# Oscilloscope "noise"?

I'm just getting a start learning about electronics, and purchased a new oscilloscope -- a DreamSource Lab DScope U2P20 4, which uses the open-source DSView software 5. I built a simple 555 timer from page 7 of the Engineers Mini-Notebook: 555 circuits 6, and I'm getting an unexpected result from the scope. (Note: I added a 0.1uF ceramic capacitor between pins 1 and 8 due to noise from my power supply.)

My question is, is this an electronics problem (my circuit needs improvement), an oscilloscope software problem (rendering poorly), a hardware problem (I chose the wrong scope), or is it normal (an education problem on my end)? What can I do to get better results / visualization of the data?

Here's what I'm seeing:

When I zoom in, I get the following:

Thanks much for helping a rookie learn!

• That's pretty normal and tbh that square wave looks pretty clean. If you zoom in to look at the faster edges you will probably see some ringing on the transitions (which is also "normal" up to a point). Nov 25, 2020 at 19:35
• I concur with the others here, nothing at all to be concerned about. I would say that you're getting pretty darn good performance considering the cost of that instrument! Nov 25, 2020 at 20:37

Look at it this way:

All signals in reality have noise. The oscilloscope samples a signal, with its inherent noise, and converts it to a digital number that has a minimum resolution. It rounds values to their nearest discrete digital representation. The green lines are spaced at the scope's sample rate. The red line is a central value, with ± 1 bit of noise. This is normal--there is always noise in the signal you're measuring and in the conversion process. The sequence of digital numbers is eventually turned into an image, and that process can exaggerate this small noise depending on how the waveform is scaled and positioned. Your square wave looks about as good as as square ever looks on a digital scope. The tiny bit of noise does not indicate any problems with your circuit.

For reference, here is a typical screenshot from a Rigol DS1054Z. There's always a little fuzz on the waveforms.

You can try decreasing the oscilloscope's bandwidth to reduce the amount of high frequency noise in the signal or increasing the sample memory depth. If you want a nice image of a periodic signal, Averaging mode is your friend.

Dave Jones from the EEVBlog has a video explaining why digital scopes appear noisy that explains this in much greater detail: https://www.youtube.com/watch?v=Znwp0pK8Tzk

Assume 1 GHZ bandwidth.

Assume low power operation, so 10Kohm Rnoise in the hundreds of FLASH ADC comparators.

What is the noise floor?

Vnoise = sqrt (4 * K * T * Bandwidth * Resistance)

and I long ago memorized this as 4 nanoVolts_RMS (for 1Kohm Rnoise),

and I scale that 4 nanoVolts up by srt(bandwidth/1Hz)

and I scale up that result by sqrt(Resistance/1Kohm)

so we have sqrt(10K ohm / 1K ohm) * sqrt(1GHz) * 4nanoVolts RM

or

sqrt(10) * sqrt(1,000,000,000 ) * 4 nanoVolts

or 3.16 * 31,600 * 4 nanoVolts RMS

or exactly 100,000 * 4 nanoVolts = 1e+5 * 4e-9 = 4e-4 = 0.4 millivolts RMS,

or about 2.5 milliVolts peak peak.

Your display has much more that 2.5 milliVolts.

hmmmm