Inconsistent measurements of encoder position using USB-6215.

[J.A.C.K.]

I am using a NI USB-6215 to get position measurements from a EMIX23 Magnetic Encoder. If I move any amount of steps in the negative or positive direction and reverse the move there is significant error. When using the manufacturer's display this process results in zero. When using the NI USB-6215 I get at least 10% to 20% error. Does anyone have any suggestions or success usin this USB-6215 product for this type of application?

[Luke_G]

Hi J.

I don't have any experience with that particular linear encoder, but I see here that it generates a TTL signal.  I'm not clear on what it means to have an output that is proportional to the speed.  Is that referring to the pulse width? The 6215 has two counters, so it should be capable of this application. 

Can you describe your program? 

Let us know how we can help.

[J.A.C.K.]

In the Linear Position Task setup that is available in SignalExpress I am not getting consistent readings. If I move the slide 200 steps and then move 200 steps back to the original position the vaalue that is read is not 0. Furthermore everytime I try this experiment the value is different. I am using a EMIX23 Magnetic encoder here is alink to the datasheet for the encoder. http://www.velmex.com/pdf/other/magnetic-encoder-data.pdf?produkt=EMIX/LMIX&id=5&z=4

I have the Green wire connected to PFI0, Yellow wire connected to PFI1, Brown wire connected to +5V and the White wire connected to DGND all on the NI USB 6215.

[Luke_G]

It sounds like your wires are connected correctly. 

Can you tell us how you have configured the acquisition step in SignalExpress?  When you configure a linear position task, what did you choose as the decoding type?  What did you set your timing settings to?

Also, when you slide the sensor forward and back, does the error seem to be proportional to the speed? 

I'm really not sure what could be causing that much error, so any clues are helpful.

Thanks,

[J.A.C.K.]

Luke,

  Thanks for your help Luke.The problem I was having was a noise issue de to the shield not being connected. Now it misses one or two counts in either direction, not really a big deal at this time. The next question I have is about the decoding type. When I look at the diagram in the manual X1 increments by 1, x2 increments by 2, and x4 increments by 4 per full cycle. The issue is that it seams as though no matter what the decoding type is, you still get the same value. Shouldn't the values differ if the decoding type changes?

Jack

[Luke_G]

Hi Jack,

Actually, I believe that the DAQmx driver compensates for the multiplied counts and returns an accurate position.  I tested this with a quadrature encoder and found the same result (360 degrees for one revolution) with each decoding type. 

I would like to point out, however, that I think the example diagram that is displayed when you configure the step is incorrect.  The diagram indicates that a count for a given number of pulses will be higher with a higher multiplicity of decoding type.  I will need to double check that, though.

Thanks,