DSO Shell frequency issues (and perhaps also DSO 138)

[eysi]

Hi

I'm a long time EEVBlog follower, first time on the forums though.
I am not an electrical engineer but like tinkering around with electronics and microprocessors etc. So I thought it would be a good thing to get a cheap "oscilloscope" to learn something before throwing in a few hundreds of $/€ for a decent one.

I got myself a DSO150 (DSO Shell) from banggood. The problem I seem to be having is that the frequency jumps around quite a lot if I change the timebase. Is it supposed to do that? For the built-in 1kHz pulse wave generator I get following readings:

 

Code: [Select]

sec/div Freq cycl (ms) PW (ms) duty
10us --- --- --- ---
20us 1,258 kHz 0,794 0,296 37,0%
50us 1 kHz 0,999 0,500 50,0%
0,1ms 1,127 kHz 0,887 0,386 43,6%
0,2ms 1,112 kHz 0,898 0,446 50,3%
0,5ms 1,026 kHz 0,974 0,487 50,0%
1ms 1,001 kHz 0,998 0,479 51,0%
2ms 1 kHz 0,999 0,519 51,9%
5ms 1 kHz 0,999 0,399 40,0%
10ms 999,020 Hz 0,999 0,399 40,0%
20ms 250,245 Hz 3,996 2,396 59,90%
50ms --- --- --- ---
The board seems to be the most recent revision, and runs the 113-15001-064 firmware (which was preinstalled). The C5 calibration capacitor didn't change a thing, unlike the C3 one.

Could someone with a similar DSO check if theirs is detecting the same values?

Thanks!

[tautech]

I got myself a DSO150 (DSO Shell) from banggood. The problem I seem to be having is that the frequency jumps around quite a lot if I change the timebase. Is it supposed to do that?

In some circumstances all scopes do so you mostly can fix it by adjusting Trigger Holdoff so there is a delay before the scope grabs another trigger point. Have a fiddle with it. 

[eysi]

In attachment a clearer view using Excel

[ataradov]

I'd say it is normal for this scope. It measures frequency based on the digitized signal. As such, it has time resolution only down to a sample period, which gets coarser with time base increase.

Real scopes have hardware frequency counters that are independent from the ADC.

[eysi]

thanks tautech & ataradov

Couldn't find much about manipulating the trigger holdoff except for a mention in the firmware changelog. In the FAQ it says something about a trigger no longer being used after firmware version 60.


And a reference in the "manual" :

About trigger state
The trigger can have three states including Holdoff, Waiting, and Trigged. They are explained below.
Holdoff: trigger is disabled until a portion of sample buffer prior to a trigger point is filled with raw data.
Waiting:  Trigger is waiting for a valid signal slope.
Trigged: A signal slope has been detected and registered.

[eysi]

I just did a few more measurements, now with the signal output from my AN8008 multimeter set to 4kHz.  Remarkably the detected frequency at 5ms and 10ms is 4 times less than what's actually generated by the multimeter. At 20ms, it shows the same 250Hz as in the 1kHz internal signal generator.

I'll try contacting jyetech to see if they have any suggestions

[ataradov]

I don't see what the problem is. You are well within expected performance of this device.  It is actually better than I would have expected.

Frequency measurement is not that great even on $5000 dollar scopes, what do you expect from a toy? Use dedicated frequency counters, if you need to measure frequency.

[tautech]

Triggering is everything to even basic scope usage.

For repetitive waveforms Holdoff is not normally required but something like a serial data stream where the period between rising edges is constantly varying .....you need it to hold the trigger back until the word has completed otherwise the trigger will grab any rising edge.
With Holdoff OFF most DSO's have a default holdoff setting, usually a few 10's of ns but there are usage cases where holdoff need be adjusted all the way into the ms range.

[jimmc]

I just did a few more measurements, now with the signal output from my AN8008 multimeter set to 4kHz.  Remarkably the detected frequency at 5ms and 10ms is 4 times less than what's actually generated by the multimeter. At 20ms, it shows the same 250Hz as in the 1kHz internal signal generator.

I'll try contacting jyetech to see if they have any suggestions

Hi eysi,
There does seem to be something strange about to 10uS/div results, I suspect that the DSO may be struggling to sample at the correct rate.

For the low sweep speed results you are seeing the effects of aliasing which affects all sampling systems.
There is a video  here https://www.tek.com/support/faqs/what-aliasing-and-how-do-i-detect-it-and-fix-it-my-oscilloscope
showing a 2GS/s Tek scope failing to display a 1MHz sinewave correctly because the timebase is set to too slow a rate.

As you increase the time per division setting the sample rate will decrease to keep the number of samples sweep sensible.
(In the video a 1mS/div sweep is used with a record length of 10,000 points giving a sample rate of (10divs * 1ms /div)/10,000 = 1sample/uS)
There are several other videos on youtube  linked at the at the end of that one, including Dave's on the Rigol.

Jim

[jimmc]

The sample rates for the DSO150 are given in page 3 of this document www.jyetech.com/Products/LcdScope/HowToUseDSO150Library.pdf (see attachment).

At 20mS/div the sample rate is 1.25kSa/sec so as you found a 1kHz ( or 4kHz) signal will give a 250Hz alias.
Similarly at 5mS/div the sample rate is 5kSa/sec giving a 1kHz alias with a 4kHz signal.
etc.

Jim

[eysi]

Thank you very much for doing the research jimmc! Didn't think of checking the library documentation.

edit: reading the DSO138 library documentation, it handles the exact same sample rates at the same time divisions.