Spikes and noise on Rigol DHO814

[jim_griff]

~17µs / ~58kHz switching frequency. The same frequency, but much higher in amplitude, when putting the probe next to the power brick. It's definitely the culprit.

[jim_griff]

~17µs / ~58kHz switching frequency. The same frequency, but much higher in amplitude, when putting the probe next to the power brick. It's definitely the culprit.

Here are measurements with two probes. One sitting on the desk (yellow trace), the other by the Rigol PSU brick (blue trace).

Yellow is 500µV/div; Blue is 100mV/div
EDIT:(20MHZ B/W limit.)

[Fungus]

The noise is definitely coming from the Rigol PSU brick.

...or through it from a dirty mains supply.

[eTobey]

The noise is definitely coming from the Rigol PSU brick.

...or through it from a dirty mains supply.

So then where in the world can this signal come from? It must be somewhere between germany and UK? Maybe its the wind turbines in the sea?? 

[shapirus]

So then where in the world can this signal come from? It must be somewhere between germany and UK? Maybe its the wind turbines in the sea?? 

Nah it must be coming from the outer space, as I'm seeing it too:



This was captured with the probe in the 10:1 switch position and the ground attachment shorted to the tip, creating a loop, lying right on top of the PSU.

[shapirus]

Definitely not the cleanest PSU. I tried a car cigarette lighter plug with USB-C outputs, and it was actually better than this. So it makes sense to look for a less noisy power supply, but the question is how do we tell them apart from each other without buying and returning a dozen or two before finding a better one.

[Fungus]

Definitely not the cleanest PSU.

Mine is perfectly clean and so is just about every other person on here.

So unless yours is defective then the spikes are coming from the environment.

Spikes like these can definitely come from the mains. Dave's done videos on tracking these problems down, I posted one above.

I tried a car cigarette lighter plug with USB-C outputs, and it was actually better than this.

In your car?

[jim_griff]

Definitely not the cleanest PSU.

Mine is perfectly clean and so is just about every other person on here.

So unless yours is defective then the spikes are coming from the environment.

Spikes like these can definitely come from the mains. Dave's done videos on tracking these problems down, I posted one above.

I tried a car cigarette lighter plug with USB-C outputs, and it was actually better than this.

In your car?

The evidence all points towards the PSU bricks of some people's devices. Occam's Razor:

1. Some people who own this scope are measuring the exact same signal, and in totally different parts of the world.
2. These signals are all being tracked down to the Rigol PSU brick (inductive coupling tests).
3. These signals are not able to be measured through the mains 120V/230V supply (unless Rigol PSU switched on).
4. Other devices and power bricks do not exhibit this exact signal and have their own switching signature, but much lower amplitude.
5. The signal disappears when using other power supplies to power the scope.

It's 100% noisy Rigol power bricks, albeit not all of them. Some have obviously been engineered better than others - different revisions, perhaps.

[csuhi17]

We have known this problem for a long time.
The early sales gave a noisy power supply, then if I remember correctly they switched to LiteOn and Lenovo.
I didn't see in the above posts which one was added for those of you who find it noisy.

I use it with a sound technician's 12V power supply, but the LiteOn that comes with it doesn't make more noise for me either.

I tested the LiteOn, I charged a solar powerbank with it, it said 20V and 66W.
I couldn't test it at 15V.

In the OP's first post, I don't see what the channels are set to. 1x 10x
I also don't know if there is something connected to CH1, CH2.

If you have the opportunity to try it with another power supply, perhaps with a car battery, you can find out that everything is fine with the oscilloscope.
If you can't solve it any other way, then return it.
Maybe write to the seller about what can be done in this case.

There are two pictures of the power supply, with 20V 66W.
And the third is just for interest, the ground wire of the probe was oxidized or dirty at the part where it connects to the probe. With a handheld scope, I only measured ~40Vpp with the same probe.
After I cleaned everything was fine.

[shapirus]

Mine is perfectly clean and so is just about every other person on here.

No switching supply is perfectly clean. Did you actually try to capture its EMI by shorting the probe's ground strap to the tip and using it as an antenna to sense noise close to the PSU?

I'll tell you more, the scope itself is also emitting like crazy. Easily sensed in the vicinity of the input power connector, where the several switching converters are located. But at least it doesn't seem to couple into the analog input lines.

So unless yours is defective then the spikes are coming from the environment.

Spikes like these can definitely come from the mains. Dave's done videos on tracking these problems down, I posted one above.

Are you kidding me? Two people thousands of kilometers apart show an identical switching noise waveform, (which in my case raises in amplitude dramatically as the ground loop antenna gets closer to the PSU brick), and you still tell that this may be coming from the environment?

I tried a car cigarette lighter plug with USB-C outputs, and it was actually better than this.

In your car?

No, on my bench. Cars aren't the only source of 12V DC power in the world.

[shapirus]

The early sales gave a noisy power supply, then if I remember correctly they switched to LiteOn and Lenovo.

Mine is LiteOn.

I didn't see in the above posts which one was added for those of you who find it noisy.

Mine isn't particularly noisy and isn't a problem in practice, as the emitted interference quickly drops down to undetectable levels as you get farther away from it, but its noise is easy to detect when you look for it.

Some other switching supplies around me are noisier. All kinds of noise can be seen when probing with an alligator clip ground strap. Using a spring attachment, or, if possible, a coaxial adapter, helps.

I use it with a sound technician's 12V power supply, but the LiteOn that comes with it doesn't make more noise for me either.

How did you capture the attached waveforms (how was the probe connected)? They don't look like typical switch mode PSU EMI spikes.

[csuhi17]

I use it with a sound technician's 12V power supply, but the LiteOn that comes with it doesn't make more noise for me either.

How did you capture the attached waveforms (how was the probe connected)? They don't look like typical switch mode PSU EMI spikes.

I charged an Allpowers 288Wh solar power bank/generator with the LiteOn power supply.
Between the two is a USB-C male female test connector.
I measured between Vcc and Gnd, with 10x.

If I have time, I'll try with an adjustable linear load, I couldn't find anything else at the moment.

[shapirus]

I charged an Allpowers 288Wh solar power bank/generator with the LiteOn power supply.
Between the two is a USB-C male female test connector.
I measured between Vcc and Gnd, with 10x.

That makes sense then.

What we've been discussing is this:



To capture the radiated interference.

[csuhi17]

To capture the radiated interference.

Everything is clear now.
I misunderstood, I thought he could see the noise from the power supply on the channels.

I checked my nearby switching power supplies using that method.
For me, they were all much noisier than the Rigol power supply.
I measured 2 power supplies of Micsig, an LED light source, a computer, the power supply for a 12V refrigerator, and even my Chinese metal housing power supply.

In all the videos I've watched so far, the power supply had a metal housing in the scopes. Maybe they could have added a properly insulated laptop charger with a metal housing, if such exists at all.

Wouldn't a thicker metal box have been better than the aluminum foil?

[shapirus]

Everything is clear now.
I misunderstood, I thought he could see the noise from the power supply on the channels.

It's possible to see this noise during normal measurements too. The signal being measured must not be too noisy itself, the ground strap instead of the ground spring must be used. Sometimes it's possible to catch it even with the ground spring.

I checked my nearby switching power supplies using that method.
For me, they were all much noisier than the Rigol power supply.
I measured 2 power supplies of Micsig, an LED light source, a computer, the power supply for a 12V refrigerator, and even my Chinese metal housing power supply.

Yeah, even if we power the scope with the cleanest PSU possible, then there still will be a lot of crap around emitting all sorts of switching noise, so really the solution is to use probing techniques that help minimize their influence, and to get rid of the worst offenders: for example, in one case, when measuring how well an RC filter on a DC power rail worked, I saw some switching noise and thought that it was coming from a switching PSU powering the circuit, until I saw that the noise remained even after switching the PSU off. It turned out that the interference was coming from a generic chinese DC/DC step-down module used in a homemade device sitting ~3 meters (!) away from the probe. That one must be really poorly designed, I'll have to replace it some day, or make one myself, or maybe just wrap it in foil.

[Fungus]

To capture the radiated interference.

Everything is clear now.
I misunderstood, I thought he could see the noise from the power supply on the channels.

Yeah, that wasn't made clear.

Simple solution: Don't do that!

[shapirus]

Well, it's not fully clear how exactly OP measured/captured that noise and in what situation it was causing trouble.

I assumed it was EMI, because I've seen that myself. With the probes disconnected this switching noise doesn't make it to the displayed waveform in my case. Maybe the OP's situation is different.

[eTobey]

I measured it on my prototype, which is powered through a cheap power supply. Maybe it came through this. But that Rigol powersupply should not spew all that noise all over the place anyway right?

[csuhi17]

Not in my case.
With the above method, I have to put the probe quite close to the power supply, I have to almost touch it. Also, there are parts and positions where less noise is detected.
If I touch the probe with my hand or its wire, the noise increases.

If you measure the noise on the cheap power supply, you will be able to determine whether it is bothering you.
The power supply I bought for my Riden is so noisy that it was also visible on the regulated output.

When I measure, I always place such power supplies as far away as possible.

In the very first post, the channel setting and the probe were set the same, both 1x or 10x?

[eTobey]

Yes