Apparently there are only specifications given for the DC output and two single point “signal” conditions in data sheet.
These are:
DC Accuracy ±(1%+2mVpp) HiZ load
Output Accuracy ±(1%+1mVpp) @ 10 kHz sine, 0 V offset
Output Amplitude flatness -0.3 to +0.3 db @ 50Ω laod, 0.5Vpp, compare to 1MHz Sine
Critically, there is no spec for offset.
The service manual “Performance Verification Test” adds more requirements for:
DC Output
AC Amplitude
Frequency Response
Each of these tests includes a range of amplitude values implying a more stringent requirement than the actual specifications, but again leaves many aspects not addressed such as offset. Also, since these are not published specs in the data sheet it’s unclear what there meaning implies but they are mentioned to be acceptance requirements upon receiving the unit.
I’ve never really questioned the output amplitude accuracy assuming with a 16-bit DAC it was probably good enough. Well, the other day I finally did notice it’s not that good! For instance, while programming a 500mVpp square wave the actual output on channel A was 544mV with a 3.5m v offset
and channel B was 576mV with a 11mV offset.
I thought this was high but was more concerned about the offset. I checked my old HP 33120A and it produced an output of 540mV with an offset < 1mV. Yeah, not that great either but at least the offset is close to 0V.
So I decided to do the complete acceptance tests described in the service manual – many results were out of spec. It’s pretty bad when using the scope the waveform obviously show an offset when everything is set to 0.
I attached my results below as a pdf file. The cal cert I got with the purchase only says it passed and has no results.
Looks like I’ll have to give Siglent a call to see if they can cough-up some calibration procedures.