Resistor values for divider for adjustable Voltage regulators

[lordvader88]

Some new parts arrived and I'm setting up 3.7V at around 100mA I think, to try powering up a smashed cellphone I found.

I'm looking at the datasheets for the LM317 and LM2596 adjustable voltage regulators and I see both of them use pretty small R1,R2 values in the voltage dividers that set the output. The LM2596 is more complicated I know, but still in the example circuits the LM317 uses a 470 Ohm for R1 and so in my case I need about 2x 470 for R2. But that sends over 2mA from the output through them to ground. Not much, but why don't they recommend much higher values if its just to set a voltage, and not waste power?
http://www.onsemi.com/pub_link/Collateral/LM317-D.PDF
on page 1


If I'm reading it right and they use conventional current, current flows out of the adjustment pin, about 50uA, so its not even that they use small resistors to allow a certain size current to flow into the adj. pin.

And the LM2596 recommends similar sized resistors in the divider......why so small, allowing more wasted current
http://www.onsemi.com/pub_link/Collateral/LM2596-D.PDF
on page 9 they use R1=1K

[Brumby]

Look for this in the 317 datasheet:

"Minimum Load Current to Maintain Regulation"

Using resistors of sufficiently low value for the reference can provide the required load as well.

If you know your load will ensure at least the minimum current, then you can increase the reference resistor values.  Just remember that if you do this and your load is one that can be disconnected/reconnected then at the times when the load is not present, your regulation won't be there.

[lordvader88]

ah ok thanks

[eblc1388]

A lot of LM317 users don't realise the 240 ohm resistor connects between Vo and Adj pin shown in typical application inside data sheet  is only applicable for LM117, which has a minimum output loading requirement of 5mA to maintain regulation. For the more common LM317, its 10mA. Therefore a 120 ohm resistor should be used.

[Benta]

But that sends over 2mA from the output through them to ground. Not much, but why don't they recommend much higher values if its just to set a voltage, and not waste power?

A linear voltage regulator needs current for the internal circuits. On the 78xx series of regulators, this current runs from the input, through the IC and back via the ground pin, which is simple enough.

The LM317 is designed to have a stable current running from the ADJ pin (50...100 microamps), otherwise good regulation is not possible. To achieve this, the supply current for the internal circuitry runs from input to output, thus the need for a certain load current (10 mA for a conservative design).

If your LM317 always supplies a load consuming more than 10 mA, you can use higher resistor values.

The 120 ohms for R1 is for cases where your load is unknown or intermittent.

[eblc1388]

The 120 ohms for R1 is for cases where your load is unknown or intermittent.

I can't put it any better. Thanks.

In case of powering a MCU as a load,  don't get caught when the MCU decide to go to sleep. 

[danadak]

The ref leg current is the the sum of the internal error amp current and the
internal Vz current, see page 8 of datasheet for internal summary picture.

http://www.ti.com/lit/ds/symlink/lm317.pdf

Regards, Dana.

[Benta]

The ref leg current is the the sum of the internal error amp current and the
internal Vz current, see page 8 of datasheet for internal summary picture.

Don't understand what you wish to say.

The ADJ pin current is I_ADJ as shown on p.8, 50...100 microamps. The absolute value is not interesting, but the fact that it is nearly constant over all load, input voltage and temperature changes.

What you also see from p.8 is that the error amp and other circuits are clearly powered from input to output.