Voltage Reglator

[adrianvon]

Hi all,

I have a 45V supply which can vary down to 20V. I want to use a 7912 IC to regulate this voltage to 12V. Which is the best way to drop the 45V so not to exceed the 35V maximum input voltage of the 7812 ?? A transformer will not work since it will drop the voltage too low when the input AC voltage goes down to 20V.
Can I use a zener diode at the input or will it decrease the efficiency?

All kind of help will be appreciated.

Thanks in advance.

 

[audioguru]

Anything in series wastes a lot of power by getting HOT. Even 20V is too high for a 7912 if the current is high at 1A. It will dissipate 8W which is a lot of heat to get rid of.
If you use a 10V zener diode then when the input drops to 20V the 7912 gets only 10V and WILL NOT WORK. But if the input is 45V and the 10V zener drops it to 35V, the poor little 7912 will dissipate (35V - 12V) x 1A= 23W which will melt it.

 

[JimB]

I want to use a 7912 IC to regulate ...
...maximum input voltage of the 7812

7812 or 7912?
7912 is a -ve voltage regulator, so I assume that this is a typo.

You could use a zener diode in the usual configuration, with a series resistor and then the zener from the resistor to 0v.
Using the zener to give a 20v supply for the 7812 will make life much easier for the 7812.

The whole thing will not be particuarly efficient and will get a bit warm.
It really all depends on how much current you need at 12 volts.

JimB

 

[Mike odom]

Normally, I would put a power resistor in front of the regulator, to remove power from it. Set the value so that at max current it drops most of the overhead voltage, only leaving how ever much (1.5V for 78xx) the regulator needs. Normally. But, like AG said, it will get hot and waste a lot of power as heat (better for the resistor to do it than the regulator to go into thermal shutdown). Normally... however, you can't even start to do calculations with a input voltage range that extreme. Is it like a transformer and has 45V when low to no current draw and 20V at max current??? That you could figure on, but 20-45V input randomly?

The only way out that I see is to use a switching regulator that will take a 20-45V input and give you the 12V output you're looking for.

Also, you wouldn't use a 10V zener, you'd use a 30V zener, to keep the max below the max of the regulator, and if the input voltage falls below the zener voltage, then it'll just follow it down and not be used.

Anything you put in front to drop voltage will make it less efficient. This is the exact kind of situation they invented switching power supplies for.

 

[NorthGuy]

You can use this module which will convert your voltage to 12V without wasting 3/4 of the power.

 

[crutschow]

What's the maximum current you need?

 

[Ian Rogers]

You can get an LM2576HVT-12 I use these on forklifts... They can withstand 60V... As some trucks are 12V some are 24V and some are 48V.... As they are switchers the efficiency is pretty good....

https://uk.rs-online.com/web/p/dc-dc-converters/7615501/

 

[adrianvon]

Hi,

Thanks for the replies. Can I use a power transistor at the input as shown in the link below (in the Example linear regulators section) ?

https://en.wikipedia.org/wiki/Voltage_regulator

 

[JimB]

Yes, but it will still get hot depending on how much current is drawn.

JimB

 

[adrianvon]

How can I calculate power in the transistor please? assuming say the load is 12V 1.5A.

Also, if I use two transistors, one to drop the voltage from 40V to 28V and another one to drop the voltage from 28V to 18V, will that be a bad design?

 

[NorthGuy]

How can I calculate power in the transistor please? assuming say the load is 12V 1.5A.

Also, if I use two transistors, one to drop the voltage from 40V to 28V and another one to drop the voltage from 28V to 18V, will that be a bad design?

To calculate the power drop, multiply the voltage drop by current. (40-12) x 1.5 = 42W. It'll happen no matter how many transistors or linear regulators you stuck up. The only way to avoid this is to build/buy a swtching converter, which is not much harder/expensive than linear.

 

[adrianvon]

Thanks for the reply.

It'll happen no matter how many transistors or linear regulators you stuck up.

Yes, but at least the power is shared between the two transistors so they will not get too hot.

 

[adrianvon]

If a transistor have a Total power Dissipation of 150W MAXIMUM, that is safe enough to use, right?

 

[NorthGuy]

It depends on the package. A transistor in TO-220 package cannot take more than 2-3W without heatsinking. 150W figure probably assumes an adequate heatsink.

 

[crutschow]

If a transistor have a Total power Dissipation of 150W MAXIMUM, that is safe enough to use, right?

That is a very liberal rating, generally assuming that the transistor case is maintained at 25C, which requires a very large heat sink or liquid cooling. Even at lower power levels you still need a heatsink to keep the transistor junction (not case) temperature below 100C for good reliability.

 

[adrianvon]

Thanks. So is using two transistors be a good option? One to drop the voltage from 40V to 28V and another one to drop the voltage from 28V to 18V ?

 

[kinarfi]

I agree with Ian Rogers, once you become familiar with LM2576HVT-12 or LM2576hvt-adj model, you'll find them much handier than linear devices and depending on how you work the feed back, they can be constant current devices. you only need to add 4 component and a pot if you use adj. and not nearly as much heat.