I am planning to use AD8230 in my new design. I would like to know what is the bandwithd it supports for.
The AD8230 is recommended for amplifying the outputs of thermocouple and mechanical sensors -- generally low-bandwidth signals. Although bandwidth does not appear in the "headline" specs of its data sheet, graphs on pages 8 and 9 provide the frequency response. For programmed gains of 100 or less, the part is flat up to 1KHz, and about 3dB down at 2KHz:
The repetitive nulls at 6KHz and above are from the part's internal 6KHz sampling frequency. Sampling theory (same as for DSP) suggests that in any case, usage be limited to below fs/2 = 3KHz. Also due to the sampling, any resistances feeding the inputs should be below 10K ohms and well-matched.
Thanks for the quick response. I would like to know your prefernce for using AD8230 In-Amp for measuring temperatures using RTD, thermistor and measuring voltage and currents.
I have selected this part because of its offset drift and gain drift.
Also I would like to know what is the setup and settling for the same.
Hi Chiranjeevi M,
I believe the AD8230 can be used in temperature sensing applications due to its low offset and offset drift, as well as its low input bias current.
Regarding the setup, are you asking how to properly measure signals from RTD and thermistors? Basically RTD and thermistors are passive sensors which means they need excitation current to provide an output voltage. You can read more details on temperature sensors on this link.
I hope this helps.
I didnt mean setup to how the RDT works. I was asking for the Setup and Settling time of AD8230. Which is not mentioned in the datasheet.
Aslo i asked whether AD8230 can be used for sensing temperature using RTD and thermistor, Voltage and current sensing. Since AD8230 has limited Bandwidth.
Hi Chiranjeevi M,
Since the AD8230's settling time is unspecified, we can estimate it from the bandwidth:
Trise = 0.34 / BW, thus with a bandwidth of 1KHz, Trise ~ 0.34mS.
Settling time to a good accuracy will be several times longer than this 10-90% rise -- I would expect it to settle in less than 2mS. This is fast enough for most temperature sensing, although you may have a specialized application in mind. For voltage and current sensing, however, this BW is barely enough at line frequency (it adds a 3 degree phase shift at 60 Hz).
The AD8230 makes an ideal thermocouple amplifier (see Figure 39 in the data sheet). Thermistors have a much greater output, so they may not need an IA at all -- depending upon the precision you need, you might get away with just a voltage divider with your thermistor and a fixed resistor, directly feeding an A/D input with suitable protection components. The resulting temperature to voltage characteristic will be nonlinear, but it's cheaper to fix that with a lookup table or algorithm in your system's microcontroller or computer.
Thanks for the info. I have constrant of time in my design. I have to get the measured (Instrumentation stage, ADC stage) value within 1msecs per channel, I have 16 Channels and at the end of 16ms I need to start scanning the 1st channel.
In that case i would like to know what is the settling time. I have go through the datasheet once again I found that the Settling time is related to the input source impedane. There is no clear indication regarding what would be the settling time.
Since I will be using the same circuit to feed different type of sensors using an analog switch. I cant bypass the Thermistor or RTD inputs. It has to go through the IN-AMP with minimum gain.
Now I see why you would need quick settling for what's usually a slow moving signal! Overshoot and settling time are now serious issues, and often only faster amplifiers include those specs in their data sheets. Also your mux's output has frequency components beyond its 1 KHz switching rate, which will alias within the AD8230 -- causing intractable errors. In short, choose a different IA for this multiplexed signal.
Generally systems of the type you're designing amplify / condition each input before the mux. I'm a bit leery about exposing a typical analog multiplexer to "off-the-board", user-connected signals in an industrial environment. The multiplexers I'm familiar with tend to latch-up on any spike that exceeds their supply rails. This may not be an issue if you have more than usual control over your inputs (for example, if all of them are enclosed sensors you have designed or specified, where the end user has no electrical access). Otherwise, the mux may end up a bigger headache than the IA.
Retrieving data ...