PSU step input damping..

[maluckey]

First off, I'm no electronic genius. I'm self taught the hard way (trial and error). My wife calls me Tim (as in TIm Taylor). I usually figure things out after a few mishaps, but this is getting expensive. I got a PSU that has been tinkered with a few times by the previous owner. I got it for free, so a few bucks spent modding it is moot. I have not yet started on this project for lack of knowledge on how to approach it.

What I am looking for is either a god go-to source (novice level) or an explanation of how I can increase step input damping on a switching PSU.

The problem is that this PSU that may be underdamped. From idle, It ramps up then overshoots and bounces above and beyond the target (12.0v) before settling down to a solid 12.038 volts. This is more pronounced as the load increases.

I can't see where this is good. I was wondering how/if I can tighten this up. My initial thought was to run smaller value, lower resistance caps, in parallel on the A/C side so that the load could drop faster. Would this work, or am I on the wrong path?

 

[BrownTown]

Do you have a cicuit diagram?

 

[maluckey]

I'm relying on visual and general layout at the moment. I can get a diagram if need be, but I can pretty much I.D. most of the layout by sight.

Something in mind?

 

[Mark R]

Find the loop compensation, and drop its bandwidth - a bigger capacitor, or higher resistance on the input to the error amp should do the job.

 

[maluckey]

I'll give it a whirl. I have LOTS of caps lying around, and a couple of VR's. One of the two is bound to work...or catch fire:laugh:

At least I'm on the right track. Now it's a matter of the right values.

 

[jmcoreymv]

Careful, if you bring up the rail too slowly, you run the risk of blowing the mosfets while they are in linear mode depending on the loading conditions.

Edit: That warning depends on the design of the supply, so it may not be an issue.

 

[maluckey]

Good warning!! I've reached my blow-up limit for the month!

Maybe I can try a VR in-line to watch the ocillation. I'm not expecting perfection, just trying to avoid large ocillations and possible bad effects as a result of them.

 

[bobsmith1492]

Put a load on the output if it overshoots at startup would be my thought... but you say that makes it worse. Did you try adding a filter cap on the output (fairly large one depending on current)?

 

[Navid]

You cannot just change the size of a capacitor and be sure what you want happens unless you want to rely on luck!
Increasing the size of a capacitor could actually make the circuit more underdamped. It all depends on where the capacitor is and the whole architecture.

You need to have a schematic diagram, analyze the circuit, and come up with a change that would make your underdamped system critically damped. But, only if there is a problem!

Just because a voltage overshoots does not mean that something is wrong! A circuit can hardly be designed to be critically damped under all possible conditions.

 

[maluckey]

No, I said that a larger load makes it worse. The less the load, the less the issue.

Good point about Filter caps. I did some internet research and found this.

Maybe relevant in this case.

 

[JohnCU]

feed back with PID controller
*drunk sorry cna't comment further beyond Kp Kd and Ki

 

[f95toli]

How many watts? I assume you know that if the PSU can deliver a lot of power (lets say everything over 50W in 12V) you need a soft start?

 

[maluckey]

It's not an issue at startup, the real issue is during normal operation. At startup, the ocillations are less than under high load.

 

[GalvanizedYankee]

maluckey, if you haven't yet, register at http://www.badcaps.net/ . Then ask in the PSU Forum. Post good close-up pics downloaded to thier server.
You will get some valuable input from several members with hands on experience.
Member yanz has modded several low-end PSUs for better performance. Living in the
third world forces some to improve what they can get.


...Galvanized

 

[bobsmith1492]

Originally posted by: maluckey
It's not an issue at startup, the real issue is during normal operation. At startup, the ocillations are less than under high load.

Wait, this disagrees with your original post: you said it settles to a solid 12.038V after a wild startup ramp. Which is it - or both?

 

[maluckey]

GalvanizedYankee,

Good thought, I'll do so and see what happens. Necessity and lack of money sometimes are a great combination.

Bobsmith1492,

To clarify: The PSU starts fine (at least not crazy spikes or dips) but after applying a heavy load quickly, it ocillates dropping to 11.90 and as high as 12.21 before settling down. At idle it's rock steady, and under load it's steady AFTER it finishes bouncing around. It just appears to me that it's a pretty big swing. I have personally (in my short time tinkering with power supplies in audio and computer equipment) never seen one get so much variation. I'm working on getting a schematic on this unit to see if anything has been modified, or merely upgraded as the previous owner said. If it's just upgraded caps my thoughts are that it should react better than average, not worse.

After all's said and done, it's a good exercise for me because I got it for FREE. I can learn something along the way...or blow it up

 

[Colt45]

are you running something mission critical? ±1.75% is close enough for me.