Voltage controlled variable miliohm resistor

[zero0d]

Dear All,

i'm looking for a miliohm range voltage controlled resistor schematic, and i tending to do it with OPAMP and mosfet but can find a good schematic to start. I looked at this thread but there is no schematic https://www.eevblog.com/forum/projects/please-help-with-an-idea-for-a-variable-miliohm-resistor/.

The thing is my buck-boost converter that senses input current it is always comparing the sensed voltage across sens resistor with 50mV and then trips the over current. And i have for different input voltages different input currents and thus cant use the same Rsns resistor for all scenarios. ( Image attached)

I have Vin ( this is the control voltage) from 9V to 27V

Vin [V]	9	12	15	18	21	24	27
Iin [A]	2.67	2	1.6	1.34	1.14	1	0.89
Pin [W]	24	24	24	24	24	24	24
Vrsns [V]	0.05	0.05	0.05	0.05	0.05	0.05	0.05
Rsns [ohm]	18.75	25	31.25	37.5	43.75	50	56.25

Is it possible to create current sense resistor from 18.75mOhm to 56.25mOhm with Opamp and mosfet or Jfet?

Best Regards.

[dietert1]

Depending on the currents involved, you can probably make an opamp controller. If you start with a discrete shunt resistor of say 10 milliohms to measure current as a small voltage and a 1 milliohm Rds mosfet in series with that shunt, you just need an amplifier of variable voltage gain to "stretch" the 10 milliohms by some factor. Errors will mainly result from the OpAmp offset voltage. So you can use a chopper amplifier like ADA4522 or OPA189 and/or adjust the OpAmp offset to zero. An amplifier of variable gain can be made with a multiplying DAC.

Regards, Dieter

[duak]

zero0d,

I can't think of any way to control the resistance of a MOSFET that is carrying an arbitrary drain current without using a sense resistor and some sort of multiplier to complete the loop. 

Is it correct to say that what you wish to do is maintain constant power by limiting current?  If so, then the circuit will have to be modified  so Iin = Pmax / Vin.  This can be done by using a multiplier to scale the current sense resistor voltage by the input voltage.  When Vin = 9 V, the scale factor is 1 allowing full current.  When Vin = 18 V, the scale factor is 2 and the current is 1/2.  When Vin = 27 V, the scale factor is 3 and the current is 1/3.

Are you using a particular controller IC?  This will affect the type of multiplier needed because if the controller works on a cycle by cycle basis, the multipiler must be fast enough to accurately handle the sensed voltage.

There is a simpler scheme that is less accurate where an offset proportional to Vin is added to the current sense voltage as Vin increases.  This can be done with a few resistors and an op-amp.  It will not give constant power though.

[Ccandrews]

For your feedback, in place of a sense resistor you could use a FET with a built in current sense lead, a FET with current mirror output, or a dedicated current sensor like the IR25750.  It all depends on the accuracy you desire.

[jbb]

Looks like you want constant power limiting. Could you try controlling the *output* current instead?

For a ‘variable resistor’ stage, how about *adding* to the sense output voltage rather than trying to multiply? This would likely be easier to implement

[zero0d]

Hi All,

first thank you for the replays.

on the sketched circuit that i attached  is a buck-boost controller that already has integrated averaging current sense limit ( LM5176), the controller is sensing with ISNS pins and when the current limit occurs the controller pulls down SS pin, but this is fixed and doesn't depend on input voltage ( or in my case constant input power).

So the idea is to replace the miliohm sense resistor RSNS with a circuit so i can adjust (the let say it so) RSNS with a DAC ( I will also measure with ADC the input voltage in order to know what max current can i have that meets the power limit, this is controlled with MCU)

1) Do maybe some one have an idea?

2) In the mean time i got a proposed circuit from TI, but it uses a lot of components and i' not shore how it works, please look attached file, and if you have time or understand how it works please comment

BR.

[Zero999]

That can't be the complete schematic. The LM5176 is more complicated than that.
http://www.ti.com/lit/ds/symlink/lm5176.pdf

Anyway, you're going about this the wrong way. Voltage controlled resistors are notoriously difficult to implement, especially low Ohm values, which work up to any decent frequency. Variable current limiting should be done by adjusting the reference voltage, rather than the sense resistor. If the controller you're using doesn't support this, then find one which does.

[zero0d]

Hi,

this is the part of the schematic that is used to limit the input power.

i'm not aware of a controller that supports input power limit out of a box.

[Kleinstein]

There is no need to really simulate a milli-ohms resistor. It is enough to modify the sense voltage, so that the controller chips sees a different voltage. For a current limit it should be enough to just add some offset and no multiplication needed. So the Ti circuit is kind of the right way but rather complicated.

For a large range one could consider switching shunts, as the dynamic range is limited: smaller shunts result in very small voltages so that noise and offset drift can become limiting. Too large a shunt would result in excessive heat and thus thermal problems.

[David Hess]

The thing is my buck-boost converter that senses input current it is always comparing the sensed voltage across sens resistor with 50mV and then trips the over current. And i have for different input voltages different input currents and thus cant use the same Rsns resistor for all scenarios. ( Image attached)

The usual solution is to add attenuation or gain, maybe jumper selectable, between the current shunt and trigger circuit.