I want to design a system wherein it can activates an LED when the ECU goes into sleep mode and deactivates when the ECU is ON. Which device should be used and where it should be placed in the following circuit?

Note- When the ECU is ON, it draws approximately 5A and 100mA while in sleep state.

 

You have a nice large current window to work with. In the interest of plane and inexpensive I would just use an inline current sensor like an ACS 712 in the 0 to 10 amp version. That requires additional circuits. Next if you want to pop $20.00 USD something like this can be used. A Google of DC Current Sensors / Switch should bring up a dozen hits from inexpensive (cheap) Chinese modules to expensive solutions.

Ron

 

The simple way will be to create a current sensing magnetic switch using a few turns of wire around a reed switch. It would take a bit of experimenting to have it switch at about 3 amps, half way thru the change. It would be easier to have the LED switch on when the ECU becomes active because most reed switches are normally open with no magnetic field.

 

The simple way will be to create a current sensing magnetic switch using a few turns of wire around a reed switch. It would take a bit of experimenting to have it switch at about 3 amps, half way thru the change. It would be easier to have the LED switch on when the ECU becomes active because most reed switches are normally open with no magnetic field.

We went over this some time back. Several turns? More like a hundred turns or more.
Reed Switch Test Procedure
is a good start. This is not as simple as winding a few or several turns of conductor around the glass envelop of the switch. Believe me I played this game developing testing of magnetic reed switches. Even with 5.0 amps you are looking at several hundred turns. Here is more on the subject. More data and information can be had by contacting the people at Coto Mfg or any reed switch manufacturer. But while wrapping several turns around the switch sounds good it won't work unless you have some heavy wire and very high current.

Ron

 

OK, it seems that there is , or maybe Was, A GREAT DEAL OF SENSITIVITY VARIATION among reed switches. Certainly a lot has changed since 1967, and so some designs have stopped being available. One thing would be certain, which is that the switches available to experimenters today are far less sensitive than those available back then.

 

<Off Topic> Funny you mention the 60s as the systems I worked with were developed in the 60s. Called a RPI (Rod Position Indicator) we used reed switches to measure the control rod position in a reactor core. Before the switches were placed on a board every switch was tested so we developed test methods. This design was used on reactors for aircraft carriers. Subs use a much different inductor system. Anyway yes, been there and done that. Yes, magnetic reed switches come in a dozen or more flavors (sensitivities) but my point here is that a few turns of wire won't work as been there and tried it. Coto is a key player in the game and has endless white papers on the subject. Me? I would just buy a turn key solution like I suggested.

<On Topic> We now resume our regular scheduled program.

Ron

 

What MCU? Maybe you can send a control signal when it enters a different state. Then use a flip flop or even a MOSFET to switch the LED.

Another option is to use a shunt resistor and an op amp / comparator with a specific threshold. I can provide an example of each.

 

What MCU?

Not MCU but rather ECU as in an Engine Control Unit is how I read it.

Note- When the ECU is ON, it draws approximately 5A and 100mA while in sleep state.

Ron

 

Not MCU but rather ECU as in an Engine Control Unit is how I read it.

Ron

Missed that one! The sun was in my eyes.

 

It seems that I will need to do some experimenting with reed switches. I am thinking of a mu-metal tube over each end of the switch, with a gap in the middle to focus the magnetic flux on the switch leaves. Is that even a valid concept??
OR, cheaper than a good current shunt, a high current diode that will provide a 0.7 volt drop, not changing much as the current increases. With the right diode then the 100mA draw in sleep mode would not trigger the switching action. But that would be tricky to adjust. maybe.

 

It seems that I will need to do some experimenting with reed switches. I am thinking of a mu-metal tube over each end of the switch, with a gap in the middle to focus the magnetic flux on the switch leaves. Is that even a valid concept??
OR, cheaper than a good current shunt, a high current diode that will provide a 0.7 volt drop, not changing much as the current increases. With the right diode then the 100mA draw in sleep mode would not trigger the switching action. But that would be tricky to adjust. maybe.

I can share this. Working from memory a long time ago. We mounted 100 switches on a board using spring terminals. Anyway as to the coil. The coil was would on maybe a 12" ID piece of Bakelite tube. Several hundred turns of AWG 16 wire. We applied an AC Voltage using an autotransformer and adjusted for 5.0 amps AC. Using a MUX network each switch was scanned. I recall a small RC network on each switch. All switches were NO (Normally Open) types. The board was placed in an environmental chamber with a shaker (shake and bake) and hot was about 300.0F and cold was 0.0F Thermal shock test ran for about 24 hours.

Again, if you get curious have at it. Last time around on this I would AWG 16.5 around an old plastic prescription bottle with the bottom cut out. That was my tube. Convenient because it was easy to wind a coil on it and drill small holes on each end to run my current carrying wires through.

Ron

 

I was really thinking to put the turns much closer to the switch.
BUT really, I like the diode scheme better because it is very reliable.

 

I was really thinking to put the turns much closer to the switch.
BUT really, I like the diode scheme better because it is very reliable.

By all means, Normal apps the glass tube is wrapped tight around the glass body.

Ron

 

Using a ten-amp rated diode in series will result in a voltage drop of about 0.75 volts at five amps. That might be excessive. Shunting the diode with a one ohm resistor should reduce the voltage drop to 100 millivolts, (0.100 volts) in the sleep mode. That voltage change should be adequate to trigger a comparator reliably, as an alternative approach. If the full load voltage drop is excessive, then a schotky type diode could be used instead.