AndrewC63

I've made several tweaks to the design and did away with the flipper sorter mechanism in favor of one that can do 6 positions. I've run 1200 pcs manually working out bugs and issues. Still trying to get someone to help me make the missing file. I found out the person I hired for a substantial amount of money only changed 3 lines of code, and I still don't have an automatic sorting machine. I also discovered that the semi auto mode for just taking pictures of sorted brass does not work either there is an issue with it trying to determine which bin to put the brass in. Will take some more pictures tomorrow after I make a few more revisions. This spiral chute method is by far the best one to get the cases flipped. I've been feeding it with my dillion case feeder it works well. I'm going to try and figure out a plywood box solution for mounting it. I'm using those red party cups as sorting bins temporarily and had 4 of the 6 positions working tonight manually.

After playing around with the turn table design I went back to the original design from the initial 223 files. I was holding off because I figured it would be hard to draw the spiral... it turned out very easy. It is the best design to be honest the least complicated the least moving parts etc. I'm working on probably an over complicated way to clamp the case for the camera at the end then it should probably work

Here's what I've came up with to take the pictures of pistol calibers I've tried three different ways based on my design skills and the fact I've bought these servos for the other design this is what I've came up with. The wheel needs a little bump on it so if the case does not drop it doesn't jam when it goes back just to lift the waiting case up a bit before the other goes back under to smooth things out and the drop hole is not in the exact right spot. If I can manage to get my idea of the rotating case flipper I should have a manual sorting machine where i have to make the distinction between the case hit a number on the keyboard and it will do the rest as we are missing the file to go to the automatic feature its not included I'm going to try it like that manually it will probably be faster and less stressful then doing it by hand. the goal was to get it auto but the person I had get the software going this much wants more money to write the file for the auto part.

I've made some massive progress, we have hit another roadblock again. I have the manual sorting feature working. We can manually sort cases potentially all the systems work for that but I've not assembled everything into a complete machined yet. We also got the Arduino moving the cases and flap for the sorter. We tried to run the automatic mode and ran into a roadblock missing files saved_model.pb and saved_model.pbtxt to move forward to the automatic sorting process. I've already invested $170 into having someone look at the code and get this far.

We'll I'm happy to say the software was delivered on the 11th but I still had some issues today we worked the bugs out and the image saving feature before the auto sorting is finally working. Hopefully I have some time tonight to try the auto sorting part of the software and see if there are any errors or bugs that need to be worked out.

How do they need to be sorted... it would be quite simple to do it by color. On the 11th my software will be delivered I've been working a bit on a different type of machine I almost got the concept going last night minus the software portion it would be simple enough to super size it for shotgun hulls. Once the software portion works all it does is match pictures so sorting by headstamp would be no different.

Working on a new design, to make one with interchangeable wheels. My idea is to add this inline with the Dillon case feeder, this will flip all cases upside down if it works then I will working on a mechanism to advance the case to the station to view it thru the camera. Should be able to make it fit all pistol calibers and 223. Might increase the size of the wheel to be able to do 308 as well but I don't do much 308 right now. I'm printing the outer housing tonight and will test it manually by hand if it works I'll add a motor and microswitch.

Well I broke down and contracted the machine software out hopefully in a week or two I should have the working software needed to make this work.

Anyone... I could really use this machine.

I'm trying to hire it out on fiver to get a solution to the code not working I can't keep waiting I have so much on the go I need to finish this and get a working unit or abandon it for now.

I gave an offering up to the brass gods. I hand sorted another bucket tonight. I could really use the working file to run the machine.

I could really use some advice on how to get the software going. I'm basically stuck here. I've resorted to hand sorting again please help it's boring as hell. I tried a different approach running spyder thru anaconda, it works slightly different but now opens an extra window that shows the picture but it just sits there with the little blue thinking circle going for ever and wont update to the next picture. I also tried matching the paths but it seems to not want to save. I really want to get this going i see the potential just don't know what to do.

Laptop

Says to use spyder in the instructions, the Arduino portion is on a Nano.

Could you look at the code, The issue I'm having is itis not saving the images. Otherwise I think I'd be up and running the other issue is the python program doesn't seem to be communicating with the Arduino the way I think it should. Tonight I'm going back to hand sorting I cant wait anymore I was hoping to have it up and running but the code is over my head. I'd hold off on printing as the files might change when Hans gets some time to upload. If you cant get the software portion running the mechanical machine parts are pretty much useless. I'm so close but so far because I've tried using my limited knowledge to fix the program part but nothing seems to work because I really don't understand what I'm doing just trying things. If you want to remote teamviewer into my computer with it running I can show you what's happening.

Any chance you could send me the updated Arduino and python code. I'm pretty sure I can figure out the rest.

HANS! Your awesome! I have someone on a python group looking at the code for me if you could help me the code running I can sort out the mechanical stuff. Like you I'm a tinker and also am going in 10 different directions with about 20 things on the go. Highlight of my day right here.

Hey GMM50, There already are these products available on the market all the way up to what the us military and commercial manufactures use. This thread is for us that don't want to paying thousands of dollars on machines and are DIY if your thinking of developing these two machines there is a limited market for it. As for your speeds most brass processors are running between 1200-3500 pcs per hour. That would be your mark 7 auto drive presses. To jump into something a bit more commercial like a camdex your at around 4400pcs per hour. My budget is as little as possible to get it done currently I hand sort but its getting boring and if i can automate it can do other things instead. On a side not do you know linux and python coding I'm stuck with the code provided not working for me or else I'd have a working machine up and running by now.

So what I still need to get his going is to figure out how to get the files to save. I've connected the arduino nano to the servos the best I can figure out his pin out diagram makes no sense. The case shuttle servo just goes back and forth and seems like it has no communication to the python program, I have the sevo connected to ground 5v power and digital input 9 as for the sorter servo I can't seem to figure out what pin the digital input goes to I'm assuming all powers and grounds are together. The case shuttle needs a linkage I tried last night with a 3" one way to long unless the servo degrees is adjusted way to much in the arduino program. I could really use some pointers from Hansimania seems to have a working unit I've got a 5 gallon bucket waiting to test this out.

So close, just have to figure out how to get them to save I changed the paths, tried to make the paths on the computer the same as the original code nothing seems to work.

Solved it. I've got to the part where the unit turns on the camera light and takes pictures but I can seem to figure out where or if its saving them. its not showing me a picture on the console. It looks like it might be working potentially.

Anyone know how to resolve this error I've been thru a bunch of them but I know literally zero about python and Linux just what I can search online.

I originally thought that is how it works but it won't needs to be assembled in the way I did it. I was up to 3:30 in the morning trying to get the software side loaded and working on my laptop ran into some errors hopefully that part will be working soon. The hose adapter for me is useless I'm in canada and I highly doubt any store here is going to have that slinky hose thing I'm either going to go with some pcv pipe or something else I've started to print a case feeder but 20 hours in I realized its probably only for pistol brass This is for sorting 223 only there is no easy way to adapt it for other calibers as far as I know. It would probably work for anything on a 223 parent case like 300 acc blackout or maybe things like 222 17 remington etc but I doubt people are sorting hundreds of thousands of those.

More progress have to get some fasteners for the servos tomorrow. The sorter mechanism is self explanatory looks like it has three positions down and left and right

So I have most of the part printed off I have the sorter scoop left and then I'm finished the camera servos and nano should be here tomorrow. I understand how it works but there are a few parts I'm not to sure of it looks like there might have been revisions to the design an then forgot about attaching the parts together. So the 223 twist rail has no mounting to the mechanism that advances the cases and case clamp I understand what its there for but don't understand how it attaches. I've already been thinking about a better way once I get this up and running and see if I can use the software provided and the best part is my idea has already been made by someone else so the concept is sound. What my plan is to make about a 8" circular pipe that has a hole in the top to drop cases in base down then inside the circular pipe there will be a wheel that captures the case and turns down 180 to orient the case upside down inside the middle the camera will look at it when the case meets the 180 degrees it will drop and then be sorted. The center wheels should be removeable and interchange able for all pistol calibers and id like to be able to do 223 and 308. It's pretty crazy that when you think about something and then stumble on it looking for videos on how this is assembled. Better way to sort different caliber brass . I'm very excited to get this going and try it out I've got a 5 gallon bucket of 223 on the way to test it out I've been hand sorting one every week. The sorter parts are pretty much self explanatory how it works and the parts attach with pins on the other parts but I'm not sure what holds everything together glue? Since there aren't to many pictures and this is pretty specialized project mostly for brass processors I've included a bunch from my parts the 223 twister I woke up to it growing "hair" it had a bunch of strings on it like hair probably from failed supports everything printed fairly well. I'm not a total 3d printing noob but I'm not an expert either. I have to re level my bed as moved the printer and noticed the skirt on the sorter hose adapters isn't adhered correctly