When I use AD7671 to convert signal (1Hz to 1kHz), I found steps in the output.
Attachment figure 1 is the waveform, where green line is input of AD7671;yellow line is the output of AD7671 (a DCA output, the DAC operated well).
Hi daydayzhq26 ,
You mentioned that the yellow line is measured on the DAC. At what throughput rate the was the AD7671 ? and what is the rate of the DAC capture? What do you mean by DCA? If it is okay to share the schematic of your application so we may have a better understanding of the problem.
AD7671 has 64kHz sample rate (see Figure 4). It operates with a MCU (see figure 3). The DAC sample rate is 64kHz as well.
The schematic is in Figure 1,2. The MCU just receive data from AD7671 and send the row data to DAC. The data has no process.
The height and period of these steps is in Figure 5.
Figure 6 is the overview of AD7671 input (green line) and DAC output (yellow line).
Could you updata my question? Give me new reply?
I found ADC have different operating range in one step, in Figure 1. The signal in circle A is the normal operating range, and the signal in circle B is the abnormal operating range. The step is form by the abnormal operating range.
But I don't know the source of the range B.
Hi Henry Zhang,
Just some clarification, where is the voltage reference coming from? The reference needs to be very stable. Unstable reference would result in huge error on conversion result. I can see that, you used a opamp to drive the reference, but where did the reference voltage come from? I would recommend using the ADR421 and AD780 as mentioned in the datasheet and it is known to work with AD7671.
I use ADR01 as the 10V reference. Resistance divider generate 2.5V reference voltage. I measured the reference pin. I think it is stable now.
I find voltage step when I change the input DC voltage, as in Figure 1. Blue line is the input DC voltage, and Green line is the voltage step of ADC output.
You mentioned that you did a voltage divider on the ADR01 to have an output of 2.5V as reference to the AD7671. Where was the resistor connected? The total impedance on the reference output can interact with the ADC required reference voltage. With the impedance and the 200uA reference current of the AD7671 would affect the stability of the reference voltage during conversion. The AD7671 technology is a switch cap dac. During transition from acquisition phase there is a large current requirement from the reference.
On one of the snapshot, the time between each level is 62khz which is almost equal to the sampling rate. Would you be able to check the result when lowering the sampling rate.
I have a buffer with 47uF capacitor after the resistant divider to drive the 2.5V reference pin. I changed the sapmling rate to 1kHz and the result had no change.
I'll outoff office for Chrismas.I'll back next week.
I found when I chang the input driver of ADC(AD7671) the steps will be different. When I use AD797 as the driver the output is in Figure 1. When I remove AD797 and drive the input directly from the generator the output is in Figure2.
I did not use AD8021 which is recommand in ADC's datasheet. I'm afraid the problem is the input driver. What's your opinion?
I think that we both have the same step problem
I have checked the specs for the recommended ADR421 and compare with ADR01 in terms of the current drive capability and they are not very far. As you suggested about the driver. The AD8031 settling time is 125ns (at 0.1%) against the AD797 800ns at 0.0015%. This means that the AD8031 can settle faster but although the AD797 is slower but has higher accuracy when it settle. Using the AD8031 as driver, as recommended since this was proven to work with AD7671.
Retrieving data ...