Touch Switch Expansion

[pebe]

Hi Adam,

As Casdsys pointed out there are simpler ways of doing this. An 8pin PIC could do the timing, either with just a pot to give a continuously variable timed pulse – repeatable to within 0.5%, or with just a switch to program in any time period. It could probably provide for the touch sensor, also. I assumed you went the 555 route because you were unfamiliar with programming a PIC – but if you prefer to go that way, just say.

I have redesigned the 555 and built and tested it (Fig 4). I have added a spare switch from the 4066 package, and the rising voltage on C1 as it charges is now fed to the 4066 CNTR pin, and that switches pin6 to V+ to reset the 555 at the end of the timed period. That isolates R1/C1 from the bits around pins2 and 6 and allows C1 to charge correctly.

There is a difference in timing now, because originally C1 reached 2/3 V+ before resetting – now it is nearer 1/2V+. So it will be necessary to increase the value of C1.

I went back to my generic toggled 555 where the charge on C2 sets or resets the 555. It works faultlessly using a push button.

Now the touch plate circuit (Fig 2). I assume the devices Casdsys mentioned work on a change of charge, and that can either be –ve or +ve, whereas the plate considered here works on the 50Hz or 60Hz static that the body picks up from the mains supply. Its amplitude will be variable depending on local conditions, hence the variable pot I put in for sensitivity adjustment. I had to increase C2/R4 to integrate the signal, and found it would work OK (with a lot of electrical stuff operating near me) when R1 slider was at the bottom end (minimum sensitivity). Maximum sensitivity is just below the voltage at which the CNRL pin turns on the switch. Beyond that, of course, IC1 OUT will stay high and keep A switched to B.

I don’t know what your end product is and whether you are dead set on using a plate, but I thought of other ways.
1. A pair of LDRs so that the hand could shade one from the light.
2. A movement detector like this.
http://www.ebay.co.uk/itm/Mini-Adjust-I ... 20ce47a9a6

Best regards

[AlexHenighan]

Hi,
Your assumption was pretty much spot on - I am inexperienced, so I went for something that I thought I could handle (ok, clearly I got that wrong! ). Anyway, I am putting together a backlit mirror and wanted a nice 'clean' method of turning the lights on / off, with a timed operation to save energy (thinking on it, I will want to have the timer to run for perhaps 30mins, so I will put a fairly large capacitor on the timing circuit). Just thinking on it, I could set up a separate switch that would allow timed or permanent on function. Would a permanent on/off, just involve bypassing the capacitor? I may be getting ahead of myself here...

As this will be indoor use, in a bathroom I would imagine that approach you have given would work well. I guess if not, then the QProx devices mentioned may be an option. I am really not fixed on any solution, I just want it to be reliable.

Anyway, I will rig your latest version up and give you feedback on how it goes. And I will keep saying it, I really do appreciate your help - I wish I knew a little more about this but I still get very excited when things start to work, even when you are doing the hard stuff .
Many thanks.

Best Regards
Alex

[AlexHenighan]

Should have said, PIR sensor would be great (providing I can reduce the sensitivity so it picks up within 1-2mm only, but I need it to be a very small area - so the large plastic cover would need to go, and I would have to extend the sensor away from the circuit board (The workings for the lighting are in the centre back of the mirror and the plate is on the outside edge). I may well buy one (for the cost of a lottery ticket!) and see what I can do with it. Anyway, in the interim I will continue on the current path.

Thanks
Alex

[AlexHenighan]

Success...almost

I have put together the circuit - first the 555 timer, which works perfectly. Everything is spot on with it.

I then put together the touch plate circuit. Now, I have a couple of components that I did not have to hand - mainly the 1uF cap (C2). I played with putting some of the caps in parallel to increase (only had enough to get to 0.6uF). Interestingly, the circuit appeared to work slightly better without the cap in place.
But the key issue is that the circuit is very sensitive, even with R1 set to max. I suspect that may be because I have wired this up with bits of wire and floating components, so lots of long leads dangling around. Next step is to wire it into a vero board and see how it goes. I will also get a 1uf (and a 4n7) capacitor to make sure I am following what you have provided.

Should sensitivity remain too high, is it a case of changing the ratio on R2, R3? Or smoothing the line with a capacitor to ground? Just trying to get my head around the circuit.

Thanks
Alex

[pebe]

But the key issue is that the circuit is very sensitive, even with R1 set to max.

But the least sensitive position is with R1 set to its min position (so its wiper is at the 0V end of its travel). If it is still too sensitive then leave out R1, and reduce R2 and connect it to ground instead of to R1.
R3 makes no difference - it is there to protect the gate from being destroyed by too much static.

The 4n7 is not fussy. I only included it in case the charge on your body was negative, in which case IC1 would not switch on. 10nF or 22nF would be OK.

You can change C2 in value if you change R4 to suit. The time constant C2 (in microfarads) x R4 (in megohms) needs to be about 0.5sec. But the value of R4 needs to be no lower than 10K, otherwise the switch cannot cope with the extra current.

As a matter of interest, I designed a light operated switch for a railroad modeller in the US. That used a PIC and a couple of LDRs. The drawing shows the total component count for your job if you went the PIC route!

[AlexHenighan]

Hi,

I finally have the circuit working (albeit I have not cracked the sensitivity issues quite) but you sparked me to some thinking on the PIR approach for the switch - It means I could hide the sensor behind a small cover which suits my need perfectly.

So, I went and got a couple of Infrared sensors (IRA-E700ST0) with the thought that I would use them to drive the switch, but again, my lack of experience is seriously limiting my progress! I looked at the various configurations that I could find, but they all appear very complex with multiple amplifiers to create the necessary gain I guess. All I am after is a very low gain (triggered by something no more than 1mm away) that drives the timer circuit. (i.e. provides the connection across A & B).

All I have really done so far is to connect the sensor using a simple config (+12v rail through a 33 ohm resistor to the Vin pin (d), Pin g to ground and the output s left open). The voltage settles on pin s to about 2.5v unimpeded and drops to 1.5v when the sensor is covered.

I would be very grateful for your support again. (either using the sensors I got or something else).

Thanks..please shout if (when) you get tired of me

Many thanks
Alex

[pebe]

Hi Alex,

I don't think much can be done, simply, using the PIR you have. It requires amplifiers to get enough useful output. Here is a suggestion that should work - though I have not tried it. It involves two small sensors; one to start the timer and turn the light on, and the other one to turn it off.

Get two ST004 LDRs like this:
http://www.bitsbox.co.uk/sensors.html
Now discard Fig 2 of my earlier circuit, and modify the circuit of Fig 4 like this:
1. Remove C2 and R6.
2. Replace R4 with an LDR. (LDR1).
3. Replace R5 with the other LDR (LDR2).

A sensor could be fixed either side of the mirror. With both sensors receiving the same amount if light, their resistances would be equal and so their junction at pins 2&6 would be about half the supply voltage. Under those conditions they will not affect the operation of the timer. If you now shade LDR1 from light its resistance will go up and pin2 voltage will fall, triggering the timer. Likewise, if you shade LDR2 the voltage on pin6 will go up - resetting the timer and turning off the light.

As I said, it should work. But if it doesn't I have a PIC program written that would do everything you want (using those sensors), and I could program a PIC for you and send it to you for the price of the PIC (£1) plus postage.