RTD - Selectable Excitation Current

[airbusboy]

Hi

I'm working on an RTD temperature sensor module that must have up to three levels of redundancy, ie, upto three modules can simultaneously monitor the one RTD sensor.

I have a circuit in mind that wiil supply the 1mA RTD excitation current (see below).

The problem I have is that, when upto two other modules are connected, I need to reduce the excitation current of each module down to either 0.5mA or 0.33mA to ensure that the current sum stays at 1mA.

I was after some advice as to whether the below circuit would be suitable if I made RREF selectable, ie, 2K5, 5K and 7K5 for 1, 0.5 and 0.33mA respectively ? Or throw it away in favour of a better suggestion ?

Any advice gratefully received

Gary

http://files.engineering.com/getfil...ed-cad17a7abf37&file=RTD_Constant_Current.JPG

 

[ScottyUK]

Any reason why you aren't using three RTDs?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[IRstuff]

Is the RTD so reliable, and the modules so unreliable that such a configuration makes sense?

Also, if the modules are that unreliable, why wouldn't they fail in such a fashion so as to kill the ability of the working modules to sense the RTD?

I agree with Scotty that the optimum configuration is 3 independent RTDs and modules . That said, there may be other constraints that you haven't revealed yet. If so, then your design needs to ensure that the single RTD isn't the pinchpoint of the system, and that failure of any of the modules doesn't affect the performance of the others.

Moreover, I would also question why all your modules are simultaneously active. Assuming some MTBF, then all 3 modules will, on average fail at the same time, which would seem to defeat the notion of redundancy. A more typical redundancy design would only power up an offline module if the primary module failed.

TTFN

FAQ731-376

 

[ScottyUK]

IRstuff,

I disagree a little on the concept of redundancy: most redundant instrumentation systems are simultaneously powered up. Often a majority vote or a comparison among peers is used to disregard a defective sensor, but this requires that all are active together.

I can remember a heated discussion a few years ago regarding MTBF and MTTF, and that mean time between failures is different to mean time to failure and can't be interchanged. Not my field some I'm not going to argue it too much.



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

Scotty is correct. Any time you see triple redundancy its for voting. You can see which unit is wrong by looking at the two correct ones and continue operation. With only two you have to stop operation because you can't know which is the correctly working one.

The RTD is the most likely failure point so they should be triplicated. Any scheme that has the RTD drivers co-connected to one RTD is bogus and not redundant in any case.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[IRstuff]

Majority voter redundancy is suitable for cases where:
> repair/replacement is difficult/impossible
> the individual components are extremely reliable to start with

If you had a system where these modules had a 1 hour MTBF, then, your system's robustness ends rather quickly with the first failure.

The OP is silent about that, but his concern about "when upto two other modules are connected" implies that he's able to switch them out at will, which is not the same as having all 3 on-line at all times, because the excitation current would then always be the same. Unless the failed module draws excess power, you might as well just leave it in and keep the excitation current unchanged.

TTFN

FAQ731-376