Fastest Pellet Sorter in the Works

[Raider03]

For those who know nothing about how to build something like this but find it extremely interesting (like me ) a tutorial:

Arduino with Load Cell and HX711 Amplifier (Digital Scale) | Random Nerd Tutorials

Create a digital scale with the Arduino using a load cell and the HX711 amplifier. First, you'll learn how to wire the load cell and the HX711 amplifier to the Arduino to build a scale.

randomnerdtutorials.com

Mlc608l High Accuracy 1mg Accuracy Electronic Balance Scale Micro Weight Sensor Load Cell 20g - Buy Micro Load Cell micro Weight Sensor load Cell 20g jewelry Scale Load Cell jewelry Scale Weight Sensor 1mg Accuracy Load Cell Product on Alibaba.com

Mlc608l High Accuracy 1mg Accuracy Electronic Balance Scale Micro Weight Sensor Load Cell 20g - Buy Micro Load Cell micro Weight Sensor load Cell 20g jewelry Scale Load Cell jewelry Scale Weight Sensor 1mg Accuracy Load Cell Product on Alibaba.com

www.alibaba.com

 

[MartinT]

That machine is so great, awesome make!

 

[bigHUN]

Oh nooo, arduino again.
This is a forth time in last ten plus years I need to re-learn the programming for a single task. Takes time to refresh and fades out quickly.
Last project was the FX radar display during the vaccine time.

 

[RScott]

I can make an auto-feeder if needed but head sizing is out of my league ATM. AFAIK head sizing would involve some swaging and is beyond the scope of this machine. Maybe in the future?

I think the poster was asking about head size measurement and sorting, as the Pelletgage is used for. I would love one of your devices for weighing (if it can reliably differentiate pellets withing 0.1 grain of each other), but weighing is much faster than determining which head size hole a pellet fits through on a Pelletgage. So if you could also come up with an automated pellet head size sorter, you'll have a few thousand benchrest and field target shooters begging for your devices.

Even with "just" automated pellet weight sorting, you have a great invention!!

 

[Kabayan]

I think the poster was asking about head size measurement and sorting, as the Pelletgage is used for. I would love one of your devices for weighing (if it can reliably differentiate pellets withing 0.1 grain of each other), but weighing is much faster than determining which head size hole a pellet fits through on a Pelletgage. So if you could also come up with an automated pellet head size sorter, you'll have a few thousand benchrest and field target shooters begging for your devices.

Even with "just" automated pellet weight sorting, you have a great invention!!

I am still struggling with "noise" at the moment, despite having a low noise amplifier installed.

The problem is really this: I am trying to get better than 0.01 grain resolution from a load cell with 1,543 grains full-scale rating, when weighing a pellet that's only 20 grains in weight. It's not gonna happen because I am just slightly above the noise floor with that pellet. I need better resolution.

To get the resolution I need, I will install a more sensitive load cell, one with 30 grains full-scale rating. This is on order and will take a few weeks to get here. I will post results as soon as I try it out.

Thanks for the heads-up on the Pelletgage. Now I understand what the requirement is: How to automate the measurement of a pellet's head diameter. An Optical Comparator can do the job but that piece of hardware is VERY EXPENSIVE. Plus the automation of singulation and the subsequent sorting.......let's stick to weight measurement for now. (BTW, swaging each pellet to size is far more easier to do than mechanical sorting using a gauge)

 

[bigHUN]

.......let's stick to weight measurement for now..... swaging each pellet to size is far more easier to do than mechanical sorting using a gauge...

This
The problem with the Pelletgauge is that the holes are "burn-cut" with a lazer, and the holes ID may be accurate enough but burrs inside walls catching the lead here and there.
I gave up sorting pellets by head size because of a lot of time consuming, but I am doing head sizing instead.

 

[Spit-n-Lead]

The auto sorter is a great Arduino task for sure!
I am really curious about this process. I just started weighing pellets, and to do so, picked up a Tree TSC 123 multi-unit balance. The Tree TSC is accurate to .001 grams with a resolution of 0.02 grains when in "grains" mode. The scale takes a few seconds to zero after a measurements, so even if I could load pellets as fast as I could, the scale cannot zero that quickly without introducing error.

What I would like to know is how you all are setting the ranges for go, no-go pellets?

I started with this: JMB .177 10.34gr, 500 pellet cans. My sort groups are under 10.30gr, 10.32-10.34gr, 10.36-10.38gr, 10.40-10.42gr, 10.44-10.46gr, and 10.48 and up. After my first full can I can tell you that only 6% were in the 10.32-10.34gr range, 27% were 10.36-10.38, and almost 41% were in 10.40-10.42gr. (notable mention: 10.44-10.46gr, 20%). My problem with the sorter is more that the strain gauge must be a capable of measuring from 0 to 100 grains or so, and do so with a very rapid measure and re-zero cycle. Can you get this system to do that? Have you checked pellets after sorting on a balance to see how well it's weighing?

 

[Kabayan]

The scale takes a few seconds to zero ........

the scale is picking up ambient noise and is using averaging to get rid of it. This is the price we pay for the sensitivity that we require. An analytical beam balance (mechanical) will likewise need some time to settle. (Somebody might come up with a clever work-around in the near future?)

This project has been shelved for the meantime. Until I can get my hands on a loadcell + amplifier that can resolve 0.05 gr at 100gr full-scale, and be reasonably fast.

 

[Spit-n-Lead]

Cool. It takes some time just to work through components. Great working model and I think it's very doable.

 

[Longerun]

the scale is picking up ambient noise and is using averaging to get rid of it. This is the price we pay for the sensitivity that we require. An analytical beam balance (mechanical) will likewise need some time to settle. (Somebody might come up with a clever work-around in the near future?)

This project has been shelved for the meantime. Until I can get my hands on a loadcell + amplifier that can resolve 0.05 gr at 100gr full-scale, and be reasonably fast.

You could add a mechanical component like a lever to amplify the force on the load cell to make use of the full resolution.

 

[Spit-n-Lead]

the scale is picking up ambient noise and is using averaging to get rid of it. This is the price we pay for the sensitivity that we require. An analytical beam balance (mechanical) will likewise need some time to settle. (Somebody might come up with a clever work-around in the near future?)

This project has been shelved for the meantime. Until I can get my hands on a loadcell + amplifier that can resolve 0.05 gr at 100gr full-scale, and be reasonably fast.

You got me thinking - are there piezoelectric sensors available that you could try? The strain gauge seems like a good option, but you need a really low mV driver to keep noise out - that's tough.

 

[beerthief]

Sorting 22cal pellets.

The LCD display shows the weight followed by the Bin # and the pellet count for that bin.

"Typical Weight" and "Tolerance" can also be seen. Both are user inputs and the test limits are generated from these 2 parameters.

Pellets are actually being sorted into bins, based on the measured weight (in grains).

At this speed the output can be up to about 4k pellets per hour.

I volunteer to test this machine for a 6 month period Free of any payment testing for longevity of parts .

Maybe I'm making a mistake, but 0.1% of 18 grains is 0.018 grains, rounded up to 0.02 grains. And 0.1% was the tolerance claim.
1% tolerance, plus or minus, is already a 2% range. That would actually be too much for an H&N production batch.

I think he has a tremendous start on a very useful machine . Maybe reserved for absolute top competition pellets at first