>200 MHz -3 dB point is awfully close to the 250 MHz Nyquist frequency for a 500 MS/s sample rate. I'd expect aliasing to be an issue.
In single channel use as there read. (but analog BW is also of course with 2 channel.)
Yes aliasing is always problem with all low price digital scopes and also with expensive scopes.
If need really scope what never have aliasing problem it can buy analog oscilloscope. I want some day find littlebit better analog (today my analog limit is around 500MHz... but if find 1GHz analog (Tek 7000 series) maybe it is solution..
More aliasing with models what have not deep full speed memory and becouse this they, with lower horizontal speed, drop samplerate to very low.
Now this measurement was also done without probes. Probe is good filter.
Single channel it use 1GS/s down to 500us/div. (With 2 channel it use 500MS/s down to 1ms/div)
1ms/div. Nice for looking 1kHz. I can not imagine how much 250MHz signal is together with typical case when looking 1kHz.
If looking fast rising pulse edge. Well yes it includes high frequencies. 1ns risetime and there it go.
In what kind of situation you mainly mean this aliasing problem (if keep mind where we are in price class and competitors)
Two channel use and signal have markable amount >250MHz frequencies if use 2ns/div -
1ms/div speed per div. If look audio signals and afraid high frequencies aliasing... well. There is 20MHz BW and around 5MHz if switch probe to 1:1. And it is good to note, 1ms/div and looking example 1kHz or 500Hz signal... and yes... if there is >250MHz doughter signals with this low freq signal they do aliases.
Where is other cheap oscilloscope what sample 500MS/s for both channels if using 1ms/div speed.
If there is other 100MHz scope what have corner freq around 100-130Mhz and samplerate is example 50MS/s in same 1ms/div situation.. well... nyquist is 25.. and analog channel is widely open up to >100MHz but becouse it is not Owon.. it is not problem even if it is famously hacked to 100 - 200MHz... now it hack is welcome... so where is real problem.
Just with these aliasing problems it is easy understand whu full speed deep capture memory is extremely good. Some peoples have tell me that they do not need so much memory. Only normal oscilloscope screen is enough. hehe... after I show what this deep and fast memory really means they are surprised. But still, it helps only some. So do not insert >250MHz with 2 channe use and not >500MHz if one channel in use, then you have high aliasing problem.
But I understand well that name Owon make some interference and it may also make aliasing problem more bad becouse this is Owon.. we can take one 100k memory Agilent and compare aliasing.... my 1980's HP digital scope is clever. It give text warning on the screen about possible aliasing situation.
Here some pictures.
(do not care memory depth becouse with 2ns/div it use ofcourse highest samplerate also with 1k just as with 10M)
1. single channel, fast risetime pulse and 1GS/s (pulse itself risetime is 1,3ns)
2. double channel, fast risetime pulse and 500MS/s (pulse itself risetime is 1,3ns (generator minimum and it is ok))
1 and 2 signal exactly same.
3. double channel, "slow" risetime pulse and 500MS/s (pulse risetime adjusted littlebit slower than
1 and 2 pic. Not know exactly but something like 2-3ns (in this generator this adjusted risetime do not work perfect - need cal or new better generator (this is HP 8161A 100MHz programmable pulse generator with options.))
Here can see that this fast risetime is too fast for 500MS/s (it includes markable amount frequencies over Nyquist and so corners "wobble")
All pictures have same persistence for looking variations enough long time.Also signal matching to scope is not perfect. (it can not be becouse feed trough 50ohm is connected to reactive input. )
So also signal risetime (nominal 1,3ns) is only in generator output connector if it is matched perfect.
Later I try find scope risetime better with Tektronix fast risetime gen. (around 70ps)
Now I do not find this output adapter connector..