Hello everyone,
As we have eluded to a few times, I've been working on a feeder system for us. Anthony Webb has been invaluable in printing/ testing/ and providing feedback/ suggestions to accelerate the design as well as Mark Harris and Jason V. for their feedback and suggestions!
The feeder system itself is based on dual lane module blocks. For starters this is 2 8mm tape lanes. The design is 3D printable but only with an EXCEPTIONALLY good printer. I am going to experiment with CNC milling some bodies in the future as it takes 10 hours to print a dual lane assembly.
Here is the model, the turn down at the front is for testing, this is still in the development phase.
The module features a single circuit board to run both lanes. The board has a ARM cortex M0 processor, twin light gates, RS485 transceiver, two toshiba motor drivers that drive 4 motors total, 4 buttons, grove connectors for daisy chaining (cheap and prevalent?). I have just made R2 of the board that now has additional connectors to allow encoder inputs. The polulu motors I am using have add on encoder boards that may turn out to be a useful option, so these new boards will be able to advance the tape by opto on the gear teeth or by a motor encoder. More to come once I resume testing.
In addition to the feeder driver boards (DRVR) I have developed a 485 transceiver board (XCVR) to interface with the computer and also integrated a high quality SMPS to power the feeder network through the grove connectors.
Here are the boards:
Here are pictures of the actual physical boards/ feeders:
(Feeders printed in clear so you can actually see though parts of it)
V1 feeder boards:
V2 feeder boards:
So the feeders are "working", in the sense that you can send them messages over RS485 and they will advance by the number of requested holes.
Here is manual feeding:
The thing we are working on is the accuracy. As you can see in the following video it does't repeat the move very well. The V1 boards had a mistake and we were not able to control the PWM of the main motor (there is a pot on the board to control the PWM of the cover tape motor). We are expecting that if we can slow down the motor the accuracy should improve. Anyways, you can see what I'm talking about here in Anthony's video (which is triggering the feeder over RS485 through the XVCR board):
I am also going to experiment with water jet cut stainless drive gears that I had made:
So there you have it, sorry I have been meaning to share this with everyone a long time ago but I've been super busy, I'm even rushing this post but I'm hoping you guys find it interesting!!
Peter.