SMPS reliability testing

Thread Starter

@android

Joined Dec 15, 2011
178
Hey guys please tell me how to validate the reliability of SMPS? Means if I have ready made SMPS then what are the major things to be checked? Are there any standard test procedures for reliability testing?? Please help.:)

Thank you!
 

R!f@@

Joined Apr 2, 2009
9,734
Welcome to the forum.

Ya there are ways.
If you have an SMPS. then next buy....

1. A digital Oscilloscope.
2. Some high power rheostats.
3. A good DMM like a fluke.

Get these and then only we can tell you how to check 'em SMPS.
Cause words just cannot show you the reliability of a hardware that we have never seen.
 

Thread Starter

@android

Joined Dec 15, 2011
178
Yeah. That is fine; but you are not helping! I'm simply asking for standard procedure to validate reliability.
 

praondevou

Joined Jul 9, 2011
2,942
I don't know about any standard procedures for reliabilty testing
.
We usually checked the following parameters:

- maximum output voltage deviation for input voltage drops of a certain length
- same thing for sudden load changes, 0% to 100% and the other way around (oscillation analysis too)
- let them run for a few hours under full load with increased ambient temperature
- overload behaviour
- behaviour with different varying loads (for example other smps)

The most important thing IMO is how it responds to load changes, if oscillations occur and for how long. Then, if it can endure the specified output current for a few hours (until component temperatures stabilize) without blowing up.
 

Thread Starter

@android

Joined Dec 15, 2011
178
I don't know about any standard procedures for reliabilty testing
.
We usually checked the following parameters:

- maximum output voltage deviation for input voltage drops of a certain length
- same thing for sudden load changes, 0% to 100% and the other way around (oscillation analysis too)
- let them run for a few hours under full load with increased ambient temperature
- overload behaviour
- behaviour with different varying loads (for example other smps)

The most important thing IMO is how it responds to load changes, if oscillations occur and for how long. Then, if it can endure the specified output current for a few hours (until component temperatures stabilize) without blowing up.
Hmmm. That will be helpful. Also I would like to know the most commonly used components & devices for building SMPS. I know the basic blocks like rectifier, filter, PWM IC, Switch mode trafo, optoisolator. Can you provide list of most commonly employed components/devices(means any standard company providing these)?
 

bountyhunter

Joined Sep 7, 2009
2,512
Yeah. That is fine; but you are not helping! I'm simply asking for standard procedure to validate reliability.
It would be the size of a New York phone book.

The reliability has to be designed in, it can't be "tested in". That's a myth I was never able to bury although I tried for 35 years.
 

Thread Starter

@android

Joined Dec 15, 2011
178
It would be the size of a New York phone book.

The reliability has to be designed in, it can't be "tested in". That's a myth I was never able to bury although I tried for 35 years.
'The reliability has to be designed in', that's true. But what if you have the SMPS circuit from your supplier and if you want to incorporate that into your product? How will you test its reliability?? Before moving from linear supply to SMPS supply I must be sure that the circuit is well designed & meets certain reliability standards. So that I can be assure that my product will not fail in the market!!
 

praondevou

Joined Jul 9, 2011
2,942
In this case you could verify all parameters that are listed in the SMPS specifications, especially those that would have a big impact if they are not met.

Depends on what your load is.

Make some tests under harsh conditions too, high low temperature changes under load for a few days...


You also need to have some confidence in the product you buy.
 

Thread Starter

@android

Joined Dec 15, 2011
178
Yeah. May be I need to define my own standards! Anyways if anybody can provide me a list of standard companies offering followings components/devices for SMPS...
MOSFETS
Switching trafo
Optoisolators
Bridge rectifier
etc.
 

Thread Starter

@android

Joined Dec 15, 2011
178
Yeah. May be I need to define my own standards! Anyways if anybody can provide me a list of standard companies offering followings components/devices for SMPS...
MOSFETS
Switching trafo
Optoisolators
Bridge rectifier
etc.
Waiting for your reply guys!!
 

Potato Pudding

Joined Jun 11, 2010
688
You know they have courses to teach you how to test equipment for safety and reliability.

What you would normally do is get a test production order of at least 100 units tear a random few apart taking measurements as baselines, and run the rest hard for weeks and tear a few more of them apart looking for signs of stress every week. Finally you have some that are tortured for several months, and the last dozen are run until they die.

Then you would require certification that there will be no changes between the test production and the main production.

It is a lot of work and a heavy commitment - which really guarantees that any weaknesses will be found ASAP just so that the testing can be stopped.

Tests needed for a SMPS.

RFI testing to show that it won't impede communications.

Output range stability to show that with changing load currents the output stays regulated. For testing you might likely try from unloaded to maxrated +25% to 100% current. It should just shut down cleanly under overload.

Short circuit testing. It should shut down cleanly. Bursting into flames would be an example of a failed test.

Input range stability. Try running it from a variety of possible input power situations.

Environmental testing. Run it at 70 Celsius under a blanket. Run it at -30 Celsius with gusting fans to try and crack it. Run it in a damp environment. Run it in a dusty environment.

Safety testing. Look for shock hazards and fire hazards.
 

Thread Starter

@android

Joined Dec 15, 2011
178
You know they have courses to teach you how to test equipment for safety and reliability.

What you would normally do is get a test production order of at least 100 units tear a random few apart taking measurements as baselines, and run the rest hard for weeks and tear a few more of them apart looking for signs of stress every week. Finally you have some that are tortured for several months, and the last dozen are run until they die.

Then you would require certification that there will be no changes between the test production and the main production.

It is a lot of work and a heavy commitment - which really guarantees that any weaknesses will be found ASAP just so that the testing can be stopped.

Tests needed for a SMPS.

RFI testing to show that it won't impede communications.

Output range stability to show that with changing load currents the output stays regulated. For testing you might likely try from unloaded to maxrated +25% to 100% current. It should just shut down cleanly under overload.

Short circuit testing. It should shut down cleanly. Bursting into flames would be an example of a failed test.

Input range stability. Try running it from a variety of possible input power situations.

Environmental testing. Run it at 70 Celsius under a blanket. Run it at -30 Celsius with gusting fans to try and crack it. Run it in a damp environment. Run it in a dusty environment.

Safety testing. Look for shock hazards and fire hazards.
'Finally someone touched the exact nerve!'
 
Top