Here first two pictures about risetime. same signal but other have infinite persitence for show that wobbling is not bad and it is amazing stabile trigged in this class of scopes.
The risetime looks slow for the frequency response you indicate, I would expect close to 1 ns for the 300 MHz bandwidth you seem to measure. It appears that something funky is going on with the frequency response, either the roll-off is shallower than -20dB/decade, or they do DSP on the signal to improve apparent frequency response. What's your impression?
And then... cursors are -3dB levels if compare to 1MHz level. (generator have enough good flatness for this test, HP8644B and 1 - 350MHz come with same level as 1MHz. (flatness (level accuracy to well matched 50ohm is inside +-0.1dB. This matching with scope is not ok, this may add some small errors but not markable for this purpose now here.)
So the ADC is sampling at just four times the bandwidth in single channel mode, and even less in dual channel mode. This means you have to watch for aliasing again unless the filter roll-off is extremely steep (something at odds with the high rise time), even at the fastest sweep speeds, just like the early DSOs. Making sure that you have no frequencies beyond the Nyquist frequency is all fine and dandy as long as you're measuring signal generators, but not so easy with a real DUT.
With this measuring accuracy and common simple thumb rule (.35 rule) for risetime this measurement is just inline. Also this risetime number is oscilloscope itself measured and with around 40% of 0-255 (8bit) signal level.
I have NOT measured 300MHz -3dB bandwith. With Tektronix originated thumb rules typical 200MHz "gaussian BW" oscilloscope have risetime 1.75ns. (yes I made also risetime measurement with fast "spike" pulse. Oscilloscope show <1ns risetime and this kind of bullshit I do not publice)
Aliasing.
1ch use. Yes if signal includes over 500MHz frequency components it have aliasing and here Owon is not alone. It depends what is over 500MHz signal total attenuation and what is over 500MHz frequency fractions level in measured signal.
2ch use just change 500MHz to 250MHz. This is natural.
But try find oscilloscope what keeps these frequency limits down to 500us/div or 1ms/div. Owon keeps 1GSa/s or 500MSa/s to these horizontal speeds.
If example take Hantek and Owon Single channel use.
Hantek Nyqyuist Nyquist
2ns/div. 500 MHz, Owon 500 MHz
20ns/div. 400 MHz, Owon 500 MHz
200ns/div 400 MHz, Owon 500 MHz
400ns/div 200 MHz, Owon 500 MHz (500ns/div)
2us/div 200 MHz, Owon 500 MHz
20us/div 200 MHz, Owon 500 MHz
40us/div 200 MHz, Owon 500 MHz (50us/div)
80us/div 100 MHz, Owon 500 MHz (100us/div)
200us/div 50 MHz, Owon 500 MHz
400us/div 50 MHz, Owon 500 MHz (500us/div)
800us/div 25 MHz, Owon 250 MHz (1000us/div)
2ms/div 1 2.5 MHz, Owon 125 MHz
Old digital scopes main problem was lack of capture memory. If old 500MSa/s oscilloscope (300MHz analog BW) have 1k capture memory. What is sample speed if it is example 1us/div speed and if it have 20div, maybe 50MSa/s and Nyquist 25MHz and front end analog part widely open up to 300 MHz without markable attenuation - yes there is really easy coming aliasing problem. What is Nyquist frequency, how about aliasing problem if compare example to Owon. Time difference is 20 year and price difference is maybe 30:1.
Imagined "old" 1k memory digital scope if it have single shot 1GSa/s:
And there is really high posibility to meet so bad aliasing that it may destroy all measurements without knowledge how to avoid or check if there is aliasing. Now if this imagined oscoilloscope have gaussian BW with (-3dB) corner frequency 100MHz we can of course draw curve and think how much there are "area" for aliasing with different horizontal speed settings.
(I assume that like typical tektronix displayed 10div and total capture is 20div)
50ns/div Nyquist 500MHz
100ns/div 250MHz
200ns/div 100MHz
500ns/div 50MHz
1us/div 25MHz
2us/div 12.5MHz
5us/div 5MHz
10us/div 2.5MHz
20us/div 1.25MHz
50us/div 500kHz
100us/div 250kHz
200us/div 125kHz
500us/div 50kHz
[/font]
Owon have not this but most of competitors have!
Then, with multi ADC interleaving system there come more problems. Example if there is 10pcs 100MSa/s interleaved ADC's for make "one" 1GSa/s ADC. IF timing is perfect and they are perfectly matched with levels, and only then, there is not problem if also they (and signal roads to them) are well matched over range. But becouse they are not in practice. Every these 100MSa/s ADC Nyquist is 50MHz. If they are not perfectly clocked and matched they produce lot of problems (becouse you do not know what is exatly time point related to other samples + level differents including whole frequency area) . Then it need heavy digital signal handling for cleaning, washing, whitening and dressing and never you exatly know what is truth. Maybe you clean out some truth but maybe not clean out some lies... then you can call it "garbage generator".Agen for aliasing:
Also this fast rising test edge includes lot of high frequency components! (original risetime on the generator output is around 70ps but with this cable and matching sure there is not these most high frequencies to Oscilloscope BNC. (They all are attenuated in this coaxial cable. If really want connect this fastest risetime to DUT, it need waveguide connection. Not common coaxial cable.
But so or so... This thinking about aliasing is one reason why I show two pictures. Single shot AND multi shot (using persistence to collect many shots together for display).
As we know aliasing do "corner wobbling" in signals what are not sinewave. Most high if signal is just fast rising square wave and specially if it include frequency components over Nyquist.
This Owon do not show high amount of corner wobbling! can you see it, can you tell that there is showed high aliasing problem now?
Home
Help
Search
Login
Register
