Relay Chatter in Load Disconnection Circuit

[Unit00]

Hey guys.

I'm designing a series pass voltage regulator with feedback. I'm trying to disconnect the load when the load current or control element temperature exceeds a value. The attached photos show my circuit design.

The temperature cutoff circuit works as is, but the current cutoff does not. Instead of a single click, the relay chatters continuously. Can anyone help me fix this?

 

[MikeMl]

Well, of course it does. Think about it. Load current slowly increasing. Eventually, it gets to the point where it gets greater than the sensor's threshold, relay pulls-in, disconnecting the load, making the current go toward zero, below the sensor's threshold, so it drops out the relay, which reconnects the load, but the load current is still over the limit, so the whole cycle repeats forever.

You need to implement a latching circuit, where the relay stays latched the first time you get over current, requiring human intervention to clear the latch by clearing the fault, and then resetting the latch...

Only other alternative is a true constant-voltage, current-limited power supply. If the current limit of such a supply is exceeded, the output voltage drops below its normal voltage.

 

[AnalogKid]

You haven't said what you want the circuit to do once it disconnects the load under fault:
1. Wait a while and then automatically reconnect (hiccup mode).
2. Keep power off until the circuit is reset manually (foldback current limiting / electronic circuit breaker mode).
3. Combination - restart automatically if it was a thermal fault and things have cooled off, but stay latched off if it was an overcurrent fault.

ak

 

[Unit00]

Well, of course it does. Think about it. Load current slowly increasing. Eventually, it gets to the point where it gets greater than the sensor's threshold, relay pulls-in, disconnecting the load, making the current go toward zero, below the sensor's threshold, so it drops out the relay, which reconnects the load, but the load current is still over the limit, so the whole cycle repeats forever.

You need to implement a latching circuit, where the relay stays latched the first time you get over current, requiring human intervention to clear the latch by clearing the fault, and then resetting the latch...

Only other alternative is a true constant-voltage, current-limited power supply. If the current limit of such a supply is exceeded, the output voltage drops below its normal voltage.

I have implemented a latching circuit; it's Figure 1. There, whenever the relay trips, the load remains disconnected; human intervention (in the form of pressing the N.C. switch) is required to re-connect the load.

I figured there would be no need for a current-limited regulator, since the load would be disconnected in any case, when a high enough current begines to flow through the load.

You haven't said what you want the circuit to do once it disconnects the load under fault:
1. Wait a while and then automatically reconnect (hiccup mode).
2. Keep power off until the circuit is reset manually (foldback current limiting / electronic circuit breaker mode).
3. Combination - restart automatically if it was a thermal fault and things have cooled off, but stay latched off if it was an overcurrent fault.

ak

Once, the load is disconnected, I've set up the circuit so that the load is not reconnected until I want it to (pressing the N.C. switch in Figure 1). So I guess this is option 2.

Suppose, though, I wanted to implement option 3. How would I go about doing that?

 

[AnalogKid]

I figured there would be no need for a current-limited regulator, since the load would be disconnected in any case, when a high enough current begines to flow through the load.

Suppose, though, I wanted to implement option 3. How would I go about doing that?

You might reconsider a current limiter, as a transistor can pop in microseconds, 1000 times faster than a relay can respond to a signal.

For option 3 you need two different inputs to the main regulator. The input from the temperature sensor inhibits the output directly, usually by shorting one of the error amplifier inputs to GND. The input from the current sensor sets a latch that inhibits the output, usually by shorting one of the error amplifier inputs to GND.

AND -
Why is one of the relay contacts presenting a dead short to the power supply output?
The collector of Q3 is grounded, so K1 is on all the time.
R3 and R4 don't do anything.
Why are there three different opamp types?

Finally, it is important to communicate your ideas in a clear and unambiguous manner. In electronics, nothing does that better than a well-drawn schematic. Your schematic would be easier to interpret if it followed standard conventions. Grounds point down. Positive rails point up. No unnecessary jogs and jumps. All node connections get dots. Never have more than three lines into any dot. Etc.

ak

 

[audioguru]

I laugh whenever I see a schematic drawn with Multisim.

 

[Unit00]

You might reconsider a current limiter, as a transistor can pop in microseconds, 1000 times faster than a relay can respond to a signal.

For option 3 you need two different inputs to the main regulator. The input from the temperature sensor inhibits the output directly, usually by shorting one of the error amplifier inputs to GND. The input from the current sensor sets a latch that inhibits the output, usually by shorting one of the error amplifier inputs to GND.

AND -
Why is one of the relay contacts presenting a dead short to the power supply output?
The collector of Q3 is grounded, so K1 is on all the time.
R3 and R4 don't do anything.
Why are there three different opamp types?

Finally, it is important to communicate your ideas in a clear and unambiguous manner. In electronics, nothing does that better than a well-drawn schematic. Your schematic would be easier to interpret if it followed standard conventions. Grounds point down. Positive rails point up. No unnecessary jogs and jumps. All node connections get dots. Never have more than three lines into any dot. Etc.

ak

I laugh whenever I see a schematic drawn with Multisim.

Thanks for the info.

Sorry, the circuit got messed up while I took out the temperature sensing circuitry; that explains why the relay was connected incorrectly. I thought Q3 and therefore K1 would only be on when there is an input to the base of Q3; that is, when the comparator saturates positively.

R3 and R4 and the subsequent buffer form the voltage sensing circuit, and I've removed that from the new schematic as well.

While designing the circuits, I looked at different op-amps, each having different qualities. Different op-amps were better suited to certain configurations due to having higher slew-rates, rail-to-rail outputs, input impedance and so forth. That's why there are multiple types of op-amps. Otherwise I probably would have used the LM741 or LM324 throughout the entire design. That's how I went about designing; is there a flaw in my methodology?

Finally, what program should I use to draw schematics, since MultiSim is apparently not that good? Other than MultiSim, I've only ever used Visio.

 

[AnalogKid]

Much better schematic. If you move Rx1 so it is between the 741 pin 4 connection go ground and diode D2's anode connection to ground, you can use sections of the LM324 for all three opamps.

ak

 

[Unit00]

Much better schematic. If you move Rx1 so it is between the 741 pin 4 connection go ground and diode D2's anode connection to ground, you can use sections of the LM324 for all three opamps.

ak

I'm not sure I understand what you're saying. Can you explain to me further?