LM393 Comparator Behavior - "Indicator Bulb Replacement"

I am building some digital electronics for an industrial application and there is one part that I would like to replace but I am struggling with the design. I will attach the schematic of my initial concept. I tried simulating it and ran into some unexpected behavior. The breadboard version worked for awhile but then I fried something. Let me explain the application first:

There is a legacy piece of custom industrial equipment that has an indicator bulb. It normally has a high impedance voltage output and then switches to a low impedance output to drive a 25W light bulb. Basically, there is always 28 VDC going to the light bulb but normally there is a resistor in series with the bulb, which limits the current such that the bulb is not on. The purpose of this is to detect a burnt-out bulb. That circuitry is essentially "sealed" but seems to require a test current of about 50 mA for the equipment to continue running. Some quick tests suggest that the maximum input impedance of my circuit must be less than 50 ohms in order to allow the required test current to flow (i.e. the output impedance is about 510 ohms).

My initial thought was to use an LM393 comparator and a current sense resistor. To prevent dissipating a continuous 15-25 watts at the board, I thought I was real clever using the comparators, a current sense resistor, an RC time delay and a MOSFET. But it doesn't seem to work as I expected. Anyone see the (probably obvious) flaw in my concept?

How I thought it would work:
Current flows from pin 2 on U8 into the node at R17 and R18. R17 is the current sense resistor (I was planning on using some 3W 10 ohm resistors because I need them elsewhere on the board).

When the bulb is on, the voltage across R17 increases above the threshold and the comparator triggers, pulling the output low. This results in the MOSFET turning off so there is only a brief, large current pulse through the resistor.

However, when the MOSFET is off, the voltage at R17(2) will go to 28V, so I put a zener in there to protect the comparator. Now the input voltage is at the zener voltage, which is still above the threshold so the "trigger" never resets.

But hey... a 393 has two comparators so I thought I'd use a slow RC ramp on one input and the reference on the other input. When the first comparator goes low, it also starts discharging the RC circuit and at some point it will trigger, pulling the input of the other comparator low, which in turn lets that comparator's output go high, resetting the circuit.

Hello tms8c8,

I draw a circuit on paper you coul check, it is very simple you´ll see.

Have a nice day.

Best regards,
Juan

Hi Juan,

Thank you for taking the time to respond. Unfortunately, I did not do a good job communicating what I was after. My electronics are to take the place of the light bulb entirely. Ideally, I would just use the bulb as a convenient power resistor but it has to go. Basically, I am forced to put a power resistor on a board and things are complicated by the existing electronics. If I go to a current sense resistor that is too small of a value, then I blow a fuse. If I go too big, I don't draw enough current to satisfy the test condition.

The circuit you drew would work perfectly ... if I could keep the bulb or put a big power resistor on the board. Since those two options are out, I was thinking along the lines of PWM - except just a single, very short "on" pulse.

There are lots of options but a conceptually straightforward one would be to change to a p-fet and move the switch to the high side. So then you'd have R7 pin 2 switched between 28V or open. When its open it will be at 0V, if current is available it will go to to 28V when the switch is closed.

There are many other options too. What you're ultimately looking for is voltage across R7.

Another option, probably a better one is to create a proper differential comparator. Create a divider from R7 pin 2 to ground and a divider from R7 pin 1 to ground. This lets you divide down the voltage on each side of R7 into a range where you can sense both pins with your 5V comparator. Now bias those points as necessary with your 5V supply to ensure your comparative only triggers when their is sufficient voltage across R7. What you're looking for is [R7 pin 2] >> [R7 pin 1]. So you'll want to bias the opposite by pulling the sense point from pin 1 up to your 5V rail. The strength of this pull-up sets your voltage/current threshold.