??? 03/01/05 14:19 Read: times |
#88779 - Yes it can but only... Responding to: ???'s previous message |
Yes it can but only if you take into account the sequence of two phase transitions like Russell has said.
If you use the digital approach such as the software sampled approach as Peter has implemented or an equivalent hardware approach without filtering then the uncertainty will accumulate always up to 1 count each time you stop on a fuzzy/buzzy edge where the fuzz is at a frequency that is higher than your sampling time but at such rate that the software or hardware sees it sometimes high and sometimes low at a different number of times than it has actually transitioned back and forth. If you need full resolution of the quadrature wavefrom it then becomes necessary to design your system in a manner that it can live with this 1 count uncertainty accumulation. If the long term implication of this accumulation is unaccaptable to your design (where long term can be defined as say one or more revolutions of your encoder) then it is necessary to use an encoder with an INDEX pulse that you can re-sync to. In the case of a linear encoder such as you may find along the slide on an X-Y table or in a linear actuator resyncing to null out the 1 count uncertainty accumulation can be more problematic since often the index pulse occurs only once along the whole encoder and is sometimes used as the HOME signal. In these cases it is necessary to get an encoder that has a TICK (my term ... but commonly known as a repeating INDEX) output that occurs at a repeatitive output say every 1024 quadrature periods. Such output does of course require extra wires and conditioning and capture circuitry at the microcontroller. If the linear encoder only offers a single extra output (TICK or single INDEX) then it would be necessary to select the TICK type encoder and implement the home position as a logical AND of a TICK pulse and an optical or proximity detector on the mechanism. Michael Karas |
Topic | Author | Date |
Digital position encoder | 01/01/70 00:00 | |
Dont do this. | 01/01/70 00:00 | |
Signals are not connected directly | 01/01/70 00:00 | |
Is assembly ok? | 01/01/70 00:00 | |
For Mehdi | 01/01/70 00:00 | |
its easy in software | 01/01/70 00:00 | |
but | 01/01/70 00:00 | |
same resolution | 01/01/70 00:00 | |
But...what if ? | 01/01/70 00:00 | |
So what is the solution? | 01/01/70 00:00 | |
absolutely no problem | 01/01/70 00:00 | |
Thank you | 01/01/70 00:00 | |
PLEASE not again | 01/01/70 00:00 | |
Sorry | 01/01/70 00:00 | |
list.hw.cz | 01/01/70 00:00 | |
thank you | 01/01/70 00:00 | |
Re: PLEASE not again | 01/01/70 00:00 | |
Re frequency | 01/01/70 00:00 | |
So what frequency do we sample at? | 01/01/70 00:00 | |
I try to explain Peter's idea again | 01/01/70 00:00 | |
We know it ! | 01/01/70 00:00 | |
Nothing ??? | 01/01/70 00:00 | |
it takes time | 01/01/70 00:00 | |
Code doesn't function | 01/01/70 00:00 | |
reverse direction ! | 01/01/70 00:00 | |
This is not a problem | 01/01/70 00:00 | |
incorrect operation? | 01/01/70 00:00 | |
right connection ? | 01/01/70 00:00 | |
I dunno | 01/01/70 00:00 | |
Now... | 01/01/70 00:00 | |
check the sample condition ! | 01/01/70 00:00 | |
clarify | 01/01/70 00:00 | |
@Erik | 01/01/70 00:00 | |
@Peter | 01/01/70 00:00 | |
Grey code | 01/01/70 00:00 | |
Grey code more | 01/01/70 00:00 | |
Grey INCREMENTAL ? | 01/01/70 00:00 | |
I guess 98 | 01/01/70 00:00 | |
Gray code | 01/01/70 00:00 | |
Not a gray code | 01/01/70 00:00 | |
Grey | 01/01/70 00:00 | |
Yes of course its Grey code | 01/01/70 00:00 | |
Absolute Vs Incremental | 01/01/70 00:00 | |
So now... | 01/01/70 00:00 | |
Yes it can but only... | 01/01/70 00:00 | |
Peter is right! | 01/01/70 00:00 | |
Lover not a fighter :-) | 01/01/70 00:00 | |
Lover of software | 01/01/70 00:00 | |
steps lost or not | 01/01/70 00:00 | |
Determining direction | 01/01/70 00:00 | |
Its not Grey, its Gray !![]() | 01/01/70 00:00 |