This is probably the reason, why the fast pulse amplitude fluctuates on DST1062B while on Rigol DS1052E you can see nice stable pulse. Frankly speaking, I don't understand why Tekway/Hantek chose this way and I don't really understand how this can work when you need to capture fast single-shot signals. What is the single-shot bandwidth then? Based on what data is the gain calculated??
It might work for sinus, but how can this work for square with fast rissing/falling eges?
I've also found out, that the varicap bias in low-pass filter is still the same for DST1062B and DS1202B unit. They must do all BW control in VGA. The low-pass filter is used only for 20MHz BW limit on/off...
The digital filter does not have any influence on waveform as long the HF component necessary to rebuild the waveform is not filtered out.
It is by default set to bw limit, so the resulting waveform on the display looks similar as with analog filter.
The VAG is used in addition to control the flatness over the whole bw (it is hard to analyze the exact functionality)
Actually Rigol/Instek are doing exact the same, the analog low-pass filter is only an addition to what the firmware is doing
(plus of course the 20MHz filter functionality)
In principle you can add analog filter to Tekway/Hantek too, the waveform looks cleaner up to 20ns/div but as soon you switch to 8/4/2ns ADCs are running with 125Mhz which is producing some "nice artifacts" (amplitude changes). This isn't really bad design, as long the DSO is calibrated properly.
Especialy the factory calibration is very important, the self-calibration data is based anway on it (+ measured data created by DAC)
I did played a bit with the factory calibration, which is without internal information from dev team a real pain. I was able to increase signal quality (and
decrease drasticly). Haven't published anything about because it is not for everybody (yet? will see).
What the diff between Rigol and HanTekway is - well, Rigol is using 5 double ADCs running with 100Mhz, which is easier to synchronize (and reduce potential conversion errors). However, if you check Instek, they using 4 double ADCs like HanTekway - with no issues at all.
So you might ask what really different or wat is exactly the reason for the "nervous" signal ?
That's actually the the FPGA firmware itself. Where Rigol and Instek are doing 800wfrm/s (more or less serial), HanTekway is doing 2500 (more or less parallel).
Set your display refresh to auto or 30Hz, set avr to 4 and the signal looks close to what on Rigol/Instek. Sure pulse signal is still jumping around (well it is doing anyway, that's not a real stable pulse) but the reason is mostly in jitter/calibration. If you do single shot for 200MHz signal on a 200Mhz enabled DSO you will see that every second peak is having same amplitude (where by 300Mhz two ok, third bad and for 400 Mhz two ok, next two bad).
This is a perfect indication for not perfect synchronized ADCs (well, it will be never really perfect due the i/o pin jitter from the FPGA)
Anyway, HanTekway did updated the FPGA design for the new hw models for reason (let's hope they will still support and improve design for old hw!).
Btw, if you look in detail the varicap (and the two series caps) together with the opamps in/out impendace, the resistors network and
wires inductance is in principle a (R)ZLC butterworth low pass filter. You can actually calculate back from the 20MHz and the total capacity
of Varicap/C combination the (R)ZL values and calculate then again the cutoff for lowest varicap capacity. As i remember it was something about 280Mhz?
To be very honest i doubt that HanTekway are having good QC and spending hours for calibration, which we of course love because every
working hr costs money and we don't want to pay too much ... that's the point.
I wish HanTekway could tell me/us how they doing the factory
calibration in detail (i can guess and track many things, but not everything without spending tons of hrs on that). Sure not everybody will be
able to re-calibrate @home, but for those who have the knowledge and equipment it could be an option.