LDO regulator with series Schottky diode

[schnell]

Hello


We are using the LM2937-3.3 linear LDO voltage regulator to get 3.3V from 5V.
Output current is 50mA maximum.

LM2937-3.3 DATASHEET:

http://www.national.com/ds/LM/LM2937-2.5.pdf

The previous engineer has placed a series Schottky diode (D12) following the LM2937 (i.e, the Schottky is immediately upstream of the output cap as in following diagram)

CIRCUIT DIAGRAM:

http://i54.tinypic.com/a2py7k.jpg

do you know why the Schottky is there?

…..the load has been represented as a resistor, but actually it’s a MC13211 RF Transceiver which has a 2V to 3.4V operating voltage with on-chip voltage regulators….(3.6V max recommended input voltage as per page 46 of the datasheet)

MC13211 DATASHEET:

http://cache.freescale.com/files/rf_...1x.pdf?pspll=1

Anyway,
I am wondering why the previous engineer added the series Schottky diode?

I noticed nothing about putting a Schottky at the output in the datasheet.

I am wondering perhaps , is it a problem if the input side is at zero volts and yet the output has some hold-up voltage lingering on the output caps.?

Or, do you know why a Scottky (D12) would be placed here?

 

[VEBill]

Perhaps he/she wanted to knock off one Schottky diode voltage drop to stay further away from the maximum voltage under some worst case situation.

Or perhaps he/she was worried about some combination of events where the voltage on the capacitor might damage the regulator on power-off. But this is not mentioned in the application notes for that regulator, so this is less likely than the other explanation.

 

[IRstuff]

Well, a diode can be used to drop voltage and rectify. However, the datasheet for the diode doesn't seem to say anything about frequency response, so I'd vote that they decided to drop the voltage for some reason.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[MacGyverS2000]

Why would you put the diode between the regulator and the output cap? Seems like a good way to introduce oscillation in the regulator...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[biff44]

I would call up that engineer and ask what's up. Seems like it is not needed though, so I would be tempted to bridge over it and see what happens.

About the only thing I could think of was that the output voltage somehow was exceeding the rated value of the load momentarily, and instead of buying an adjustable regulator chip, the designer just put in the shottky.

http://www.MaguffinMicrowave.com

Maguffin Microwave wireless design consulting

 

[btrueblood]

A possible (guess) answer - the older style LM317 or LM117 series regulators had a very low reverse voltage withstand (-0.6 volts, compared to the -15 volts min for the OP's LM2937). If it was ever possible to crowbar (short) the input voltage terminal to ground, the output cap. (now at positive some volts above the input pin) would fry the regulator. I vaguely recall an app note, or an old RAP article, somewhere, that described this failure mode, and advised a diode on the output leg to prevent the problem.

 

[VEBill]

btrueblood: "very low reverse voltage withstand"

That's the one I was thinking about (my less likely explanation #2). But now I seem to recall the solution was a 'backwards' diode in parallel output to input. The diode would allow the output capacitor to discharge around the regulator.

So I'll put a dollar on 'shaving-off' some output voltage.

 

[btrueblood]

Uh oh. Now we've got a fight on our hands VE1BLL! First one to find the app note wins, loser buys a round.

Seriously, you may be right, and a quick glance through the datasheets for the LM317 series failed to locate the circuit diagram...ah well. The problem with your solution would be the possible zapping of the diode if the reverse current could get high enough, the problem with a forward diode is the voltage drop, thus using a Schottky with its lower Vf...

 

[VEBill]

Not an app note, but I found this (image below) with a quick Google Image search using the obvious keywords.

 

[btrueblood]

I owe you one!

 

[itsmoked]

I have always seen the diode as depicted in VE1BLL's diagram.

I vote shaving voltage too.

Keith Cress
kcress -

http://www.flaminsystems.com