Highest sensor read question

[Veganode]

Hello this is my first post, my problem as following:

1. I've 4 analog thermometer sensors and 4 relay switches
2. each switch will be controlled by only one sensor
3. I want to open the relay switch controlled by the sensor reading the highest temperature among the 4 sensors

how can I achieve that?

 

[Gryd3]

Can you link part numbers?

Will all of the components be centrally located?
Only the 'highest' temperature sensor will open the relay, all others will stay closed.
Will the relay switches cause a change in temperature that could result in two sensors 'fighting' to stay open by toggling back and forth, or will temperature change be gradual?
On the rare chance that all 4 sensors are the same, what do you want to happen? (Current thoughts indicate that no matter what... one relay switch will be open)

Can you give us a little detail on how fast you want it to respond, and what kind of environment it will be in?

 

[Veganode]

Will all of the components be centrally located?
Yes

Only the 'highest' temperature sensor will open the relay, all others will stay closed.
Yes, all other will be closed

Will the relay switches cause a change in temperature that could result in two sensors 'fighting' to stay open by toggling back and forth, or will temperature change be gradual?
No

On the rare chance that all 4 sensors are the same, what do you want to happen? (Current thoughts indicate that no matter what... one relay switch will be open)
They are the same and sensing different objects within the same location.

 

[Gryd3]

On the rare chance that all 4 sensors are the same, what do you want to happen? (Current thoughts indicate that no matter what... one relay switch will be open)
They are the same and sensing different objects within the same location.

I should have been clearer, by the 'same' I mean same temperature.

 

[Veganode]

I should have been clearer, by the 'same' I mean same temperature.

Yes sometimes they will have same temp, at this time nothing should change and all switches should be closed

 

[shrtrnd]

A variation on a 'sample and hold' circuit?

 

[BobK]

Sounds like a job for a microcontroller to me.

Bob

 

[Gryd3]

Sounds like a job for a microcontroller to me.

Bob

I've been suggesting mcu's a lot lately... figured I'd wait for another to say it.
Poll each one at a time. Would allow configuring the device to open change it's output on the 'highest' sensor read IF it is higher than the other sensors by a certain threshold.

 

[BobK]

And quite easy to tweak the behavior, unlike and analog circuit.

Bob

 

[Gryd3]

And quite easy to tweak the behavior, unlike and analog circuit.

Bob

For completeness.. do you have ideas on how this would work in an analogue circuit?
I've been trying to think of how I could possible compare more than two values at the same time. My first thought is the use of 6 comparators and some and gates... but that would never allow for all 4 outputs to be 0 if the sensors were the same.

(6 Comparators. Each comparator would monitor a pair of sensors. 3 inputs to each of the 4 AND gates would determine if all 3 comparators connected to a sensor are high and trigger the output. Adding an additional sensor would require 4 input AND gates, and 10 comparators)
I feel my analogue solution is an incredibly poor design, but I am not yet familiar with many of the tricks of the trade.

 

[KrisBlueNZ]

In case you're interested...



That's a design that I came up with for someone on a different thread who wanted to compare the received signal strength indication (RSSI) voltages from four radio receivers and choose the highest one.

The op-amps used with small MOSFETs as separate voltage followers. They are connected by the "IHR" (instantaneous highest RSSI) rail, which follows the highest input value. Whichever input has the highest value will cause its op-amp to drive its MOSFET to pull the IHR rail up to that voltage. Voltage followers with input voltages that are lower than the IHR rail voltage will "drop out" and will not activate their MOSFETs.

The IHR rail has a resistive load to 0V and this causes current to flow in the drain circuit of whichever MOSFET is active. This is sensed across a resistor and used to turn on a PNP transistor, which provides an indication that this input has the highest voltage.

The circuit is not perfect; if inputs are very close in voltage, it's possible that it could activate two or more outputs simultaneously, or possibly even no outputs at all! Some logic post-processing could be used to avoid those problems. A better way would be to modify each stage so it only activates if its input is somewhat higher than the IHR voltage. I did this originally then threw the drawing away :-(

 

[Gryd3]

Thanks Kris!
I knew there were other ideas out there for this. It was nice being able to see one.
(My short google search did not produce immidiate juicy fruit... so I did not look much)

 

[Veganode]

Many thanks to all of you, I'll research ur suggestions