Adjustable current sink/load

[donavan]

Hello,
first of all i tried to search this topic but the only results i was getting was for similar circuits but with alot more current.

i try to draw 2-30mA from this 12V source. the current drawn from this 12V source needs to be limited to 50mA.
the 0-5V on the ninv opv input are adjusted with a dac i got spare channels left.

so my questions are,
will this circuit work as expected?
is there a better suited circuit or even ic for this purpose?
any special opv requirements except single supply and rail to rail capability?
what mosfet is suited for this, id guess one with low Rdson and Logic level Ugs?

thanks in forward for your input on this

[madires]

The circuit has a problem with driving the MOSFET. At 30mA the voltage drop across R1 would be 220 Ohms * 30mA = 6.6V. And that's also the voltage at the MOSFET's source. The problem is that the MOSFET needs some voltage (value based on the MOSFET) between gate and source to pass current. The 5V of the OPamp  aren't able to drive the MOSFET to pass 30mA because the gate voltage would be lower than the source voltage. Please breadboard the circuit and check the maximum current possible. Then replace R1 with 10 Ohms for example and try again.

[Jay_Diddy_B]

Hi,

Try changing your circuit to this:



And then tell us how it works 

Regards,

Jay_Diddy_B

[donavan]

thank you, this circuit seems to work realy fine in simulation. what mosfet can i use to breadboard this? will BUZ11 give me similiar results as the fds6912 which would fit quite well because i need more than one mosfet anyways.
as opv ill use lm324.

[Jay_Diddy_B]

thank you, this circuit seems to work realy fine in simulation. what mosfet can i use to breadboard this? will BUZ11 give me similiar results as the fds6912 which would fit quite well because i need more than one mosfet anyways.
as opv ill use lm324.

You really need a Logic level MOSFET if you are going to power the circuit from 5V. The gate threshold, the vgs voltage that the MOSFET starts to conduct is 3 volts typ, 4V max.

The LM324 will work with its input and output at ground, but the high level output is Vcc-1.5V. So with a 5V power supply the maximum output of the op-amp is 3.5V. With 0.5V on the source of the MOSFET it only leave 3V for Vgs, which might not be enough.

The LMV324 will work better. This is a low voltage version with Rail-rail output.

Jay_Diddy_B

[Thilo78]

You really need a Logic level MOSFET if you are going to power the circuit from 5V. The gate threshold, the vgs voltage that the MOSFET starts to conduct is 3 volts typ, 4V max.
...

One question in between: What exactly does "logic level MOSFET" mean?
The ones I have here have a typical threshold voltage UGS of 2..4 V DC and work fine with an opamp at ~8V DC power.
What would logic level be?

[Jay_Diddy_B]

One question in between: What exactly does "logic level MOSFET" mean?
The ones I have here have a typical threshold voltage UGS of 2..4 V DC and work fine with an opamp at ~8V DC power.
What would logic level be?

Logic Level as oppose to 'standard' MOSFETs is a poorly defined term. About 20 years ago most MOSFETs required more than 5V Vgs to get reasonable drain current. In those days 'logic' meant 5V so a class of MOSFETs evolved that were marketed as 'logic level' MOSFETs.

The advantage was that you drive these using a 5V supply.

Now it simple means the parts have a low gate threshold voltage. You really need to look at the transfer curves.

Here is a standard MOSFET, the IRF540:



And here is the Logic Level version, IRL540:



Most of the time the lower gate threshold is better, but not always.

I hope that this helps.

Jay_Diddy_B

[Thilo78]

...
I hope that this helps.

Jay_Diddy_B

Yes, I think I get the idea.
Thanks for the heads up! 

[rxxbxyhxxb]

Hi,

Try changing your circuit to this:



And then tell us how it works 

Regards,

Jay_Diddy_B

Hi, Jay_Diddy_B

About the current sink, it's difficult to understand the R2 R3 C1 in the feedback loop for me.

Would youd mind commenting more? (How to determine the value?)

I only know they've improved the loop's stability otherwise it could oscillate.

Thanks.

[donavan]

thats how i understood this circuit

R2 because opvs behave badly when they only have to drive a capacitor (MOSFET Gate).
the C1 makes this circuit either an integrator or its for phase compensation my opv knowledge doesnt go that deep.
R3 10k is propably needed for the C1. otherwise C1 would lose its charge on the R1 10 ohm resistor. tau=R*C

[Jay_Diddy_B]

Hi,
Donavan is right.

You can measure the loop gain by introducing a disturbance source, V4, into the LTspice model. You then run AC analysis to sweep the frequency.



I have named two nodes A and B.

The loop gain is displayed by selecting V(a)/V(b). You get a Bode plot like this:




You can then experiment to see the effect of the components. The general idea is to make the circuit performance dominated by the C1 and R3. With values shown the loop bandwidth is around 30kHz and the phase margin is 90 degrees.

There may be some other considerations. In this thread https://www.eevblog.com/forum/projects/dynamic-load-bode-plot-using-hp-35665a-dsa/msg309192/#msg309192 I talked about how lead inductance can effect the control loop behaviour.


Regards,

Jay_Diddy_B

[rxxbxyhxxb]

Hi,
Donavan is right.

You can measure the loop gain by introducing a disturbance source, V4, into the LTspice model. You then run AC analysis to sweep the frequency.
...

Thank you for your explanation.
I think I need find some books about op to supplement...