Multi-voltage switch sensing

[itsmoked]

A regular problem in industrial controls is the range of voltages one must sense. Users may have 12Vdc, 24Vac, 120Vac, or 240Vac signals that need to be sensed.

Classically one uses two 1/2W resistors(240V), one in each leg feeding the input of an AC opto coupler. Then on the other other side of the coupler you have your micro looking at the level.

It's a pretty fine line to walk between 12V and 240V. It often results in a lot of power dissipation because the resistor values have to be low enough to work for the 12V and are cooking at the 240V level.

Does anyone have ideas for robust general sensing input?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[VEBill]

If one was designing a single universal hardware isolation module for such a wide range of inputs, then perhaps simply provide several sets of input pins (and different resistors inside) along with installation instructions. The micro would have to be 'informed' of the calibration, but it's an engineered installtion so not a big issue. Each voltage range would thus have appropriate value, cool running, resistors.

Another option might be to reduce the duty cycle. Just turn on the voltage when the micro wants to make a measurement. I'd use a one shot to make it fail safe. Micro emits a transition, then has say 100 ms to make a measurement. Needs an optoisolated switch and is much more complicated.

How about a nonlinear device? Some Log type response where the current is limited, and the resultant signal is compressed and nonlinear.

Or just reduce the power, much smaller signal, but use more bits (16 should be plenty). Beware noise floor. Might need shielding etc.

Use a wee heater and measure temperature rise. Built in RMS.

 

[Paulusgnome]

With a bit of thought I reckon I could come up with a reasonably simple electronic circuit to do this. I am assuming that you must want digital inputs ie 0 or 1, not to actually measure the voltage. I am thinking of a circuit that might use a couple of diodes connected in anti-parallel to cut whatever input voltage was used back to +/- 0/7V, and an op-amp or two to convert the input into the required output. This is just the bare-bones of it though, it would take more thought than I have yet given it to come up with a circuit that will work. Entirely possible, but.

 

[VEBill]

My suggestions were aimed at sensing the voltage value, not just Off or On. I probably misunderstood the question.

The multipin input concept is still applicable.

The universal input On/Off detection circuit could be a bridge rectifier with filter cap, and a constant current source driving the optocoupler LED.

The problem with the overall concept would be that the limit would be the same for all inputs. So if the 240vac fell to 8vac, it might still register as being On. That's where a multipin input might be better.

 

[OperaHouse]

If you actually have micro access just perform auto ranging. Use something like 300K or higher sense resistor. Limit max voltage with a zener, then switch in additional shunt resistors to change range. Shouldn't take more than two outputs connected to resistors.

 

[itsmoked]

If one was designing a single universal hardware isolation module for such a wide range of inputs, then perhaps simply provide several sets of input pins (and different resistors inside) along with installation instructions. The micro would have to be 'informed' of the calibration, but it's an engineered installtion so not a big issue. Each voltage range would thus have appropriate value, cool running, resistors.

This probably won't fly for several reasons.
1) Every screw terminal is another expense and incurs assembler "hand-time".
2) It requires decisional thinking (a modern disaster)
3) It makes a product a lot bigger.

Another option might be to reduce the duty cycle. Just turn on the voltage when the micro wants to make a measurement. I'd use a one shot to make it fail safe. Micro emits a transition, then has say 100 ms to make a measurement. Needs an optoisolated switch and is much more complicated.

Hmm. This might work but does seem a bit over-the-top. May consume precious CPU pins.

How about a nonlinear device? Some Log type response where the current is limited, and the resultant signal is compressed and nonlinear.

I like this! Maybe need a current limiter that limits to something low like 1mA..

Or just reduce the power, much smaller signal, but use more bits (16 should be plenty). Beware noise floor. Might need shielding etc.

Null and void as I'm only seeking a solution for ON/OFF detection.

Use a wee heater and measure temperature rise. Built in RMS.

Not a bad idea since we can get pretty dang cheap chip thermistors now. Isolation might be a problem.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

With a bit of thought I reckon I could come up with a reasonably simple electronic circuit to do this. I am assuming that you must want digital inputs ie 0 or 1, not to actually measure the voltage.

Correct-toe-mundo.

I am thinking of a circuit that might use a couple of diodes connected in anti-parallel to cut whatever input voltage was used back to +/- 0/7V, and an op-amp or two to convert the input into the required output. This is just the bare-bones of it though, it would take more thought than I have yet given it to come up with a circuit that will work. Entirely possible, but.

Not seeing it since the diodes would still need to have current limited to them over the big span 12V to 339V(240V 0-pk)

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

If you actually have micro access just perform auto ranging. Use something like 300K or higher sense resistor. Limit max voltage with a zener, then switch in additional shunt resistors to change range. Shouldn't take more than two outputs connected to resistors.

I have to think about this.. This is sort of like VE1BLL's duty cycling scheme where the input isn't ON long enough to thermally matter.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[IRstuff]

What level of accuracy are they looking for?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[itsmoked]

ON/OFF. Voltage present/not present.
Accuracy? I'm not following your inquiry.

Keith Cress
kcress -

http://www.flaminsystems.com