CCS with no load - op amp warms up

[atferrari]

I've just completed the smoke test of these two CCS. They work as expected until I leave them unloaded and the voltage control is in any position other than fully turned OFF; in few seconds the respective op amp starts to warm up.

After some thinking I understand that by not having load, the opamp sinks / sources current from / to its corresponding Darlington trying to balance both inputs what never happens.

What could be the solution? For a moment I thought to use kind of a bleeding load similar as in voltage regulators but I discarded th idea. If the resistor is too high I will be in a situation similar to what I describe and, if too low, too much of my current wil be going through there.

For a moment I considered a base resistor but it seems useless as well. Or tell the user (me) to keep them fully turned OFF when not loaded? This last seems typical of a bad design.

Suggestions or comments appreciated.

 

[MikeMl]

How about a resistor between the opamp output and the base of the Darlington input transistor? This will limit the max output current that the opamp sources or sinks...

Or replace the Darlington with a power FET. It is hard to pump much current into the gate of a FET.

 

[atferrari]

How about a resistor between the opamp output and the base of the Darlington input transistor? This will limit the max output current that the opamp sources or sinks...

Yes! Now wondering why this morning I thought this:

For a moment I considered a base resistor but it seems useless as well.

Let us see with what value I come out.

 

[ChrisP58]

Mosfets are the preferred component for current source/sink applications, because the zero gate current does not affect the load current.

With a bipolar transistor, the base current will cause a mismatch between the current in the sense resistor and that of the load.

 

[crutschow]

Mosfets are the preferred component for current source/sink applications, because the zero gate current does not affect the load current.

With a bipolar transistor, the base current will cause a mismatch between the current in the sense resistor and that of the load.

True. But the error is quite small when using a Darlington for that purpose.

 

[atferrari]

Hola Mike

I know I could try values until I get something that works but I would like to know how to reasonably estimate a value to start testing a base resistor. For that I measured the voltages for five conditions including that of "no load".

Honestly I do not know where to start. Stumped this time.

Measurements belong to the first circuit.

/Edit to add:

I forgot to tell that for "No load" I measured the drop across Rsense allowing to calculate about 27 mA. Is it possible that the op amp is sucking that much of current?

/Edit

 

[ronsimpson]

I forgot to tell that for "No load" I measured the drop across Rsense allowing to calculate about 27 mA. Is it possible that the op amp is sucking that much of current?

That might be because of the input offset of the op amp.

 

[crutschow]

You don't need measurements, you just need a calculation to determine the resistor size.
Divide the maximum Darlington collector current by its minimum Beta to determine the maximum base current you need.
Then just select a reasonable voltage drop across the resistor for that current in normal operation (say a volt or two) and calculate the resistor value from that.

 

[ronv]

Hola Mike

I know I could try values until I get something that works but I would like to know how to reasonably estimate a value to start testing a base resistor. For that I measured the voltages for five conditions including that of "no load".

Honestly I do not know where to start. Stumped this time.

Measurements belong to the first circuit.

/Edit to add:

I forgot to tell that for "No load" I measured the drop across Rsense allowing to calculate about 27 mA. Is it possible that the op amp is sucking that much of current?

/Edit

Without the load resistor the darlington just looks like 2 diodes in series so the op amp is trying to drive the sense resistor directly.
To be really safe you could use a gain of 10 for each transistor for a total gain of 100. So to get full current you would need about 4 ma of base current.

 

[tomizett]

I assume that you've calculated the op-amp shouldn't dissipate much power even under full load. If the op-amp is heating with the circuit unloaded then I'd suspect that it's oscillating - have you/are you able to check this?
You may need to provide some loop compensation by placing a cap from out to inverting in of the op-amp and an additional resistor between the sense resistor and the op-amp.

 

[atferrari]

I assume that you've calculated the op-amp shouldn't dissipate much power even under full load. If the op-amp is heating with the circuit unloaded then I'd suspect that it's oscillating - have you/are you able to check this?
You may need to provide some loop compensation by placing a cap from out to inverting in of the op-amp and an additional resistor between the sense resistor and the op-amp.

Before realizing my mistake of discarding the base limiter option, I looked for oscillations. None.

More I think of it, is the op amp trying to equalize both inputs. Modifying the circuit now to insert a base resistor.

 

[tomizett]

Ah... yes. I see what you mean - I'd missed that. So with no collector current path the op-amp is trying to drive all the programmed load current through the sense resistor.
As you say, a resistor in the base should sort that right out.

 

[atferrari]

Problem solved.

Inserted two 10 K resistors. Op amps taking 1,2mA and 860 uA, respectively with no load.