Help:Remote sensing in Programmable Power Supply

[abicash]

Hi

We are using a TDK GEN-60-40 Programmable Power Supply for testing few parameters for a DC motor.
As per conventional usage , the remote sensing terminals of the power supply (S+ , S-) are supposed to be connected to the actual load, so that the power supply will compensate for the cable loss.
Our usage is a bit modified wherein the motor is not directly connected to Supply+ or Supply- but through a MOSFET (to turn the motor on/off) and a small shunt resistor (2 ohm) to measure the current through it.
With this scenario , there is a voltage drop in the shunt and we want to compensate this dropped voltage.
But as soon as we connect the S+, S- to Motor+, Motor- respectively , the Power supply runs in OVP (Overvoltage protection) mode.

What are we doing wrong?
Is the power supply NOT to be used in such a way?

Kindly advise.

 

[KlausST]

Hi,

Show a sketch of your circuit

Klaus

 

[abicash]

Hi,

Show a sketch of your circuit

Klaus

--- Updated Sep 17, 2022 ---

Hi Klaus

Please find attached

 

[KlausST]

Hi,

S+ wiring is O.K.

but regarding S-:
I´d swap resistor an MOSFET.
and connect S- to the node of MOSFET_Drain (upper of shunt)

Klaus

 

[abicash]

Like this?
If so , what is the benefit of this placement?
Since the voltage remains same irrespective of MOSFET and R placement.
Kindly explain

 

[KlausST]

Hi,

sorry my mistake: I Meant Source of MOSFET.

Klaus

 

[abicash]

That also won't help since then the voltage across the entire branch (excluding the R) will be measured.

Why does the Power supply run into OVP?
Any explanation you realise?

 

[KlausST]

Hi,

For sure, it does exactly what you want.
With the S- at the MOSFET source it doesn´t run into OVP.

Klaus

 

[d123]

Hi,

Motor turn-on spike triggering OVP of PSU. Try capacitor across motor terminals, might help. That or put RC LPF in path of OVP sensing input so it ignores transient spikes.

 

[abicash]

Hi,

Motor turn-on spike triggering OVP of PSU. Try capacitor across motor terminals, might help. That or put RC LPF in path of OVP sensing input so it ignores transient spikes.

Interestingly , when I keep the MOSFET on before turning on the motor (by providing duty cycle update) , the PSU functions as it should.
So i think there is some sense in what you have said.

One more intriguing point is that I have nowhere noticed the scheme which we are using.
I mean the remote sense compensation is advised to be used as S+ --> M+ , S- --> M- , and not have a floating ground or Positive supply.
Perhaps the manufacturers of the supply (TDK) do not want the user to compensate for a loss through an extra circuit or components.

 

[KlausST]

Hi,

providing duty cycle

In post#1 you used the MOSFET for ON/OFF of the motor. Now it seems you do PWM control.
Please clarify.

I mean the remote sense compensation is advised to be used as S+ --> M+ , S- --> M-

Generally true .. but never with a wsitching device (MOSEFT) inbetween.
So between
* +Out and +S and
* -Out and -S
there may be the shunt (non switching), but not the MOSFET.

***
In post#1 you wrote you want to compensate th voltage drop of the shunt.
I gave you the information how to do it.

But what you try to do is to include the MOSFET into the feedback loop.
As soon as he MOSFET is OPEN, the motor voltage drops to zero ... and the supply tries to maintain the motor voltage. It increases the voltage ... until OVP ... because it has no chance to compensate for the voltage drop across the OPEN MOSFET.

--> Do how I told you. and get waht you requested for.

Klaus

 

[abicash]

Hi,


In post#1 you used the MOSFET for ON/OFF of the motor. Now it seems you do PWM control.
Please clarify.


Generally true .. but never with a wsitching device (MOSEFT) inbetween.
So between
* +Out and +S and
* -Out and -S
there may be the shunt (non switching), but not the MOSFET.

***
In post#1 you wrote you want to compensate th voltage drop of the shunt.
I gave you the information how to do it.

But what you try to do is to include the MOSFET into the feedback loop.
As soon as he MOSFET is OPEN, the motor voltage drops to zero ... and the supply tries to maintain the motor voltage. It increases the voltage ... until OVP ... because it has no chance to compensate for the voltage drop across the OPEN MOSFET.

--> Do how I told you. and get waht you requested for.

Klaus

Hi Klaus

I have understood your point. I was missing the fact that a switching device in a feedabck loop will sabotage the loop.

Secondly regarding the PWM control : Yes the Motor speed is controlled by PWM which is separately provided and is separate from the ON/OFF contol of the MOSFET.

 

[KlausST]

Hi

I have understood your point. I was missing the fact that a switching device in a feedabck loop will sabotage the loop.

Imagine this:
Try to pull on a rope (like a crane lifting stones) .. with constant force. Now someone cuts the rope. What will happen?

PWM:
I don´t think it´s a good idea. You can´t expect the power supply to compensate for high speed (switching edge of the PWM) steps in voltage drop of the shunt.
The SENSE feedback is meant for compensating constant or slowly changing voltage drop.
For long wires it (as already mentined) try to add a capacitor at the motor side of the long cabling.
Again: the capacitor not across the motor, but to stabilize the power supply "before" the switching MOSFET.

***
I have done a lot of measurement and regulation tasks. From this experience I ask myself how much the feedback improves the performance at all.

So tell us:
* why you need that accurate voltage at all
* what´s the voltage range of your application
* what´s the current range of your application
* what´s the expected shunt value
* what´s the expected overall accuracy (error) of your circuit

***
PWM and shunt:
Usually the shunt is for current measurement. Low value. And for WMing the motor you need a free wheeling diode.
The free wheeling diode carries current during PWM. This current flows through the diode and the motor, but in most cases not through the shunt. So shunt current differs from motor current. This causes a measurement error. usually bigger than the voltage drop across the shunt.
So if you want to keep accuracy high you need to look and cure the "biggest" errors first.

All depends on your applications parameters ... I asked above.

Klaus

 

[abicash]

Hi



Imagine this:
Try to pull on a rope (like a crane lifting stones) .. with constant force. Now someone cuts the rope. What will happen?

PWM:
I don´t think it´s a good idea. You can´t expect the power supply to compensate for high speed (switching edge of the PWM) steps in voltage drop of the shunt.
The SENSE feedback is meant for compensating constant or slowly changing voltage drop.
For long wires it (as already mentined) try to add a capacitor at the motor side of the long cabling.
Again: the capacitor not across the motor, but to stabilize the power supply "before" the switching MOSFET.

***
I have done a lot of measurement and regulation tasks. From this experience I ask myself how much the feedback improves the performance at all.

So tell us:
* why you need that accurate voltage at all
* what´s the voltage range of your application
* what´s the current range of your application
* what´s the expected shunt value
* what´s the expected overall accuracy (error) of your circuit

***
PWM and shunt:
Usually the shunt is for current measurement. Low value. And for WMing the motor you need a free wheeling diode.
The free wheeling diode carries current during PWM. This current flows through the diode and the motor, but in most cases not through the shunt. So shunt current differs from motor current. This causes a measurement error. usually bigger than the voltage drop across the shunt.
So if you want to keep accuracy high you need to look and cure the "biggest" errors first.

All depends on your applications parameters ... I asked above.

Klaus

Hi and thanks for such an elaborate reply . Appreciated!

This project is about testing few parameters of a BLDC motor like sleep current , active current and current profile of the motor.
The PWM control is an inherent control available on the motor and it only expects a few Hz at varying duty which varies the speed of the motor.
The scheme which I have shown in my OP is about providing the DC terminal voltage to the motor and it is isolated from the PWM control.
Thus , the main current which flows through the MOSFET , shunt and the motor is what I have to measure and has 3 ranges ,
a) up to 100 uA, b) 100uA upto 200mA , c) 200mA to Max current (around 100A) with flexible accuracy in each range.
I have 3 separate MOSFET brances with different shunts to measure these currents. One shunt is actually a Hall sensor.
The measurement happens on a NI card.

I have a FW Diode at the Power Supply leads for protection.