Understanding encoder logic

[karivelm]

Grafcet can be implemented into ladder logic? I would just like to have a visual idea of what needs to be programmed.

 

[Peter Nachtwey]

Basically what I mean by "OK", is that I am looking to control the position of a motorized launder by 1degree through an HMI. I wanted to make sure that I could obtain that degree of precision...

That isn't much
I think the big problem is exceeding the count rate. Given the RPM you were talking about.

I ask first because the combination of 800 kHz pulse frequency and only a 16-bit accumulator sounds like a gross mismatch to me

I don't understand. Most motion controllers have 16 bit hardware counters and can count up to 10 MHz. A 16 bit encoder will do just fine since the maximum number of counts in 10 millisecond, a PLC scan, is 8000 counts. A 16 bit counter will handle rates up to 32767 counts per scan. A 32 bit counter may be more convenient but it isn't necessary. One only needs to look at the change in counts each scan and add that to a DINT or DWORD.

 

[kamenges]

Originally posted by Peter Nachtwey:

I think the big problem is exceeding the count rate. Given the RPM you were talking about.

His max pulse rate is only 150 kHz. His card can handle 800 kHz. At 150 kHz he shouldn't even need to worry about cable capacitance too much assuming he is staying under 200 feet or so.

Originally posted by Peter Nachtwey:

I don't understand....A 32 bit counter may be more convenient but it isn't necessary. One only needs to look at the change in counts each scan and add that to a DINT or DWORD.

I suspect that didn't even occur to him. That doesn't seem to occur to most people when it comes to position devices, even though they will happliy use the same technique in other applications. Oh well.

Keith

 

[Steve Bailey]

Peter, you have a tendency to view everything through the lens of a motion control application. High speed counter modules in PLCs get used for a lot more applications than that.

My point about the mismatch between pulse frequency and the number of bits in the counter's accumulator register is that at maximum frequency of 800 counts per millisecond, the counter can only track the encoder for 81.9 milliseconds before overflowing. That's not too good if you're using the HSC in a programmable limit switch application.

 

[Peter Nachtwey]

Guilty as charged.

Peter, you have a tendency to view everything through the lens of a motion control application.

Guilty.

My point about the mismatch between pulse frequency and the number of bits in the counter's accumulator register is that at maximum frequency of 800 counts per millisecond, the counter can only track the encoder for 81.9 milliseconds before overflowing. That's not too good if you're using the HSC in a programmable limit switch application.

So this is how a motion controller guy approaches this.
Each scan
DifEncoder16:=NewEncoder16-OldEncoder16;
OldEncoder16:=NewEncoder16;
Counter32:=Counter32+INT_TO_DINT(DifEncoder16);
If the PLC scan is 10 milliseconds then DifEncoder will be +/- 8000 counts or less. Three simple rungs. I see no problem.

 

[ndzied1]

I'm wondering bow much backlash this over 4000:1 gearbox has? I used a 5 stage once that was only around 2600:1 so I'm guessing this could be 5 or 6 stages. Depending on the quality of the gearing, the whole tolerance could be eaten up in backlash and measuring the output to +/-1 deg with the motor encoder could be fruitless.

Have you considered putting an encoder on the shaft you are actually worried about, i.e. The output shaft. (speed is no issue there...)

 

[kamenges]

Originally posted by ndzied1:

I'm wondering bow much backlash this over 4000:1 gearbox has?...Depending on the quality of the gearing, the whole tolerance could be eaten up in backlash and measuring the output to +/-1 deg with the motor encoder could be fruitless.

Holy gear clearance, Batman!!! You have actually seen a gearcase without a serious defect where you could spin the input shaft more than 10 TIMES without the output shaft turning? That is HIGHLY impressive.

Keith

 

[ndzied1]

Holy gear clearance, Batman!!! You have actually seen a gearcase without a serious defect where you could spin the input shaft more than 10 TIMES without the output shaft turning? That is HIGHLY impressive.

Keith

You have a point there. I was tired when typing and had it in my head that the backlash of each stage in the gearbox got multiplied. I have to do a little more research to really figure out how backlash works across multiple stages or a gear train.