8/21/2015 11:18 AM  
Posts: 1529 Rating: (202) 
I've noticed that Siemens is still using encoders with resolutions that are powers of 2, such as 1024, 2048, 4096 pulse/rev. I know this was done years ago to make the math easier with 8 bit processors, so the calculations could be done with bit shifts instead of using time consuming integer or floating point processes. The question is, is this still relevant? Is it better if I select a 2048 p/rev encoder instead of a 2000? This is purely a theoretical question since I know they both work fine. I still select those with powers of two, but one of my engineers asked this question and I could not give him a good answer. Any thoughts? 
8/21/2015 3:21 PM  
Joined: 10/11/2006 Last visit: 5/23/2024 Posts: 5852 Rating: (793)

Hello, meanwhile it is not more important to use encoders with 2^x ppr resolution. Encoders with e. g. 2000/5000ppr etc. will work without any disadvantages in actual drive systems. 
With best regards, 

8/22/2015 12:37 AM  
Posts: 1529 Rating: (202) 
Thank you for the reply. I suspected as much. I suppose that encoder manufactures do this to support legacy applications. 
8/22/2015 10:02 AM  
Joined: 9/8/2009 Last visit: 7/17/2023 Posts: 1410 Rating: (150) 
I do always tend to get an power of 2 encoder, for calculation reasons. Let we have an encoder with 1024 pulses (400 hex) and we want to calculate the difference between setpoint and actual position:
Difference can be looked as signed double integer in this case the result will be the shortest path. If we look interpret Difference as unsigned integer then the result will be the positive path. In both results, no problems regarding overflow or turnover will happen. 
Follow us on