FEELER: Continuously Adjustable Test Load

[lasersbee]

We seem to be jumping the gun again... Wolfman hasn't even got
the thing running or tested it yet...

I'd wait until he shows us something actually working before
commenting on the problems that may or may not be present
with his design....:beer:


Jerry

You can contact us at any time on our Website: J.BAUER Electronics

 

[Wolfman29]

While yes, we are jumping the gun, this is simply an alternation to a known, working design and several sims have shown it should work. Again, this thread was to see if there was enough interest, and, assuming it works properly (i.e. how does the signal compare to that of a normal test load), I have a feeling that this will be popular enough for me to continue the design.

 

[Wolfman29]

Hmm. So interesting results. I am getting REALLY oscillatory waveforms for some input voltages (to the BlitzBuck), and when I am at other input voltages, I am getting a stable waveform on the test load. It adjusts as expected, but the issue now is that the transistor seems to be oscillating very quickly. Does anyone have any idea how to make the transistor stay in the unsaturated zone so that it acts entirely as a variable resistor instead of switching on and off?

 

[Cyparagon]

That depends entirely on the circuit you're using. We can't help you very well if you're trying to keep this a secret to monetize it

 

[Wolfman29]

I thought it was mentioned earlier in the thread that I was using a mosfet/adjustable zener to do this? Well, I am, anyway.

Typical voltage regulator - MOSFET in parallel with a zener - easy peazy.

Anyone have any ideas?

 

[Hiemal]

You're oscillating because of how CC supplies work.

When the gate has enough voltage it turns on right?

Well the driver will attempt to adjust until it outputs enough current. In the meantime, since it's supplying MORE voltage to the gate, it turns on MORE.

Then the voltage drop will decrease.

Driver keeps attempting to adjust for this and then oscillates, as you've seen. I've attempted something similar, using an NPN transistor and a potentiometer and had the same issues too.

I think you can stop it two ways; adding capacitance to the gate, to help slow down the changes, or creating a voltage reference for the gate itself. Adding a voltage reference to the gate itself is the most reliable way, as it removes the driver from having to supply the power to the gate too.

It would for sure remove the oscillations. I'm not sure about the capacitor idea though.

 

[Wolfman29]

I'm sorry, I meant a PNP - not a MOSFET. And I should have been more clear - the zener is in parallel with the gate of the diode, separated from the positive rail by a 100k resistor.

So there is a reference on the gate... hmm....

 

[benmwv]

I don't think a typical voltage regulator design is going to work here. You can't regulate voltage and current on a fixed load. Or maybe I've misunderstood your approach.

Anyways, I would use an opamp driving the gate of a big heatsinked mosfet comparing a reference adjusted by pot with a voltage divider across the mosfet. So basically you set the voltage that will be dropped on the mosfet and it will change resistance according to current. But personally I think it seems counter intuitive to use active voltage regulation when testing current.

 

[Wolfman29]

Hmm. That's the sort of design I was thinking of. I was just referencing a datasheet, really. The idea is that a perfect "voltage sink" shouldn't care about current, it should just draw as much current as it wants for whatever voltage. Then, the driver is forced to match that voltage while outputting the current. That was the idea, anyway.

 

[benmwv]

Here's what I was talking about in a sim:

I even put in a 0.1 ohm shunt for measuring the current. Thats a 100k pot and a 1v reference. With the 90k:10K divider on output we get a range of 0-10v. Works great in sim, but no guarantees for real life

 

[Wolfman29]

What worries me is that that is basically what I am doing, but instead of the opamp, I just am having the value of the zener changing, so no need for the op-amp. >.< My circuit worked perfect in sim too.

 

[benmwv]

Yeah, it will probably end up oscillating. Once I tried to make a super simple linear driver very similar to that circuit and it oscillated really bad. You could hear it squeal but there weren't any inductors. I'm not sure what component was actually making the noise.

I just went back and looked at cyp's circuit, its nearly the exact same thing as mine haha.