Low ripple Power supply Design

Thread Starter

anukalp

Joined Jul 28, 2018
158
Hello

I am trying to design power supply SMPS for a project. I know I can get ready made SMPS in market but I don’t want to buy it. I also understand designing my own SMPS is big challenge but I am going to do it

Specification

Input:
230V AC, 1.0A, 50/60 Hz

Output:
24V DC, 3 A
12V DC, 2 A
5V DC, 2A

This is my attempt to make 12 V and 5 V DC

upload_2019-3-5_15-0-36.png

I need suggestion to make low ripple smps.

What should be the rating of the transformer for 24 V DC?

I would appreciate any help
 
Last edited:

ericgibbs

Joined Jan 29, 2010
18,766
OK,
Looking at your specification
Output:
24V DC, 3 A
12V DC, 2 A
5V DC, 2A

That transformer will not be suitable.
If you calculate the output power alone, that would be:
72W+24W+10W = 106Watts, obviously the psu will not 100% efficient, so you are requiring at least 150W to 200W.
If your your mains supply is say 230Vac, you need a 200VA to 250VA transformer as a rough estimate.

E
Look thru this PDF for SMPS help.
 

Attachments

Thread Starter

anukalp

Joined Jul 28, 2018
158
OK,
Looking at your specification
Look thru this PDF for SMPS help.
I know that's why I asked for a transformer rating in post #1

I understand to make power supply we should have transformer, diode's, regulator IC's and component for filter circuit

as shown circuit for regulated supply but how to design SMPS for given requirement
 

ericgibbs

Joined Jan 29, 2010
18,766
OK.
As you can imagine a 250VA transformer will be big, heavy and expensive, this is one reason why SMPS are preferable, as they use smaller, lighter transformers.
They are more complex than a linear PSU's, but IMO more suitable for your application.
Lets know how get along with that PDF.
E
 
Last edited:

recklessrog

Joined May 23, 2013
985
Here.......,.,.,..https://www.youtube.com/watch?reload=9&v=wopmEyZKnYo......is a method I have used many times for reduced ripple.
Back in the '60's, large value capacitors were expensive and this method was used a lot. There are many refinements and variations, but the basic idea remains the same. (you may notice that in this video, he does mix up his polarities)
Also another method that is even better using more components, is ripple cancelling.

https://pdfs.semanticscholar.org/2dd4/1c749d90e4a470947afcdb78d2d58b18152e.pdf

Look up active ripple filters etc as there are many articles on how to do it.


.
 
Last edited:

mvas

Joined Jun 19, 2017
539
OK.
As you can imagine a 250VA transformer will be big, heavy and expensive, this is one reason why SMPS are preferable, as they use smaller, lighter transformers.
They are more complex than a linear PSU's, but IMO more suitable for your application.
Lets know how get along with that PDF.
E
Won't we still need a large transformer to isolate from the mains?
 

recklessrog

Joined May 23, 2013
985
The article that Eric gave you in post #6 is an excellent reference for anyone wanting to understand switch mode power supplies. It will give you a pretty good understanding of SMP operation.
Also, anyone who is trying to repair smp's will find it usefull. I have used it as a teaching aid as it is comprehensive and fairly easy to understand.
Do however, bear in mind the safety precautions needed when working on them as it is only safe to do so with an isolating transformer to help avoid electrocution. For safety, without specialist training and considerable knowledge and experience, you should NEVER attempt to work on live mains equipment without an isolating transformer.
The reason that the normal smp's in use do not have a large isolating transformer is because they are usually either in sealed plastic cases, have optical and galvanic isolation internally (beware, many cheap ones are not fully isolated) and are not meant to be worked on by non proffessional engineers without the correct safety equipment.
To isolate 50 or 60 Hz mains requires large iron cored transformers. SMP's operate at much higher frequencies ranging from 20 kHz up to a few MHz. At those frequencies, small internal ferrite cored transformers are used.
To avoid any confusion, remember that the saftey isolating transformer is on the INPUT of the smp and is only required when uncased work involving repairing, modifiying, or designing them.
Almost at least as necessary, is a variac on the output of the isolating transformer.

Last year I posted a series on working with smp's, Type this into the search or click on the link if you want to see them........
Switch Mode Power supplies, My experience
https://forum.allaboutcircuits.com/threads/switch-mode-power...



Here is another useful article.
 

Attachments

Last edited:

ericgibbs

Joined Jan 29, 2010
18,766
Won't we still need a large transformer to isolate from the mains?
hi @mvas
In many SMPS, often the type used for Laptops, the mains supply is rectified, smoothed to a 'high' DC voltage.
This DC is high frequency 'chopped' on the high side primary of a small isolation transformer, designed for say 40kHz operation.
The isolated low side secondary of the this transformer is rectified and smoothed, to say approx 19Vdc at 3Amps.

This is an example of the type of SMPS I am suggesting the TS designs, the secondary of the switching transformer would have say three windings to suit the the voltages he requires.
It's a fairly major challenge, but the TS wants to learn from his design work.

E
AA1 06-Mar-19 09.06.gif
 
Last edited:

recklessrog

Joined May 23, 2013
985
Check the spec sheet for each regulator. You should have a capacitor on the input and output of each regulator right up close to the pins.. Be careful not to exceed the dissipation in them as you have to lose 19 Volts at whatever current is drawn in the 7805 and 12 volts in the 7812 so at least a large heatsink will be required. Allow at least 50% greater thermal capacity than the minimum you calculate othewise the heat will cause them to shut down. Even with no load, a few milliamps flow from the input to the ground pin that may add 2-3watts to the calculation.
Also check the output for ripple and noise as per your first post requirement. Even though the spec sheet shows ripple rejection exceeds 50 db, it rarely is that good in practice.Thats when you may consider to add the extra ripple reduction if it is necessary.
 
Last edited:

recklessrog

Joined May 23, 2013
985
In addition to my previous post, depending on the quality of the SMP you use, there is often a lot of spurious radiation from the poorly screened ones that can get picked up in the wiring of the regulators, Judicious use of ferrite suppressors on the leads fom the smp can help reduce the directly injected noise, and an earthed metal screen around the smp may be required.
As I am an amateur radio operator, I have a dislike for smp's because of the wideband noise they spew into the ether so I try to use linear power supplies whenever possible.
 
the 7805 is good for 1A and the power dissipation would be too high. You want about 9V DC on the input side.
You would have (2A)*(24-5) Watts to dissipate as heat. Not (1A)(9-5) or about 3W for a normal 7805 design.
You would generally have multiple secondaries for multiple outputs.

SMPS generally starts with either half or full-wave mains (120/240). That's the DC value it starts with. It's also a couple of diode drops away from ground of he rest of the circuit. Troubleshooting tends to blow stuff up.
 

AnalogKid

Joined Aug 1, 2013
10,986
Also, C1 needs to be at least 10,000,000 times larger. Really. Once you change the regulators to something that can handle 2 A output current each, that's a *lot* of ripple current in the main filter capacitor. A 10,000 uF cap will have approx. 3.3 V of ripple. That's ok with a 24 V input, but there will be a small amount of ripple on the outputs. To reduce that, increase C1.

Also, the 5 V regulator will dissipate almost 40 W of heat at 2 A output current. You will heed a medium sized heatsink and a small fan.

ak
 
Top