replace adjustable regulator with fixed?

[KevinLee]

Hello,

Is it possible to replace an adjustable regulator (like a LM317 ) with a fixed regulator?

The LM317 in circuit has +6v on the output, I'd like to drop in a +6v Dexa.

NewClassD UWB Regulator

I cannot think of any reason why this would not work, I just want to make sure I am not missing anything.

cheers
Kevin

 

[sofaspud]

The LM317 has external components not required by fixed regs. You'll need to remove those after determining that they aren't needed for anything else. And the LM317 has a different pinout than its fixed output cousins.

 

[AudioSan]

you need to flip the fixed ones 180degree.

you need to remove the trimmer and a resistor or 2.

 

[Conrad Hoffman]

The LM317 is a really good performer if properly bypassed, better than its fixed cousins. No doubt one can do better with a discrete design, but that's a lot of money to replace a 317. It's debatable if it will have an actual measurable benefit, but after spending that much, I'd certainly be convinced that it did.

 

[jackinnj]

The NewClass D performed better (sonically) than the LM317 -- see the new issue of "Linear Audio"

 

[RJM1]

The NewClass D is only rated at 630mA where the LM317 is rated at 1A. Make sure you don't draw too much current.

 

[AudioSan]

The NewClass D is only rated at 630mA where the LM317 is rated at 1A. Make sure you don't draw too much current.

it's the LM78xx that is rated at 1A. LM317 is rated at 1.5A.
but they will work with higher currents to if one have som "bigish" sinks and
close rail voltage. i run close to 2A out of a LM7812. with 15V in.

 

[sofaspud]

you need to flip the fixed ones 180degree

That wouldn't account for the center pin, which is Ground for the fixed types and Out for the LM317.

 

[AudioSan]

yes you are right

 

[AndrewT]

i run close to 2A out of a LM7812. with 15V in.

What's the specification for an LM7812?

Are you suggesting that others copy you and operate devices outside the manufacturers specification?

 

[sofaspud]

Doubtful it's continuous unless there are some heatsinking secrets to tell, but the current-limit curves do show it possible to pull a couple of amps through the regulator.

 

[AudioSan]

the heatsinks has little to do with the current passing through it.
but more the voltagedrop. if one run 25V in and 12V out. you will need some "big sinks" to pull any current. but with 15V in and 12V out, thats a whole other story.
lets say 1A current. at 25V in. thats 13W dissipation. at 15Vin its only 3W.

 

[sofaspud]

I'll agree, but there is much to do with junction temperature (via the in-out differential, ie power dissipated in the reg, ie Tj), so heatsinking is quite relevant. There's current limiting and thermal shutdown to consider, so AndrewT is correct - it's treading into no-man's land.

 

[AudioSan]

there is. but the difference between 1A and 2A is kids candy.
the difference from 3Vdrop to13-18V drop is a much bigger issue.
all diveces reach a maximum. but the main reason that a regulator fails, is voltage drop(dissipation). not the current that it draws.

 

[AudioSan]

lets say a worst case. LM317 with 35V in and 5V out. at 1A. hell. thats 30W dissipation
not a easy task for a to-220die

 

[AndrewT]

Audio
What pre regulation do you have for the implementation you have described in post7?

How much does the Vin vary?
How much could Vin vary for worst case operation?

Does Vin, with some ripple, drop low enough for regulator drop out to occur?

 

[AudioSan]

Vin is pretty darn stable.
worst case operation? do you think of in regards to heatsink?

 

[AndrewT]

Well, guess what, I disagree with "pretty darn stable".

For 15Vdc @ 2Adc, I'd use 230:12Vac, 100VA transformer operating from 240Vac, feeding a bridge rectifier and 4u7F of smoothing.

Apply a -6% mains voltage variation. i.e. 225.6Vac for a check on whether regulator drop out occurs.
Input to regulator becomes a worst case low of about 14.8Vdc +-200mV of ripple. Giving a Vin from 14.6Vdc to 15Vdc.

Now apply a +6% mains voltage variation, i.e. 254.4Vac. to check for maximum Pd.
Input to regulator becomes a worst case high of about 16.72Vdc +-200mV of ripple.

For a 2A output the transformer is running at 50% o maximum continuous output and the regulator Pd varies from ~5.6W to ~9.4W

That is not "pretty darn stable".

The questioner should do something along the same lines as I have shown and insert his own data to arrive at his operational Pd and how much it can vary for his/her mains voltage variation.

 

[AudioSan]

i don't know about you. but over here i have never had more then about 2-3% variation on mains.

 

[jackinnj]

i don't know about you. but over here i have never had more then about 2-3% variation on mains.

Over here the line voltage can dip into the mid 90's and run as high as 128 during the summer == i had our electric company put a data logger on the main panel.

We had a competitor whose a.c. power was cut in the summer of 2007 -- their thermal curing ovens were powered by natural gas and didn't cut out quickly enough during the initial stage of a brownout -- what a mess!

If it's a low voltage application, I would try replacing the LM317 with one of the fixed voltage regulators in the LT1963A series -- there are 4 fixed voltage choices.