Computer crashes for inductive load

[rk_t]

Hello,

I my application, I am trying to test a 12V DC motor through a labview program i created.

Connection diagram is as shown in the figure attached.

I am using NI USB 6002's analog output 0 to generate a 0-5v signal which is then given to a differential amplifier. This is a unity gain amplifier. The output of this amplifier is then connected to a SMPS based 0-30V power supply as 'remote voltage' input. This power supply in turn generates 0-30v out put in proporting to the 'remote voltage' applied.

When there is no load or resistive load connected, the system works fine.

But then I connect a DC motor, the computer system just crashes.

Can anyone help me with this.

Thanks!

[Norbert_B]

What do you mean with "system just crashes"? Do you get a crash dump?

Without further information i guess that there is a feedback over the USB which in- or decreases voltage levels inside of the computer as such that the mainboard detects a malefunction and making the computer to reboot.

[rk_t]

yes, crash dump! the one with the blue screen

[Norbert_B]

If you have a full crash dump (not the mini one), i recommend you to contact your local NI branch directly.

However, please make sure that you do not have voltage/current feedback to the computer. The best way is to include a galvanic separation between the measurement device and the motor.

[pallen]

Check your Output current draw.

http://www.ni.com/pdf/manuals/374371a.pdf

The specs indicate this device is capable of 5mA only.  You want to be sure you're sourcing you're drive current from somewhere other than the USB device. 

An over current condition could be causing your crash.

[AnalogKid2DigitalMan]

What is the make/model of the motor?

My hunch is that the inductive component of the motor or possibly commutation noise is generating voltage spikes that are conducted back through the power supply/op amp/DAQ chain and into the PC causing the crash.

Any way to provide electrical isolation in the chanin of devices?

-AK2DM

[Ben]

I will not speculate on the reason but I do want to point out there may be a safety issue here.

Galvanic Issolation!

Ben

[rk_t]

I have already tried to electrically insulate the motor from my circuit using a differential amplifier. That did not work.

So then I tried to narrow down the reasons and found out that it is indeed the noise from the inductive motors causing this.

So then I did 2 things:

1. Used an inductors in series with my both positive and negative of the NI analog output(which acts as remote voltage to power supply) and the powersupply which in turn gives me the acutal voltage and current required to run the DC motor. I tested this for around 600 cycles. Worked fine.

2. I shorted the negative or the output of the powersupply(which is SMPS based) with Earth. This also seems to work. I have tested this for about 400 cycles and counting.

Now I wanna know which one is the best solution. Is the second option really recommended? I have heard it might give problem to other machines on the factory floor.

[johnsold]

1. The biggest issue with filters in the analog output lines is whether the outputs ever need to change faster than the filters will allow and whether the analog output circuits are stable with an inductive load. Apparently neither is a problem for your situation.

2. Earthing (or grounding) is a complex subject. First and foremost be certain to comply with the electrical safety code for your location. Some places call for connections to earth while others do not. The next consideration is whether the output circuit of the SMPS power supply is isolated from the mains supply. Again, both isolated and non-isolated designs exist.

The general principle for measurements is that every device should be connected to the system voltage reference (which may or may not be earth) by exactly one conductor and that no other device may share that conductor. "Device" includes power supplies, motors, the computer, the DAQ device, the signal conditioner (differential amplifier), and so on.  Unfortunately, many ssytems cannot easily follow this principle because some components were not designed with that principle in mind. For example if the computer chassis is connected to ground through the power cord, then the USB ground is probably connected to ground through the USB cable. Connecting the USB DAQ device AI ground to a different ground connection may result in a ground loop.

The fix often requires careful analysis of where currents are flowing through earthing conductors.

Lynn

[Henrik_Volkers]

The isolation/differential amplifier should do the job.... 

BUT : Will the system crash if you run the motor/inductive load without the AO connected? 

Just the GND/neg pole connected?

(using a pot or a switch to control the load....)