400 kHz DC ripple signal of 80 mV, need to see diff. of 1%

Which product is right for your exact requirements
Post Reply
Posts: 1
Joined: Tue Jul 24, 2018 3:59 am
Location: Germany

400 kHz DC ripple signal of 80 mV, need to see diff. of 1%

Post by linuxfan » Tue Jul 24, 2018 4:39 am

I use a flicker meter than detects up to 400 kHz ripple flicker and tells the ripple percentage. It has an oscilloscope output to BNC to further analyse the voltage. I want to do that and I am not sure which oscilloscope would fit my needs.
The maximum output voltage is 2 V, but in reality it is much lower. I want to measure display flicker and I often get values of just 80 mV or lower. If it is important I will test how low they really get in the worst cases. I need an oscilloscope that has no interfering noise, so I can clearly see differences in amplitude of 1% in the lowest available input ranges of +-100 mV, +-50 mV, and maybe even +-20 mV (if I really need that little - not sure at the moment). This is critical. I suppose a ripple of 1% may be hard to spot. Maybe the software can automatically detect such small changes in amplitude for added comfort?

My current USB oscilloscope, the admittedly cheap Hantek 6022BE, has much noise, and even worse, it does NOT have full 8 bit vertical resolution in 20 mV/div mode. I suppose any PicoScope will be of much higher quality and deliver full 8 bit resolution even at its lowest available input range?

This is a screenshot of my Hantek USB oscilloscope to show you what trouble I currently have. You can clearly see the noise and the less-than-8-bits resolution. I wonder if a PicoScope will work much better and let me recognize a 1% ripple at low voltages and frequencies up to at least 400 kHz. In the picture the ripple percentage it is somewhere around 25%. With all the noise and low vertical resolution, 1% ripple would be impossible to spot:
AMOLED 60 Hz flicker

Posts: 494
Joined: Mon Aug 11, 2014 11:14 am

Re: 400 kHz DC ripple signal of 80 mV, need to see diff. of

Post by Gerry » Wed Jul 25, 2018 2:27 pm

Hi linuxfan,

In the image that you posted the noise traverses up to 2 step sizes, so the smallest ripple that you could reliably detect would be one that traverses 4-bits, which means that you have a worst case of 6-bits of effective resolution. To detect a 1% ripple you would need 7-bits of effective resolution. Using one of our 2000 series PicoScopes (see here: https://www.picotech.com/oscilloscope/2 ... 0-overview) you would have similar resolution, but you can increase the resolution of the displayed waveform by using our Resolution Enhancement feature in our PicoScope 6 software (see here: https://www.picotech.com/library/oscill ... nhancement). This uses dithering and moving average filtering to reduce the noise and, at the same time, increase the number of resolution steps that you have. To perform the filtering you would have to sacrifice some bandwidth of the signal. So, for instance, to get 2-bits of Resolution Enhancement your bandwidth would reduce by a factor of 16. If the PicoScope has a bandwidth of 10MHz, e.g. our PicoScope 2204A, this would give you at least 8-bits of Effective resolution, and reduce the bandwidth to 625kHz (which would be fine for your requirement of measuring up to 400kHz of ripple). If you need to see even more detail then you could increase the resolution enhancement and choose a faster PicoScope.

What you need to bear in mind here, is that Resolution Enhancement only affects the data being displayed on the screen, not the raw sample data that you have captured. So, whether Resolution Enhancement would be enough for you depends upon whether or not you need to do anything with the data (specifically the ripple) other than analyze it in PicoScope 6 (e.g. would you want to export it as a CSV file and then import it into Excel for further calculations?). If you need the raw sample data to clearly show the ripple (without any clever software tricks) then you would need a 12-bit PicoScope such as our PicoScope 4224 (see here: https://www.picotech.com/oscilloscope/4 ... 4-overview) which would give you the resolution that you need, and the bandwidth that you need.

You also need to decide how long you would need to be capturing the data for, as the 2204A, for instance only has a 4kS buffer for storing the data, and a 400kS bandwidth (which would require a 2MS/s sample rate) would only give you 2ms of capture time (so if your looking at longer captures you would need a PicoScope model with a bigger buffer).


Technical Specialist

Post Reply