How to switch on a circuit with a 12v feed....

[ljcox]

Deleted - I was on the wrong tram.

 

[ljcox]

I became confused while answering post #19 & missed one of the points I wanted to make.

A capacitor of 6800 uF will take about 6800 * 0.01 * 5 = 340 seconds to fully discharge via the 10k resistor.

So if you were to turn off the ignition after a short period (say < 60 sec) & then turn is on again within a few seconds, the relay would not operate since the capacitor would be still partially charged.

 

[benweston]

As I wrote above, your figures don't add up.

To put it in other words, if you measured 4 Volt between A & B then Vb - Va should equal 4 Volt.

Right, I've taken it direct to the USB ground wire (rather than a pad marked 'GND' on the PCB) and I got readings of:

Pin A - 0.75V
Pin B - 5.00V

Across the button pins, still reads 4.03V which still doesn't add up though. However, the GND pin of the USB socket MUST be the best place to ground?!

The Voltages remain consistent before and after the button is pressed but voltage ACROSS the button pins obviously zeroes out WHILE the button is pressed.

Sorry this is a bit painstaking - I imagine this is slightly frustrating for you!!

 

[ljcox]

That is much better agreement than before.

The difference is not worth chasing.

The best place to put the negative probe of your meter is on a ground point inside the GPS.

The ground pin of the USB socket at the GPS end would be a good choice.

It is less desirable to use the other end of the USB cable because there will be a voltage drop across the ground conductor in the cable if there is a current flowing through it (Ohm's Law).

Such a voltage drop will cause an error in any reading you make if the negartive probe of the meter is at the supply end of the USB cable.

No I don't find it frustrating, I just want to be sure that what I design for you will work.

I'll Post a circuit for you later today.

 

[ronv]

Thanks for the suggestion but it has a few issues:-

1. the minimum switching current shown in the relay data sheet is 10 mA & the OP measured about 4 mA. See my comments in an earlier post (post #10) re the contact current issues.

The life might go down to 10 million cycles at 1 ma at 5 volts, but it will still be chugging long after the beamer.

2. I don't know if the pulse can be applied immediately the ignition is turned on. A short delay may be necesary.

Seems easy enough to check.

3. It would be necessary to put a diode across the coil in order to limit the back EMF to about 0.7 Volt.

Yes, good idea.

Regarding the drop out time: Don't forget the load from the GPS is still in circuit on the other side of the cap. Now this may be a problem if this turns out to be the GPS battery voltage. Something that needs to be checked unless you want to use the 12 volts before the power supply.

 

[ljcox]

The life might go down to 10 million cycles at 1 ma at 5 volts, but it will still be chugging long after the beamer. Perhaps, but the data sheet states 10 mA min, so I would not want to make assumptions.



Seems easy enough to check. Is it? The only way I can see to test it is to build the circuit. How else could you simulate the effect?



Yes, good idea.

Regarding the drop out time: Don't forget the load from the GPS is still in circuit on the other side of the cap. Now this may be a problem if this turns out to be the GPS battery voltage. Something that needs to be checked unless you want to use the 12 volts before the power supply.

My thought was to put a diode in series between the 5 Volt & the R & C so the cap charge is blocked when the ing is turned off.

I'm about the post my suggestion, so the OP can choose which idea he wants to try.

EDIT.
I neglected to make the main point above about the contact issue.

The question is not the life time, it is a question of contact wetting.

If the manufacturer recommends a minimum current of 10 mA, then it is unlikely that 4 mA will provide sufficient contact wetting for reliable operation - as I alluded to in my post #10.

 

[ljcox]

I have attached my suggestion for you.

However, you may like to try ronv's idea if you have a suitable relay.

Circuit description:-
The 5 Volt supply starts when the ignition is switched on.

Capacitor C1 starts charging via R1.

When the voltage of C1 reaches the upper threshold of the Schmitt Trigger 1a (approximately 5 seconds), the output of U1a goes low so the output of U1b goes high.

This starts C2 charging via R2. Thus the input of U1c (pin 5) is pulled high - since there is no initial charge on C2.

Therefore output of U1c goes low which makes the outputs of U1d, e & f all go high thus turning on the LED in the opto coupler U2.

The light from the LED turns the transistor in U2 on thus applying a short circuit between A and B.

When the voltage across R2 falls (because C2 is charging) to the lower threshold of U1c, its output goes high & therefore the outputs of U1d, e & f all go low.

This terminates the pulse. The pulse length will be approximately 100 ms.

When the ignition is switched off, C1 discharges rapidly via D1 and R3.

R3 may not be necessary: it depends upon what else is connected to the 5 Volt line.

I have taken a guess that a delay of about 5 sec will be adequate. If necessary it can be increased or decreased by changing C1 to a different value.

I have taken a guess that a pulse of about 100 ms will be adequate. If necessary, change C2.

The delay and pulse timings of are very approximate since the threshold levels of the Schmitt Trigger have wide tolerances & because I have assumed that the times will be approximately equal to the time constants.

 

[ronv]

I understand your concern over the current. They can't publish all the possible combinations of voltage and current as it pertains to life as it is not a simple matter. But, have a peek at this one which does specify several levels as well as the often mentioned 10 ma min. https://www.mouser.com/ProductDetai...=sGAEpiMZZMv4tz1TW/ArE4mGdvW5WwrPkzLj6m0m3UI=
Turns out the best performance is at 5 volts, 1 ma.

As far as the instant on I would just hold down the GPS button then plug it in.

Having said all that I like your circuit.

 

[ljcox]

I understand your concern over the current. They can't publish all the possible combinations of voltage and current as it pertains to life as it is not a simple matter. But, have a peek at this one which does specify several levels as well as the often mentioned 10 ma min. https://www.mouser.com/ProductDetai...=sGAEpiMZZMv4tz1TW/ArE4mGdvW5WwrPkzLj6m0m3UI=
Turns out the best performance is at 5 volts, 1 ma. See below

As far as the instant on I would just hold down the GPS button then plug it in. You may be right. But I have been assuming that it may be necessary to allow some time for the car systems to stabilise before starting the GPS.

Having said all that I like your circuit. Thanks

I looked at the data sheet you pointed me to and I could not see 5 V, 1 mA mentioned. Where did you find this info?

These figures are interesting because, many years ago, I worked as a design engineer designing complex relay systems for electromechanical telephone exchanges.

The rule of thumb we used was that - for adequate contact wetting - you needed at least 5 Volt across the contacts when open & at least 1 mA through them when closed.

The contacts we used were an alloy of silver & some other metal that I don't now recall.

EM telephone exchanges had thousands of relays & hundreds of thousands of contacts (a relay could have up to 24 springs) so the relays & their contacts had to be VERY reliable.

In cases where voice signals were switched via relay contacts, we used DC to wet the contacts because the voice signal voltages & currents were too small to reliably do the wetting.

I recall an earlier EM technology that did not use DC wetting & the maintainence staff had to often clean the contacts with a special contact cleaning tool (basically a soft file) in order to correct the faults that occurred due to the lack of DC wetting.

 

[benweston]

Len, that's great! You are right in presuming it would need a delay before starting. A ~2 second delay before starting and a ~1 second pulse would be absolutely ideal - what would you replace C1 and C2 with, with this in mind? Otherwise, that's fantastic and very easy to follow too - thank you

Incidentally, I shall be posting a video of the finished result on the car forum so I shall link you to it too so you can see it in action

 

[ronv]

You have to open the actual data sheet. Then it's about 7/8 down the first page on the left.
I think most are now sputtered with Ruthenium because it is inert.

 

[ljcox]

Len, that's great! You are right in presuming it would need a delay before starting. A ~2 second delay before starting and a ~1 second pulse would be absolutely ideal - what would you replace C1 and C2 with, with this in mind? Otherwise, that's fantastic and very easy to follow too - thank you You're welcome.

Incidentally, I shall be posting a video of the finished result on the car forum so I shall link you to it too so you can see it in action That will be good to see.

The approximate formula is:-

T = RC. So if R is in Meg & C is in uF, T will be in seconds.

For T1 = 2 sec, try C1 = 0.22 uF.

For the pulse T2 = 1 sec, try C2 = 0.1 uF.

I mentioned the reasons why it is approx, in a previous post, so I won't repeat them here.

You can also change R if you wish, but it is probably easier to change C only.

If you need the fine tune, you can connect caps in parallel or in series to give you "in between" values.

Eg. 0.1 uF in parallel with 0.022 uF will give you approx 0.122 uF.

And if you connected say two 0.33 caps in series it would give you 0.33/2 = 0.165 uF approx.

The formula for caps in series is C = C1 * C2/(C1 + C2).

So if you connected a 0.1 in series with a 0.68, the resultant capacitance would be

0.1 * 0.68/0.78 = 0.087 uF approx.

 

[ljcox]

You have to open the actual data sheet. Then it's about 7/8 down the first page on the left.
I think most are now sputtered with Ruthenium because it is inert.

Thanks, I missed that.

I'm puzzled as to why they state:=

 Contact material Ruthenium

Min. recommended contact load 10mV, 10mA

Minimum switching voltage 10mV
Initial contact resistance 200mΩ max. at 10mA, 6VDC

But then seemingly contradict themselves in the part you mentioned:-

 Electrical endurance
form A (NO) and form B (NC), resistive load, +25°C
20VDC, 500mA 1x106 ops.
20VDC, 250mA 20x106 ops.

5VDC, 1mA 100x106 ops. form C (CO) contact, resistive load, +25°C

10VDC, 500mA 1x106 ops.
10VDC, 250mA 20x106 ops.

5VDC, 1mA 100x106 ops.

Rather strange!

 

[audioguru]

I always used gold plated contacts for switching audio. They didn't cost more than silver plated contacts that corroded and did not pass dry audio.

 

[ljcox]

I always used gold plated contacts for switching audio. They didn't cost more than silver plated contacts that corroded and did not pass dry audio.

In an EM telephone exchange, most of the audio paths have a DC current for signalling reasons anyway. So the extra cost of gold contacts could not be justified.

Besides, most of the contacts were used in the logic circuits that worked on 50 Volt DC & generally had currents of several mA flowing so wetting was not an issue.

In cases where we needed relays that did not have any contact bounce, we used relays with mercury wetted contacts. There were 2 small pools of mercury inside the relay case & these were moved together by the armature somehow.

Once the pools met, they closed the circuit without any bounce.

 

[benweston]

As I'm planning to get the PCB made up (as it's not a lot of money), I've designed the PCB on their software. Does this look right to you? I think it is! Excuse the messy design, I was more concerned with making sure everything was in the right order!!

View attachment 61085

 

[ljcox]

The PCB looks good. Some suggestions:-

1. Have you checked whether you can buy the components & their costs?

2. Personally, I would remove the (+5V) & +(0.75V) beside B & A as they are misleading.
The reason I put them on the circuit was to indicate that B had to be the more positive side of the button.

3. I would like C3 to be closer to U1.
You could achieve this by extending the track from U2/2 up toward U1/14 & mount C3 near U1/14.

 

[ljcox]

I just noticed another issue with the PCB.

You have several right angle corners in some tracks.

My understanding is that these should be avoided. I think it is something to do with the etching process.

I avoid 90° bends by inserting a short section of track at 45°, ie. two 45° bends.