[MattTheNoob]

Ok electrical gurus!

I'm setting up my trunion table (heavy, German), and the limit switches on it have me stumped. They are 2 wire, and the signal is <1mA off, and >3mA on. So the 8Vdc + and - get hooked to the two leads and I need some way to detect that tiny current and get it out to a relay contact that I can use.

The other useful part is a marking that says Ri=1K Ohm.

[keebler303]

I doubt you have limit switches with a current output. Where are you measuring this current? Do you have a picture of the switch? I think you might not be connecting it correctly, which is giving you funny results. Post up some more details.

Matt

[MattTheNoob]

I would doubt me too. The switch is a Pepperl+Fuchs 83/2022X. Metal, threaded barrel about 3/4" long and 3/16" dia (smallest prox I've ever seen), with potted cable. Only two wires, blue and brown. On the cable is printed "8V-" and BN+ and BU-.

I tried measuring voltage in series (thinking that it might be a mag switch)... nothing, read the same voltage with or without proximity to target.

Then I did some Google-Fu, and while I couldn't find a spec sheet for the switch itself, I did find it as an option in a flow control valve. If you follow the link and search for 2022x, you'll find the specs...

Inductive limit switches,
Proximity switches SJ 3.5 SN, CENELEC intrinsically safe,
EEx ia IIC T6, PTB No. Ex-83/2022X,
Limits freely adjustable between 0 and 100 % of stroke,
Current circuit to DIN 19234 with 8 V DC control voltage
Control current < 1 mA = logical “0“
Control current > 3 mA = logical “1“

http://www05.abb.com/global/scot/scot211.nsf/veritydisplay/8a255238f1aa2d2dc12574cc0038b263/$File/13_832se.pdf

So, I hooked it up in series with my amp meter, and sure enough got the current change with proximity.

[Al_The_Man]

Generally two wire prox switches are used principally with a low current level logic input to I/O logic devices such as PLC's.
If you want to switch a relay, a 3 wire type with transistor output is usually recommended.
Al.

[keebler303]

That is strange, I have never seen a switch like that.

Do you know about Ohm's Law? I think you need to put a resistor in series with the switch and measure the voltage drop across the resistor. That voltage drop can be fed into an op amp comparator which will give you a 5V logic level output based on the state of the switch. Do a little more google-fu and you should end up with a working circuit for about $1.

Matt

[MattTheNoob]

No doubt, and I'd buy something along those lines if I were to start from scratch. That being said, I know there must be a way to cobble a few transistors and optocouplers together to get me where I want to go.

(I only picked those two parts because that's what I needed to build my gecko monitor circuit which was similar... having to detect the 5v without pulling any current.)

[Al_The_Man]

For 2 wire examples Check out post #252 here.
http://www.cnczone.com/forums/open_s...e_easy-21.html
There many 2 wire that will go up to 100ma, did you try and see if it will support a current burden of ~50ma by trying a 160ohm resistor in series as suggested and check the volt drop across the res? If it does then you could sub the res for a relay.
Al.

[irving2008]

Here's one possible solution that uses minimal components...

This wont drive a relay direct (why do you need to drive one?) but its easily adapted...

[irving2008]

For 2 wire examples Check out post #252 here.
http://www.cnczone.com/forums/open_s...e_easy-21.html
There many 2 wire that will go up to 100ma, did you try and see if it will support a current burden of ~50ma by trying a 160ohm resistor in series as suggested and check the volt drop across the res? If it does then you could sub the res for a relay.
Al.

It'll fry if you do... I've seen those ones before, the internal chip won't handle it...

[MattTheNoob]

COOL!

Now on that graphic, the resistors are shown as 1k0 and 4k7. Are those supposed to be 10k and 47k? Also, a quick check of my electronics store (Ebay) shows 74C14J's, N's, and D's. Do I worry about that?

[Al_The_Man]

It'll fry if you do... I've seen those ones before, the internal chip won't handle it...

Do you have the spec sheet on it? It would help in making an informed decision.
The Pepper & Fuchs catalog I have, most of the low voltage two wire go up to 100ma rating.
I don't see much point in them using such a low rated one in that application?
Al.

[irving2008]

COOL!

Now on that graphic, the resistors are shown as 1k0 and 4k7. Are those supposed to be 10k and 47k? Also, a quick check of my electronics store (Ebay) shows 74C14J's, N's, and D's. Do I worry about that?

They are 1k and 4.7k, that is the international standard way of showing resistances avoiding the issue of different decimal seperators (some countries outside the US use commas instead of decimal points)

The J, N, D usually designate different packaging and temperature standards... N i think is the one you'd probably want - normal 0 - 85degC plastic DIP package, but check.

[irving2008]

Do you have the spec sheet on it? It would help in making an informed decision.
The Pepper & Fuchs catalog I have, most of the low voltage two wire go up to 100ma rating.
I don't see much point in them using such a low rated one in that application?
Al.

No I don't Al, just from experience... and since they are expensive beasts and he wants to reuse whats there I wouldn't risk it personally.

[MattTheNoob]

I'm just getting smarter by the second, thanks to you guys.

Unfortunately I don't have the spec sheet, and couldn't find one on line. I think I found one in Turkish, but couldn't make heads or tails of it. I'd guess that these sensors are a mid-80's vintage.

[irving2008]

I couldnt find the data sheet for the switch itself, but there are plenty of products offering it as a protection mechanism, all of which seem to spec a 'secondary switching amplifier' with the switch (SJ 3.5 SN) as a 'switching initiator'. On that basis I'd be surprised it could handle any significant current. There are other variants (e.g. SJ 3.5 E2) that seem to have a built in amplifier.

-edit-... this data sheet http://www.acculex.com/pdfs/Protection%20Housing.pdf lists a functional equivalent and some more info and specifically states that it has no integral amplifier....

[Al_The_Man]

A nice series for interfacing is the ULN2xxx series http://www.st.com/stonline/products/...79/uln2001.pdf
Or the 2n7000 at ~25c.
Capable of switching 24vdc relay.
Al.

[mike_Kilroy]

u originally had a spec Ri=1k.....

I've run into some two wire prox switches in the past where, IIRC, put a 1k resistor in series with the two wires (polarity may matter so if it doesnt work one way, swap the two wires) and simply feed this into your logic input!

ie., put say 10v on one side of a 1k resistor then wire your 2 switch wires in series to the 10v common. Your switched output is the junction of the 1k resistor and the wire attached to it. no decoding required. no ICs. said yet another way, your output is across the two switch wires and provides a 9v or 1 volt output from 10v, similar for 24v (if it can handle 24vdc).

In one state it should draw the <1ma so the output voltage will be 10v minus the resistor drop (1k*1ma=1v) so 9v output. At other state of the two wire switch it will draw more than 3ma (like the min value where normal is probably around 10ma) so the output voltage drops down to 10-1k*10ma=0v.

Did you try it this way? Unless you found the real spec sheet on it and it is just the same as the Namur sheet someone listed before this. If not, remember to reverse the 2 switch leads and try again.

I'd venture if it does not work this way your switch has been blown by having direct voltage applied across it, shorting it out.

[MattTheNoob]

Irving2008, I'm pawing over that schematic and pondering the thoughts of Mr. Kilroy, and I'm hung up on that schematic. (In case you haven't picked it up, I'm pretty weak with this electronics stuff... but I like to think I'm a quick study)

Question #1. With the 1k0 resistor on the downwind side of the prox, it seems to me like the input to the schmidtt trigger would always be seeing the same voltage... but wait, maybe not... [this is the part where I may answer my own question, but I'll babble for everyone's amusement]. If the switch regulates the current, it must do it my changing its effective resistance. (tap-tap-tap... a little ohm's law work). So the input to the schmidt sees 1v when the prox is off and 3v when the prox is on.

So, is the schmidtt invert? The additional juice come from the power input on pin 1?

Question #2. 3v3 Zener diode will provide 4.7v to pin 1, but how did you size the 4k7 resistor to gound?

[irving2008]

Irving2008, I'm pawing over that schematic and pondering the thoughts of Mr. Kilroy, and I'm hung up on that schematic. (In case you haven't picked it up, I'm pretty weak with this electronics stuff... but I like to think I'm a quick study)

Question #1. With the 1k0 resistor on the downwind side of the prox, it seems to me like the input to the schmidtt trigger would always be seeing the same voltage... but wait, maybe not... [this is the part where I may answer my own question, but I'll babble for everyone's amusement]. If the switch regulates the current, it must do it my changing its effective resistance. (tap-tap-tap... a little ohm's law work). So the input to the schmidt sees 1v when the prox is off and 3v when the prox is on.

So, is the schmidtt invert? The additional juice come from the power input on pin 1?

Exactly so, the input voltage to the schmitt will follow the current flow, so 1mA = 1V and 3mA = 3v. The schmitt output is LOW (logic low = 0.4v) when the input is above 3v and HIGH (Logic high = 3.5v) when the input is below 1.4v. The schmitt's circuitry is powered from pin 14

Question #2. 3v3 Zener diode will provide 4.7v to pin 1, but how did you size the 4k7 resistor to gound?

Simple, it just needs to sink enough current to keep the zener in its breakdown zone when the schmitt isnt drawing any current (i.e. all outputs LOW). 1mA is sufficient so by ohms law 4.7v/.001A = 4700ohm. To be honest any value from 1k to 4k7 would do.

For your further education: If all six outputs of the schmitt are driving their rated output of 3.6mA then the zener is passing 6 x 3.6 + 1 = 22.6mA, which means @ 3.3v is is dissipating .023 * 3.3 = .076W. This is a bit close to 100mW so I spec'd a 500mW diode.

A small caveat: The switching spec for the schmitt input is 3.0 - 4.3v on the high side and 0.7 - 2.0v on the low side. This means that a particular combination of schmitt and proximity sensor might not work and will need a resistance change. It might be better to use 1k2 resistors instead of 1k0 to give better tolerances, tho you might need to go as high as 1k5 or 1k8. Increasing the voltage to the proximity switches to 10 or 12v would possibly help too. The Zener would need to be changed accordingly, 5v1 or 6v8.

[mike_Kilroy]

Seems to me you should get REAL data sheet from F&P? for your switches.

But the most common 2 wire prox switches work as I say - no need for external ICs and stuff!

Here is common data on a common 2 wire prox switch:

http://www.ia.omron.com/data_pdf/dat...eet_csm448.pdf

Look closely at the specs and you see they match your device: <1ma in one state, >3ma in the other. And polarity does not matter. If you run yours on 24v be sure to use a larger resistor than 1k - I suspect the Ri means MINIMUM resistance to put in series with it. Most should be good for use on 24vdc, and most are likely good upto 100ma.

Again, if you stick a 1k resistor in series with the two wires, measure across the 2 wires and they do not change from near 0v to near the voltage you supply, then the switch is either bad or is indeed one of those rare ones with half the circuitry missing inside.

Keep in mind that in either case, if you EVER - even once very quickly - put a voltage directly across the two wires it is 99% sure blown.