marine navigation light problems

Hi Dianne,
Yes, you are right and indeed it is not the current what gives you the ultimate answer, but the light intensity. But Dianne, I have learned in life, that one can make it as fancy as possible, there are still loopholes and we still have to be on our guard. An example is if a bird does its thing on the glass front, or we moving into a fog situation. The fancy measuring of light intensity is gone on the blink. Measuring light intensity is a costly affair. extra wiring, and expensive globes. But your point is very valid.
Hi Ellawendy,
While driving to and back from Cape Town, I came up with the best solution for you and for myself. I still have to test it, as I need it for my own project also. But I would, if it works satisfactory, use solution no B or C together with a piece of copper wire of 10 meter extended in between the ground wire to the globe. I would then measure the Voltage drop over the 10 meter copper wire, ( 0.01 Ohm, rolled up) O.K, O.K. we could use an expensive 0.01 Ohm 10 watt ceramic resistor.
A Microchip with a 10 bit Analogue to Digital converter gives me 1024 steps ( 10 bit) Thus the 5 Volt is in steps of 0.00488 Volt broken down. I would measure the current (Voltage drop over the 10 meter copper wire) , store this in memory and then compare every 100 Milli Second in a loop. Should the Voltage drop suddenly vary, it means that a bulb is gone on the blink.
Advantage: You can immediately see which globes are faulty and switched on BEFORE you switch other globes on. Then with the added Microchip (cheap +/- 10 dollar) you measure and set the alarm when a globe packs up during the night.
Procedure :
You have first checked whether all globes are working (Solution B or C, thread 28)
Example, you have ONLY 7 out of the 13 globes switched on. You press a button connected to the microchip, and the current of those 7 globes are stored in memory, thereafter verified whether the current is still the same during the night, if not >>>> set the alarm or buzzer or bell or whatever.
The advantage is, that if different globes are used or a mix with LED lights, you just measure the current before night falls and the Microchip thereafter verifies that the current is staying the same.
Before somebody can shoot me down, I first have to create the circuit and test it for:
a) Is a 10 bit ADC sufficient or should it be a 12 or 14 bit ADC
b) How many steps do I have for the tolerance i.e. One 40 Watt globe is 1,666 Ampere i.e. a Voltage drop of 0.0166 Volt over the 0.01 Ohm copper wire >>> versus one 10 bit ADC step of 0.00488 Volt. Probably maximum 3 steps for tolerance. I need this should the battery voltage drop, or that the temperature drops drastically and has an influence on the copper and globe filament.
c) How stable and secure could we make it. I have totally faith in Microchip, but little faith in temperature fluctuations and battery voltage fluctuations.
d) Do I need a complex temperature and battery voltage compensation circuitry.
e) Give me some time to make the program and circuit.
Bert

 

God Bertku this bloke lives in a third world country ( donky carts on streets) and you suggest a microprocessor .......get a few reed relays and keep it KISS ...

 

You must be joking. So has South Africa, but we had 2 years before Australia fully Internet banking, ATM's and secure systems in place and even Europe only started a few years ago.
I had some private e-mails with him and know what his requirement is. Do you know that Cairo has 40 million people? But I can assure you, not only with donkey carts. Yes, their buildings look different then yours. But how do you think that the students started the revolution? With landline connections or with WiFi? Yes, their economy got a tremendous blow.
Bert

 

Thats always the problem when you conduct half the conversation re the problem off site leave the rest of us out of the frame ..... lengths of copper wire ....you could use some bits of electric fire element ......(nichrome wire) . dont follow this ..one minute you are keeping the cost down by not using 10w resistors and he next its microprocessors ???

dont reply got bored

 

I was in Cape Town remember. 400 km from home. By the way, it is more than 10 million people in Cairo and not 40. For testing it is better to use just a piece of copper wire. If it all works satisfactory, one can start searching who is making stable 0.01 Ohm resistors and improve the circuit.

 

Let me explain this. I can e-mail on my cell-phone (mobile phone) right through South Africa, by quick stopping along the road and knock a few buttons on my mobile (cell) phone., but don't feel like in depth discussions about the circuit via Boatdesign.net on the phone. Why don't you help me to quickly write the software for him? .
Bert

 

so now it comes down to the temperature stability of a resistor .....suppose we must stabilise the supply for the bulbs in case he starts the engine or the fridge comes on and the surge drops the current ......boring boring get a reed relay ..clunk click every trip .

Unsubscribed bye bye

 

Yes, I agree with you very boring. Get a reed relay and then a different globe is put in the holder and it does not work anymore. Or the coil must have sooooooo many windings that the contacts start sticking after a few years.

How far is your programming? Although I measure every 100 mS, the program will only kick the alarm/buzzer etc in, after the drop in current is measured over an extended periode of time, let say 10 seconds.

Does it matter, if you get woken up 10 seconds later? Or do yo think the spikes are 10 seconds long? I did some experiments with reed-relays. It has too many disadvantages. He should use magnetic sensors or resistors.

What do you have against One diode 1N4007, 2 IC's (7805 + PIC16F877) + 2 Elco's + 4 resistors 1/4 watt and a push button? ( we use the internal oscillator).

Bert

 

Pistnbroke, the challenge is on.

You make a system with 13 reed-relays. Like you said, the 13 LED's are connected to the reed-switch and then it latches the LED. The globe is removed and the reed-switch of the relay stay on.

O.K. O.K. we putting now a (break) push button in line with all the LED's.

What now, NO alarm. You need thus to design some form of electronics to drive the alarm/buzzer. OP Amps? 13 diodes and a transistor?

Now me.

I make a system with 13 magneto sensors. Need a 7805 for the sensors.
2 elco's for the 7805. The 7805 can handle up to 30 Volt. Maybe to protect the 7805, I use a small coil and a capacitor for spike protection.
The sensor switches the LED with a resistor to the 5 Volt line. If 13 globes are switched on, I need between 65 and 130 mA extra. Thus we need to put an heatsink on the 7805.

Now I start scratching my head on how to make this all working when only one globe packs up. Also how do I make sure that the Alarm is not going off, as soon I need to know whether all globes are working ? Do I switch all 13 globes on to see which one is faulty?

Is a Microchip not much easier??
One panel with 13 LED's (system B or C) straight away you see which one is faulty or switched on. Then with a Microchip measuring the consumption and compare ? two push buttons for the procedure, with all the flexibility?
Bert

Come on Pistnbroke , take the challence

 

Pistnbroke, I even post you 13 brand new reed relays and 13 faulty ones, which I took out of products from customers. At no cost. Both 13 are still working. But I don't trust the replaced ones anymore. Do you?
I even pay the postal cost for you.
Bert

 

Yes, indeed I could use a piece of Nichrome wire, provided we don't burn the boat down, also one could buy a resistor in Australia. Unfortunately none are available here, I phoned around, except If I order 1000 pieces.
Bert

 

Hi there,

Is there anybody who knows where to buy a 0.01 Ohm 10 watt resistor off the shelve, which has NO temperature and therefore NO resistance changes at 1 Ampere up to 25 Ampere? Your help is greatly appreciated.
Bert

 

Not possible as far as I know. Whoever needs that makes their own from constantan wire (copper, nickel and a dash of manganese). Buying the wire should be no problem.

 

Thank you CDK, It is thus easier for me to buy 10 meters of copper wire with shellack and wind it nicely in a roll, then to drive 400 km to Cape Town for getting special materials. In this part of the world, I only have to drive a few kilometers to a motor re-winder to get copper wire. For the rest it is a journey of hundreds of kilometers.

Pistnbroke, You have a case, I made a "spider web" with multiple globes and 2 reed switches, not reed-relays. Put 1,5 mm2 wire 5 times around the reed switch with the contact

I had 13 x LED's and 13 resistors 4700 Ohm to the base of a PNP transistor (and a resistor between base and emitter) . I had only 2 reed switches, thus one has to imagine the other 11.
As soon the globe is switched on, the reed-switch is made, the PNP transistor conducts and a second PNP transistor is shortened between the base and the collecter. This second PNP transistor let a buzzer sounds, if a globe goes, the switch opens up, the LED goes off, the first PNP is no longer shortening the second PNP and the buzzer sounds .

The only disadvantage is that one has to switch all the lights on, to see which one is faulty and if all the light switches are off, the bell goes, thus one need a toggle switch to the plus suply of the transistors to switch the alarm off. Also if all 13 globes are on, the base current through the base is 66 milli Ampere, and the PNP must be able to handle this.

Ellewendy has now a few options to choose from.
Pistnbroke , your idea with reed-switches was thus not a bad one.
Bert .

 

just put a led in series with the reed contacts and junk all the rest or use a change over reed and sound the alarm with that .....or a visual led and an optical coupled transistor to work the alrarm