Limitations of HSC and SETSYSTEM 4

[plc_user]

Hello again,  after about five years of use of the MD888 in my facility, its still going strong.  The only problem I have with the unit is the HSC is not keeping a consistent count.  I gain/lose pulses all the time.  I have commented on this problem in the past here in this forum, trying the suggested solutions with no improvement.  

I have the 888 connected to a HMI using the Omron protocol through the RS232.  The HMI displays a scaled value of the HSC that relates to the position of a log relative to a saw blade.  I can set the HSC on the fly after verifying the position with a tape measure.  Ie, the tape reads 12" so I enter 12" on the HMI and scale that input back to the correct HSC value.  By the end of the day this is off by several hundred counts.  My feedback device is a Quadrature Encoder.  I have used several different encoders with no improvement.

I know the knee jerk reaction is electrical noise.  I have eliminated this thought by 1.  O-scope inspection.  2.  Ran a test program that sets the HSC inputs to interrupts and seen if there was any noise tossing the inputs around.  No interrupts were fired.  I am confident that noise isn't an issue.

I did have some limited relief using this  http://www.beiied.com/PDFs2/AntiDither.pdf
which leads me to believe that maybe under certain circumstances the SETSYSTEM 4 command fails.  I am not sure if at certain frequencies the encoder transitions sent to the HSC are not being handled correctly.  
I am not sure if anyone else has had these issues with the unit.  I really like the PLC, they are awesome in my opinion, but I cannot get this problem resolved.  

I am willing to try another solution if all else fails, but to me its quite extreme.  I am considering sending the encoder channels to schmitt trigger RC low pass filter then to a D Flipflop(for each channel) that is clocked with a much lower frequency than 10khz(spec of the PLC), but higher than my max output frequency of 1.3khz; putting the encoder channels on the D inputs and only sending the channels out to the HSC inputs on the clock transitions.   I really don't want to do this.

If it helps anyone, I have used both open collector output encoders and line driver outputs of A/A' B/B' and none of this has helped.  

I will give a better description about all of this, if I see an interest in resolving this problem with everyone here.  All ideas are welcome.


Regards

[support]

cumulative error may have something to do with the jiggling effect when changing direction. I presume your system would be moving fore and back hundreds or thousands of time day and the error was accumulated?

Do you have a home position or known position where you can set the HSCPV[n] counter to zero or to a fixed value whenever the home sensor is activated? This can eliminate the cumulative error quite easily.

[plc_user]

Yes, the direction changes very often and I am sure that at some instances one channel transitions(jiggles) many times due to vibration, but I thought that the setsystem 4 statement resolved this issue?

I can install a home position sensor, but I was hoping that we would not have to do this.  I have seen other saw control systems built by others using omron and idec plcs without the use of a home sensor.  I was kinda hoping the same.

Considering you mentioned the home sensor are you implying a plc limitation or firmware issue?  I really like your PLC, they are easy to program and fun to play with.  For the life of me, I just cannot understand what is happening.

[support]

Thank you for your feedback.

Does your program run any save_eep or load_EEP statements. On T100MD+ these commands shut off the interrupt momentarily and can have adverse impact on the performance of the interrupt based HSC firmware.

Also when the HSC reverses direction, the width of pulses produced could be much narrower than the frequency response of the HSC which resulted in lost counts. The limitation could be due to the slower CPU on the T100MD+. The new F-series PLCs with 5 time faster CPU and even shorter interrupt latency would be less susceptible to such losses.

We suggest a home sensor to eliminate the cumulative error in HSC because it is relatively simple to implement and generally works very well. Hopefully this will work out for you too.

[plc_user]

Thanks for the response, no I am not writing to or reading from the EEPROM.

I would entertain using the F-series, if I got one for free.  
Until then, I guess I will have to implement the home sensor.

Thanks for your time, thanks for the great product, I really find these controller's incredible.

Regards