The problem is not the encoder, it's the code. You need a better internet.
I don't know what the CH A/B "PGT" columns are supposed to mean, I'd have guessed from the code that they *should* mean something about whether the channel had a state transition, but the spreadsheet data isn't consistent with that interpretation.
Further, it isn't clear why there would be separate columns for counts on channels A & B. That's not how quadrature decoding should be done.
Generally the code looks sorta approximately reasonable, but the overall form seems wrong. My understanding is that the correct way to handle quadrature decode includes recognizing valid vs. invalid transitions in combined A,B state, *NOT* trying to first interpret each channel individually and then combining the result.
There are 4 quadrature states - {(0,0), (0,1), (1,1), (1,0)}. Direction of motion is determined by whether your state transition moved you by one element from left to right in that list or right to left. You can't move by two elements at a time -- these are illegal transitions for quadrature decoding, suggesting that both A & B changed simultaneously. The quadrature count either increments or decrements by 1 each time a valid quadrature state transition happens. Invalid state transitions should cause a fault.
...(time passes). Does "PGT" stand for "positive going transition"? If so, there's definitely a problem with the code. Quadrature decode needs to care about both rising *AND* falling edges on both A and B. Any algorithm that ignores falling edges cannot do a proper quadrature decode.
EDIT: please note that in the discussion below, I'm talking from the point of view of standard x4 quadrature decoding. It appears you're looking to do x1 decoding. x4 decoding is actually more straightforward b/c there's no special rules. x1 decoding kinda requires you to designate some specific pair of states as the "master transition" pair -- your count only changes when transitioning between that specific pair of states in one direction or the other.
You can unit test your algorithm with some test scenarios to prove it out.
1. a stream of (A,B) states that represent continuous positive rotation. Each state transition should cause a +1 increment in quadrature count.
2. a stream of (A,B) states that represent continuous negative rotation. Each state transition should cause a -1 decrement in quadrature count.
3. a set of 4 streams of (A,B) states. Each starts at a different one of the 4 possible quadrature states. The rest of the stream follows by first toggling CH A back and forth several times, ending on the original quad state. You should always end back at your original count (probably 0).
4. same as before but toggle CH B several times. You should again end at 0.
It appears to me that your code would fail steps 3 & 4. I didn't work through steps 1 & 2 carefully enough to comment either way.
-Kevin P
ALERT! LabVIEW's subscription-only policy came to an end (finally!). Unfortunately, pricing favors the captured and committed over new adopters -- so tread carefully.
