Voltage Limiter Circuit

[Jdo300]

Hello All,

I'm just curious to know if there is a way to rig up an adjustable voltage regulator so that it will output whatever voltage is applied to it up to a point?

I have a digital circuit that is powered from a large capacitor. In the vast majority of cases, the voltage on the capacitor may vary from 5 to 15 volts, which I can pass right to the ICs that it is powering (MOSFET drivers that can handle up to 18V). However sometimes, the capacitor voltage could rise as high as 20-24V which would obviously kill my ICs. My initial thought is that I could just put a fixed regulator of, say 9 or 12V on there but my driver ICs will switch faster at the higher voltages so I don't want to simply waste the extra voltage if it's within the supply range of the ICs. So I'm wondering if there is a way that I can rig up a voltage regulator to let all voltages through up to the 18V limit that the ICs can handle and then regulate voltages that go higher than that down to 18V?

Thanks,
Jason O

 

[Averagesupernova]

How about a low dropout 18 volt 3 terminal regulator?

 

[Jdo300]

That wouldn't work for my application because if the voltage goes below 18V then it will just shut off completely. I don't mind if the input voltage is below 18V but I want to regulate it only if it goes above 18V.

 

[NoTime]

Lots of possibilities.
Choices get limited somewhat by just what is causing the voltage change.

Pass thru - at the simplest a transistor resistor and zener.
Crowbar (variable load)
Both minimal effect below set voltage but will be dissipating power at 24 volts.

Switching variants of pass thru.
Buck boost.

 

[zeitghost]

That wouldn't work for my application because if the voltage goes below 18V then it will just shut off completely. I don't mind if the input voltage is below 18V but I want to regulate it only if it goes above 18V.

I've found in the past that low drop regulators don't turn off completely as the input voltage decreases, but remain conducting (but essentially unregulated) down to quite low voltages.

The gnd pin current goes up a lot as the series element is turned on as hard as possible.

You may like to investigate this further.

 

[finenil]

i just give the idea...
all you need is number of zener diodes man.so that their total sum of reverse voltages is 18 volts.
this is the simplest limiter. for Vin>18 the zeners will turn on and sink current and Vout reamins at 18 volts.for Vin<18 the zeners are just not enough reverse biased and Vout=Vin . .
you should certainly consider the effect of loading and amount of current it sinks on your zenner circuit design

 

[Averagesupernova]

Go with the low dropout regulator. I've yet to find a 3 terminal regulator that simply drops the output voltage to zero when the input goes under the rated output voltage. I think you are mistaken on the assumption that it does.

 

[deakie]

Thinking about this, i wondered the following...could be worth the design...

arrange a variable regulator to have a feedback circuit to the adjust pin.
Sample the input voltage and do a compare against the max voltage and use it to control a current mirror on the adjust pin.
this way, there is a continual matching with the input excepting when the voltage matches or exceeds the max voltage you want to pass.
I envisage that and opamp could act as your comparator with a long tailed pair of transistors as your current mirror...a maximum current resistor could be the tail resistor of the long tail...
basically, the opamp replaces a feedback resistor...so there has to be some interacting with the output voltage as well, though there isn't strictly a need for such a resistor on most three pin regs.
anyone else thinks that this is feasable?

 

[Averagesupernova]

I really don't see the point since a voltage regulator IS a feedback device to start with.

 

[klystronman]

maybe a zeener diode would work

maybe a zeener diode in series with a resistor / you would have to calculate the values / seems to me the would limit the voltage and the resistor would limit the current. i do not know. check it out first. good luck

 

[berkeman]

Go with the low dropout regulator. I've yet to find a 3 terminal regulator that simply drops the output voltage to zero when the input goes under the rated output voltage. I think you are mistaken on the assumption that it does.

I agree with ASN. That's how all linear regulators that I know of work. Just look at the equivalent circuit shown in the datasheets.


EDIT -- Keep in mind, though that the Zout of the regulator + output capacitor changes when Vin is below Vout+Vdropout. When the linear regulator is working properly, the Zout is kept very low by the feedback loop. When Vin is to low for the regulator to stay in regulation, the Zout is only kept low by the output capacitor.

 

[finenil]

I am asking this for my own ...
what's the advantage of the dropout regulator if it drops the Vout to zero when the input goes down?
He wants to maintain the Vout same to Vin if Vin goes under some threshold.So Why a simple zener regulator is not enough?

 

[berkeman]

I am asking this for my own ...
what's the advantage of the dropout regulator if it drops the Vout to zero when the input goes down?
He wants to maintain the Vout same to Vin if Vin goes under some threshold.So Why a simple zener regulator is not enough?

A simple zener regulator would give a similar behavior, as you suggest. Zener regulators, which are shunt regulators, are very inefficient, however. That's why they are seldom used. Series regulators with low-current control circuitry are generally preferred, and usually are not much more expensive than shunt Zener regulators.

 

[ws0619]

A variable resistor,can it be useful in this area?To lower down the voltage.

 

[NoTime]

A variable resistor,can it be useful in this area?To lower down the voltage.

Most of the regulators mentioned are basically variable resistors.
What changes is how the resistance is controlled.

 

[Mike Phan]

Your question: Need a regulated 18V limit. Wheras the input voltage range is from 5V to 24V

Answer: Yes, possible. However,

we cannot have a fixed voltage of 18V because the input migh be lower than your expected output voltage (18V) if we use a buck converter.
For buck, input = 24V, output = 18V. input < 18V, output <18V
If we use a boost converter, input<18V, output=18V. input >18V, output>18V

Final answer: we need a buck-boost converter which is combination of both type of converters mentioned above.

 

[NoTime]

Final answer: we need a buck-boost converter which is combination of both type of converters mentioned above.

There is always the problem that the source may not provide enough power to run the load at 18 volts.

 

[Averagesupernova]

I honestly can't believe we continue to beat this dead horse. The OP said he doesn't care what the voltage does as long as it is prevented from going above 18 volts. My original suggestion is still the answer.