Univseral Auto-Trickler

[Fliers]

Adam's auto-trickler/auto-thrower is pretty cool, but it's limited to just a few high-end scales.

The Dandy trickler works on beam-scales only and uses light sensor. https://dandyproductsllc.com/products/dandy-auto-trickler

I'm tempted to just by Adam's solution and be done with it, but figure a universal solution for beams and any digital scale would be cool, fun to create and could potentially help out other shooters that can't/won't dump $750+ on Adam's setup.

It could also bring back a few RCBS or others were they no longer trickle correctly, but the scales still work.

Just looking for feedback to gauge interest in this type of solution. Note, it would be slower than Adam's, but perhaps not too much slower.

 

[mousegunner]

I have the Omega trickler only (same thing) and wondering if they could add the sensor to it. I'll have to ask. I like my Omega and it has made things go just a little faster when using my 505.

 

[Hari]

I've bought a Kern PCB100-3 back in the days, which is relatively cheap (was around 240 EUR a few years ago). The RS232 port on the back always tempted me, and one day I did write some esp8266 micro controller code to control a vibration motor from an old phone, shaking out kernels from a pipe (made from an old antenna). A bit like the omega powder trickler:

In the end it worked only ok'ish, relatively slow and +/- 0.12gr at worst, +/- 0.02 or so at best. It was very slow as I always had to pause trickling to get a stable reading from the sale, and then iterate. Unstable values had too much delay to allow a reliable control loop. You can see how long it takes to dispense 5gr in these videos:

It later even got a nice web interface for settings, but I was unhappy with the powder throughput and hence got the autotrickler hardware parts from Adam. That finally revealed the bottleneck being the scale, it takes quite a while to get stable, and has a lot of delay in the readings. Also - even having an option to turn off averaging - it didn't detect small additional amounts properly. To drive the stepper motors in Adam's hardware, I swapped the esp8266 with a arduino mega, as it has more hardware serial ports. It uses trinamic stepper drivers which can be controlled with serial commands and are completely silent (no whining anymore from the motors). This is a best case run with the Kern, which took me some times to record to have a "show off" video:

Now compare that to the repeating performance of the A&D FX120i:

So yeah, the FX120i is not the cheapest scale on the market, but you get what you pay for. Anyway, if you want to give a cheaper scale a try, you can look at the current Powder'O'Matic code, I've uploaded all code and designs to GitHub.

Anyway, if you want to give a cheaper scale a try, you can look at the current Powder'O'Matic code, I've uploaded all code and designs to GitHub.

 

[Hari]

That's how it currently looks hardware-wise:

The code is on GitHub (search for Powder'O'Matic, I somehow can't link that here).
If you already have an RCBS trickler, you can fit some brass bearings and use a 3d printed NEMA motor adapter (also on GitHub):

Long story short, the money is well spent on Adam's solution if you want something plug and play that works fast and reliable. Go for that if you want a great solution for a fair price.
If you're up for some DIY, maybe the pointers I gave will help. Compared to Adam's PnP solution, Powder'O'Matic Step has a few benefits:
- silent operation of motors
- beep when ready, error beep on overshoot
- switch to toggle automatic mode
- rotary encoder and display to toggle settings and to enter the target value

br Hari

 

[earlcurtis67]

I bought the Dandy rwo speed trickier because the manual ( Hornady) trickler would tend to over charge more times than not. I did not buy the sensing beam attachment and auto charge, only the two speed trickler. Ive been using my 10-10 long enough as to throw and weigh charges fairly quicky so i didnt see the need for the auto throw and beam.
I unboxed the Dandy this morning to check it, looks like its gonna do exactly what i need. Its easy to adjust and the high and low speeds work great, the low speed can drop a single kernnel of RL23 no problem. You must have the tube fairly close to the pan though to prevent kernnels from bouncing out of the pan. I’ll try to do some reloading sometime this week and get a better idea of just how well it works.

 

[Fliers]

Nice videos. Yeah, clearly digital scales will often have the issue of detecting trickles, but there are ways to help it some, including support for 'tap' resets, but I do see people trickling on gempro's, etc. This would work much better I think with beam scales.
Either way, it's an automated replacement for whatever trickle method people currently use. It should be fairly cheap as well, all in, parts around $35 I'm guessing. Less in volume.

 

[Fliers]

I'll check out your code, but what was the trick to reduce the hum noise? I've done a stepper motor control before, but even on low revolutions with ramp up/down, it still hummed.
I'll have to check which IC I was using, may be related to that or the library I was using. Don't recall if it was emulating digital changes with PWM. It's been awhile

 

[Hari]

@Fliers the trick is to use stepper drivers from trinamic. I'm using the TMC2208 as they feature serial port control. The trinamic drivers have a feature called "StealthChop": https://www.trinamic.com/technology/adv-technologies/stealthchop/

 

[Ammolytics]

Hey folks, a friend mentioned I should share my most recent project on this thread since you might find it interesting.
http://forum.accurateshooter.com/threads/open-trickler-the-diy-smart-powder-trickler.3978138/

All of the software is fully open source, which you can see here:
https://github.com/ammolytics/projects/tree/master/trickler