Lowish DO regulator

[MooseFET]

I have a 12V lead acid battery or maybe someone will connect this to a
voltage up to 35V. I need to make 8.3V. Some moron may connect the
power backwards.

Currently I am using a circuit with a P channel MOSFET to do the
reverse protection and an LM78XX type regulator. Does someone know of
a 78XX like regulator with a lower overhead number.

The load current is about 200mA max. I have a large heat sink /
chassis available.


BTW: Did you notice that in the debates nobody ever mentions the need
for better voltage regulators and op-amps. They aren't dealling with
any of the issues that really matter.

 

[Leon]

I have a 12V lead acid battery or maybe someone will connect this to a
voltage up to 35V.  I need to make 8.3V.  Some moron may connect the
power backwards.

Currently I am using a circuit with a P channel MOSFET to do the
reverse protection and an LM78XX type regulator.  Does someone know of
a 78XX like regulator with a lower overhead number.

The load current is about 200mA max.  I have a large heat sink /
chassis available.

BTW:  Did you notice that in the debates nobody ever mentions the need
for better voltage regulators and op-amps.  They aren't dealling with
any of the issues that really matter.  

How much current?

Leon

 

[Leon]

I have a 12V lead acid battery or maybe someone will connect this to a
voltage up to 35V.  I need to make 8.3V.  Some moron may connect the
power backwards.

Currently I am using a circuit with a P channel MOSFET to do the
reverse protection and an LM78XX type regulator.  Does someone know of
a 78XX like regulator with a lower overhead number.

The load current is about 200mA max.  I have a large heat sink /
chassis available.

BTW:  Did you notice that in the debates nobody ever mentions the need
for better voltage regulators and op-amps.  They aren't dealling with
any of the issues that really matter.  

Sorry, I didn't see the 200 mA.

I like the D-PAK LM1117.

Leon

 

[Jan Panteltje]

I have a 12V lead acid battery or maybe someone will connect this to a
voltage up to 35V. I need to make 8.3V. Some moron may connect the
power backwards.

Currently I am using a circuit with a P channel MOSFET to do the
reverse protection

If you have 12V, even if it drops to 10.7, then a simple diode for reverse protection would do.
Below that you better disconnect that battery anyways to protect it.

and an LM78XX type regulator. Does someone know of
a 78XX like regulator with a lower overhead number.

In cases like this I would use the MOSFET as low dropout pass, or
a nomal BJT,

As you are not playing switchmode, then a simple discrete stabilizer would work,
get a cheap reference fro ma low power LM317,
let's see:




PNP power
+10 to + 35 V ---- a diode k ------------------------e c---------------- + 8.3
| | b T1 | |
=== | | | |
10u --- | | | R3 5k9
tant. | | | | |
/// | about +2.4 c c |
LM317L-------------- b b----|
| | | e e |
| R1 150 | | T2 | T3 |
| | | |_______| R4 2k4
|---- | | |
| === 10u R2 ///
R2 150 --- tant. | 150
| | ///
/// /// 2 x NPN

The load current is about 200mA max. I have a large heat sink /
chassis available.


BTW: Did you notice that in the debates nobody ever mentions the need
for better voltage regulators and op-amps.

No, you forgot to specify how accurate the 8.3V needs to be

They aren't dealling with
any of the issues that really matter.

Maybe the above circuit, something I solder together on the desk when I need a voltage,
will work for you.
All component values are guessed.
Use a trimpot for R4.

The way it works:
The LM317L creates a reference voltage, very stable, of about 2.4xx volt.
When output voltage is low, T2 conducts, and base current of T1 is limited by (2.4 - 0.7) / R2.
Then when the output voltage rises, T3 starts conducting when its base get to 2.4 V, and
lifts the emitter of T2, decreasing drive currrent in T1.

The circuit has a low dropout of .2 V, the saturation voltage of T1.

 

[Joerg]

I have a 12V lead acid battery or maybe someone will connect this to a
voltage up to 35V. I need to make 8.3V. Some moron may connect the
power backwards.

Currently I am using a circuit with a P channel MOSFET to do the
reverse protection and an LM78XX type regulator. Does someone know of
a 78XX like regulator with a lower overhead number.

The load current is about 200mA max. I have a large heat sink /
chassis available.

I'd consider a switcher here. 35V minus 8.3V at 200mA is over 5W.
Personally I stay away from LDOs, seen too many not so well or not at
all documented instabilities. The topper was one that began rather wild
oscillations once the battery impedance rose above a certain level. Of
course none of this was mentioned in the datasheet. I did not like that
LDO in the circuit but a client's engineer wanted it. Call to the mfg,
more and more guys came into the office over there, lots of mumbling,
shoveling of papers, schematics etc. Suddenly someone at the mfg
exclaimed "Oh drat!". Then I knew we were in hot water.

BTW: Did you notice that in the debates nobody ever mentions the need
for better voltage regulators and op-amps. They aren't dealling with
any of the issues that really matter.

I'd vote for Prop 8888: A ban on LDOs

 

[neon]

Yes use an lm117 by all means forget about controling the voltage batteries could care less as long it is there higher then the battery use the LM117 AS CURRENT REGULATOR just one resistor and the regulator voltage out is 1.2v none other. It is fixed and finaly do add a diode from battery to input as follows cathode to input anode to battery. the use is obvious.

 

[Nico Coesel]

I'd consider a switcher here. 35V minus 8.3V at 200mA is over 5W.

My thoughts exactly. Some switchers stop switching and act as a pass
element when the input is equal or less than the desired output.

 

[Joerg]

The nice thing about rolling your own LDO is that you can properly
compensate the loop instead of "relying" on getting the right
dissipation factor in a capacitor.

I'm sure it was marketing guys who conjured up that crap in the first
place. Now everyone is stuck with it.

I have a nice LDO in a recent (2007) chip design... it's only 1.8V out
at 15mA, but stable as a rock with ON-CHIP compensation.

Probably the only way to get it would be to buy some big GPS device
manufacturer or something like that ;-)

Usually I need more than 15mA. But bottomline I do not trust these
things anymore unless you, I or someone I trust has designed it.
Especially since manufacturers have fallen into the habit of not
disclosing circuit innards anymore. My "consumption" of ICs in designs
has actually dropped since that started. The design I am wrapping up
this week would drive tears into the eyes of semiconductor execs. Almost
600 parts, except for two chips all jelly bean stuff. Dirt cheap.

 

[MooseFET]

Sorry, I didn't see the 200 mA.

I like the D-PAK LM1117.

It doesn't withstand the 35V max. Bummer.

 

[MooseFET]

I'd consider a switcher here. 35V minus 8.3V at 200mA is over 5W.
Personally I stay away from LDOs, seen too many not so well or not at
all documented instabilities. The topper was one that began rather wild
oscillations once the battery impedance rose above a certain level. Of
course none of this was mentioned in the datasheet. I did not like that
LDO in the circuit but a client's engineer wanted it. Call to the mfg,
more and more guys came into the office over there, lots of mumbling,
shoveling of papers, schematics etc. Suddenly someone at the mfg
exclaimed "Oh drat!". Then I knew we were in hot water.

I want nice quiet non-EMI making linear.

10.2 - 8.3 - 0.5 = 1.4V so this really isn't LDO land.

 

[MooseFET]

Why not roll your own LDO out of discrete components?

For less parts, I can use a non-LDO if I use a P-MOSFET for the
inverse voltage protection.

 

[Joerg]

I want nice quiet non-EMI making linear.

10.2 - 8.3 - 0.5 = 1.4V so this really isn't LDO land.

That is LDO land, or at least pretty darn close. But I'd roll my own and
not trust chip manufacturers.

 

[[email protected]]

I'd consider a switcher here. 35V minus 8.3V at 200mA is over 5W.
Personally I stay away from LDOs, seen too many not so well or not at
all documented instabilities. The topper was one that began rather wild
oscillations once the battery impedance rose above a certain level. Of
course none of this was mentioned in the datasheet. I did not like that
LDO in the circuit but a client's engineer wanted it. Call to the mfg,
more and more guys came into the office over there, lots of mumbling,
shoveling of papers, schematics etc. Suddenly someone at the mfg
exclaimed "Oh drat!". Then I knew we were in hot water.




I'd vote for Prop 8888: A ban on LDOs

There is a lot more that can go wrong in a DC/DC than a LDO. [I've
designed both. In fact, most switcher chips have an intergrated LDO
for internal use, but I've done stand alone LDOs as well.] I think you
are in the minority on this one, er, not that there is anything wrong
with that. ;-)

My voters guide is just slightly less in weight than a National semi
catalog. Oh wait, do they still publish catalogs?

 

[[email protected]]

Sure. But then I might as well roll my own switcher out of jelly bean
parts. Which I usually do

I just did two more of those, neither one will make any PWM chip vendor
happy because there ain't none in there. Ok, one contains a PLL but I
could only bring myself to using a 15c logic part.

I find it hard to believe once you integrate all the start-up and
protection circuits that this is cost effective.

 

[MooseFET]

I want nice quiet non-EMI making linear.

10.2 - 8.3 - 0.5 = 1.4V  so this really isn't LDO land.

[snip]

I presume the "10.2" is your nearly dead 12V battery?

But what's the "0.5"... your disconnect device shouldn't be dropping
that much... unless you're just using a diode???

I would like to use a diode for the inverse protection and not the P
channel MOSFET circuit.

 

[Joerg]

I find it hard to believe once you integrate all the start-up and
protection circuits that this is cost effective.

That sometimes surprises clients as well. At first there is some gasping
when they look at the schematic. But once they see the BOM and learn
what it would take to coax a PWM chip to fulfill their particular
requirements that changes. For example most PWM chips feature only one
feedback path yet you usually have to curb power not just after reaching
target voltage but also for over-current, over-temp, external parameters
and such. Sure, chips like the TL494 offer two FB inputs but their
output drive, frequency and other parameters, well, let's just say we
can do better ;-)

However, the predominant benefit is freedom from a sole source. The
first such switcher design in my career was what sales guys call a
"design-out" of a PWM chip. A vendor (M....) just could not deliver
production quantities and my client had a hard and financially painful
line stop. That never, ever happened again. It was a very long time ago
and this design actually survived the company owner :-(

 

[neon]

On a sunny day (Wed, 8 Oct 2008 06:41:16 -0700 (PDT)) it happened MooseFET
<[email protected]> wrote in
<9c3e7c02-1ceb-4736-b737-9305d8aacffe@k36g2000pri.googlegroups.com>:

>
>I have a 12V lead acid battery or maybe someone will connect this to a
>voltage up to 35V. I need to make 8.3V. Some moron may connect the
>power backwards.
>
>Currently I am using a circuit with a P channel MOSFET to do the
>reverse protection


If you have 12V, even if it drops to 10.7, then a simple diode for reverse protection would do.
Below that you better disconnect that battery anyways to protect it.


>and an LM78XX type regulator. Does someone know of
>a 78XX like regulator with a lower overhead number.

In cases like this I would use the MOSFET as low dropout pass, or
a nomal BJT,

As you are not playing switchmode, then a simple discrete stabilizer would work,
get a cheap reference fro ma low power LM317,
let's see:




PNP power
+10 to + 35 V ---- a diode k ------------------------e c---------------- + 8.3
| | b T1 | |
=== | | | |
10u --- | | | R3 5k9
tant. | | | | |
/// | about +2.4 c c |
LM317L-------------- b b----|
| | | e e |
| R1 150 | | T2 | T3 |
| | | |_______| R4 2k4
|---- | | |
| === 10u R2 ///
R2 150 --- tant. | 150
| | ///
/// /// 2 x NPN



>The load current is about 200mA max. I have a large heat sink /
>chassis available.
>
>
>BTW: Did you notice that in the debates nobody ever mentions the need
>for better voltage regulators and op-amps.

No, you forgot to specify how accurate the 8.3V needs to be

>They aren't dealling with
>any of the issues that really matter.

Maybe the above circuit, something I solder together on the desk when I need a voltage,
will work for you.
All component values are guessed.
Use a trimpot for R4.

The way it works:
The LM317L creates a reference voltage, very stable, of about 2.4xx volt.
When output voltage is low, T2 conducts, and base current of T1 is limited by (2.4 - 0.7) / R2.
Then when the output voltage rises, T3 starts conducting when its base get to 2.4 V, and
lifts the emitter of T2, decreasing drive currrent in T1.

The circuit has a low dropout of .2 V, the saturation voltage of T1.

I THINK MAYBE 1.2V NOT 2.4V

 

[neon]

If you are going to use lm117 forget about voltage battery only care about current . So a current source for the LM117 will be 6 ohm from adj to out for 200ma. add a diode from out to input for protection and make sure your source is allways 15v. now you got it.

 

[[email protected]]

That sometimes surprises clients as well. At first there is some gasping
when they look at the schematic. But once they see the BOM and learn
what it would take to coax a PWM chip to fulfill their particular
requirements that changes. For example most PWM chips feature only one
feedback path yet you usually have to curb power not just after reaching
target voltage but also for over-current, over-temp, external parameters
and such. Sure, chips like the TL494 offer two FB inputs but their
output drive, frequency and other parameters, well, let's just say we
can do better ;-)

However, the predominant benefit is freedom from a sole source. The
first such switcher design in my career was what sales guys call a
"design-out" of a PWM chip. A vendor (M....) just could not deliver
production quantities and my client had a hard and financially painful
line stop. That never, ever happened again. It was a very long time ago
and this design actually survived the company owner :-(

Well maybe, but I've seen home brew switchers fail in the field, or
worse yet, do funny things like whack the battery when starting up.
Dell USED to design their own switchers. Note the emphasis on used.

If I don't use a LDO I designed myself, I just buy from TI. TI seems
to make cheap stuff that works. It's like National, but with quality.
Granted, stuff for your personal use is different than making a
product to sell.

I would need a damn good reason to result to using a switcher. Now I
do agree that 5W is really too much for a LDO. If the original poster
kept it reasonable (say 16V), I'd go for the LDO.

 

[MooseFET]

Well maybe, but I've seen home brew switchers fail in the field, or
worse yet, do funny things like whack the battery when starting up.
Dell USED to design their own switchers. Note the emphasis on used.

If I don't use a LDO I designed myself, I just buy from TI. TI seems
to make cheap stuff that works. It's like National, but with quality.
Granted, stuff for your personal use is different than making a
product to sell.

I would need a damn good reason to result to using a switcher.  Now I
do agree that 5W is really too much for a LDO. If the original poster
kept it reasonable (say 16V), I'd go for the LDO.

Note the wording on the 35V requirement "or maybe someone will connect
it". The requirement isn't unreasonable because it is a real problem
in the intended product. I am very willing to do the needed heat
sinking and / or have a thermal shutdown trip to handle the case. I
was hoping that someone knew of a regulator that would have a lower
drop out than the LM78XX or LM240-XX types of regulators. My current
design has to have a P-MOSFET reverse protection circuit because I
can't stand the drop of a Schottky. I was hoping to make a simpler
design with a different chip.