Fast Pulse Counter, Micro850

This might give you some information although it only mentions using quadrature i.e. two inputs.

https://twcontrols.com/lessons/alle...eed-counter-configuration-for-encoders-in-ccw

I have set up now as an HSC_Mode0 Instruction.
It seems to work well with Input 0 exactly as I wanted it.

I have a question about filtering Inputs though.

In my test set up, once the HSC instruction is working, I tried to add filtering. If I set it to the lowest level of 5uF the input will increment many times upon turning the input on once. Its just a push button to experiment with, but a single press gives me 20 counts on the HSC. If I set it to 1ms it seems to catch the pulses as fast and accurate as I can with simply using a push button.

I'm just trying to understand the filtering and how to use it to my benefit without it adding counts to my value.

The manual says its the time it takes the controller to detect the change of state from an external input voltage that reaches an on or off state. Its a bit confusing to see it make several counts with a lower filter that all. How do you normally figure a starting point for the Input filter?

Thanks

Yeah never use push buttons it needs to be an electronic generated pulse as contacts tend to give quite high pulses on press & release.
If you can gete a proximity connected & just use something to energise it like a screwdriver blade. Have you checked you can use the other input ?

I just watched this video.
I had seen it in the past but was a good refresher.
He didn't talk a lot about the filtering but actually mentioned doing another video specifically about that topic.

Thanks for this!

parky said:

Yeah never use push buttons it needs to be an electronic generated pulse as contacts tend to give quite high pulses on press & release.
If you can gete a proximity connected & just use something to energise it like a screwdriver blade. Have you checked you can use the other input ?

Click to expand...

Yes, the other input works perfectly!

Ill give this a try with he proxy to see if it allows me to run a smaller input filter without gaining counts.

Thanks

Ok. Right you are.
The prox switch allows me to flag once, extended or quickly, and see one count on my HSC with my Input filter set to the finest 5uF.

I think I have something I can work with here now. Confident that I will catch all my pulses.

I appreciate the responses, the conversation and the education.

The only reason for false triggers using a prox will be if the notches are not clear & you get double pulses I had this once detecting bolts on a shaft, we ended up rounding them off.
This was due to the bolt heads having writing on them enough to give occasional double reads on the bolt heads, increasing the filter a bit might overcome this.

There is one other possiblity. If your "dimple" has a rounded profile you could be looking at a hysteresis problem. According to SICK's website your selected sensor has a hysteresis value of between 5-15% and a switching value of 4Khz not 10Khz as you stated in your intial post.

If you find that you're still having problems even after switching to HSC function, you might want to mill slots into your tubing stock to give yourself a more reliable state change on the sensor.

Hey folks.
I wanted to just give thanks for the help and suggestions here. The machine works amazing now! I'm pumped. It took a little messing with the code to reset the counter when I wanted it to, but now this thing is cutting logs to within 25mm every time! Before this change, I was getting anywhere from 4 "to 10 " of inaccuracy, not repeatable and completely erratic behavior as the whole machine depends on keeping position.
I spent a few hours yesterday processing and it never missed a beat!

I have a lot of hours into this machine. Its been a work in progress for a couple years now, on the side..... Its been completely overhauled. Solenoids, wiring, hydraulic lines, sensors, computer and the even the frame has been completely disassembled and a lot of it re-fabricated. Furthermore, I have never run a harvester before this, so I had to trial and error my controls and my whole theory of operation and very time consuming as well, since its sitting back in the woods and laptop battery time was an issue among other challenges. I watched a lot of videos of other harvesters running just to get my sequencing correct. 40 years of logging experience from my old man helped a lot too!

Its very gratifying to see something Iv been stumbling on for so long finally work flawlessly. Not to mention, some of the other folks around that have logging processors, take one look and figure I'm crazy for not just going and giving an OEM $$$K's for a new proprietary control system! I'm in it now (control system) for only a Micro850 and a 7" Weintek...


Thankyou!

When do we get to see pictures?

All I have here is a picture of it just hanging in front of the garage before the massive overhaul.
I do intend on getting some videos of it in operation and getting them on you tube. Iv had a few folks ask me for some videos but have yet to take one of it since its been working properly.

Hopefully on the weekend Ill have a chance to get back to it again while its running. I need to be there when the operator is to take one. By myself its impossible.

This picture is small, but this is a big machine. From the ground its easily 6' tall and requires a ladder to reach the coupling. Its mounted on a 15 ton Daewoo Excavator.
It handles up to a 20" tree across the bottom.

If you want to see a couple other ones in operation then Google Arbro1000 and MC Power Equipment. Also watch the Arbro400 as its the same, just smaller frame and fewer solenoid valves.

:site:

I guess the picture is a lot larger once posted than I thought......

https://www.youtube.com/watch?v=yh89snfBRFs

Wow, my jaw is on the floor.

I love the piece of line holding up the extra bight of hose or wire.


So was the final fix going to the HSC, getting the filtering right, and having the HSC internal count stop the motion (not in a PLC scan)?


And the dimples did not need to be munged at all?

Wow.

drbitboy said:

https://www.youtube.com/watch?v=yh89snfBRFs

Wow, my jaw is on the floor.

I love the piece of line holding up the extra bight of hose or wire.


So was the final fix going to the HSC, getting the filtering right, and having the HSC internal count stop the motion (not in a PLC scan)?


And the dimples did not need to be munged at all?

Wow.

Click to expand...

This video is exactly like my machine. The only difference is my excavator is orange! They really are cool machines. It had my interest, really is the only reason I got into them. Aside form the fact that we have been cutting with power saws to do our harvesting till now. I have a few in the neighborhood now that guys get me to help troubleshoot with. None are the 1000 model though, usually smaller ones. Normally its just failed sensors or broken wire and sometimes solenoid coils fry, but usually these things are bullet proof.

The dimples were not touched at all. In my original post, I stated that there was 14mm between the dimples, that was a typo, there is only 6mm. So On for 6, off for 14mm. It works impressively now!

"I love the piece of line holding up the extra bight of hose or wire."
That is a bungee cord that keeps the set of lines attached to the stroke head pulled back behind the machine. If they get in front of the machine, the tree limbs will brake them off. We replace a lot of lines!

drbitboy said:

So was the final fix going to the HSC, getting the filtering right, and having the HSC internal count stop the motion (not in a PLC scan)?

Click to expand...

This is all true, except for the last part. I'm simply moving the value of the HSC counter to a routine that uses it to make a value in MM. That calculated value is what I use to measure. It adds every time the stroke IN happens and adds to the previous total until the target length is reached.
Of course it only accumulates in auto feeding mode, as sometimes with larger trees you will need to take over from the auto sequence and stroke in and out several times to smash off larger limbs, then continue auto sequencing where you left off. So long as the bottom grippers don't move on the log while manually stroking, it will keep its current log length accurately.

I suppose by doing as you suggest, using the internal count to stop motion on the final approach to the target could make it a bit shaper in design. But I'm very happy with it at this point.

Actually.
If you look close in this video, you will see between the top and bottom set of grippers, there is a small wheel that is rolling along the log. This is a newer version that is using an encoder to measure the length. The bonus with the encoder riding on the log, is that if the log slips, you don't accumulate length. On my set up, if the log slips while delimbing, you need to start again or it will be too short as the machine will not notice that it slips. That is up to the operator. Mind you, if the grippers are good and sharp, slipping is not really a problem.

PLC Pie Guy said:

I suppose by doing as you suggest, using the internal count to stop motion on the final approach to the target could make it a bit shaper in design. But I'm very happy with it at this point.

Click to expand...

Or, if the PLC could drop the speed when it gets close to the target, then no counts would be missed in the main program. But like you say, you got it where you want it.


Thanks!