Input impedance matching ADS131 (or other if better) with TIPD118

Ireneusz Magiera

Other Parts Discussed in Thread: ADS131E04, OPA376, TIPD118, ADS130E08, OPA735, OPA335

Hello,

I've found very useful circuit design in your resources: http://www.ti.com/tool/TIPD118#1

I have a prototype with ADS131E04, single supply AVDD=5V, internal Vref=4V also taken out from Vref+ and divided by 2 and buffered to provide CM of single ended input.

U1=OPA376, R1=36k, R2=14,3k, C2=1200p, R4=62, C9=100n

I usually notice about -7 value when nothing is connected to input circuit. I assume it is caused by not matched input impedances, am I right? How should I calculate IN- branch? I have used same 62R + 100n as low pass filter but it's sourced from "+" terminal of U1, which is sourced from buffered Vref/2 via 1k + 1uF.

This effect can be seen regardless of OFFSETCAL opcode using so I assume this command is executed with disconnected IN+/IN- inputs and can't compensate that.

I'm mostly interested in signals from range 10..2000Hz but I usually use 8ksps, 16bit is enough.

Or maybe there is another ADS better suited for 4 single ended signals?

over 9 years ago

0 Brian Pisani over 9 years ago

TI__Mastermind 38745 points

Hey Irenseusz,

If the inputs to the ADS131E04 are floating, the output is not a determined value so I wouldn't worry about the output of the device in that circumstance. If you only need 16 bits, I would advise using the ADS130E04. It is a 4-channel, 16-bit delta sigma ADC. What is your application?

Regards,
Brian Pisani

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

Hi Brian,

thank you for your reply. I didn't mean that inputs of ADS131E04 are floating. They are driven by TIPD118 circuit but nothing is connected to J1 input. I know that -7 is not much when opposed to full-range 16 or 24 bits but if it could be fixed by better design (with matched input impedances) it would be great to make such improvement. When I connect something to J1 this offset is also present and IIRC it slightly depends on signal amplitude.

Thank you for pointing out this ADS130 - I will take a look. Is this just ADS131 with lower resolution or should I expect more differences?

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

I have checked ADS130E08 (E04 doesn't exist, I think) and I will stay with ADS131 because of higher data rates available. Usually I use 8ksps which is ok with ADS130 but in other applications it would be nice to have ability to increase data rate if needed.

My current application is vibration monitoring.

About impedance matching I have found these documents:
www.ti.com/.../slyt310.pdf
www.ti.com/.../sloa054d.pdf
but it probably will take a long time to dig through without additional help.

PGA implementation shown on Figure 42 in www.ti.com/.../ads131e04.pdf differs from circuits presented in slyt310 document - signal is connected to positive terminal (CMOS) without feedback resistor RF. Input bias current specified in datasheet is 5nA, so I wonder if it really generates my problem or maybe it's something else?

0 Brian Pisani over 9 years ago in reply to Ireneusz Magiera

TI__Mastermind 38745 points

Hey Ireneusz,

If you look at Figure 12 in the ADS131E04 datasheet, you will see that the offset you are seeing is within the typical values expected for this device. However if you said you ran OFFSETCAL and the output remains exactly the same, then you're right it could have something to do with an offset external to the device. Looking at the datasheet for the OP376, it looks like a 5 uV offset would be considered typical. You could confirm this by probing from input to output of the amplifier. Otherwise, what are the resistors on the other input? Could the 5 nA leakage current be causing that much voltage difference between the inputs?

Regards,
Brian Pisani

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

Unfortunately at the moment I don't have access to good enough instruments to measure that directly.

IN- input is connected to buffered (with another OPA376) Vref/2 via 1k+1uF (connected to U1 "+" as Vcm) and 62R+100nF, so theoretically it could generate 5,31uV. But I'm not sure if it won't be compensated by U1 because voltage drop on 1k resistor also changes Vcm

IN+ input is connected to TIPD118 U1 output via 62R+100nF.

0 Brian Pisani over 9 years ago in reply to Ireneusz Magiera

TI__Mastermind 38745 points

Hey Irenseusz,

I'm having a tough time envisioning your setup. Could you provide a schematic to show the connections?

Brian

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

Here is the circuit:

0 Brian Pisani over 9 years ago in reply to Ireneusz Magiera

TI__Mastermind 38745 points

Hey Ireneusz,

I do not think the offset is coming from the current across the series resistors with the input. Like you mentioned, the feedback look for U2 falls outside of R7 so theoretically that input's voltage is set directly by the amplifier. The voltage drop across R8 has the potential to be 5 nA * 62 Ohms = 0.31 uV which is not large enough. However, like I mentioned earlier, the OPA376 as a typical offset of about 5 uV. This seems like a prime suspect for being responsible for the offset you are seeing.

Brian

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

Maybe you are right Brian. Out of curiosity I'd like to check another opamp as U2, with lower offset specified. Could you offer something for such application? Single 5V supply, unity-gain stable (also for Vref buffering), SOT23-5 package, rail-to-rail (or GND to 4V output).

0 Brian Pisani over 9 years ago in reply to Ireneusz Magiera

TI__Mastermind 38745 points

Hey Irenseusz,

Both the OPA735 and OPA335 are specified to have 1 uV typical offsets with a 5 uV maximum. In addition, both devices come in two channel variants.

Brian

0 Ireneusz Magiera over 9 years ago in reply to Brian Pisani

Intellectual 335 points

Ok, thank you very much for your help.

Data converters

Data converters forum

Input impedance matching ADS131 (or other if better) with TIPD118