My own version of a pick and place machine. Made from mechanical standard and 3D printed parts.
Already have an account? Log in.
To make the experience fit your profile, pick a username and tell us what interests you.
First overview of the installed components.
The main components are installed on a slidabble baseplate. This way It's possible to access the components under the main "pick and place"-plate.
Assembly of the NEMA 8 stepper that is used for rotating the picked components.
Pick head components "knolling":
Most of the mechanical parts finished (15.01.2021):
Support:
If you want to support me, you can do so by a like this project page, on Instagram or by donating a coffee. :)
I think it's time for an update again. :)
In the meantime I haven't quite managed to get the PNP up and running. The board that I had initially planned to use does not seem to be compatible with the current smothieware and thus also with the current openpnp version. That's why I'm now waiting for a new board.
In the meantime, however, I've been working on the following details.
Nozzle holder:
The tips with which the components are later sucked on are available in different sizes. Depending on the component size, these can then be changed automatically. To do this, the nozzle to be changed is inserted horizontally into the corresponding holder and then pulled down from the Z-axis stepper motor. Picking up a new nozzle then works the same way. Just the other way around. You can see the first two iterations in the pictures. As soon as I have found the perfect slot width I will print a nozzle holder in which all nozzles can be accommodated at the same time. That makes the assembly a little easier.
Vacuum sensor adapter board:
To connect the vacuum sensors to the mainboard of the pick and place machine, I built a small provisional adapter board on a breadboard.
First movements:
It doesn't look like much, but it was quite a bit of work. :D The controller boards I have used so far are all based on the Marlin firmware. The controller board of my pick and place machine is based on the smoothieware. The latter uses a completely different concept for configuration. A couple of timing belts have suffered a bit, but now it's moving.
Wiring:
In the last few days I have stripped, shortened and crimped a lot of cables. That means slowly the first real test of the pick and place machine is getting closer and closer. The lines are all laid and connected. For the (negative) pressure sensor I built a small provisional board on a breadboard. So two of the sensors (one for each receptacle) can be connected to the mainboard. I replaced the old valves with relatively cheap but sufficient 12V valves. Both the adapter for the pressure sensors and the valves are installed on the base plate via a 3D printed adapter. The MDF board on which the circuit boards are to lie later is already cut to size and is just waiting to be installed. So the picking and placing will start soon.
Recently I learned a lot about how pneumatic valves work - or don't work. Originally I had planned to use a 3/2 way valve for the pcknplc machine. This has the advantage that in one switching state I can switch the vacuum through to the pick head and in the other switching state the pressure line to the pick head is vented.
This is important because otherwise it can happen that the sucked component "sticks" due to the unbalanced negative pressure in the hose line to the pick head. At least that’s the plan.
But a (dry) test of the valve did not even deliver the satisfactory "click" that is known from relays. After I was pretty frustrated I took the complete valve apart and found the fault. (He sat in front of the valve): These valves work with the help of a pilot valve. This means that the 12V coil or its magnetic field do not move the main valve piston directly but a smaller/lighter piston that uses the air pressure at the pressure connection to control the main valve piston.
This results in two problems for my application: In complete: Epic fail, but learned a lot. 🙃😂
Usually I'm one of those people who starts wiring and at the same time thinks about what to connect to what. That can work, but in the past it has often only partly worked. That's why I decided to do everything differently this time.
Before I started wiring, I made a plan. For this I used the tool "wireViz" (link below), which was presented a few weeks / months ago on Hackaday.com.
In the beginning it was a bit tricky and it took me a while to understand the system. But over time it all made sense and new wiring was quickly planned.
So I also planned the pneumatic connections and the USB connections (between the two cameras, control board and USB hub). The current status (see below) is not quite ready yet. In some cases, the color codes are not yet congruent with reality and some additional information is also missing. But hopefully the basic scheme can already be seen. :)
At this point, many thanks again to Daniel Rojas (https://hackaday.io/project/173240-wireviz) for this really helpful tool. (For which, by the way, I haven't found a free and comparably simple alternative.)
I redesigned the bracket for the ring light of the cameras over the past few days. So a diffuser made of milky acrylic glass can also be clipped into the LED ring. Hopefully this will scatter the light a bit and avoid glare.
After I finsihed this build somebody told me that commercial PNPs use cone shaped led rings to avoid this kind of reflections. So I decided to test this design first. If needed I will have to redesign this. Lets see. :)
Some pictures of it:
A small important component for my "pick and place" machine #pcknplc is the pressure sensor, which can be used to detect whether a component has been successfully picked up or not.
Because if a component "clogs" the pick up tip, the pressure in the vacuum line between the pump and the tip drops noticeably. With the help of a suitable pressure sensor, it is therefore easy to determine whether a "pick" process was successful.
Since I was a bit overzealous when I bought a sensor that didn't fit when building my vacuum chamber (it could only measure a "positive" pressure) I ordered a few candidates to test this time. The sensor shown seems to fit nicely. Now all I have to do is print a suitable bracket and at the weekend I will finally start wiring the individual components. 🙃
In the last days I finished most of the mechanical parts.
All axis are working and can be controlled by the steppers (as soon as they are connected of course). Also there is only one endstop sensor für the y-axis missing. Now I will continue with the wiring and some minor mechanical improvements.
Here a little video including some close ups of the actual state:
After some (about three) iterations I reached the final design of the pick head. It is a very central piece since there are the following components mounted to it:
Some impressions... :)
Fabian
Hyna
Mika Heikkinen
MasterOfNull
Norbert Heinz