generator?

 

The "Zener diode" cannot be a shunt switch mode device as it will have to clamp the generator's voltage continuously, not switch on and off, so the efficiency you are looking for can't be done that way. To make use of the generator's internal resistance to drop the excess voltage, current will have to flow, and there will be power dissipated. So, a power transistor and Zenner on a heat sink would be a way to go. Or just go with a buck/boost switch mode regulator to control the voltage to the load and not worry about the actual generator voltage.
What is the ratings of the generator?

 

The energy gets shunted back to the generator. Theoretically the only energy loss is the inefficiency of the buck regulator itself and the rectifier the generator interfaces with the regulator through . Nothing else in the circuit is absorbing energy. Assuming wire resistance is not in play.

 

The energy gets shunted back to the generator. Theoretically the only energy loss is the inefficiency of the buck regulator itself and the rectifier the generator interfaces with the regulator through . Nothing else in the circuit is absorbing energy. Assuming wire resistance is not in play.

If you don't want to do anything with the energy, then what's the point? Just let it spin - the energy will never leave the generator.

 

if..

The energy gets shunted back to the generator.

that will raise the voltage of the generator. Tha actual thing you are trying to stop.

 

Guys it is a permanent magnet generator. Look these up please. They create the same H field whether or not the generator is connected to anything. If the current doesn't have a path from one end of the stator winding to the other the voltage across the coil (theoretically) will be infinite. If the coil is shorted (one end connected to the other) the voltage across it will be zero and the current through the coil will be an equivalent to the magnitude of the changing H field.

1. You can't just isolate the stator after it passes through the rectifier the voltage it will produce will be DC and INFINITE

2. Connecting or "shunting" the current back to the generator will reduce the voltage across the generator not increase it. Voltage need resistance in order to form. Essentially shorting the stator will produce 0V across it.

3.The point of this thread is to design a switched mode shunting regulator. Before another person says this is impossible... this is already done with SCR circuits in motorcycle rectifiers. I'm trying to make a more efficient version of one.