protection circuit

[ElectroNub]

I am working on a 5A dummy load project and I need some guidance.

My circuit (below) works pretty good except for one thing. When I power my dummy load first, then hook up my power supply, it starts at 5 amps and quickly throttles down. On the other hand, it starts smoothly if I power up the dummy load after hooking up power.

Not a big deal at first while testing with batteries and my 12v supply. Then I tried my 24v supply and poof! My op-amp blew the output on the gate! I thought I melted the n-mos at first but some probing found no output on the op-amp. I suspect a spike from the gate killed it from the full power start on the 24v supply. I think the power supply was amping up while the op-amp was throttling down and in that fraction of a second was melted down. I whacked in a new op-amp and it is working again. I just have to be careful of what order I power up.

Have I diagnosed the problem correctly? Can I put a diode between the op amp and filter to prevent it from happening again? How will it affect the filtering? I can definitely see myself 'oopsing' on this in the future.

I also googled various protection circuits but they all seem geared toward being on the power supply side of things. The problem is here as well when I crank up power until the supply protection kicks in. It gets stuck in a loop powering up and down repeatedly because the circuit starts at full load after things are heated up.


I would like to be able to leave the dummy load on and switch between power sources at will. So I was thinking of some way to detect voltage on the inputs and use that to turn on the feedback circuit after the power supply or battery is connected. What kind of ways can I go about creating a circuit for this?

Thanks in advance. Hope my nub makes sense

(p.s. r2 is actually 1k)

[ElectroNub]

sorry...

'hooking up power supply' means to wire in the DUT.

'powering dummy load' means starting the feedback circuit from its own supply.

The dummy load is powered from 12v wall pack.

V and A are volt/amp meters.

[ajb]

If your load is powered on with no supply connected, the opamp is going to drive the transistor into saturation to try to meet its current setpoint--which it can't do, because there's no power supply to source the current.  When you then hotplug your power supply it's going to see as damn near a dead short as your pass transistor and current sense resistor can make.

In normal operation you should always have a certain amount of voltage drop across the pass transistor.  You can monitor this voltage, and when it drops below a minimum threshold you can be pretty sure that the load is disconnected and inhibit the drive signal to the pass transistor to prevent it from saturating.  When you then hook up your supply, the the transistor will be fully open, the saturation detector will deactivate, and your current loop will take over.  This will also effectively indicate when the compliance voltage required by your dummy load exceeds the voltage of the power supply which could be handy.

[Kalvin]

Make an additional circuit which will detect whether the load is connected or not. If the load is not connected the op amp should be driven for zero output.

I thought something like this: Add a two 10K resistors in series from the potentiometer output to op amp's + terminal. Add a suitable capacitor between the mid point of the two resistors and the ground, which will make the adjustment a bit smoother and will filter any potentiometer contact noise. After that make a voltage detector which will detect whether the load is connected and powered up. If the load voltage is not detected, the detector should short the capacitor in effect driving the op amp output to zero. When the load is connected and the voltage detector senses load voltage, the capacitor is allowed to charge in a controlled manner into set voltage.

[carveone]

Dave's original posting on the dummy load (#102) would have the same problem I suppose. Here's his circuit, drawn by someone on stackexchange:



Thus with no load, the - input of IC1B is grounded and its output is ramped to Vcc. Is the solution as simple as a FET across R2? Or a P-FET just before R1? Feels a bit brutal somehow !


I built something similar for an LED based on this:



which worked fine but you could see the small surge when the switch closed. The output drifted with temperature and applied voltage - not surprising since the reference is a B-E junction! If you want a precise current source, I suppose the opamp reference voltage should be precise too; like a 1.25V shunt voltage reference...

[bobcat]

Put a high value resistor (100K ish) between the NFET gate and ground. This will keep the NFET off until the opamp is powered up.

[Liv]

Electronic loads have programmable parameter "Von". The easiest way to get it is shown in sch.

[Kalvin]

Electronic loads have programmable parameter "Von". The easiest way to get it is shown in sch.

Nice

[carveone]

Nice indeed! It feels so obvious somehow when experienced engineers points out how to do things