I'm using MCP4725 (datasheet) module, VCC is connected to a 4096V reference voltage. there's a steady 23mV offset:
0 - 23mV
50 - 73mV
100 - 123mV
...
How can I zero this offset error?
From the datasheet:
\$\begingroup\$
\$\endgroup\$
3
asked Dec 15, 2019 at 16:19
-
1\$\begingroup\$ Also see electronics.stackexchange.com/questions/406431/… \$\endgroup\$– Elliot AldersonCommented Dec 15, 2019 at 22:46
-
1\$\begingroup\$ Or this one electronics.stackexchange.com/questions/275337/… \$\endgroup\$– Elliot AldersonCommented Dec 15, 2019 at 22:47
-
\$\begingroup\$ @ElliotAlderson thanks for the links. \$\endgroup\$– ElectronSurfCommented Dec 15, 2019 at 22:51
Add a comment
|
2 Answers
\$\begingroup\$
\$\endgroup\$
5
At least two ways come to mind.
1) Put a summing amplifier after your DAC adding in a -23 mV offset to the signal. Need to be careful and use precision resistors so that you don't introduce more error than the offset you're trying to correct for.
2) Do it in software, which was suggested by the data sheet and Elliott Alderson. Just subtract the digital equivalent of 23 mV from the desired voltage value before writing it to the DAC. You would have to do some sort of limiting or saturation operation on the low end of the range since the DAC won't handle a negative value properly, to prevent wrap-around of the DAC output.
-
1\$\begingroup\$ It's a DAC, not an ADC. \$\endgroup\$ Commented Dec 15, 2019 at 19:08
-
\$\begingroup\$ Your right. Thanks for the correction. In that case then just change "before" to "after" in my response. \$\endgroup\$– SteveShCommented Dec 15, 2019 at 22:35
-
\$\begingroup\$ @SteveSh can you please add an schematic of summing amplifier which can trim offset error? \$\endgroup\$ Commented Dec 15, 2019 at 22:39
-
\$\begingroup\$ @Hamed - not quickly. There are lots of references for precision summing amps out there. If you Google "summing amplifier" or something similar, I'm sure get lots of hits to various app notes from TI, Maxim, etc. \$\endgroup\$– SteveShCommented Dec 15, 2019 at 22:51
-
\$\begingroup\$ @Hamed - One I have in my collection is AN-31 from National Semiconductor. For a historical perspective look for AN-4, another one from NSC. This one was authored by Bob Widlar, one of the early IC op amp gurus. \$\endgroup\$– SteveShCommented Dec 15, 2019 at 22:57
\$\begingroup\$
\$\endgroup\$
5
In software, you could subtract 23 from the digital value before sending it to the DAC.
If I am reading the datasheet correctly then this offset value is within the specified maximum \$V_{OS}\$ for the converter, which is 0.75% of full scale. If the full-scale output is 4.096V then the maximum offset should be less than \$\pm 30.72\,\text{mV}\$.
However, you should make sure that this really is a constant offset value. The gain error and non-linearity errors may be more significant over the entire range of output values.
answered Dec 15, 2019 at 18:57
-
\$\begingroup\$ But I need to to trim this offset in hardware because the output of DAC is going to be connected to an op amp... \$\endgroup\$ Commented Dec 15, 2019 at 19:27
-
\$\begingroup\$ I'm sending 0 in software and DAC output 23mV, how can I subtract that? \$\endgroup\$ Commented Dec 15, 2019 at 20:06
-
\$\begingroup\$ OK, if the value you want to output is 23 or greater then subtract 23 from it, if is less than 23 then output zero. That's really all you can do in software. The DAC has an offset...there is no method in software that will cause this device to output a voltage less than 23mV so the best you can do is remove the offset for larger data values and voltages. If that's not good enough then you will need to add additional hardware to deal with the offset. \$\endgroup\$ Commented Dec 15, 2019 at 22:02
-
\$\begingroup\$ That's the question; What additional hardware can trim this offset error? can it be done with an op amp? how? \$\endgroup\$ Commented Dec 15, 2019 at 22:37
-
1\$\begingroup\$ Well, your question was not clear about this. Yes, it can be done using an op amp, but you will need a precise and accurate voltage reference. Then just use a summing amplifier approach...that question has been answered here before. \$\endgroup\$ Commented Dec 15, 2019 at 22:42