RE:Connecting Thermocouples to the ADS1148

Hi Dhiren,

I split your post from the previous discussion as the topic changed from digital to analog.  There is a big difference between an RTD measurement and a thermocouple (TC) measurement.  The RTD measurement requires some form of excitation whereas the TC does not.  The TC requires cold junction compensation (CJC) at the point where the TC connects to the system, whereas the RTD does not.  Both devices require that the input be within the correct common mode input range for the ADS1148.

In the RTD case, the correct common mode is established by the reference resistor.  In the more normal TC case, either VBIAS is used on the ADS1148 to place the TC into the correct common mode, or external biasing resistors are used.  However, this may not be easily done in your situation as the analog ground may be the same potential as the grounded TC.  So you must make sure that the TC when being measured is in the correct common mode range for the ADS1148.  If the analog ground and the TC grounding point is at the same potential, then the better course of action would be to use bipolar supplies (+/-2.5V for example) where then the analog ground is well within the common mode input range (mid analog supply).

You will need to have a cold junction where the TC connects, which is an isothermal block where the temperature remains consistent across the input terminals.  This temperature is measured and used to correct the actual TC measurement.  Right now your circuit shows no means to measure the temperature of the CJ.  Also I noticed, and I think I commented on another post, that the differential cap values at the inputs should be at least 10 times greater in capacitance value than the common mode caps of your analog input filters.  As far as any additional filtering, this really is determined by the environment the TC will be in as well as any possible transients that may occur.

Best regards,

Bob B

Hello Mr.BOB. Thank you for your support. I have changed my circuit of input in my schematic as follows. Here I have changed the supply AVDD = 2.5V and AVSS = -2.5V as according to you dual supply is required for grounded thermocouples. I have kept the schematic common for RTD type  as I want to make the circuit common for T/C &  RTD. Please check whether the circuit schematic as attached 7382.CPU.pdfis OK.?  Please refer sheet 2 and sheet 3 where through jumper S14 and S17 I have given 2.5V and -2.5V to AVDD and AVSS respectively.

Also I have taken care of CJC (temperature ) using external current source LM334 given to internal ADC using a stack connected as shown in sheet1 of my schematic via REL-3pin. The internal ADC code would be converted into equivalent mV and added with the mV obtained from ADS1148. I wanted to know whether the internal sensor of ADS1148 could be used for CJC calculation instead of LM334?

The differential cap would be 1UF instead of 0.1UF as shown in the schematic.

In my software as I have mentioned earlier the settings for thermocouple is as follows:

eg: channel1:
MUX0 = 41H

VBIAS = 02H

MUX1 = 30H

PGA = 50H

Please check whether the above settings are applicable for the new schematic and whether VBIAS would be required?

Best Regards,

Dhiren

Hi Dhiren,

I haven't forgotten you, I've just been really busy.  I will take a look at this in more detail soon and hopefully get back to you tomorrow.

Best regards,

Bob B

Hello Mr.BOB. I understand you have been very busy. It would be very helpful if you can give us your feedback to my queries related to my schematic so that I can incorporate changes if any in my schematic as my PCB designing is in progress. You have been giving the best support due to which we were successful in making 1 product which has been successfully working on field. 

          Waiting for your reply.

Regards,

Dhiren

Hi Dhiren,

I am very sorry for my delayed response.  Last week I had a family emergency which has delayed all my responses.  I will look at your information as soon as I am able.

Best regards,
Bob B

HI Dhiren,

You can use the BCS to check if the TC is open, just don't leave it on when you take a temperature measurement as this will add error.  In other words, take two separate measurements.

As far as the TC measurement itself, you must be within the correct common mode for the ADC.  As long as the TC contact connection is within the correct common mode range the measurement is valid.  So as I have stated in an earlier post, you need to make sure the potentials are referenced appropriately.  Analog ground may not be at the same potential as the TC connection.  Turning on VBIAS will establish a common potential, but there still may be issues relative to the power origination and the TC contact point causing a large potential difference.  If the potential difference is large enough, high currents can run through the ADS1148 through either the ESD cells or the VBIAS if used.  These high currents may damage the ADS1148.  Make sure that you do not exceed the absolute maximum specifications for the ADS1148.

Best regards,

Bob B

Hello Mr. BOB . Whether clamping diodes would do to protect the ADS input to not deviate from the common mode range? The clamping diode used by us is BAT54S. If clamping diodes are to be used whether it should be directly connected to the input of ADS1148 after filter circuit including the common mode capacitor?.

With regards,

Dhiren

I have just asked my PCB designer about placing the BAT54S at the input but he is finding it difficult to place BAT54S clamping diodes after the RC filteration circuit at the ADS input. He can place the clamping diode before the RC filteration circuit just after the input terminals. Whether the clamping diode can be placed just after the input terminal with a series resistor and then the signal may go to the ADS via RC filteration circuit?

Hello Mr. Bob,

I am now testing thermocouple input with ADS1148 for 8 channels on my PCB. The circuit design is as follows. Refer sheet2 and sheet34846.7382.CPU.pdfWhen I am connecting mV source signal individually to every input then my ADC code is not fluctuating and stable. But when I am connecting common mv source signal to all my inputs then I am seeing fluctuations in my ADC code. Also when mv source signal is changed the ADC code changes slowly. For connecting the mV source common to all inputs I have made VBIAS = 0. I have used single supply in the circuit using appropriate jumpers(S7,S11,S13 and S18) in the circuit. Please suggest whether any changes are required or where I am going wrong.

With Best regards,

Dhiren

Hello Mr. Bob, I have now added bipolar supplies (+/-2.5V) and kept Vbias = 0. Now I am not get fluctuations in ADC code and in turn the temperature display is stable when input from mVsource is looped to all inputs. AGND is still not connected to the input. As you have mentioned in your earlier post, if sensor is connected to AGND then Vbias should be OFF but when I turned ON Vbias I obtained some fluctuations which was not observed when Vbias = 0.

Please confirm whether Vbias should be turned ON or OFF when all my inputs are shorted but not connected to AGND. Please refer the circuit diagram  5826.4846.7382.CPU.pdf where jumpers S12, S9 and S8 is used to get bipolar supplies.

Also since it is observed that when input channels are shorted Vbias should be OFF and when the input is given individually to each channel Vbias should be ON and since on field both kinds of situation exists, 2 kinds of calibration(with and without Vbias) would be required. Whether there is a better solution than this as here I would have to store both the calibration constants and then use one of them after detecting what kind of input is given?

I would also like to confirm whether my schematic with bipolar supply is OK for RTD input also, since I have not used bipolar supplies with RTD input.

With warm regards,

Dhiren