Per the datasheet, the device has an 8Ohm ouput resistance.
I have a customer using the DAC driving a 2k load. The output error is 120 counts at 0xFFF. Based on the 8Ohm output impedance spec, the maximum it should be off is 16 counts.
Is there something I'm overlooking?
Hi Gregory,
I am not sure where you get the 8Ω number. There are few things that contribute to a DAC's loading error. If you are in the linear region of the output (supply headroom and footroom) for the output, then the output impedance can be used to calculate the error. But if you are operating at the max output, then your error could be worse depending on the supply headroom. Seeing as this DAC has VA = VREF, there is really no headroom when the output is at full-scale. This means that the output will start to collapse when you draw more current.
You can test it as well. I bet if you look at code =0xF00 the error will be much better than code =0xFFF. The DAC121S101 is a bit older device so we do have the relevant plots for this, but we can look at the DAC60501 for an example of this:
In figure 34 we see the relatively flat output when the DAC is loaded with ±25mA, this flatness is the linear region, where the only slope on that line is due to the output impedance. Once the load gets too great, the device enters short circuit mode.
Figure 32 we see the headroom required for the output to source certain currents. If you look at the red line you can see that after about 15mA, the error starts to increase. This is with 200mV of headroom. Now imagine what this line would look like if there was 0V headroom.
Hope this helps,
Thanks,
Paul
