IR2130 Current Sensing Problem

[losturcos]

Hello,
I am driving motor with IR2130(3phase) and controlling the IR2130 with PIC. I am using 5 12v/7a batteries.
The system is working like this:
First I am starting the motor to run to one direction for 2 seconds. I am measuring the current. Then I am
making the motor to run in reverse direction for two seconds. And I am measuring the current again.
I am checking which current measurement is lower, and then I am starting the motor to run to the lower
current direction which should be the easier direction.
While I am making tests at a motor which is not carrying load, it works as I think. But when I am trying this
in another motor which is carrying load, it happens reverse. I mean the current measurement lower in the direction
which is harder to run.
Do you have any idea about why this is happening? The IR2130 schematic is as below:


Thanks

 

[losturcos]

No idea? I couldnt find any reason yet.

 

[losturcos]

Nobody has an idea?

 

[iop95]

It's possible that current to decrease when load increase; Induction motor at no load looks like an transformer w/o load, so major component current is reactive. When add load, reactive component lower and active increase; so, rms value of current may decrease when load increase (up to a point, where active component became dominant and surely current increase with load). In this case I think load is small (maybe 10-15% of nominal).

 

[losturcos]

Thanks for your help.
Do you have an idea abouy how I should detect the easier turning direction?

 

[iop95]

May you give more details about application / requirments?

 

[losturcos]

I am driving 3-phase asyncronous motor by using IR2130 and IRFP250N from PIC microcontroller with 5 batteries. When needed, I must drive the motor to the easier side.

 

[pilko]

@ losturcos,

Can you please post the full schematic.

Thanks

pilko

 

[losturcos]

Driving circuit is as below:

 

[iop95]

Without no details about application/requirments I try to propose a solution. Problem is to detect when IM have more load (for what direction). If at no load (or light loads) power factor is so weak that is hard to detect higher load situation, may be a solution, much complex, to compute power factor. It's known that power factor increase with load, so a higher PF mean a higher load. But, to compute power factor, need to sample voltage also, detect maxim or zero crossing for both (current and voltage) and compute time difference, PF respectively.

 

[losturcos]

can you give a little more detail about calculating PF? A formula for eg.

 

[iop95]

PF=cos(fi) where fi is angle displacement between voltage and current. So, when displacement between voltage and current go to zero, PF go to 1, it's case for resistive circuits; when displacement angle increase, PF decrease and go to zero when displacement angle go to 90 degree (pi/2). In this case, I think is easier to detect when voltage and current from same phase cross zero (in same direction, from negative to positive for example). Compute difference time between these zero crossings and lower value mean higher PF. You don't need absolute value, only detect direction with higher PF.

 

[mtwieg]

How are you measuring current? Do you have a separate measurement method, or are you using the IR2130's current sensing?

I've never used the IR2130 in particular, but I think your current sensing circuitry has some problems. The datasheet indicates that the Vs0 pin should connect directly to the Sources of the lower FETs, while the other end of the current sense resistor should connect directly to Vss. And I'm not sure what the purpose of the diode in your circuit it. It seems you're trying to put some kind of a lowpass filter on the Vs0 pin, but that will likely interfere with the lower gate drivers, since they use that pin as a return path for gate current. If you want to filter the current sense signal, then filtering the output of the current amplifier, which you're already doing, seems fine.

 

[losturcos]

Thanks for your replies.
@iop95: I am checking how I can measure PF. I think I will need to make changes at the circuit for that.
@mtwieg: Do you think that diode can cause that kind of problem? I will try by removing the diode.

 

[iop95]

mtwieg right, Vso pin must be conected direct to FETs sources; in your schematic, lower side FETs are not corectly driven but much important, non-inverting input voltage of current OA is disturbed by spikes induced from lower gates divers return. Modify this according datasheet. Regarding PF compute, need to sense voltage and current crossing; for accurate I thing must use minim 100 samples/period. Also, take care about IM voltage frequency variation in these start-up test sequences. As I told, if can provide more details may be a simple way comparing this PF compute.

 

[losturcos]

What kind of details do you need? I will try to supply.

---------- Post added at 20:23 ---------- Previous post was at 20:19 ----------

By the way, "GNDB" is coming from the low side of MOSFETs. The diode is not between vs0 and mosfets. There is only 220R between the vs0 and mosfets low side.
Does that resistor effect?

 

[iop95]

Details about motor (speed, power), load type (value, inertial), load is changed for each direction? Why need to find direction with higer load in such why? Maybe is a more simple solution. Any other application requirment.

---------- Post added at 16:45 ---------- Previous post was at 16:40 ----------

Yes, resistor affect current sense circuit because voltage across it from drivers FETs modify non-inverting OA input. Vso must be connected like in datasheet.