Maker Pro
Maker Pro

measuring AC ripple on 5 volt PS

UnionTech

Mar 5, 2012
17
Joined
Mar 5, 2012
Messages
17
Hi guys I'm looking for some advice on the best way to take a measurement. I recieved a notice from our DCS manufacture that noise on our dc power supplies can cause NVRAM failurea if the ripple is over 200mV. The paper also recommends replacing PS ever 5 years due to wear on filter caps and using a 500 MHz Oscope for taking measurements. I ended up using a 200 MHz fluke since it was the best I could get my hands on and my results were inconclusive.
First I had a hard time setting up the digital scope so I grabbed an old techtronics anaolg scope and got a reading of 700 mV. I tried to recreate the reading with the faster fluke with no luck untill I switched the probe. I got very little ripple with the 10x 200 MHz probe that came with the fluke but when I used a straight coax with bnc I matched results. So to get to the point should I be using a 1x probe for to measure low mV ripple or should the 10x be just as acurate at that low voltage. The specs on the 1x probe I used are unknown it was sitting in a drawer for awhile probely was in use around the same time as the older analog scopes.
Thanks for taking the time to read this long post lol
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Did you multiply the reading by 10 when you used the 10X probe? I.e. if you are using a 10x probe and you read 1V the actual reading is 10V.

Bob
 

UnionTech

Mar 5, 2012
17
Joined
Mar 5, 2012
Messages
17
Im pretty sure the fluke digital scope does the math when u select 10x probe from the menu. For example when I did a probe compensation cal it read 5 Volts which was the test voltage. I guess my main concern is if the coax or connector on the 1x probe was rated way lower then what i needed could that have screwed up the reading? Or if the 10x probe has a capacitor in there not rated for the speeds I needed could it have filtered the noise? When I got the 700 mV reading my time base was around 10uS to get a good wave form. I thought I was seeing the switching of the chopper transistor on the PS.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Oh, Okay. My digital scope does this as well, IF i remember to set it correctly. I was thinking of my analog scope, which does not,

Bob
 

UnionTech

Mar 5, 2012
17
Joined
Mar 5, 2012
Messages
17
Here is a picture of the waveform I was viewing. Any thoughts if that looks real or not?
 

Attachments

  • 2012-05-16 10.17.00.jpg
    2012-05-16 10.17.00.jpg
    137.2 KB · Views: 229

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
What I would do is connect a low-ESR cap in parallel with a 0.1uF ceramic and a (say) 1k resistor (all 3 in parallel) across the power supply output and see how the signal changes.

The 10x probe will have a lower capacitance and will thus not attenuate the signal so much, HOWEVER from a power supply I would expect that the filter capacitors (or even a 0.1uF bypass cap) would swamp the probe's capacitance.

Are you measuring this with the power supply loaded or unloaded? Does it change as the load changes? At the moment it looks to me like it's very lightly loaded and running in intermittent mode -- and ringing like a bell.

I would expect some change in the ripple as the power supply goes from intermittent to continuous mode.

I would also agree that checking the filter capacitors might be a good thing (or maybe add a couple of low impedance ceramics to attenuate the high frequency noise.
 

UnionTech

Mar 5, 2012
17
Joined
Mar 5, 2012
Messages
17
The power supplies r running with a load, and it should be pretty light since the setup are two seperate power supplies put in parrell for redundancy. What Steve suggested is acually the recommended fix from the manufacture but of course they want to sell it as part of a noise reducing kit.
Unfortunatly the higher ups decided not to worry about the noise and just run everything till failure. Hopefully when it happens it doesnt take down the power plant.
Thanks for the input guys its been fun using a scope for this job. I still have a lot to learn on probe selection though, I will have to read some more and hopefull use the scope some more.
Thanks again
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Gee, I hope whatever fails costs less than a couple of caps and a resistor.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
How much did that famous O-ring cost?

Bob
 

JMW

Jan 30, 2012
90
Joined
Jan 30, 2012
Messages
90
7 Lives plus untold hours of rework, and a loss of a lot of jobs.
As for the OP, due to the critical nature of the measurement, I suggest he get a more knowledgeable user to assist him. Calling Fluke or Electronic may find a local sales rep or similar person who can demonstrate the proper set up.
Using an open forum while helpful, is not the proper procedure.
 
Top