RTD/TC i2c ADC 4 to 8 channels device

Hi Almir,

Generally high channel count systems don't make a lot of sense when using I2C as the communication transfer rate becomes too slow to be useful when changing mux and device configuration.  At this time time we do not have high channel count ADCs in the Delta-Sigma portfolio, although there are some SAR ADCs that have a higher channel count.  Using a SAR will require an external excitation method for the RTD.

Another possibility is to use a micro that has an integrated ADC and in this case will require neither I2C or SPI to collect the data as the hardware is integrated.

Best regards,

Bob B

Hi Bob.

first thanks for reply and explanation about high ADC channel numbers and it's speed. To be honest I don't need many channels, up to 8 is more than enough. Four channel is almost perfect for now. About speed over i2c, well my goal is to use rtd or thermocouple and at this point speed is not something crucial in my design. As you know temperature is very slow process (in my case this is particularly emphasized because process I want to controll is stove that uses pellets as fuel mass) and I don't  need data every microsecond. I read temperature from 4 sensors in about few second.

Just reading MCP3424 datasheet, and this is what I need (but want to use TI devices, if possible), not too many external components, and footprints is excellent. I'm reading ADS1015-Q1 and ADS7828 datasheets and it looks like those devices can be used in my design.
I have to check in detail what these devices can do, but if they can monitor voltage change for all channels then I can not see any problem to use them for temperature monitoring (at least temp. change is slower than voltage/current change).

My biggest concern right now is footprint, I have limited space and I want device which can help me to solve space problem, speed, like already said, is not problem, and this device will be only device on bus.

Just 4-8 sensors connected to ADC and one link to the microcontroller.

Will read more carefully those datasheets and come back to see is anyone used these devices to monitor temperature and transfer data over i2c.

When I complete scheme with devices I find suits my needs, I will post it here and any help/opinion would be welcome.

Sorry for my poor English, hope it's good enough to explain what I want to achive with these devices.

Best regards,

Almir

Hello Almir,

You can consider using ADS7828 (8 CH) or ADS7924 (4 CH). Let me know if you have any queries on these.

Regards,

Lokesh

Hi Almir,

A couple of things you will also need to consider when using RTDs or TCs.  With RTDs there needs to be some means of excitation.  The preferred method would be to use a constant current source and measure the voltage directly.  You could use a voltage source and a voltage divider to calculate the temperature based on the ratio of the RTD to the fixed resistor.  In this case the fixed resistor needs to be a precision/low drift resistor.  For the TC case, you need to measure the cold junction temperature of where the TC connects to your system.

Also, TI has a great portfolio of MSP430 micros with embedded ADCs.

Best regards,
Bob B

Edit: small error in scheme corrected (cathode of protection diodes D2 and D4 should be connected to +5V).

Hi to all,
below I will reply to both of you (Bob and Lokesh) and thank you for support.
Yes, i'm familiar with that...in theory.
When it comes to practical then I have some problems.
There is many documents on TI site regard this measurements, but I didn't find to many application note files to use it for measurements with thermocouples (so scheme is not completed for temperature measurements using thermocouple).
Document which seems good for my design is located at:
http://www.ti.com/lit/an/snoa663b/snoa663b.pdf
On figure 3 I saw good example how to do cold junction compensation.
My question is can I use this method and directly connect OUTPUT to ADS7828 input?
And is there any other devices which allow me to use lower voltages (my design use 12V max), much lower than 15V?
By datasheets it wouldn't be problem if I use lower voltages (even 9V should be fine).
Am I right?
If you look to the channels 5 and 6 you can spot that I connect positive end of TC to the channel 5. Would it matter if I connect positive end of thermocouple to the pin 6 of ADS7828?
I'm sure that application note for this type measuremens is available, maybe not for ADS7828 but I'm not only person who would like to build some kind of universal solution for temperature measurements.
If you can direct me to files where I can find more details about using ADS7828 with thermocouples I will be thankful.
One thing I note when reading RTD's v.s. theromocuple is that even thermocouple is chepaer that doesn't mean whole design would be cheaper compared to RTD oriented design.
If I decide to use thermocouple then I need extra components, mainly potentiometers which is expensive.
So what is your thoughts about my idea to make universal solution for temperature measurements?
Should I discard thermocouples and stich with RTDs?
Right now I can not replace micro controller, but thanks for tip, will consider it for projects in future.
I have read manuals for ADS7828 and decide to use it, ADS7924 is not easy to solder and this project is just started, so right now I don't want qfn packages in design.
When we go to production then I will consider using qfn type chips.
I attach part of scheme, with ADC module. Can you look at it and give me some tips about design?
I decide to use current source, it is more accurate than simple voltage devider.
My questions about RTD inputs is related to LM334.
1. Is this adequate for this type of measurements (I found REF200 also, but LM334 is very easy to find, relatively cheap and have small footprint) because I used to use one current source for 2 RTD's or should I use LM334 for all RTD's separately?
Using one current source for all RTD's I will save some space (and spent another channel, but when I need more channels it won't be hard to add LM334 for all RTD's and use all 8 channels).
Other concern about LM334 is initial accuracy about 6%.
Accuracy I would like to achive is about 0.5-1 degree, and with LM334 I'm within those boundaries).
Using LM334 can I expect to achieve this accuracy?
Other questions is about current for RTD excitation.
I decide to go far below 1mA, and current I choose is 100uA.
Is this ok for ADS input or should I increase escitation current?
About 1N457, used for improving temperature haracteristics of LM334, may I use 1N4148 (which I have in SMD and from datasheets and cross reference I read 1N4148 should be adequate replacement)?
Another questions is about protecting ADS7828 inputs.
Like you can see in scheme I put two 1N4148 diodes for one inputs, and because I used inputs in differential mode I think both inputs should be protected.
What is your thoughts about this type of protection?
Is this overprotection or good design?
If we take look for thermocouple as input we can see that there is no protection at all, so what protection method should be used for thermocouple input (can same 1N457 be used)?
Or both protection methods can be ommited if we take care about other protection method for whole circuit (reading datasheet I see that voltage at inputs can be about Vdd voltage high so my thoughts are  that protection could be omitted if we take care about protection of whole design)?
Note: I will use good grounding for all components, so RTD boddy (two wire) will be grounded too.

Now something about ADS7828.
Searching web I found some applications where pins A0 and A1 are connected to teh ground directly.
If ADS is only device on bus can I do the same (want to get rid of resistors connected to the A0 and A1 if possible)?

Sorry for long message but this is what I researched past 10 days and what is bothering me with this design.
Waiting for your reply.

Thanks again for all help.
Kindly regards,
Almir