next topic »

[Reply #495 on:]

If you do multiple runs do you get the same readings each time for each carrier?

That is a much larger variation than expected due to construction tolerences, can you think of any reason why the range is so large? Have you tried to measure the carriers with a conventional scale to see if the problem is with the way the load cell is measuring the weight?

[Reply #496 on:]

If you do multiple runs do you get the same readings each time for each carrier?

Unfortunately, it does not show the same readings.
I also thought the problem was in the bad weight sensor installation so, I tried to modify the way I installed it in all possible ways, such as adjusting the height, angle, etc...
But, after the numerous testing I could find the cause.
I've been keep rotating the conveyor with the carriers attached on it,
The weight of the carriers all vary because the guide piston mechanism I came up with has significant friction, causing the bad readings. As I explained before, I used this guide piston under each carrier. Through below link, you can see how the mechanism is constructed.
http://www.flickr.com/photos/34955071@N02/?saved=1
The guide piston will move only up and down. but the metal angle(yellow thing) in the picture makes friction with the PVC plate which is attached horizontally to guide piston(This prevents the guide piston not to rotate while attached to the carrier, otherwise, the carriers would spin during the conveyor moving)
If I increase the gap between the metal angle and the PVC plate, the carrier will tilt and this causes the carrier to contact with adjacent carrier. The guide piston also makes fricktion internally inside the cylinder. This friction varies.
 I couldn't think of any way to simplify free up and down motion of the guide piston. I couldn't get much help from engineering forums on the internet. Getting the right time weight data from the weight sensor using the photo detector works great as you've helped me. But, with this guide piston mechanism, the weight readings are bad.
oh..by the way,  the weight sensor seem to be working ok. because I am checking with calibrated 100g, 200g, 500g weight object. I bought this when buying the weight sensor.

Below weight readings(my memo) are from the first time, second, and third time test trial.
This will easy to compare. The values are from the serial monitor.

Code:

           #1      #2     #3
carrier 1      93      108      142
carrier 2      210      197      176
carrier 3      338      302      202
carrier 4      152      157      116
carrier 5      144      145      171
carrier 6      160      123      161
carrier 7      76      175      115
carrier 8      202      198      189
carrier 9      178      189      187
carrier 10      195      194      199
carrier 11      139      144      151
carrier 12      221      199      236
carrier 13      139      150      164
carrier 14      175      177      169
carrier 15      209      199      164
carrier 16      143      158      161
carrier 17      193      197      181
carrier 18      167      127      107
carrier 19      187      171      152
carrier 20      146      135      125
carrier 21      170      169      198
carrier 22      189      162      165
carrier 23      197      216      168
carrier 24      228      206      313
carrier 25      116      129      157
carrier 26      161      146      104
carrier 27      464      446      247
carrier 28      215      215      142
carrier 29      179      258      247
carrier 30      141      144      180
carrier 31      177      151      149
carrier 32      230      229      213
carrier 33      218      218      223
carrier 34      347      248      66
carrier 35      208      213      151
carrier 36      177      196      214
carrier 37      173      155      150
carrier 38      155      176      153
carrier 39      184      196      248

[Reply #497 on:]

It does look like the variation in weights is due to friction affecting the readings.

Have you done any investigation on how commercial sorters solve this problem?

Is it acceptable to stop the belt for a second each time a carrier is over the sensor?

[Reply #498 on:]

Yes, I've seen it long before.
There are many farm houses posessing sorting machines around my home because my home is in farmming area where agricultural industry is very well advanced. I grow some fruits at the back yard, but not much.
Anyway, what I've seen was a plastic moulding product and there is a free arm that moves up and down freely. The arm has a very small contacting point(5mm diameter) to the weight sensor. I think this is to avoid large friction area. Literally, the machine is equiped with expensive, precisely processed materials. If I were able to use such materials, I would definitely follow their method but as I explained, the design of the carrier is based on the materials that I have, so...

Is it acceptable to stop the belt for a second each time a carrier is over the sensor?

This is a great idea. I believe we once had a discussion over this issue.
The speed of the conveyor doesn't matter. It can run as slow as it can be according to the motor performance.
The motor I am using is a broken one but I repaired it for the conveyor but the motor does not fit for the conveyor. It is just too fast.
That's why I am rotating the conveyor by my hand at the moment. It is very hard.
Also, this motor runs only when full AC220 supply is reached to it.
So, I think using just DC variable speed motor would be suitable for the coveyor, but the price is killing me.
It costs almost 300 US dollars(including an inverter).
But, I will get one anyway. I got this far. There is no stop.
Will using DC motor do to stop the belt for a second each time a carrier is over the sensor? I believe this can be worked with arduino using a relay.

[Reply #499 on:]

A dc motor should work, perhaps you could try to use the DC motor and gearbox from a battery operated power drill. These are not so expensive and you can often find used ones very cheap as the batteries usually die while there is still plenty of life left in the motor.

Try to find one that you can power from your existing high current power supply. You can test with a relay but a high power transistor or FET will be more reliable in the long term.

[Reply #500 on:]

Thanks a lot.
I will look for them.
Have a good day.

[Reply #501 on:]

Hi, mem. how are you?
To improve the sensitivity of the loads on the carriers, I came up with this kind of free up and down mecanism.
http://www.flickr.com/photos/34955071@N02/4036159749/sizes/l/
As you can see, I used two pieces of pipes. Those two pipes have screws at the end of its side. I used this to fix it to the carrier base.
So, basically, I used one for the carrier and the other for the chain.
In between them, I used some kind of arms. I don't know how to call this. This is one of accessaries of aluminum profile.
I secured this mechanism under the carrier like this.
http://www.flickr.com/photos/34955071@N02/4036159755/sizes/l/
http://www.flickr.com/photos/34955071@N02/4036159751/sizes/l/
This has increased the sensitivity dramatically. The tolerence of the weight value is now only + or - 1g ~1.5g.
I will keep testing and continue to comming back.

[Reply #502 on:]

Sounds like you are making great progress. If its not too much trouble it would help me visualize what you have if you could post a short video on youtube.

[Reply #503 on:]

Hi, mem.
oh.. I am sorry, I just checked your last post.
I thought the recent free up and down mechanism is the last thing that I have to modify. But it is not. I just encountered with more mechanical problems. At first, I attached one carrier with the recent up and down mechanism attached to it. when I test, the reading was perfect. But, when I attached 10 of the carrier attached with the mechanism, I could find the problem.
The chain leans side to side in an irregular pattern. On some certain chain link, it does not lean. but strangely, most of the chain link leans to side when the chain slides on the chain guide.
To cure it, I increased the height of chain guide recently. The chain guide is composed of still bars(12mm x 12mm x 300mm). This is used to sustain the alignment of the chain from one sprocket to the other sprocket. That leaning causes the free up and down mechanism to lean, as a result, causing friction in the joint of the mechanism because the joint stucks when the chain leans.
I realized that this happens because there is no mechanism for chain tension!!!
What a genius. I built a conveyor without a mechanism for chain tension adjustment. I was angry with me so much that I felt dizzy when I discovered this problem.
Anyway, that's why I tried to increase the height of chain guide.
So, I took everything apart from the conveyor except the structure itself in order to move the chain guide.
But, adjusting the height of chain guide is not enough. The space is limited. I pulled up the chain guide as much as I can.
but this also causes another error.
The two sprockets which are used to drive the chain is designed for the original height of the chain guide. So, if I increase the height of the chain guide, the carrier goes to the sprocket will be crushed because, the spinning range of the sprocket. When the chain guide is in the original height which was lower than the teeth of sprocket, carrier actually goes up a little bit without sliding on the rollers and spins with the sprocket.
But, now, I pulled up the height of the chain guide higher than the teeth of the sprocket, therefore, the carrier still in contact with the roller when it reached to the sprocket for the rotation.

I don't have camera with me right now, but this drawing will show it.
The sprocket is installed like this.
http://www.flickr.com/photos/34955071@N02/4057559361/sizes/o/
So, due to that eage of the conveyor frame the carrier can not rotate with the sprocket while making contact with rollers.

[Reply #504 on:]

Hi, mem. I prepared some pictures.
I somehow managed to increase the tension of the chain by adjusting the position of the sprocket.
This time, the circled area does not move. Before it was wobbling due to the lack of chain tension.
http://www.flickr.com/photos/34955071@N02/4068506728/
Now, the whole chain leans to the side.
http://www.flickr.com/photos/34955071@N02/?saved=1
I used just 5mm bolt to drag the carrier for a test to see how it seriously leans.
The conveyor seems to need more tension.