frequency divider 2:1 reed contact & electro magnet

[alec_t]

Can the LED work w/o D3 & C4? It will be very short pulse then but I think will be visible, no?

Worth a try. Not sure if 20mS is long enough for everyone to see, but I guess it depends on an individual's vision.

The Q- on U1a is not connected in the circuit?

It's unconnected.

Maybe there is a solution with a coil that centers it magnetic field all at the top or 1 point

If you can't have the reed inside the coil then I think you will need a coil with a [ -shaped ferrromagnetic core having two pole pieces, one at each end region of the reed, so that half the reed sees a N pole and the other half sees a S pole. I don't know of any commercially-available coil like that; you'd probably have to make one yourself.

I found BC547B (instead of the 2N3904) and 1N4007 (instead of 1N4148) is that OK?

They'll do.

If you want to use a 3V supply then only R4 needs changing, to 100 Ohms.
If you don't like 470uF you could change it to 2.2uF and add a transistor stage as in the attached
View attachment 65546

Edit: As you say, before doing anything else it would be best to find a coil and see how much current is needed to operate the old reed.

 

[kael]

I saw this video

the reed is about 20mm away from the coil here!
Strange thing that this matter is not very common.
I'm not sure how to make one my own, I don't understand completely your instructions.
Maybe I should try a separate post on this subject in a hope to find a ready avialable solution, afterall it would be a shame if my project would stall on this matter.

 

[alec_t]

If you know that the video isn't a spoof (too many online ones are) and if the reed switch is several cm to the right of the coil (as it appears from the blurry image) then I'm surprised that setup works. That coil, btw, looks like one from a ding-dong door chime and its core seems to be the chime hammer.
Here's how I'd use a DIY coil. No guarantee it would work, though.

Edit: Personally, I'd prefer to use the Q1 transistor in my circuit, or something like that, to drive the speedo directly, rather than try to operate the old reed. Does the old reed just plug into the speedo? If so, can you use a multimeter to establish if one terminal of the plug connects directly to the + or - rail of the speedo circuit?

 

[kael]

a spoof, really? This same user has a second video with 2 coils, there you can see the core moving. Yes possible a door chime coil. I have one here (Friedland) but it is 8V AC, although I see a diode inside, maybe it will work on 9VDC also? It looks more like a solenoid the coil is 25x15mm, the core moves about 10mm, it makes me thinking, why not stick a small permanent magnet to the end of this core and let this do the triggering? But then there is still the amperage, I think such a coil needs 1A or so? The 9V will not last long then.
Like you say maybe it is better to go beyond the reed, I have another look at it. Why establishing the + or - rail on the reed terminal? To know how to connect it in the transistor circuit? If this is not possible maybe I can use the contact of a reed relais then, parallel over the old reed switch?

 

[alec_t]

Why establishing the + or - rail on the reed terminal? To know how to connect it in the transistor circuit?

Yes. I would guess (I could be wrong!) the reed simply connects an input point to ground (-) or to the positive rail. So a transistor could do that instead of the reed.

maybe I can use the contact of a reed relais then, parallel over the old reed switch?

Sorry, I don't understand. Won't that feed the original pulses to the speedo?

 

[kael]

No, since the reed relais would be connected in the 2:1 circuit. The reed relay his internal contact is then connected to the leads of the old reedswitch so that the 'new' plus is going to speedo and the old reed switch remains unused.
But this would only be necessary when it not can be established how the internal reed is now connected to the speedo, otherwise the diirect transistor route would be preferable.

 

[alec_t]

the old reed switch remains unused.

I still don't understand. Won't the old reed still be needed to supply pulses to the 2:1 circuit? Or are you proposing to have (1) the old reed, unused, (2) a new reed and (3) a reed relay?

Edit: Is there a plug/socket to disconnect the old reed from the speedo?

 

[kael]

The new reed is the one in your drawing on the left. The old reed supply pulses to the speedo when using the electromagnet way, or can be overbridged in parallel with the contact from the reedrelais when an electromagnet is not possible. But then I have to change the wiring on the speedo circuit. I had another look it's all sealed and watertight. So the way with and electromagnet still looks my best solution.
I saw these **broken link removed**
It's IP65, kinda small 18x12mm, powerfull 49N (w/o gap), 1,4W... what do you think is it worth a try? It's 12VDC, not sure what it will do on 9V?
There are others too from Intertec, cheaper but more power and not so strong...
PS: the poster from the video replied to my questions on youtube.

 

[alec_t]

Seems like a lot of money for what it is. I guess on 9V it should operate a reed, but it will draw ~90mA.
Why are you reluctant to cut the wire to the speedo? You could always re-join it with a waterproofed connector.

 

[kael]

Yes the price is kinda high but it looks very good quality to me IP65, low power, etc.., if it will do the trick I can live with it.
Cut the wire will void the warranty.
The current of 90mA is about 3x more than original, can the circuit (transistor) handle this? I thinking of using a 9V Lithium battery, gives 1500mAh!
OTR: what is that with the board here, my cursor seems to dissappear while typing now and then, on 2 PC's?

Thanks for all the help so far.

 

[alec_t]

Cut the wire will void the warranty.

Understood. Just don't spend more on the magnet than the price of the speedo .

The current of 90mA is about 3x more than original, can the circuit (transistor) handle this? I thinking of using a 9V Lithium battery, gives 1500mAh!

The base current of the coil-drive transistor may not be enough, so the single transistor may need to be replaced by a Darlington pair or a FET.

OTR: what is that with the board here, my cursor seems to dissappear while typing now and then, on 2 PC's?

Probably Google adverts or something butting in. If you use Firefox (dunno about IE) you can set it to prevent a page redirecting automatically to another page. That might help.

 

[kael]

Yes the BC 547 is maybe not strong enough. The datasheet says 100mA, so maybe it is worth a try. If not can I place 2 of these in Darlington with the same circuit?
I think then only 9V -0.65V -0.65V (voltage drop over the 2 NPNs) = 7.7V is there for the 12V coil...

 

[alec_t]

the BC 547 is maybe not strong enough. The datasheet says 100mA, so maybe it is worth a try

I wouldn't recommend it. It's good practice to allow a wide margin (say 50%), so 90mA through a 100mA rated device is pushing your luck; but as the 90mA is only intermittent you may be ok. If arranged as a Darlington pair only one will be passing 90mA of course.

 

[kael]

I made some progress!
The coil from relay I tried on 11/july do seems to work afterall. On my first test I didn't remove the attached switch part, so all the flux went through this metal part instead the of the reed I was holding against it. So now with the switch removed from the relais, the flux searches is way through the reed and is closing alright.
The resistance of the coil is 192 Ohm & taking about 40mA on 9V; so that looks the right spec's for me even with the BC547. And indeed I did a practical test on the original reed contact and it is working!

 

[alec_t]

That's good news!

 

[kael]

So finally I've got some results. It seems to work.... sort of.
The 2:1 divider seems to be erractic somehow. Sometimes I get a 1:1 pulse, the next two 2 turns a 2:1, then even 3:1, then again 1:1 for 2 times, again 2:1, ..... it constantly changes in these ranges.
I don't understand completely the working of 4013 chip but can it be something with RC circuit?
Any ideas what to try?
I didn't use C4 and D3 and the LED seems visible enough with the short pulse. I used a 470 Ohm for led resistor R4 instead of 1K (9V -1.9V for the led -0.6V for the transistor / 0.02A = 325 Ohm)

 

[alec_t]

Any ideas what to try?

Could be interference from the coil switching (or other source) getting back into the circuit and causing false triggering. Is the circuit built on a breadboard or a printed circuit board? Straggly wires on a breadboard are notorious for picking up interference. The circuit should also be screened by a metal enclosure, connected to circuit ground (battery negative).
Another possibility is a drop in battery voltage when the relay draws current. You could try wiring a 100uF capacitor across the battery terminals to see if that helps.

 

[kael]

No not on a breadboard but soldered on a pre-drilled PCB board about 6x4cm in size. Not easy to screen such a board w/o making short circuit. I was thinking to put it a small box but there is not much space for it, also the coil and reed are best left outside for optimal 'contact'. I have some small ABS and alloy boxes but then I still don't have a screen. My other option was to give the circuit some coating w/o a case but I don't have any experiences with this. And the battery still needs to be replacable.
But first I need to get it stable, yes perhaps the voltage drop, I started with a 9V from my multimeter, but I have fresh Lithium 9V available now.
Maybe interference the chip and transistor are only 1cm away from the coil. And some wiring is passing in between. May I can screen the coil with metalpaper on the other sides?

 

[alec_t]

Try disconnecting the coil and check the circuit operation using just the LED as load/indicator. If all's well then it's likely the coil was the source of interference.

I have some small ABS and alloy boxes but then I still don't have a screen.

An alloy box makes a good screen. You just need a connection from box to circuit ground. If the coil is outside the box and the circuit is inside you're fine. I wouldn't think additional screening of the coil itself would be necessary.

 

[kael]

That's a very good idea to disconnect the coil to try, why didn't I think of that?
With the coil outside the alloy box, I still need to protect it from the wheater elements any tips fot that?
About the circuit, does it matter how long the income pulse take? I mean if the reed is closed for 20ms, 500ms or even continously what happens with the output? Does is still gives 20ms pulses to the coil every 2:1 income pulse? Is it at the make or brake of the reed contact?
When the reed stays closed (wheel with magnet parked just in front of it) the coil is not constant on then?