Confused about oscilloscope ADC resolution and accuracy

[elepo]

I'm planning to buy an oscilloscope for the first time and I'm confused about the accuracy (so to speak) of an oscilloscope when measuring voltages.
So it turns out that most scopes have an 8bit ADC. Does that mean if I'm going to measure a 12v signal I will get a maximum accuracy of 12/256 = 46mv?
That's be insane because a scope is supposed to be super accurate right?

I also see "Vertical Scale: 1 mV/div to 10 V/div" in Rigol DS1054Z datasheet. What does this one mean? Is it specifying the accuracy?

[Fungus]

I'm planning to buy an oscilloscope for the first time and I'm confused about the accuracy (so to speak) of an oscilloscope when measuring voltages.
So it turns out that most scopes have an 8bit ADC. Does that mean if I'm going to measure a 12v signal I will get a maximum accuracy of 12/256 = 46mv?

Probably less than that. Most 'scopes are only about 3% accurate.

That's be insane because a scope is supposed to be super accurate right?

Nope. Who says that? The 8-bit ADC can never be very accurate.

I also see "Vertical Scale: 1 mV/div to 10 V/div" in Rigol DS1054Z datasheet. What does this one mean? Is it specifying the accuracy?

No, it's the vertical scale. It's telling you how many volts each square on the screen represents.

[JimLev]

Here’s all the specs for that scope.
https://beyondmeasure.rigoltech.com/acton/attachment/1579/f-0564/1/-/-/-/-/DS1000Z%20MSO1000Z%20Performance%20Verification%20Guide.pdf

[tunk]

Does that mean if I'm going to measure a 12v signal I will get a maximum accuracy of 12/256 = 46mv?

If it has 10 vertical divisions, you have to set it to 5V/div, i.e. the resolution
is 50V/256=0.195V. And as Fungus said, on top of that you have the accuracy.

[elepo]

Nope. Who says that? The 8-bit ADC can never be very accurate.

If it has 10 vertical divisions, you have to set it to 5V/div, i.e. the resolution
is 50V/256=0.195V. And as Fungus said, on top of that you have the accuracy.

I don't understand.
Let's say I have a 12v power supply and I want to measure its ripple. Are you telling me the maximum ripple I can measure is ~50mV?! That can't be right. I've seen people measure 1-2mV ripple with an oscilloscope.

No, it's the vertical scale. It's telling you how many volts each square on the screen represents.

But if the resolution is 50mV how can I get 1mV/div?

Basically I want to know if I can measure 1mV variations in voltage in a signal. And if yes, how is it possible with an 8bit ADC?

[Grandchuck]

Use AC coupling when measuring ripple on a DC supply.

[RoGeorge]

Does that mean if I'm going to measure a 12v signal I will get a maximum accuracy of 12/256 = 46mv?
That's be insane because a scope is supposed to be super accurate right?

- 46mV would be the RESOLUTION, and that is unrelated with the accuracy
- oscilloscopes are NOT accurate instruments, and they were never meant to be so.  Oscilloscopes are mostly to show the SHAPE of a waveform of a measured signal, with just good enough approximation of that signal.


Resolution, accuracy and precision are 3 different things, they are not the same and they are not interchangeable.
- resolution is about how many different levels the instrument can distinguish
- accuracy is about how close the instrument is from the real value
- precision is about how well grouped are the results when the same measurement is repeated many times





Source:  https://www.maxbotix.com/articles/understanding-range-readings.htm

[elepo]

Use AC coupling when measuring ripple on a DC supply.

I know that. I've seen videos of it. But I just wanna make sure I can measure small voltages before spending 400$ on an scope.
I've always seen people just zoom in and measure small amplitudes in the mV range.
But after realizing there's a 8bit ADC I'm no so sure about it anymore.

@RoGeorge
Thanks that was useful.
But I still need to know how/if is it possible to measure very small voltages in the mV range with an scope before I spend 400$ on it.

[RoGeorge]

Yes, it can measure small signals, too.  The "zoom in" is unrelated with the resolution.  Before the ADC, there is usually  an analog amplifier/attenuator, and that is what Volts/div knob is changing to "zoom in".

It can even measure down to tens of microvolts, or tens of millioms, too, if you know how.  Here's an example:  https://www.eevblog.com/forum/projects/oscilloscope-with-trace-averaging-as-a-lock-in-amplifier-(rigol-ds1054z)/

$400 for a new oscilloscope is very, very cheap.  A handheld DMM alone can cost much more than that.

[Brumby]

I had some other comments - but others have beat me to the punch ... and have done it better.

As stated above - scopes are NOT known for their measurement accuracy.  Their main purpose is to show you the shape of the signal.  Any other features are a bonus - but are still limited by the capabilities of the circuitry.  In the case of digital scopes, the ADC resolution is the most obvious limiting factor.

But I still need to know how/if is it possible to measure very small voltages in the mV range with an scope before I spend 400$ on it.

It entirely depends on how accurate you want those measurements.  The primary constraint is the resolution of the ADC - that is, the smallest difference which can be discerned.  The higher the number of bits, the more steps and the greater the resolution.  Second to that is how well the device has been calibrated.  AC coupling allows you to essentially ignore any DC component and focus only on the ripple, so all those bits of the ADC are only spread across the ripple signal.

Scope measurements are more a convenience than metrology class.

[elepo]

Yes, it can measure small signals, too.  The "zoom in" is unrelated with the resolution.  Before the ADC, there is usually  an analog amplifier/attenuator, and that is what Volts/div knob is changing to "zoom in".

It can even measure down to tens of microvolts, or tens of millioms, too, if you know how.  Here's an example:  https://www.eevblog.com/forum/projects/oscilloscope-with-trace-averaging-as-a-lock-in-amplifier-(rigol-ds1054z)/

Ah that's relieving to hear.
Actually I just came across this blog post: http://www.effectivebits.net/2011/03/oscilloscopes-exposed-volts-per.html
Which explains how you can have 1mV/div with an 8bit ADC. But I'm still not sure how that works with a large signal such as 12v because when you zoom in to see the 1mV ripple the 12V is also being amplified right?

$400 for a new oscilloscope is very, very cheap.  A handheld DMM alone can cost much more than that.

400$ is actually very, very expensive for a hobby thing where I live. (It's 3 months of my salary.  )

As stated above - scopes are NOT known for their measurement accuracy.  Their main purpose is to show you the shape of the signal.  Any other features are a bonus - but are still limited by the capabilities of the circuitry.  In the case of digital scopes, the ADC resolution is the most obvious limiting factor.

Everyone keeps saying that and it's very surprising to me.
Let's say you want to measure ACCURATELY how the voltage across a capacitor varies with time in a circuit. What instrument do you use?

[Fungus]

Basically I want to know if I can measure 1mV variations in voltage in a signal. And if yes, how is it possible with an 8bit ADC?

There's a programmable amplifier in the input to make the signal bigger for the 8-bit ADC to be able to measure.

[elepo]

There's a programmable amplifier in the input to make the signal bigger for the 8-bit ADC to be able to measure.

Thanks.
What spec in the datasheet should I look for to find out what the minimum voltage it can measure is?
The DS1054 datasheet says:
Vertical Scale(Probe ratio is 1X): 1 mV/div to 10 V/div
So what if I use 10X?
Also what does 1mV to 10mV mean? Is it based on frequency of the signal, etc?

Ah I'm sorry I misread it as 1mV to 10mV and I got confused because I knew it could do more than 10mV/div.
Disregard this post and thanks again.

[bdunham7]

Let's say you want to measure ACCURATELY how the voltage across a capacitor varies with time in a circuit. What instrument do you use?

Setting up a scope and getting the information you want is not a trivial matter at first.  In the case of a voltage that varies with time, an oscilloscope is what you need.  You have to realize that there are limits on how accurately and how quickly a voltage can be measured--an accurate 6.5 digit DMM will take a single, accurate reading but integrated over a fairly long time, sometimes more than 1 second.  So the number has a high resolution in voltage, but not in time.  A standard digital oscilloscope like the DS1054Z will have a much lower resolution in voltage--about 2.5 digits worth, but will take a billion readings per second.  To get a better resolution in voltage (10, 12 or even 16 bits) and still keep the high sample rate costs much more money.  So which instrument you use will be determined by what you actually need as far as precision in voltage and time.  Try not to use the word 'accurately' unless you are prepared to be very specific.  For the vast majority of applications where you want a high resolution in time and thus need an oscilloscope, the 8-bit resolution is usually good enough. 

For measuring ripple as you stated, here is how it is done.  This is two screenshots of a scope somewhat similar to the DS1054Z and a 12V power supply, one showing the 12 volts directly at 5V/div and you can see that there is no chance of seeing small ripple because the resolution at 8 bits and 5V/div is 40V/256 = 156mV.  The second is showing the same signal with AC coupling--which blocks the 12VDC and only lets the ripple through--at 10mV/div and there you can clearly see the ripple with a resolution of 80mV/256 = 312uV.  The small signal and noise performance of the DS1054Z won't be this good, but the same principle applies.

[RoGeorge]

Let's say you want to measure ACCURATELY how the voltage across a capacitor varies with time in a circuit. What instrument do you use?

Sounds like the goal is to learn about electronic circuits.  For that you don't need high accuracy, or high precision, or high resolution.  An 8 bit oscilloscope would be more than enough.  If you need more, maybe use a sound card input instead, that would give you 16-24bits almost for free.

For learning, you can always use LTspice, a simulator that is free and gives better accuracy than any real instrument can give.

On this forum people are thinking at least 3-4 decimal places, if not 7-8 digits when its about metrology, so asking for high accuracy or high precision, or high resolution would mean way, way more expensive instruments than $400.


Let's switch from words to numbers.  Commit to a maximum budged you plan to spend.  Give some examples of measurements you would like to be able to make with whatever instrument you would choose.
- How many decimals would be enough?
- What range of voltages, currents, whatever you plan to measure?
- What kind of electronic applications you would measure?

[elepo]

Thank you @bdunham7 and everyone who has answered so far. I've learned a lot.

Sounds like the goal is to learn about electronic circuits.  For that you don't need high accuracy, or high precision, or high resolution.  An 8 bit oscilloscope would be more than enough.  If you need more, maybe use a sound card input instead, that would give you 16-24bits almost for free.

For learning, you can always use LTspice, a simulator that is free and gives better accuracy than any real instrument can give.

On this forum people are thinking at least 3-4 decimal places, if not 7-8 digits when its about metrology, so asking for high accuracy or high precision, or high resolution would mean way, way more expensive instruments than $400.


Let's switch from words to numbers.  Commit to a maximum budged you plan to spend.  Give some examples of measurements you would like to be able to make with whatever instrument you would choose.
- How many decimals would be enough?
- What range of voltages, currents, whatever you plan to measure?
- What kind of electronic applications you would measure?

The goal, besides learning, is also being able to troubleshoot/analyze circuits.
The matter at hand is this power supply I have repaired (tbh just replaced a couple components). But I wanna check its output ripple voltage and maybe the switching circuitry to see if it's working alright.
Occasionally I want to troubleshoot a circuit that doesn't work as expected and a scope can come in very handy.
Or just checking stuff out of curiosity, like checking the output of power supplies I buy, or see if an LED light actually has flicker, etc.
The very cheap and useful logic analyzer I bought has reduced my need for a scope but I have to admit that seeing a scope in almost any electronics video, and just the sciency look and the awesomeness of having one is also strongly driving me towards buying one.

Besides I've already saved the money and if I don't buy one I'm gonna have to buy crypto with it or something haha

[dave j]

Actually I just came across this blog post: http://www.effectivebits.net/2011/03/oscilloscopes-exposed-volts-per.html
Which explains how you can have 1mV/div with an 8bit ADC. But I'm still not sure how that works with a large signal such as 12v because when you zoom in to see the 1mV ripple the 12V is also being amplified right?

That blog post has one important bit that I suspect is contributing to your confusion. It says "Ignoring offset". Offset is very important. In the diagram from that page:



The variable pre-amp also applies an offset to the input signal before it amplifies it. This offset is adjusted by a vertical position knob. So, in your 12V example, you could set the vertical position/offset to -12V and zoom in on what remains after the amplifier has subtracted 12V from the signal. This is all in DC coupled mode.

Examining the ripple of supposedly steady voltages is so common that oscilloscopes have a special mode for it - AC coupling. This blocks the DC component of the signal (your example's 12V) so it is centred around zero and you don't have to apply an offset.

It's a lot quicker to switch between DC and AC coupling than to repeatedly adjust the position and then zoom in a bit until you have a 12V signal at 20mV/div in DC mode.

[RoGeorge]

The goal, besides learning, is also being able to troubleshoot/analyze circuits.
The matter at hand is this power supply I have repaired (tbh just replaced a couple components). But I wanna check its output ripple voltage and maybe the switching circuitry to see if it's working alright.
Occasionally I want to troubleshoot a circuit that doesn't work as expected and a scope can come in very handy.
Or just checking stuff out of curiosity, like checking the output of power supplies I buy, or see if an LED light actually has flicker, etc.
The very cheap and useful logic analyzer I bought has reduced my need for a scope but I have to admit that seeing a scope in almost any electronics video, and just the sciency look and the awesomeness of having one is also strongly driving me towards buying one.

Besides I've already saved the money and if I don't buy one I'm gonna have to buy crypto with it or something haha

Looks like an oscilloscope would be helpful.

For the kind of goals you are describing, any oscilloscope would be more than enough.  Those "only 8 bits" won't stay in the way of doing any of the projects you just described.

[Fungus]

For the kind of goals you are describing, any oscilloscope would be more than enough.  Those "only 8 bits" won't stay in the way of doing any of the projects you just described.

Yep. Don't worry about "8-bits" - we're all using 8 bits. It works. You'll figure it out right away when the 'scope is in front of you.

[bdunham7]

The variable pre-amp also applies an offset to the input signal before it amplifies it. This offset is adjusted by a vertical position knob. So, in your 12V example, you could set the vertical position/offset to -12V and zoom in on what remains after the amplifier has subtracted 12V from the signal. This is all in DC coupled mode.

The actual design usually applies the offset at an intermediate stage, from what I've seen only some active probes and very few oscilloscopes apply it right at the input.  In any case, there is also a limitation as to how much offset can be applied--in the case of the SDS1104X-E scope I have used here, you cannot use a 12V offset in any range below 200mV/div.  So to look at ripple at 10mV/div, AC coupling is the only way.

[Bud]

You do not need high accuracy or high resolution to measure power supply ripple. 10mv vs 15mv ripple is 50% difference but it will make no difference to 99% of downstream circuits.

[elepo]

The variable pre-amp also applies an offset to the input signal before it amplifies it. This offset is adjusted by a vertical position knob. So, in your 12V example, you could set the vertical position/offset to -12V and zoom in on what remains after the amplifier has subtracted 12V from the signal. This is all in DC coupled mode.

Ah now that makes sense!
Thank you.

Looks like an oscilloscope would be helpful.

For the kind of goals you are describing, any oscilloscope would be more than enough.  Those "only 8 bits" won't stay in the way of doing any of the projects you just described.

Yep. Don't worry about "8-bits" - we're all using 8 bits. It works. You'll figure it out right away when the 'scope is in front of you.

Glad to hear that 

[elepo]

Thanks everyone for your answers. My confusion is resolved now.