High Speed Counter Misoperation

[sid 0]

I am using c238 counter(at input X3) to count total no. pulses from 360ppr encoder. Another counter C251(Bi phase counter at x0 & X1) is used to count servo feedback pulse from MRE Servo encoder for position feedback. Input X0 is tied to X2 which is used to count total no. of pulses from servo encoder using C237. Sometimes counter C237 starts to counts even when machine is in stop condition (Servo is in lock condition). Control wires are shielded and shieldings are grounded at each ends. I am not able to understand that this is because of noise or any hardware fault. *(s/s is connected to 24VDC)

[BobLfoot 302]

I am no mitsi expert so take this with a grain of salt,but two thoughts come to mind. 1. Servo Lock does not mean no motion is simply means the servo is constantly moving minute distances to keep a reading at zero. Usually to fight gravity or vibration. So these counts may be valid. 2. Shields grounded at both ends. The schooling I had on shield gronding was always one end only to prevent ground loop induced noise onto your control lines. I'd try lifting the shield gronds at one end of the cable. Others may have better answers. You get what you get at 4 am.

[sid 0]

Counter(c237 when misoperating) is counting at fast rates i.e 5 or 10 pulses/sec, and also my output servo encoder pulses are set to 1000pulses/rev. So i think this is not because of servo lock. Earlier i had grounded shieled at one end but later because of these problem i tied both ends to GND.

[Sergei Troizky 26]

Even being stopped, a servo system performs the feedback loop which may result in oscillations. Being a single-phase counter, the C237 will count every encoder pulse, even if this is just oscillation around the same encoder position.

[sid 0]

Its necessary to use c237, so what is the solution? I think i have to switch over to proportional control mode of servo whenever servo is at rest.

[Colin Carpenter 10]

I had a problem with flow meters sending pulses when I knew that there was nothing going through them. Turns out that sometimes they do it when moist air is passing through rather than liquid. Solution was to put a contact prior to the high speed counter instruction that only allowed the counter to count pulses when the PLC was expecting them to be there .... in other words when the relevant pump was running. Is it possible to do that with your system?

[sid 0]

In my application, it's not possible to put a contact in series with high speed counter. Is there any other alternative?

[Colin Carpenter 10]

I have known one instance of a rogue input on a FX PLC where the input would come on and go off at will even though all grounding was improved and PLC S/S terminal was correctly connected. Have a feeling it was X5. In this instance, solution was to just move it to another input as it wasn't being used for a high speed counter. Never did understand why that one input did it .... but it would come on even though no wires were connected after swapping the cabling. Also, I've known the high speed counter inputs X0 - X7? have problems coping with a rough (unsmooothed) 24VDC supply that all the other inputs on the PLC could quite happily "see". Seem to recall tech support telling me that it was possible to adjust threshold voltage levels for these inputs, though never pursued it to find out how to do it ..... just improved the 24 VDC. Not much help, I know, but you never know ......

[panic mode 246]

i think you should read more carefully what Sergey Troizky posted, you need to understand how high speed counters work... if you have just single input (such as only X2 or only X3), you sure can count pulses but can't tell in which direction (forward or reverse). this is why quadrature encoders are used for positioning (two inputs per counter). with two signals it is possible to tell direction as well, based on phase difference between two signals (such as X0 and X1). btw. overall number of pulses received by X0 and X1 is almost the same (just one pulse diff. when direction is changed) but because of the phase difference, it is possible to tell if counter should count up or down. even when servo is in lock you will get few pulses forward then few reverse etc, this is how locked condition is maintained (servo performs continuous corrections even though it may look that it's not moving at all). thanks to direction information, counters with quadrature inputs will count little bit up then little bit down but remain around target. single input counters however (pulse counters), will not know the difference, they will blindly count every pulse regardless of movement direction. they will keep accumulating pulses and drift away from target...

[P Daniil 0]

I can only contribute on the hardware aspects of this problem. There are a number of hardware related causes for the symptoms you describe. (Sergei, above, pointed to one of these). Our paper "Eliminating Phantom Movement in Encoder Applications" goes over these and may be of help/interest. Good luck

[Sergei Troizky 26]

Actually, not clear why read the servo encoder feedback into PLC. Especially if there is another encoder for the machine position reading. Servo encoders are usually of high resolution which may result in excess counting frequency on PLC inputs. A servo drive in position control mode will perform a feedback loop by itself and emulate a stepmotor for the PLC. Some of them even provide their encoder passthru with user adjustable frequency dividing ratio. Even in your configuration, if the servo encoder is quadrature, you can have two AB counters (C251, C253, one per encoder) and also one extra single-phase counter C237. Just redistribute the inputs and verify for top frequencies. If the servo encoder is single phase, there is no simple solution, IMO. The concept should be reconsidered. Why, by the way? Edited 14 Nov 2006 by Sergei Troizky

[sid 0]

I am attaching application details with diagrams Application_Details.pdf