lm338 voltage regulator

[maicael]

Hello
I wanted to power a lot of IC's
I wanted to use the 7812 and parallel them but I changed my mind.
The lm338 can give out up to 5A so it should be OK for my IC's
Now if a input a 24vdc into the lm338 and set an output of 12vdc
If the input supply drops to say 20v will the output still stay at 12vdc?
By the way I am powering a lot of IC's but just enough for the 338 to cover.

 

[Ian Rogers]

I don't think people think too much about the old linear power supplies!! They do need quite a bit of thought.

If a LMxxx states that an output of 3A is possible, then the conditions specified must be met.. Imagine 24v converted to 12v delivering 3A... That's quite a voltage drop for that kind of current... If the same convertor has to supply 12V at 3A but has a much lower voltage... say 15V. Then the convertor will stand more of a chance.

If the max current was specified being 3A.. I doubt it will run too good with an input voltage of 24V..

I had an LT1129 3-3 running off a 12v supply... Even with a decent heatsink it was burning up.. I bought a 6v adapter and the little device cooled down somewhat!! Spec says 700mA... I was only pulling 350mA and it nearly fried at 12v...

 

[crutschow]

Note that the power dissipated in the regulator is the voltage drop times the current. Thus the drop is 24v - 12V =12V and the power dissipated for 5A of current is 12V * 5A = 60W .
That will require a very large heatsink for the regulator.

It takes a heck of a lot of ICs to require 5A of current.

 

[maicael]

Note that the power dissipated in the regulator is the voltage drop times the current. Thus the drop is 24v - 12V =12V and the power dissipated for 5A of current is 12V * 5A = 60W .
That will require a very large heatsink for the regulator.

It takes a heck of a lot of ICs to require 5A of current.

Well I don't think the IC's will consume even up to 2A. 5A is the peak outpt of the IC.its just that I don't want to parallel 7812's as each may slightly differ and current sharing becomes unequal.
This is why I want to use this alternative IC lm338.
What about the question about if the input drops to 20v will it still maintain a 12v output

 

[audioguru]

The LM338 datasheet shows that it has excellent voltage regulation. If the input voltage changes then the output voltage changes a tiny amount. If the load current changes the output voltage changes a tiny amount.
If the output is 12V then it will "dropout" when the input drops below about 14V, so an input of 20V is fine.

 

[maicael]

The LM338 datasheet shows that it has excellent voltage regulation. If the input voltage changes then the output voltage changes a tiny amount. If the load current changes the output voltage changes a tiny amount.
If the output is 12V then it will "dropout" when the input drops below about 14V, so an input of 20V is fine.

Hello
But what about mr crutschow assertion about power dissipated.
Although I don't expect the IC's consuming all that amount of current.
Aslo if going by his assertion then if I used a 7812 powered from an 24vdc source delivering 1A current that does mean that power dissipation is 24-12=12
12*1 = 12W right?

 

[granddad]

What are these 12v IC 's only ones I can think of are opamps or audio amps , but they need quite different psu, cmos can run at 12v , but a 100 would only consume give or take 1 A.

 

[alec_t]

The planned current draw is surprisingly high. Approximately how many ICs are you planning to supply, and what type?

 

[audioguru]

We do not know why you want to make a lot of heat in a 12V regulator by powering it from a voltage that is too high at 24V. You need only about 14V or 15V for the input of the 12V regulator, then with a 1A load the heating of the regulator will be only 2W or 3W.

 

[maicael]

Hello
Well approximately the IC's will be 15 in number ranging from comparator's to oscillators to 555 timer.
I think i will use a lower input voltage to make things easier

 

[audioguru]

Many circuits do not need voltage regulation of their power supply.

 

[maicael]

Many circuits do not need voltage regulation of their power supply.

So this means that with just proper filtering or bypassing the IC's can be powered directly.
Well this makes sense as I just realised that if I powered an IC like ir2110 for example that can source up to 2A from a regulator then at 24 V input to 12v out I will waste as much as 20watts of power.
Not ideal so I power directly from battery

 

[audioguru]

If the input Vcc to the IR2110 is 24V (or a little more if it is a lead-acid battery) then its gate drive output is also 24V or a little more that will destroy many Mosfets that have an absolute maximum gate-source voltage rating of only 20V or less.
The IR2110 sources 2A only for the very short durations that it is charging the gate-source capacitance of a Mosfet. Its datasheet shows that its current from Vcc is typically only 180uA (340uA max) most of the time.