6074BD - offset errors

gf1

Dear Support,

on my 6074BD I notice significant offset errors.
Utility -> Calibration does not eliminate them.
The following image was captured immediately after running "Calibration":



As you can see, CH1 has an offset error of more than 1 div (and at 2mV/div, the error becomes even 2 div).
CH2 and CH4 have some offset error as well. Only CH3 happens to be almost OK (but still a bit off).

Btw, all inputs are terminated with 50 Ohm resistors - no signal present.


Turning the 20MHz bandwidth limit on or off also moves the trace a little but up or down.

The AC/DC switch seems not to affect the offset

After calibrating all 4 channels, I turned off CH2, CH3, CH4 and switched to a faster timebase of 500 and 200 ns/div. The offset of CH1 did change to than 5 div:



Sotware version 20160628
FPGA version D001
PCB version V1.06

[ I had installed the software from the official product page. Is there any newer versionavaiable? ]

gf1

gf1

I do not see a statistically significant difference with open inputs and 50 Ohm terminated inputs. The resulting offsets vary anyway when I run the calibration multiple times in sequence. And after a calibration run, the residual offset is not invariant either, for each channel, but varies with the set of enabled channels, V/div, time/div and bandwidth limit setting.


[ It were certainly not the best idea, though, to keep the probe connected and to short the ground lead to the probe tip, since this "loop antenna" may pickup noise from the environment. ]

gf1

gf1

When I look at the PCB photos available in the internet, then V/div attenuators, ADC driver, channel multiplexer look quite different in the 2000 and 6000BD series. One common sub-circuit is obviously the dual-path input buffer amplifier, which uses a discrete FET source follower on the fast AC path, and an AD9510 on the DC path. This sub-circuit seems to be present - with small variations - in several scopes (IIRC also in some Rigol ones).

gf1

gf1

OASJ2YSeeEhBQo1 replied at 2019-3-24 12:15
I got in contact with Hantek by e-mail and they replied with one possible solution.

Try this: conne ...

Thank you, I will try...

gf1

Thank you! Works much better now