ADS1100: looking for matching precision voltage reference

Daniel Ka

Part Number: ADS1100
Other Parts Discussed in Thread: REF02, ADS1110, ADS1118, ADS1115

Hello there,

my question is directed at AD converters ADS1100 that draw their reference voltage from Vdd. In other words, supply voltage is the reference voltage and as that should be precise and stable.

I'm working on an existing design where one MAX6195AESA+ voltage reference (Vin= 12 V, Vout = 5V, load current max. 0.5 mA, temp coeff vout = 5 ppm/°C, ripple rejection 72 db, load regulation 0.7mV/mA) powers two ADS1100 in single ended configuration (15 bit). Now the datasheet of MAX6195AESA+ states that it is stable with capacitive loads up to 2.2nF. However, the datasheet of the ADS1100 recommends to place 100 nF bypass capacitors between Vcc and GND: when converting, the device draws current in short spikes. The bypass capacitor supplies the momentary bursts of extra current needed from the supply.

So in the existing design the MAX6195 sees 200 nF capacity at the output which obviously does not meet its specification.

I'd like to improve this design. FYI, the application does not rely on measuring absolute values precisely but relative values, i.e., measured values should be stable without drifting over time or temperature.

Could you please recommend a voltage reference that is sufficiently precise and stable to drive two ADS1100. For example, is a REF02 stable enough for maximum AD precision (within one LSB)?

Alternatively, which AD converter would you recommend to replace the ADS1100 which includes a voltage reference (and is at least as precise as the ADS1100).

Thanks.

Best regards,

Dan

over 5 years ago

0 Joseph Wu over 5 years ago

TI__Guru* 98235 points

Dan,

I have two recommendations that you could try. First, you could try the ADS1110. This device is basically the ADS1100 with an internal reference. The register map is similar to the ADS1100 and you could just remove the existing device and replace it with the ADS1110 because the pinout is the same. I would note that the internal reference of the ADS1110 isn't accessible from any pin and the input range is ±2.048V max. Because the reference isn't brought out to the outside world, the reference accuracy is rolled into the gain error specification. It's an older device (like the ADS1100), but I suggested the ADS1110 because it's a simple replacement for your board to try.

Similarly, the ADS1115 could be a very good choice for your application. The ADC also has an internal reference (also not accessible from the outside world). This is a newer version of the ADS1110, with an internal reference that has better accuracy. It also has an input multiplexer, so you could reduce your two ADS1100s to one of these devices. Note that this device has a different set of configuration bits and registers. Because of this, the pinout and communication will be different, but the device does use I2C. Another similar device would be the ADS1118, which uses SPI.

Let me know if these work for you, or if you need other suggestions.

Joseph Wu

0 Daniel Ka over 5 years ago in reply to Joseph Wu

Intellectual 637 points

Hi Joseph,

Thank you for your recommendations.

Regarding ADS1110: I think it's worth trying this device as you assert that is very similar to the ADS1100. I just checked and figured that our relevant input range is from 0 - 1000 mV, so instead of PG=4 I'll just set PG=2 with the ADS1110. That should do it.

Regarding the ADS1115: the ADS1100 was used as a single channel ADC avoids the challenges that an input multiplexer can bring to a design. So I'm hesitant to employ a multichannel adc for the time being.

I'd appreciate if you could suggest a matching voltage reference to drive two ADS1100?

Thanks again for your swift support.

Dan.

0 Joseph Wu over 5 years ago in reply to Daniel Ka

TI__Guru* 98235 points

Dan,

I think a good reference would be the REF5050 (or REF5040 depending on what voltage you want for the ADC). These references have high accuracy and stability. It also has a large cap load drive and should be able to supply current for the two ADS1100. If you look in the REF5050 datasheet, there are guidelines for the load capacitance:

"A 1-μF to 50-μF output capacitor (CL ) must be connected from VOUT to GND. The equivalent series resistance (ESR) value of CL must be less than or equal to 1.5 Ω to ensure output stability. To minimize noise, the recommended ESR of CL is from 1 Ω and 1.5 Ω."

If you have questions about the reference, I would post in the appropriate forum. They will have more detailed knowledge about reference usage.

Going back to the ADS1115, I think this would a good device for your application. While we currently don't have a good EVM for this device, you can find a breakout board from Adafruit to help test this application.

https://www.adafruit.com/product/1085

I would use the ADS1115 in single-shot conversion mode, alternating between the channels that you want to measure.

Joseph Wu

0 Daniel Ka over 5 years ago in reply to Joseph Wu

Intellectual 637 points

Hi Joseph,

That's helpful. The REF5050 (Vin range up to 18 V, load current max. 10 mA, temp coeff vout = 3 ppm/°C, load regulation 30ppm/mA) exceeds the REF02 specifications and also looks like a better match than the MAX6195.

In a new design I'd also evaluate the ADS1115. However, in an existing design, rules of change management apply and smaller changes are more likely accepted than bigger ones. So I tend not to change the converter configuration (i.e. single channel to multi channel) and instead change only the voltage reference (i.e. different external reference or ADC with internal reference) using the same single channel ADC configuration.

Thanks for your comments.

Dan

Data converters

Data converters forum

ADS1100: looking for matching precision voltage reference