10 Bit absolute encoder

manmeetvirdi · Jun 21, 2007

Hi there,

We are using the absolute encoder from the SICK, model :ARS 60-FAA-01024.
(Single turn shaft/hollow shaft featuring core tech)

1024 implies the disc of the encoder is divided into 1024 steps giving resolution of 0.351 degrees.
For 1024 steps the digitals input (DI) signals needed to be used is 10 (From G0 to G9). Now due to limitation in DI of PLC, I can use only 8 DI to fit 4 encoders like this. If I use only 8 DI then I think I can measure angle up to 90 degree only through this encoder. Right?

The ZERO adjustment of the ecoder is possible by pressing a button located under a cap at the rear end of the encoder but they have not said anything about adjusting the resolution of the encoder.Is this because its not possible in case of absolute encoder? or this is possible but this facility is not available in this model of the encoder?

Manual of the encoder: http://www.stegmann.com/product/absolute/index.html

Regards,
Manmeet

bernie_carlton · Jun 21, 2007

If you use the 8 least significant lines (G0 to G7) then you will have 256 increments 4 times per revolution. So you would still have the accuracy but it would only be absolute for 1/4 turn.

If you use the 8 most significant lines (G2 - G9) then you will have 256 increments once per revolution. You will have 1/4 of the accuracy but it will be absolute for a full turn.

manmeetvirdi · Jun 21, 2007

Hi Bernie

Got your point, thanks very much. Since the maximum angle which machine is expected to move is only 45 degrees, well within the absolute range I can go ahead using only 8 inputs of this encoder.

If I use G0 to G7 as you said then in full 360 degree I will get the reading something like this:

0---90---0----90-----0,right?

If I use G2 to G9 then in full 360 degree I will get 0—90 only once.

panic mode · Jun 22, 2007

almost but not quite, in second case you will not get 0-90 but 0-360 and - rounded...
this means it will be of lower resolution (for example in decimal system 0.0..360.0 is 10x lower resolution than 0.00...360.00;
range is same 360degree but later one can show more detail or
in other words, it has better resolution).

bernie_carlton · Jun 22, 2007

I couldn't tell from the documentation but this may be a Grey Encoded output. In that case you would use G1 - 10 for a shortened 8 bit interface and perform the appropriate grey to Binary decoding. This would give you 256 counts for a full turn. There would be no way, as far as I can tell, to get a 1/4 turn hookup with Grey code.

bernie_carlton · Jun 22, 2007

An update on my previous post - I verified that the Sick part number supplied by manmeetvirdi is a Grey coded output. Thus my previous post stands. You will have to use G1 through G10 and perform the Grey To Bi nary transformation. This will give you 256 readings for 1 full revolution. 45 degrees will only have 32 different codes. I hope that will be sufficient.

manmeetvirdi · Jun 22, 2007

Hi Bernie

Iam using G0 to G7 (8DI) only.

I am simulating the program , and when absolute encoder will come in a position such that it gives the following output:10000000 binary equivalent will be 255.

This multiplied with true resolution of the encoder(.35) gives approx 90 degrees of angle.

Now from this position, if encoder moves again by 0.35 degrees then the output of the encoder will still be :10000000 (because Iam reading first 8 bits only) but in actual the output will be:110000000.

Gray code will start mirroring it up after this value.Right?

this is a good link to understand mirroring porperty of gray code: http://www.diycalculator.com/docs/puzzle-gray-code.pdf

bernie_carlton · Jun 22, 2007

Using only the low 8 bits you will see (after Gray Code conversion)for one full rotation:

0 - 255 then
255 - 0 then
0 - 255 then
255 - 0

(This illustrates the 'mirroring' of the lower bits of the Gray Code.)

And you will not be able to tell which quadrant you are in. But if you can constrain the revolution to only 1 of the 90 degree sections you would be ok. You mentioned that it would be only 45 degrees.

Using G2 - G9 (ignore the 'G1 - G10' suggestion in the previous posts) would give you 256 unique codes for 1 full revolution.

manmeetvirdi · Jun 23, 2007

bernie_carlton said:
Using G2 - G9 (ignore the 'G1 - G10' suggestion in the previous posts) would give you 256 unique codes for 1 full revolution.

Hi
Using G2-G9 is a good idea.
As operator may bring the machine to diffrent quadrent in manual mode and then in auto its function will be unpredictable.

10 Bit absolute encoder

manmeetvirdi

Member

bernie_carlton

Lifetime Supporting Member + Moderator

manmeetvirdi

Member

panic mode

Member

bernie_carlton

Lifetime Supporting Member + Moderator

bernie_carlton

Lifetime Supporting Member + Moderator

manmeetvirdi

Member

bernie_carlton

Lifetime Supporting Member + Moderator

manmeetvirdi

Member

Similar Topics