Low-dropout regulator

[chenjk09]

Hi,

Can anyone tell me what dose Low-dropout regulator do?
https://en.wikipedia.org/wiki/Low-dropout_regulator

I know the Lowers the better why?
Example: 1 V dropout and 0.5V dropout?

 

[ronsimpson]

If you want 5V out of a normal regulator you need to sent 7 volts or more into the regulator.
With a low drop out regulator you need only 6v or 5.5v into the regulator.

Drop out voltage = (input voltage - output voltage) where the regulator stops working well because the input voltage is too low.

 

[4pyros]

Its used alot on battery powered stuff so it will last longer as the battery dies.

 

[chenjk09]

Hi,
Thank
Now I understand.
The lower drop out voltage the better so can last longer as the battery life.

Example: shown in the link below
https://sg.element14.com/national-s...-0/ic-v-reg-ldo-5-0v-2937-to-220-3/dp/9486038
Primary Input Voltage: 10V ?
Can I still connect with a 9V battery?

 

[bountyhunter]

If you want 5V out of a normal regulator you need to sent 7 volts or more into the regulator.
With a low drop out regulator you need only 6v or 5.5v into the regulator.

Drop out voltage = (input voltage - output voltage) where the regulator stops working well because the input voltage is too low.

I always wished they had let me write that app note back at national Semi explaining what dropout actually is...... so many really well versed people believe the above definition, but it is not true among semi makers.

For LDO regulators, dropout is defined as the input-output voltage differential when the output has dropped 100 mV below the nominal voltage. So, it's not the point where the regulator no longer works "well", it is WAY PAST the point where the regulator works AT ALL.

I would wager 99% of the people in the electronics industry don't know this... and I ended up talking to most of them. When a regulator is in dropout, it has no loop gain, no ripple rejection, no nothing..... it just looks like a resistor from input to output.

But an even more interesting fact: the regulator stops working "well" at input-output voltage differences a lot higher than dropout. For example: a "typical" LDO whose "dropout spec" is 0.5V might have 70 dB of loop gain at 100 Hz with an input-output voltage differential of 2.5V. At 1V differential (twice dropout spec) the gain may have dropped to as low as 30 dB. So, it's still a voltage regulator, it's just a poor one.

I once tried to publish some performance curves showing a family of curves for both ripple rejection and loop gain at different input-output voltages approaching dropout..... and was threatened with my life. Semi makers prefer to perpetuate the myth that regulators "work" right up until dropout... they don't actually. They gradually collapse.

 

[4pyros]

Primary Input Voltage: 10V ?

I dont know where that spec came from but yes it should work off 9 volts.

 

[3v0]

Great post bountyhunter, thanks.

 

[ronsimpson]

I use LDO regulators:
To take a well regulated 5V down to 3.3V.
To make 1.5V and 2.5V from a quite regulated 3.3v.

In analog circuits where noise is very important:
I do not use LDO. We keep the input voltage higher than normal. We read data sheets and do experiments to determine at what conditions does the regulator have the best noise.

All this is too much data for a simple question.

 

[ronsimpson]

As long as we are talking about details of LDOs:
Most (not all) LDO regulators are very picky about the output capacitance. READ THE DATA SHEET! Use the capacitor range in the data sheet. Too much or too little capacitor and the LDO regulator will not be happy. You may have to choose a capacitor with the correct internal series resistance.

 

[bountyhunter]

As long as we are talking about details of LDOs:
Most (not all) LDO regulators are very picky about the output capacitance. READ THE DATA SHEET! Use the capacitor range in the data sheet. Too much or too little capacitor and the LDO regulator will not be happy. You may have to choose a capacitor with the correct internal series resistance.

It's actually the ESR (equivalent series resistance) of the output cap that causes most of the oscillation problems in customer usage. Every true LDO has restrictions on the output cap both in minimum capacitance and ESR. The other problem is when aluminum electrolytics are used because their ESR goes straight up at colder temps. Most LDOs are intended for use with Tantalum output caps, but many of the new series LDO's are designed for use with ceramics and are not stable with Tantalums.

Always read the data sheet.

 

[Boncuk]

Hi bountyhunter,

also thanks from me for your excellent post.

Just one thought, but might be worth consideration when designing voltage regulator circuits.

My friend (wkrug) and me designed a DDS function generator lately working at a frequency range from 1Hz to 75MHz (triangle and square) and (planned) 35MHz of pure sine wave aiming for least amount of artefacts in the sine wave shape. (D/A conversion).

Capacitors are very critical in the design as there are are lots of them and a pretty high number of inductors for a 7-pole filter switching automatically with increasing frequency for a clean sine wave output.

(Present stage: 25MHz clean sine wave, still calling for improvement)

Tantalum caps could be used for the power supply, but not for the HF-part of the generator, since they can't stand opposite charge!

Low ESR electrolytic caps are much too large for the planned board size (I want to keep it down to EURO-board dimensions of 160X100mm).

My final choice was using multilayer ceramic caps. They are available for relatively high voltages (25V) and small body size (SMT). They are not polarized but can act as such with no risk of false polarity. (no ringing!)

Example: 100μF/16V-package 1206

May be you can use them for experiments with LDO and VLDO voltage regulators.

Regards

Boncuk


P.S. The DDS-function generator will be an open source project and be published at ETO soon. Everyone is welcome to improve the design and additionally write windows software for remote control via USB.

 

[bountyhunter]

Every true LDO regulator actually needs the ESR of the output capacitor to be stable because it forms a zero in the loop response to partially offset some of the phase shift from one of the poles which is caused by:

fp = 1 / 2 pi x RL x COUT

This pole is inherent and can not be designed out.

Tantalums are usually ideal for LDO's because the ESR drops the pole in somewhere around 10 kHz to 50 kHz which is just about right.

You can't use ceramics with most LDOs since their ESR is so low the zero is way out in the Mega Hertz range and does no good. The LDOs we released which were stable with creamics had another zero built in to replace the one normally added by the tantalum's ESR..... however, they were then not stable with tantalums since that would add a second zero and make the loop bandwidth too wide. So it was an "either or" proposition: stable with ceramics or stable with Tants, not both.

Good reference app note:

https://www.electro-tech-online.com/custompdfs/2012/07/snva167.pdf

 

[4pyros]

Some good stuff here guys.