Detecting a closed circuit with a few complications.

[system24seven]

Hi All,

First of all, thanks for the replies, this has been very educational.
Apologies for not replying sooner, I have been stuck in transit with the snow.

Contactor in use is:
Schneider LC2K1210B7
https://www.schneider-electric.us/en/product/LC2K1210B7/tesys-k-reversing-contactor---3p---ac-3-%3C=-440-v-12-a---1-no---24-v-ac-coil

Minimum switching voltage 17V.
4.5VA holding power to keep the contactor coil energised.


I did just realise that the contactor coil is in fact AC, and the power supply in the unit is not rectified. Fail.

I only need to monitor the limit switch state, as correctly pointed out, the limit switch/contactor is part of a hoist assembly.

It sounds like I need to use a two part approach and OR the results of voltage and current sensing.
I will be taking output from the detection as a logic signal, so the optocouplers are a great solution for isolation, and would have likely been part of the signal path in the final design.

The next query is whether changing to AC will require more complexity. I assume that I could rectify the output signal somehow, as I don't need more than 0.1 Second detection accuracy, so plenty of time to debounce etc.

In reality, I have a cable with two limit switch circuits sharing a common (one for up, one for down) and also a separate ground, and I would have two identical circuits monitoring these.

[BrianHG]

You can use an optocoupler to detect the current through the GO-switch. The circuit assumes that the 24V solenoid will work at 21V due to voltage drop across the optocoupler. Using the optocoupler will also provide nice galvanic isolation from the circuit you are monitoring. The string of 1N4001s are there to limit the maximum voltage/current across/through the optocoupler.

This is the same as my relay solution above: https://www.eevblog.com/forum/beginners/detecting-a-closed-circuit-with-a-few-complications/msg1441675/#msg1441675

I used 1N4007s as an example. One can select whatever diode is suitable for the job. I decided to post my suggestion as one might have an optocoupler and suitable diodes at hand.

Ok, you are correct.  But, I did the searching and math, as in my above posted, sitting withing the safe margins of the stratup current and startup voltage for the contactor with a little additional safe margin, the cheapest diode combination which I used was only 2x MUR540's.  However, at startup, the total voltage drop was 1.6v, once the contactor was sealed, this voltage dropped to 1.3v.  Unfortunately, 1.3v isn't good enough with a series protection resistor to turn on an optocoupler reliably, especially when you get that 1.6v or more startup current.  My original cheap 2$ 1.5v relay's 98ma draw and 1.3-1.6v coil operating voltage was the most rugged and gauranteed to function design with the available information.

+, the relay is good if all the op wants a large warning lamp, or buzzer/siren, even operated from 120v/240v.
Except your 1N4007's will explode when your switch on that 3.75 amp contactor coil and, with 3 of them, even at full 1 amp load, you get a 1v drop across each equaling 3v total meaning the contactor coil will get 21v.  According to the contactor data sheet I posted above, the absolute minimum DCV drive is 85% of 24v, or, 20.4v.  So, if the op's powersupply and length of wiring is only down by 0.6v in any of his machines, the contactor isn't guaranteed to turn on properly.  This is a super close margin for a 24v system, especially if the OP 24v source isn't regulated.

[system24seven]

Page 21 for those interested for the device schematic.

http://www.cm-et.com/Public/45552/CMET%20Lodestar%20Manual%2000000999%20REV%20AC-lowres.pdf

[BrianHG]

Hi All,

First of all, thanks for the replies, this has been very educational.
Apologies for not replying sooner, I have been stuck in transit with the snow.

Contactor in use is:
Schneider LC2K1210B7
https://www.schneider-electric.us/en/product/LC2K1210B7/tesys-k-reversing-contactor---3p---ac-3-%3C=-440-v-12-a---1-no---24-v-ac-coil

Minimum switching voltage 17V.
4.5VA holding power to keep the contactor coil energised.


I did just realise that the contactor coil is in fact AC, and the power supply in the unit is not rectified. Fail.

I only need to monitor the limit switch state, as correctly pointed out, the limit switch/contactor is part of a hoist assembly.

It sounds like I need to use a two part approach and OR the results of voltage and current sensing.
I will be taking output from the detection as a logic signal, so the optocouplers are a great solution for isolation, and would have likely been part of the signal path in the final design.

The next query is whether changing to AC will require more complexity. I assume that I could rectify the output signal somehow, as I don't need more than 0.1 Second detection accuracy, so plenty of time to debounce etc.

In reality, I have a cable with two limit switch circuits sharing a common (one for up, one for down) and also a separate ground, and I would have two identical circuits monitoring these.

WAIT A SECOND, your contactor has an AC coil, but, in your circuit,, you are using a DC supply.  Is this correct?

Ok, if your circuit is DC just change the 2x 'MUR420' diodes in my 2 design with 2x or 3x Diodes Incorporated '10A01-T'.
However, you need to be sure about your true turn on voltage of 17v, TeSys K contactor does not specify this.  From what I have seen, 85% of the coil specified voltage, ie 20.5v DC to guarantee turn on of the contactor.  Where did you get this 17V?  Is you 24v DC supply regulated, or, just rectified AC-DC with a 100hz/120hz pulse in it?

[system24seven]

WAIT A SECOND, your contactor has an AC coil, but, in your circuit,, you are using a DC supply.  Is this correct?

Ok, if your circuit is DC just change the 2x 'MUR420' diodes in my 2 design with 2x or 3x Diodes Incorporated '10A01-T'.
However, you need to be sure about your true turn on voltage of 17v, TeSys K contactor does not specify this.  From what I have seen, 85% of the coil specified voltage, ie 20.5v DC to guarantee turn on of the contactor.  Where did you get this 17V?  Is you 24v DC supply regulated, or, just rectified AC-DC with a 100hz/120hz pulse in it?

My supply is inside the hoist, which has a rectifier for another purpose (DC coil brake), and it has a conventional transformer, so the circuit is not using DC at all. That was my big mistake.

As far as the 17v, if you scroll waaaaaaay down the page, there is another table with "Complementary" title.
There is where the extra signal circuit details are listed.

[BrianHG]

  AC, this changes quite a bit.  Yes, the AC coils have a lower turn voltage range because of the peaks in the swing of the AC.

Also, taken from the datasheet:
Minimum switching current 5 mA signalling circuit ***  This usually means once switched on, this is the amount of power which can hold the relay in place.
Minimum switching voltage 17 V signalling circuit

But read just above:

Control circuit voltage limits     0.2...0.75 Uc at <= 122 °F (50 °C) drop-out
                                                0.8...1.15 Uc at <= 122 °F (50 °C) operational  ***This is 19.2v

Inrush power in VA 30 VA at 68 °F (20 °C)  ***This means around 1.25amps to switch.
Hold-in power consumption in VA 4.5 VA at 68 °F (20 °C)  ***This means around 188ma to hold.  These numbers make a lot more sense...

0.80, or 80% of 24v = 19.2v at 122 degrees.  The lower 17v figure is probably minimum at room temperature, not a good choice for guaranteed operation at all specified temperatures.

Next question, will an optocoupler output be sufficient for your detector.  What do you need to switch, an AC or DC device and at how many volts?

[system24seven]

I will be taking multiple of these detector circuits in using MCP23017 IO expanders, so an optocoupler would be ideal, as I could separate my +5v rail from the AC side of the system. Originally I was going to use an LM339 Comparator then chain in an optocoupler after that.

[BrianHG]

New schematic coming...

[Ian.M]

As its AC, the best option for the current sensing would be current transformers. That gives you isolation and minimal voltage drop. As the holding current is about 190mA with a 1.25A surge at pull-in, there's plenty of drive available.   Each current transformer secondary would feed a bridge rectifier (1N4148 signal diodes would be suitable as long as the transformer ratio is high enough) and then be loaded by a burden resistor and a Zener clamp to limit it to an appropriate logic level, with a capacitor to smooth the output enough so it stays at logic '1' between half cycles.   Choose the burden resistor so it reaches the logic '1' threshold at 2/3 the holding current.    For a one-off you could easily wind your own on pot cores.  I expect a 100:1 ratio would be satisfactory.

The voltage sensing will obviously use an AC input optocoupler, with a pair of suitable limiting resistors in series, with Zener clamping (2x 24V back to back, or a 24V bidirectional TVS diode) across the opto LED + the lower resistor to handle the back-EMF from the contactor coil when the Go button opens.

Its purely a matter of personal preference if you take the voltage and current inputs direct to separate MCP23017 inputs and OR them in software or if you use external OR gates to combine them before the MCP23017.

[system24seven]

As its AC, the best option for the current sensing would be current transformers.
...
 I expect a 100:1 ratio would be satisfactory.
...

Having never used current transformers, I had a search through the offerings at RS, they seem to only have large units >30A that install onto cables.
I assume there are tiny <2A PCB mount offerings available?

[BrianHG]

Ok, this circuit will work with AC or DC.
It is 100% isolated from your lift's wiring.  This means no powering, filtering, and protecting any analog comparators/op-amps on the AC side or within your MCU side since there is no connected link.
It meets all the min and max values for your contactor's coil surge - on - minimum hold for off - off currents with some good clearance.

You may parallel the transistor outs of the optocoupler if you want just the status of the limit switch, otherwise, having separate outputs may be useful to you.  I recommend a 10-100k pullup on the OC's transistor out with a 1uf to GND to filter out the 120/100hz from the AC cycling transitions in the circuit which the optocoupler is fast enough to pass through.



The contactor coil will get 20.8vac during power-up surge 21.4vac while kept on.  When the GO switch is open, the contactor coil will only see 2ma.  This circuit will withstand a 4amp continuous.

Note you can switch the 8 diodes with a current sensing transformer.

[Ian.M]

That's why I was suggesting pot cores.   Try winding 200 turns of fine magnet wire directly on the former for the secondary and a 2 turn heavy gauge primary wound on top of it.  Eventually you'll use enamelled solid copper wire for the primary, and protect the secondary with tape, but while you are messing around with it to get a suitable turns count and turns ratio, its easiest to use stranded hookup wire for the primary, so you can easily get to the secondary to add or remove turns.    Use a Variac feeding a low voltage transformer and a fixed resistor to get a suitable current through its primary for testing. 

Its also possible to build them on ring cores, but they are a PITA to hand wind so don't do that unless  you are planning on volume production.

If you want to use BrianHG's circuit,  you could replace the eight diodes with two bridge rectifiers.  Short each one from + to - and connect them in series via their AC terminals.
Brian's circuit is likely to be the cheapest and most compact, but you do need to be able to tolerate about 2V voltage drop, and we don't know how much is already taken up in wiring losses.

[BrianHG]

You can eliminate D4 and D8 in my design, however, during operation, the optocoupler LED will only see 2v going thru the 220 ohm resistor, 2.5v max during the current in-rush.  The nice thing here is regardless of some small losses in cables, these voltages are fairly closely constant as they represent the loop current and current across the diodes.

And yes, using a bridge rectifier with the +&- shorted is a cheap way to get 2 diode drop, but you will need 2 of them.

[system24seven]

You can eliminate D4 and D8 in my design, however, during operation, the optocoupler LED will only see 2v going thru the 220 ohm resistor, 2.5v max during the current in-rush.  The nice thing here is regardless of some small losses in cables, these voltages are fairly closely constant as they represent the loop current and current across the diodes.

And yes, using a bridge rectifier with the +&- shorted is a cheap way to get 2 diode drop, but you will need 2 of them.

In the context this circuit exists in, the cables between this detector and the contactor are normally 1mm diameter copper of up to 100m.
That equates roughly 4v drop in the line.

I am worried that this would be possibly unreliable at distance.

If I was to make a gain stage out of a couple of transistors and a few resistors, place this where the optocoupler is, and feed the opto with the output of the transistors, could I decrease the number of diodes to 2?
 
One forward, one reverse in parallel
~1V drop on the 24v line.

I would need to have a low voltage rail to switch with the transistor, but thats ok in context.

[BrianHG]

Yes you can.  I just checked the optocoupler datasheet.  Using only diodes D1,D2 & D5,D6, you will see a 1.35v drop (voltage loss across relay @ 200ma holding contactor on) and the optocoupler needs typically 1.2v to turn on.  During the 1.25 amp surge, you voltage loss at the contactor will now be 1.6v according to the MUR240 datasheet.

Going down to 1 diode, you are now talking about 0.65v across the diodes.  Even with transistor amplifiers, this is getting messy.  Having single diodes back-to-back, and using a signal transformer to send to an op-amp on your PCB or MCU to sample will be the only cheap isolated choice here.  In other words, for when the switch is open, use a single optocoupler.  When the switch is closed, use the output of the transformer which you will need to amplify with an opamp or comparator to feed a MCU digital input.

Using this transformer across 2 parallel back-back silicon diodes diodes for that 0.75v drop, you can now go to a schottky diode making your voltage drop go down to around 0.3v.

Xformer #1: https://www.digikey.com/product-detail/en/tamura/TTC-5023/MT7238CT-ND/674136
Xformer #2: https://www.digikey.com/product-detail/en/triad-magnetics/TY-400P/237-1133-ND/242655
Schottky : https://www.digikey.com/product-detail/en/vishay-semiconductor-diodes-division/SB540-E3-73/SB540-E3-73GICT-ND/3711864


However, the output of the transformer will be a 50/60hz square wave at 0.2v instead of the optocoupler transistor output of 5v.

[system24seven]

I think that gives me some serious tinkering to do now, ill build it and show my results. Could take a bit of time though.
Thanks all.

[BrianHG]

Using the 4 diode total scheme, only D1,D2,D5,D6 you may need to lower the 220ohm series resistor to 100 ohm, or even 47 ohm.

[Brumby]

You could also feed the meter voltage via a resistive divider to an analogue input of a micro and display it however you want.

... ... ... ... ...

I don't think the op will be able to get a good measurement electronics out of that except at the level of a good DVM.

* The analogue meter was just to illustrate the four conditions.  Placement of the dots was indicative (actual positions will be dependent on the math).  Try doing the same illustration with a digital one.
* Implementing a digital processing solution was always going to be the practical path.
* The concept is simple and valid
* The AC twist adds a minor complexity - resolvable in at least 3 simple ways that came to mind as I was typing this.

... but considering the direction this thread is going, I'll leave you to it.

[drussell]

In the context this circuit exists in, the cables between this detector and the contactor are normally 1mm diameter copper of up to 100m.
That equates roughly 4v drop in the line.

Have you measured the actual voltage in typical operation?  "24 V" is often more like 25.6 V to start with, for one thing....  Can you check the actual voltage at the contactor in a typical sample unit and get a real idea for how much additional voltage drop you're going to be able to tolerate?

[Zero999]

Lots of elabourate solutions have been proposed, involving opto-couplers.

Why not just use a current monitoring relay? It goes in series with the load and closes a set of switch contacts, when the current exceeds a certain threshold. The downside is, most current monitoring relays I've found are quite expensive. As this application is AC, a current transformer could be used to switch an ordinary AC voltage relay.
https://www.digikey.com/product-detail/en/littelfuse-inc/TCSH2A/F10659-ND/7931381

[Brumby]

Why not just use a current monitoring relay?

Perhaps because the state of the limit switch is required to be known when current is flowing ... and when it's not.

[Damianos]

Continuing to play on this subject, I attached one more idea!
Note: LED2 and U2 input must be connected as shown. An AC input optocoupler does not work in this circuit.

BTW: signaling minimum levels, that referred in datasheet, are relative to auxiliary contact (terminals 13 & 14), to keep it clean from corrosion.

[Kalvin]

Here is another one: Put a resistor in series with the GO-switch and select a resistor so that it gets heated to 60C when the solenoid is active. Then glue the resistor to Attiny85 and use the Attiny85 to measure the temperature. The temperature is "high" when solenoid is "on" and temperature is "low" when solenoid is "off". Alternatively you can use a diode or NTC/PTC as temperature sensor for the resistor and use Attiny85 to measure the voltage change across the diode/NTC/PTC. Of course you can build analog counterpart from a comparator or an op amp, but that would be too simple and old school.