Ultra stable voltage regulation

Thread Starter

nogrow

Joined Dec 9, 2010
3
Is it possible to accomplish 5V (20 to 200 mA) voltage regulation that varies less than 0.5% over the current and surrounding temperature range (20 to 40 °C).
An initial stabilization period is not an issue.
If you think it is possible how would you go about it?
As small as possible footprint is desirable.
 

R!f@@

Joined Apr 2, 2009
9,918
Welcome to AAC
Yes.. Even 0.1% is possible.
You might need a DC to DC converter.
Small footprint means Surface Mount Devices.

Ur info is limited.
1. What is the Input Voltage ?
2. Is the load Transient ?
3. Can you solder SMD's ?
 

Thread Starter

nogrow

Joined Dec 9, 2010
3
Blocco,
So you are thinking something lika a voltage reference + voltage following op-amp able to output 200 mA?
 

blocco a spirale

Joined Jun 18, 2008
1,546
Blocco,
So you are thinking something lika a voltage reference + voltage following op-amp able to output 200 mA?
Yes, that is essentially all you need. The PCB layout will also be a factor in determining the performance since 200mA is a significant current. So, be careful will 0V paths and the point from which negative feedback is taken.
 

ci139

Joined Jul 11, 2016
1,898
usually the 10kHz to 100 kHz load variation at OUTPUT side is . . . a "bug" most likely weak spot for load rejection to overcome (smooth out , filter out , reject)
_Draft-DC2DC-SPL-10F.png
(the peak near 10k)
 
Last edited:

MrAl

Joined Jun 17, 2014
11,389
Is it possible to accomplish 5V (20 to 200 mA) voltage regulation that varies less than 0.5% over the current and surrounding temperature range (20 to 40 °C).
An initial stabilization period is not an issue.
If you think it is possible how would you go about it?
As small as possible footprint is desirable.
Hi,

As you know now a stable reference is the first step. Also, keeping it away from the heat producing devices is also a good idea to help keep the temperature of the packages stable. A counter example is the LM317, which typically has a TO220 package and the voltage reference is mounted right in the same package as the pass transistor(s) so when they generate lots of heat the temperature of the internal reference goes up too, which makes the temperature stability not very good at all for decent load currents. I cant remember what i measured now, but it's nothing like 1 percent. It WOULD be something like 1 percent because the reference is precise, but because it heats up the reference too the stability is much worse. It's still good in many applications, but not when you want a very stable temperature characteristic.

So external reference, heat generating components away from it. If you have any problem after that you may want to go to an oven or get a reference that has an oven on chip if they still make them.

The rest is just feedback, and make sure you have fast enough response to react to load and line transients.
 
Top