Hi James,
Thank you for the interest in the ADS1248. See my answers below in red.
Best regards,
Bob B
[quote user="James Anderson"]
Hi,
I am looking to update an existing wireless sensor design to be able to use 3-wire RTDs. In this case I discovered the ADS1248 chip.
I have a couple of questions:
1. With the ADS1248 using 3-wire RTD config, from what I understand from the datsheet I should be able to use a total of 4 RTDs. Is this correct ? BB> In theory this will work, but you may have an issue relative to analog input filtering. I would suspect that you will need some heavy filtering relative to noise (RFI) that may exist on the board. If the input filter resistance is too high, then you may end up with an issue with IDAC compliance, or an issue relative to component mismatch resulting in a difference voltage at the inputs.
2. My existing design has a single 3.3v supply from a tps61029 regulator. I see that for avdd if my supply is as low as 3v3 I may loose some of my pt100 range depending on the gain that I use with the PGA. Is this correct ? Is it worth changing the design to use a 5v supply such as the tps61026 for avdd ? How do I seperate the Avdd from the dvdd nicely ? What regulator would you suggest for getting 3v3 to the dvdd side ? BB> I usually generate the 5V and then connect the 5V output to the 3.3V regulator input. There may also be a dual regulator in a single package, but I'm not that familiar with the power components.
3. I would like to use the internal referenece of 2.048 volts if possible so as to keep my external part count as low as possible. What is the downside to this vs using an external one ? I would like to be able to get the full range of -200 to +850 of the pt100. BB> The internal reference is required to generate the IDAC sources. In turn, the best approach is to make the measurement ratiometric where the transfer function reduces to the ratio of the RTD to the reference resistor. This is accomplished by using the same excitation source for both the reference and the RTD. In this way the noise and drift will cancel.
4. Is it worth using the full range of the pt100 or should I take the effort adjust the range to be specific for each temperature monitoring application ? I would really like a one solution to fit all, but not at the expense of accuracy. BB> The answer to this question really depends on the desired temperature resolution.
5. This is going to be on a pcb for a wireless sensor that will use a nordic nrf905 433mhz transceiver. Are their any special shielding or ground plane concerns for the input pins with regards to my transceiver chip ? BB> Yes. Basically you have a low level signal effected by the environment. You need to prevent the noise of the transceiver from getting into the analog inputs. The degree of protection will depend a lot on how compact your circuit needs to be.
6. The datasheet is a bit confusing with regards to calibration. From what I can tell there are 3 kinds, system, gain, and offset. It says that for the gain calibration I must have an external full scale value, and for the offset I must have a 0 differential input. Now it also says that "it is recommended that calibration be performed after power on, a change in temperature, a change of PGA and in some cases a change in channel." To me this means that I in theory need a full scale input voltage as well as a 0 scale voltage. I am uncertain how this would be possible in an application like mine where the RTDs are always connected. Could you please elaborate on how I would actually perform the calibration via software on startup or when any of these condition changes that requires calibration occurrs. BB> The most important calibration is the offset cal. This can be issued using the SELFCAL command. Gain is trimmed for each gain setting for the ADS1248, so this may not be required.
7. The 833 ohm bias resistor. It is used in many of your examples. However this does not seem to be a standard value. Should I make this up by mixing and matching various resistors ? How important is the tolerance of this resistance value. Should I look at 0.1% or is 1% still ok ? This document http://www.ti.com/lit/an/sbaa201/sbaa201.pdf seems to suggest in section 2.1 that 1% resistors should be fine as long as the value is less than 20k. Can I trust this going forward ? I am thinking of 590E + 243E 0603 E96 resistors BB> The actual value is not that important as long as the range of the RTD remains within the common mode input range for the ADS1248 at the desired PGA gain. The tolerance of the resistor is not as important as having a low temperature coefficient for drift purposes. I mentioned the ratiometric measurement earlier where the result code is a ratio of the RTD to the reference resistor. If you do not want to calibrate your system, then you will only be as accurate as the tolerance of the resistor chosen.
8. I am from South Africa. Who is your distributor here? Do you have local tech support here ? At the moment we are buying from RS Components, however their tech support doesnt continue much further than general sales questions. BB> Unfortunately you know more than I do. Perhaps someone else will jump in and give a specific answer.
Thanks and Regards,
James Anderson.
[/quote]