Analog Discovery lost factory cal

[HK2]

Hello -

My Analog Discovery (first version) shows several offset on a analog input channel. I tried to reload the factory defaults but it doesn't change anything. In the last tab one can see many parameters for the factory calibration. They are all zero. Is this normal? The board has a good looking serial number and otherwise works great. I think the EEPROM should healthy but it is maybe partly zeroed. What can I do other than doing a manual calibration cycle? Is this warrantied by Digilent?

Thanks!

- Henry

[JColvin]

Hi Henry,

I'm not the most familiar with the original Analog Discovery, but from my understanding the factory settings should be good to go. You can go through a calibration yourself though on the initial WaveForms page in case some offset has occurred in your hardware. An engineer more experienced with the Analog Discovery commonly checks this sub section of the Forum and will be able to provide more advice/correct what I've said if needbe.

Thanks,
JColvin

[attila]

Hello,

The currently used calibration parameters are shown in the "Parameters" tab.
The last "References" tab only shows the current calibration process reference values, for the calibration done by the user since this dialog is opened.

What offsets levels do you see?
In case you want better precision than it comes from factory, you could try recalibrating the components you want. In case something goes wrong you can turn back to factory parameters any time.

 

[HK2]

9 hours ago, attila said:

 

Hi -

The channel 2 have 54mV offset. If the front multiplier circuit switches on, it reads 540mV. Activating factory parameters it doen't change.

[D@n]

@HK2,

Just a quick question: what are you using for a ground?

Dan

[HK2]

The ground from the Analog Discovery module.

[D@n]

@HK2,

And ... how is the voltage you are trying to measure related to that ground?  Are the grounds common between what  you are measuring and the AD module?  A "floating" ground, where the two "grounds" are not equal, could cause ... all kinds of measurement troubles that would look like calibration problems.

Dan

[HK2]

Dan, I understand what you meant. A OpAmp needs input current to establish a correct input voltage sensing. Right.

Channel 1 doesn't have the offset problem. And because all 4 analog input lines do have input resistive dividers there is no need to relate the input to ground. If plus and minus input lines from one channel are connected together, it shows the offset voltage. And this value is wrong on channel 2 !

Here are some pictures. Is it right that all values in the parameter tab are zero? Even if I say use factory values.

 

 Thanks!

[attila]

Hello,

I was thinking that you are using WaveForms 2015. The reference and parameter tabs are in different order in this obsolete WaveForms version.
I would recommend to use the newer WF 2015. https://reference.digilentinc.com/reference/software/waveforms/waveforms-3/start
You can also have a the same time installed the last version of the legacy WF: https://reference.digilentinc.com/reference/software/waveforms/waveforms-2/start

The offset could be because of the (un)calibration or the differential inputs hit the common mode limitation. 
In case you see this offset even when the p/n inputs are tied to device ground (or floating) it is a calibration offset.

The scope differential inputs are not floating these have certain common mode limitations relative to ground, see:
https://reference.digilentinc.com/reference/instrumentation/analog-discovery-2/reference-manual#scope_signal_scaling

Are you really sure that after pressing Load Factory under Reset menu you get only zeros?
It is strange because there is no access to erase or overwrite this from the WaveForms application.

[HK2]

Hello -

There is no difference. If both inputs are grounded the offset remains (As theory predicts). I use the old and the new software from time to time because I like the older GUI more than the new one. Anyway, the offset is exactly the same in the new WF Version 3.

Yes, I get only zeros after pressing Load Factory! I loosly remember that I've seen months ago there are many different numbers. So I think the internal EEPROM lost some data.

Thanks.

 

BTW: When do you expect a new software version? I miss the Impedance Analyzer in WF 3. And there is a bug: The phase plot is exactly inverse polarized. I would like the option to disable phase plot. There is no transparency 'color' so I cannot disable the phase plot.

Thanks for the otherwise great software.

[attila]

Hello,

This means the offset is due to the calibration. Try calibrating the device.

It is strange that you lost factory calibration. I have looked again the source of the applications and I see no way to erase this, and the eeprom data retention is +200 years.

I have compared the phase plot between WF v2 and WF v3 2015 and these are identical.

The phase plot is disabled when Channel 1 "Use as Reference" is unchecked or can be hidden by unchecking it on the right side panel.

In WF 2015 you can find Custom One/Two plots and load Example scripts for different impedance analyzer functions, like series capacitance, see the application help for more details.

4.7nF 

[HK2]

Hallo Attila -

I tried it on another PC and there is no difference, the same problem persists. Is a user calibration than written to the EEPROM?

I meant the hidden Impedance Analyzer and not the Network Analyzer. In the Network Analyzer I see no option for compensating probes.

Thanks!

- Henry

[attila]

Hello,

The user calibration is written to the same device eeprom as the factory calibration is, just in different areas.

There is no impedance analyzer in WF 2015, yet, just this network analyzer custom script.