How to stop supply current going to zero?

[treez]

hello,

the input supply current into the following circuit from the ~13V source (V1) must never go to zero.

if it goes to zero then a fault detector at the source thinks that the circuitry has failed. -we don't want this to happen.

Thing is, when there is a transient in the supply from 14V to 12 V , the circuits supply current does indeed go to zero for a few microseconds........this is due to the EMC filter oscillating.

How can we stop this?

That is, how can we stop the supply current from V1 going through zero when there is a transient in V1?

Here is an example of a 14v to 12v transient.............

Here shows the input current which does indeed go to zero for a few microseconds when the transient occurs...........

=============================
FYI, we are in Nebraska doing some LED light stuff off car batteries.

there are many led lights and they are all fed from the same supply, -only when a different main light goes out do we want failure indication, therefore the above circuit is rigged up to actually draw current through the resistor in the npn's collector when this lamp fails.............that way the sensed current from the supply will remain high..........and not give a main lamp failure...........until the main lamp itself fails!

anyway, how do we stop the current going to zero...?

or is this just a feature of the front and CLC filter that it resonates when a supply step occurs?

 

[waross]

Add a little time delay to your monitoring circuit. How important is it to know within micro seconds that an LED failed?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[djs]

What is zero current? Place a non-inductive resistor across the input sized to draw the min current?

 

[treez]

djs...well the above circuit must always draw at least 40mA.

waross, -i've tried a time daley here........the problem is that it makes no difference when a transient voltage input step occurs.....the CLC filter resonates and its current goes through zero.

the resistor across the supply sounds good but we wnt to avoid this due to dissipation as this circuit can go down to 4v input voltage for a bit.

 

[djs]

I think what waross was refering to was the amount of time the current stays below a setpoint. For example have your monitoring circuit only trigger when the current is below 20 ma for at least 2 seconds.

 

[treez]

djs,

the problem is the monitoring circuit is done and dusted now and can't be changed.

we cant accept any zero currenrt for any time period

 

[itsmoked]

Add a cap fed thru a diode to the input would be the standard way to prevent this on a vehicle powered device.

This also protects from reverse jump starts. And after that first resistor you also want a TVS to protect the circuit from load dumps.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

As well as zero Volts, you may get 100 Volts from a load dump. Use the TVS that itsmoked suggests.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[treez]

itsmoked,

thankyou i appreciate adding that cap would work, but i forgot to mention that the input power battery has already been sold to the customer and we are not able to add a cap to it.

We can only work with our end of the circuit, which has the CLC EMC filter at its front end, and which makes the input current go to zero when a load step occurs such as 14V to 12V in a few microseconds.

I have done this circuit which "promises" to solve the problem as it gives a positive pulse at Q3 collector when the 14_to_12V load step occurs.........

here is the whole circuit.......

and here is the pulse which happens as soon as the input voltage steps from 14 to 12v.......

......now "all" i need to do is make the above pulse turn on a transistor with its load for about 20 microseconds, then have that transistor turn off again.

and ideas how i can do this?

i thought of a 555 monostable but think it will not respond quickly enough since the pulse is only about 2us wide.

 

[zappedagain]

Send your turn on pulse through a diode and resistor in parallel, and connect this network in series with a capacitor to ground. The out put is the resistor capacitor node. The output pulse will have a fast turn on (ESR of forward biased diode and capacitor define timing) and a slow turn off (Resistor and capacitor define timing). As long as your ESR of your diode is much less than your resistor value your output will rise much faster than it falls. Thus you can turn a transistor on in 0.2 uS and turn off in 20 uS.