Bringing up SMPS..Low impedance input filter needed?

[Flyback]

Hi,
Do you agree, that when bringing up an SMPS, the bear minimum needed is an input source voltage which remains flat-in-voltage no matter if the SMPS goes unstable?
Do you agree, without that, you cannot even start your SMPS bring-up testing?

[EDIT....we are testing a 90w boost SMPS, 5vin, 24vout, 350khz, xover ~2kHz, RHPZ~7kHz. END OF EDIT]

 

[rjenkinsgb]

I'd consider that any switched mode PSU design that cannot tolerate random voltage variations at least a few percent beyond its designed input range, is fundamentally flawed and not worth pursuing!

Input supply variation tolerance is one of the most fundamental requirements.

 

[Papabravo]

Hi,
Do you agree, that when bringing up an SMPS, the bear minimum needed is an input source voltage which remains flat-in-voltage no matter if the SMPS goes unstable?
Do you agree, without that, you cannot even start your SMPS bring-up testing?

Absolutely NOT. In practice, an input voltage source with a near zero rise time is a practical impossibility.

 

[Flyback]

Thanks, but i apologise, as both of your good selves are not on the same page as me....and its my fault....i am saying if youre testing a smps which requires 5vin, and you set 5vin.....then vin goes from 2v5 to 5v5 at 1khz when you connect the smps , then that input source is not suitable for the testing in the first place. Do you agree?

 

[rjenkinsgb]

If the power source significantly exceeds the tolerances of the PSU being tested, it's obviously unsuitable.

However that's a very different thing to your original
"an input source voltage which remains flat-in-voltage"

Switched mode supplies generally have a fairly wide input range; the power source should not exceed that significantly, but it must be tested outside the specified limits to guarantee correct operation to those limits.

Any SMPSU that needs a tightly-toleranced input voltage is a very strange and obscure device, that I'd not want anything to do with or would include in any design or product!

 

[Papabravo]

Thanks, but i apologise, as both of your good selves are not on the same page as me....and its my fault....i am saying if youre testing a smps which requires 5vin, and you set 5vin.....then vin goes from 2v5 to 5v5 at 1khz when you connect the smps , then that input source is not suitable for the testing in the first place. Do you agree?

If your input source is doing this, then it is not behaving as a voltage source or there is a fault in the SMPS that it is trying to power. To test the input voltage source connect it to a fixed resistance and see if you get the expected current.

You can also try using a battery. That should tell you something.

 

[Flyback]

You can also try using a battery. That should tell you something.

...Yes, good point..i now suspect, from what you kindly say, that the feedback loop of the bench power supply (CPX400D in this case) , is interacting (fighting) with the feedback loop of the Boost SMPS...and this is the cause of the problem.

 

[Papabravo]

...Yes, good point..i now suspect, from what you kindly say, that the feedback loop of the bench power supply (CPX400D in this case) , is interacting (fighting) with the feedback loop of the Boost SMPS...and this is the cause of the problem.

One thing for sure -- batteries do not attempt to regulate the output.

 

[Papabravo]

The other thing you might want to try is a filter made from a differential mode choke and a common mode choke with Y-capacitors in between.