I work on designing condition monitoring boards with 24bit resolution with sampling rate of 50 ksps. I have some problems to reduce noise when I want to get analog signals and convert to digital signal. I use ADS127L01 that were made by TI. I use REF6033 as reference voltage in ADC and I need a noise level well below 100 microvolt but I have 2mV noise on ref_ voltage. How can I reject this noise?
Apparently, something is wrong with your application. I mean, the TI datasheet specifies 5 uV total noise, you report 2 mV. Points to problems with your application. This could be:
- Power supply issue
- Decoupling capacitor or output capacitor issue (wrong value,wrong type, resulting in incorrect ESR and non-stable amplifier)
- Dont know. Recommend you read the datasheet again, and analyse your circuit.
Henri
Henri Cloetens Thank you for your reply.
I considered your answer and I found that the problem is in power supply. Do you know any EMI filter that can help me? I have experience with Schaffner's EMI filter but I used them in AC supply in other project. This project is instrumental system and so sensitive about noise.
Dear Sir,
What noise do you have ?. Where is it ?.
You should do measurements with an oscilloscope, and debug it.
Best,
Henri.
Dear Henri Cloetens
Thank you for your concern.
When I measure the reference voltage between "REFP , REFN" ADC's pin with multimeter , I observed that varies randomly between "3.003-3.005V". It means a variation of 2mV and we call this noise. Because this random value is added to our ADC output. The ADC I assume compares its measurement (Ain) with this reference value to give us an accurate number. Please see the following, copied from the data sheet.
"The ADS127L01 measures the differential input signal VIN = (VAINP – VAINN) against the differential reference VREF = (VREFP – VREFN)."
My question is how to get rid of this 2mV variation, or it is not important?
Thank you again for your assistance.
Hossein