EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: iBeta on April 16, 2014, 10:00:44 am
-
Have just got my first DSO, a Rigol DS2072A to replace a 25yr old analogue scope.
I did wonder about the measurements recorded when the probe tip is connected to ground and the voltage scale is increased. For example at 10V scale get a reading of 400mV Max and 800mVpp. At 100V scale this becomes 4Vmax and 8vpp.
Rigol have said that this is within the 2% fullscreen error limit.
But 8Vpp noise just seems strange when the tip is grounded.
-
Watch Dave's video:
https://www.eevblog.com/forum/blog/eevblog-601-why-digital-oscilloscopes-appear-noisy-download/ (https://www.eevblog.com/forum/blog/eevblog-601-why-digital-oscilloscopes-appear-noisy-download/)
-
Hey mate, i have the exact same scope DS2072A, 70MHz
Anything you want to know, i can Probably tell you
i can also guide you through the upgrade very easily if you haven't already figured it out.
but as for your question
it's not Noise
your Analogue Scope didn't have the Memory Depth and Resolution that this DSO does,
as per the previous comment
Watch Daves Video on By DSO's Appear Noisy
it explains everything
Be Cool
-
it's not Noise
your Analogue Scope didn't have the Memory Depth and Resolution that this DSO does,
Of course it is noise. :) However, you're correct that an analog scope will not expose this, due to all the reasons that Dave explains.
Bandwidth limiting, short sample memory selection, and Averaging can help control that, to a significant extent.
-
Have just got my first DSO, a Rigol DS2072A to replace a 25yr old analogue scope.
Congrats! You will find there are some new things to get used to, since a digital scope has both strengths and weaknesses, when compared to analog. Since an analog scope doesn't quantize, there is no quantization error.
At 100V scale this becomes 4Vmax and 8vpp.
Rigol have said that this is within the 2% fullscreen error limit.
But 8Vpp noise just seems strange when the tip is grounded.
I agree that can seem strange. It may help to recognize that fundamental to any digital sampling device, there has to be quantization error, due to the number of bits, that an analog scope doesn't have to deal with.
These DSOs sample at just 8-bits, but without oversampling, the ENOB is closer to 7-bits, most of the time (sometimes less). With that bottom bit of uncertainty, flipping around somewhat randomly, even if there were no actual 'noise' at all, that's 0.4% of fluctuation. Usually contributions from front-end noise will about triple that (depending on scope quality... Hamegs do better, cheapies do worse). Current Rigols are on the better end of that quality spectrum.
At very large scales (e.g. 100V/div), that can indeed look pretty big. You're just bumping the resolution limit of the device. You're noticing absolutes, when what you have to focus on here is relatives.
-
on the 10 volt/dev, full scale is 80 Volts,
8 bit sampling where 200 levels are visible on screen, ( 8 div * 25)
so the smallest possible voltage change on the screen is 80/200 is 0,4 Volt
and that is what you are seeing.
-
Thanks for all the replies.
Quantization error of course. I think I must of panicked at the thought of having to send it back.
Quote from Mark_O:
These DSOs sample at just 8-bits, but without oversampling, the ENOB is closer to 7-bits, most of the time (sometimes less).
The DS200 data sheet states : 12 bit of resolution When >=5 us/div @ 1 GSa/s (or >=10 us/div @ 500 MSa/s).
If that were the case, then the error I'm getting could not be completely explained by quantisation, could it?
BTO - was planning to use the DS2000A Upgrade Utility - which seems to be pretty automated so hopefully that will go ok. Will still panick insanely while its upgrading tho' ;)
Still, thanks for the offer of DS2072A specific info. If something comes up, how should I contact you?
-
Oh good grief. Now I can't even read a datasheet properly.
The vertical resolution is 8-bits - so please ignore the 12-bit drivel above.