Measuring power supply ripple with Siglent 1202X-E

[MarioBros69]

Hello, I have tried to measure the ripple of a Geti 3005E power supply, according to the specifications the ripple should be 1mV (RMS) 20mV (p-p)

The two pictures below are almost the same, but one is with the power supply on and the other is with the power supply off

How is this possible?

The 1202X-E is not valid for measuring the ripple of this source or is it that the noise exceeds the ripple and I can't see it?

" alt="" class="bbc_img" />
" alt="" class="bbc_img" />

[Grandchuck]

Noise from the environment.  Not the power supply.

[pope]

Better measure with the bandwidth set to 20MHz (not full as in your pictures)

[Martin72]

Hi,

First, real screenshots from the scope would be better.
Then:
What we see there is only noise, not ripple - Because the setup you made is not the way it should be.
Ripple has to be measured as near as possible, so you should measure it directly at the output, to minimize unwanted noise.

[MarioBros69]

The truth is that I have tried connecting the probe directly to the source and there is no big difference

With 20 Mhz it removes some noise but I still can't see the ripple and the strangest thing is that with the source off it shows almost the same

With this oscilloscope and the default probe, should I be able to see it?

Below a closer shot

" alt="" class="bbc_img" />

[tautech]

The truth is that I have tried connecting the probe directly to the source and there is no big difference

This is a closer shot

(Attachment Link) " alt="" class="bbc_img" />

Is the PSU loaded in any way ? It needs some otherwise noise makes measurements incorrect.
BW limit and 1x probe should be used for PSU measurements.

Also you need take screenshots to USB with the blue Print button. < much much smaller than photos and better detail.

[MarioBros69]

I only had it for a few days to test it and my intention is to buy a 1104X-E but I was a little disappointed not to be able to measure the source ripple
I seem to remember that I also tried x1 and with 20Mhz activated without a big difference

" alt="" class="bbc_img" />

[MarioBros69]

The truth is that I have tried connecting the probe directly to the source and there is no big difference

This is a closer shot

(Attachment Link) " alt="" class="bbc_img" />

Is the PSU loaded in any way ? It needs some otherwise noise makes measurements incorrect.
BW limit and 1x probe should be used for PSU measurements.

Also you need take screenshots to USB with the blue Print button. < much much smaller than photos and better detail.

I seem to remember that I also tried x1 and with 20Mhz activated without a big difference
In some of my tests I put a 1K resistor as a load, but in these photos it is not included

[tautech]

I seem to remember that I also tried x1 and with 20Mhz activated without a big difference
In some of my tests I put a 1K resistor as a load, but in these photos it is not included

Firstly, ~15us period indicates some 50kHz, is that expected ?

Load the PSU more and check for local interference sources, wallwarts and such.

Don't blame a sensitive scope for the noise, it's what's there !

[Electro Fan]

These two videos will help you measure signals that can be sensitive to how your scope and ground connection are set up.





If you want to see what the power supply really can do on the ripple measurement try using something like these:

https://www.pomonaelectronics.com/products/adapters/bnc-f-double-stacking-banana-plug

https://www.mouser.com/ProductDetail/Cal-Test/CT3655?qs=vgjKjNJexTj1hwPidYnr1Q%3D%3D

They come in different sizes - make sure to get one that will fit the oscilloscope probe you are using.

https://www.farnell.com/cad/2702874.pdf

This 3.5mm adapter fits a Rigol RP3300A.  Some of the older Tektronix probes I think require a slightly larger adapter.  Not sure what size your Siglent probe requires - tautech can probably help you determine that.

[tautech]

Not sure what size your Siglent probe requires - tautech can probably help you determine that.

Mario has a SDS1202X-E supplied with PP215 probes and probe to BNC adapters.
These are a nominal 5mm tipped probe and these should fit (select 5mm size)
https://www.aliexpress.us/item/2255800926120810.html?

[Caliaxy]

Have you tried turning your (presumably LED or fluorescent…) lights off?

[BillyO]

BW limit .. should be used for PSU measurements.

Why?

This is one of those ancient things from a bygone era when the scopes got better than 20MHz BW but the equipment under test could not react to noise at higher frequencies.  This is a different time and ignoring HF PS noise can lead you to endless grief.


To the OP:
Turn off your PS and measure the noise.  See if any of that is being caused by something else.

[tautech]

BW limit .. should be used for PSU measurements.

Why?

I said:
BW limit and 1x probe should be used for PSU measurements.

The modern 1x probe is itself a BW filter to some 6-10MHz as its additional tip capacitance helps reduce noise.
Of course probing technique itself further determines what we might see as does if the PSU is linear or SMPS.

The OP's screenshots photos indicate something at 50+kHz plus lots of other random noise, some of which is expected at higher sensitivity settings when probing technique is suboptimal.
For his 20mV/div 10x probe setting the actual scope setting is 2mV/div and well within the region where all manner of noise gets displayed.

His PSU appears linear with a ripple and noise CV <1mV rating so at these levels probing technique is everything so to not get bogus results and with us timebase settings instead of ms settings for a linear PSU it's no wonder the man is confused when using a highly sensitive scope with a fine antenna probe based measuring system.

[MarioBros69]

Thank you for your answers, when I receive another device I will put all your advice to the test, measure in X1 with 20Mhz and the source with a load and direct connection to the source terminals

The power supply is linear and I have tried turning off all the LED lights, I have even taken the power supply and the oscilloscope to the kitchen and I got the same result, it is as if that noise were in the electrical grid and will be amplified within supply transformers

The strangest thing is that with the power supply off I get the same noise level

[BillyO]

The strangest thing is that with the power supply off I get the same noise level

That's where you need to start.  Get that noise dealt with first.

[tautech]

The strangest thing is that with the power supply off I get the same noise level

That's where you need to start.  Get that noise dealt with first.

Yep, and it could be anywhere nearby.

A customer with brand new 4ch X-E called, convinced the new scope we supplied was faulty as there was unexplained 24kHz interference in his circuits......a wallwart behind a curtain across the room was the culprit.

[Kean]

a wallwart behind a curtain across the room was the culprit.

Always pay attention to the wallwart behind the curtain!

[tautech]

a wallwart behind a curtain across the room was the culprit.

Always pay attention to the wallwart behind the curtain!

It had the poor chap really stumped after checking adjacent rooms and turning OFF most appliances in his house then to find a wallwart that belonged to whom he did not know.

It's only when you purposely go looking for RFI with something as simple as the Reference (Gnd) lead clipped to a probe grabber which makes a useful RFI pickup that you realise how many appliances are noisy, either their circuits or PSU's.
So useful is a probe used in this manner I even use it for setting peak RPM in a chainsaw.
12000rpm/60 = 200Hz.

[BillyO]

I said:
BW limit and 1x probe should be used for PSU measurements.

The modern 1x probe is itself a BW filter to some 6-10MHz as its additional tip capacitance helps reduce noise.
Of course probing technique itself further determines what we might see as does if the PSU is linear or SMPS.

All agreed.

My point though was to not use any kind of filtering.  It just amounts to sweeping the dirt under the rug.

[dophuc]

Here is how I tested the ripple & noise of DC power with Scope:

- The DC power supply and the Scope are powered by an isolating transformer to ensure that they are not affected by AC mains. If I want to be more precise, I will separate the power supply for the oscilloscope from the AC mains, while the DUT is powered by an isolation transformer.
- Coax cable attached to the Siglent SDS1104X-E.
- At the output of the DC source, I connect 0.1uf+10uf (50V) ceramic capacitor in parallel, and use a 0.01uf capacitor to block the DC signal. I tried replacing the 10uf ceramic capacitor with a 10uf Tantalum capacitor but the results didn't make much of a difference.
- Ripple & noise about 3.36 mVp-p.
Compared to using probe with GND wire ripple & noise about 8.96 mVp-p

Another example compares the ripple & noise of 65W GaN charging when using the oscilloscope probe far from the source with using a coax cable near the output of the source, the results will be different.

[markus_jlrb]

BTW,

the KP18X introduce own noise,

se thread

https://www.eevblog.com/forum/testgear/an-investigation-about-ripple-current-of-kunkin-kp184/

Take a plain resistor as load not to mix noise of other devices in your test.

Markus

[BillyO]

Take a plain resistor as load not to mix noise of other devices in your test.

Well, that's the trick isn't it?

[u666sa]

How is this possible?

You doing it wrong!

The manufacturer: making all conditions ideal.
YOU: not doing it like the manufacturer 

A) You gotta load your PSU. Load with 1 amp through your multimeter.
B) You gotta install two caps in parallel
C) You gotta use probe without ground wire, directly with a paper clip or spring
D) You gotta turn off your lights and all other electronics in the room, including computer
E) You gotta use 20 Mhz bandwidth limit
F) You gotta use 1x probe
G) You gotta use AC coupling

...and even after all of this, you still would go out of spec, because you measuring at the output terminals of your PSU and not on your PSU's board.   



Basically, in repair, if it's out of wack it won't work, so if you see it out of wack too much and it don't work, assume it is out of wack.

[sonpul]

Delta, MeanWell...e.t.c. Use for the technical report such scheme.