prox on wheel teeth only one direction

Ken Roach said:

Is the chain on the sprocket fully clear of the prox switch; could it be drooping toward the switch in one direction only ?

Is the conveyor under a greater load in the direction where the prox switch stays on ? Is it possible that induced magnetic fields from a VFD are causing the switch to stay on ?

I stopped using proxes on sprockets when one of the operators I trained showed me how hard it was to get his work gloves on over his three remaining fingers. Especially with modern rugged contactless magnetic encoders, the risk and adjustment challenges of a prox on a sprocket just aren't worth the cost savings.

Click to expand...

sorry I forgot to say that the gear is driven by hydraulic two way valves.

Thank you so much for all the inputs so far. I will have to look more into the physical alignment issue. I will update as soon as I know.

What size is it (physically) 12mm 18mm what is the sensing distance (mm) and what is the switching frequency, do you have a part # we can look specs up, Typically the bigger the prox (round) the larger the sensing distance can be, doesn't necessarily mean it is.
You say you replaced it with same size, was it the same part number?

Could you provide part number of prox? Maybe it’s a reed switch instead of a prox. Two or three wires?

If it is indeed a reed switch that would explain why it works for only a day or so.

Maybe switch to Pulse Position Indicator encoder like photo craft where you can adjust the encoder ticks by dip switches. Can the belt on the conveyor be reached by mounting a wheeled encoder underneath? Or are you able to drill a hole in the shaft idler pulley to mount an encoder on the shaft?

I have not used the prox method in years due to the above concerns by other posters.

The manufacturer below has different mounting and options depending on your application.
Example of a manufacturer I have used above:
https://www.ttco.com/encoders/

have you tested old sensors?

1)have you tested any of the old proximity switches to see if they are indeed bad?
2) most sensors today have built in overload and short circuit protection, they can be shorted with no damage.
Any chance you have got a shorted signal wire to input and with the short it stays on until short removed???
3) is the input the only thing wired to prox signal wire, thinking they only handle small current and again being pulled down to protect itself from short or overcurrent??

A couple of other thoughts

Did you replace the prox switch with the same exact part number?

Is it a barrel prox or a rectangular housing? Is it mounted in the same way?

Already asked but things like part numbers, number of teeth of gear, fast and slow speeds would go a long way to figuring out what is going on.

Please excuse my ignorance on post 19

Proximities have a dome shaped detection field, it could be that in one direction it is seeing the gap of the toothed wheel but not in the other I had a similar experience where as a quick mod a toothed wheel was fitted to a conveyor drive so that we could position pallets with some accuracy, when we first fitted it, it seemed to work, however, we did find that when reversing the conveyor we got varying results in the position (need to add that the system measured the pallet length, detected types of runners & blocks so that the system could run the pallet into the strapper, decide using PE's if the strap swords could pass through the pallet and move it back/forward to find the gaps between the blocks). It turned out that the "Dome" of the sensing field appeared to be mis-shaped in other words When going forward the leading edge of the sensor was far enough away that the field was not strong enough to trigger the sensor, but running backwards the sensor field would detect the preceding tooth. We tried adjusting the distance but this gave inconsistent results in both directions, after trying various sixed Proximities we ended up having the teeth on the wheel ground to a smaller width, this cured our problem. A little hard to explain but when the leading edge of the proximity came in view of the next tooth the trailing tooth was on the trailing edge of the prox, this on occasion gave no change in state, it was impossible to reduce the number of teeth to increase the gap between so the sensor was clear of both teeth, so reducing the size of the teeth made enough gap so the sensor would not stay triggered. Ideally, a sensor diameter smaller would have done the trick but we could not get one. We did try different sensors and some seem to work better than others, for future machines they used fork type sensors but because of the nature of large conveyors & tolerances on the drive these were only partially successful (often the fork would touch the toothed wheel & damage the sensor). Before anyone says it, yes it was mainly due to pallets getting stuck & hairy A%$% staff jemmying them to get them moving that caused the brackets to get bent.

parky said:

Proximities have a dome shaped detection field, it could be that in one direction it is seeing the gap of the toothed wheel but not in the other....

Click to expand...

Yes but they are working for 1-2 days then stops working in one direction, you would think if this was the case it would not work from day one... or give some bad counts

Caveman said:

Please excuse my ignorance on post 19

Click to expand...

Does this now mean I need to start apologizing for WAGs that were not thought through?

May seem weird I know, but we had exactly that, it worked for a week or two, we even thought that the proximity may have picked up some filings but a clean & inspection did not solve the problem, perhaps clutching at straws but who knows. One other thought occurred to me, is this to a high speed counter input, and is the logic relying on internal code to enable the count direction by using the cylinder output to determine the direction ?. Just ideas, I have seen some really weird faults that could not be explained & those miraculously cured themselves without even finding a fault or what caused them in the first place.

parky said:

I have seen some really weird faults that could not be explained

Click to expand...

I had a light not working in my house once... it was on a 4 way switch, to this day I still have no idea what happened, it just stopped working and I could not find power anywhere, I had one of the guys that work for me go under the house and check for power and we tried to trace it back, then one day (a week later) it started working and has not missed a beat since and that was 10 years ago, but yes I agree sometimes odd things can happen.

I had to install a multi counter/relay/timer once that was on a high of a lobe cam because the signal coming from the would light the input LED but it was to fast for the input to register, so we used the relay to give the drive a longer input time (delay off) it was not a PLC it was in a drive that had I/O and the programming was in statement, they did not offer any HS cards, it was Control Techniques drive.

unsaint33, you may want to look at slowing it down to see if it will read then... if yes try one of these https://www.automationdirect.com/ad...ers/timer_relays_-z-_counters_-z-_tachometers

geniusintraining said:

Does this now mean I need to start apologizing for WAGs that were not thought through?

Click to expand...

I hope you're not thinking of starting a trend here (said with my foot completely out of my mouth this time)

Anyway - WAG is an official maintenance and engineering term often used by the highest level of engineers.

unsaint, you originally mentioned reading ones and zeroes. And then always staying on the one. What indicator are you actually looking at? Do you have an led indicator on the input? If the adjustment is backed off, does the indicator go to zero? I other words, what does the one mean? Could you possibly just be too close?

Proximities have a hysteresis so at a particular distance the output will be on and off at a larger distance, however, as I said before the sensing field is cone shaped so a very small target will need to have the PX closer and will trigger at different points across it's face. There could be many reasons for the px to "stop working" 1. The hysteresis is changing slightly over time i.e. heat generated through the moving components causing the sensor or target to "move" slightly, 2. the difference of the on/off sensing hysteresis is so small that any changes in temperature or position is enough to "lock" the sensor on if moved back away from target it de-energises but will not trigger again until moved back into correct position. We had systems with a chain conveyor that had dogs on the chain to position trays, on the return leg the chain was kept in position by a S/S guide, the dogs sat either side bet were only 2mm from this guide, the index prox detected these dogs but as they were so close to the guide only one type/make of sensor would work consistently. Even then the position of the sensor was very critical, wear on the chain dogs or slight damage could cause either latch or miss of the dogs.