Again, another feeder design

[mrandt]

Hi Karl,

thanks for the photos. Looking at the pictures I guess the base could easily be printed from ABS.

As you already pointed out, the gears and levers should probably be milled from aluminium - especially the teeth look too delicate for ABS, let alone PLA. Maybe Nylon could work, but then they would also flex a bit which I assume is not desirable.

Transparent cover could be laser cut or milled from acrylic.

I have bought an Inventables' X-Carve kit a while back but have not found time to assemble it yet - finally a good reason to build that mill.

Looking forward to your CAD files

By the way: Juki shafts / holders seem to be much harder to find than nozzles. On Aliexpress / Alibaba there is just one (expensive) offer for a shaft. Do you think it is possible to make your own coupler for those? As the spring is in the nozzle, I think it could be possible...

Otherwise I might go for Samsung or Fuji nozzles, for which the couplers are much easier to get.

Best regards
Malte

[Spikee]

The Samsung nozzles and holders are pretty available.
--
I have a nice CNC , I could try the design next weekend if the files are available.

You can get the real sprocket of samsung and other feeders on ebay. maybe this is a good alternative for this ?

Could also cnc a mould and pour polyurethane to make the gear / mechanics. You can get this in different hardness.

[JuKu]

I would think one wants several of these, like 20-30. With that quantity, laser cutting the parts from stainless steel is feasible.

[mrandt]

I would think one wants several of these, like 20-30. With that quantity, laser cutting the parts from stainless steel is feasible.

That's also what I thought, but the company I order all laser cuts from cannot (or does not want to) produce such small parts

Have you asked your manufacturer? Maybe you could order a few hundred sets of gears and sell them

On the other hand - once I have my CNC mill ready, it could to that work for me - assuming aluminium is stiff enough.

[gtrplaya101]

Such a great solution! I think waterjet might also work for these guys

I think we could easily do a group buy of several hundred at a time and get everyone going. I would be happy to do some research on what vendors would charge here in the USA. Any chance we could get the cad files for the metal parts?
Curt

[Knas]

I think we could easily do a group buy of several hundred at a time and get everyone going. I would be happy to do some research on what vendors would charge here in the USA. Any chance we could get the cad files for the metal parts?

Guys, guys, i'm super happy and humbled over your enthusiasm but you've gotta hold your horses!

I haven't yet verified that it's spot on and you don't want to spend money cutting stuff that isn't gonna work perfectly in the end. I cut two more housings yesterday and i'm trying to do two sets of hardware today, hopefully i'll be able to test it before the day is over. I think i should have at least 3 feeders before saying it's OK. It might also be fertile to wait for the 2 & 3 hole spacing designs to be verified so everything can truly be made in one go.

Transparent cover could be laser cut or milled from acrylic.

I'd actually suggest making the entire housing from HDPE, i only did it acrylic so i could verify & show operations. The HDPE sides i've made have a much nicer play together with the hardware.

By the way: Juki shafts / holders seem to be much harder to find than nozzles. On Aliexpress / Alibaba there is just one (expensive) offer for a shaft. Do you think it is possible to make your own coupler for those? As the spring is in the nozzle, I think it could be possible...

The shafts are only $100 on ebay and a guy on the OpenPNP list just made his own for even less, check it out. The *huge* benefit with the Juki nozzles as i see it is that the spring action inside also yields a tube that presses up into the shaft, i use this as my zero detection switch instead (picture attached) and it works much better than having the whole shaft move as it a) doesn't impeded movement/change tension of the A-belt and b) the push is a *lot* softer, i had issues with the original version sometimes pushing the components hard into the tape during probing making the vacuum not be able to pull them out.

Karl

[mrandt]

The *huge* benefit with the Juki nozzles as i see it is that the spring action inside also yields a tube that presses up into the shaft, i use this as my zero detection switch instead (picture attached)

Wow - did not realize that. It is indeed a *huge* benefit. Samsung's nozzles rely on a springy mechanism in the holder itself but I have not had a smart idea yet to detect indentation of that mechanism as the coupler has to be able to rotate:

samsung_nozzle_cp40.png (152.87 KiB) Viewed 12995 times

How does your switch work? Do you check for electrical conductance from needle to the pin shown in your photos? Or is there a microswitch mounted behind the white plastic (hidden in the picture)?

[...] b) the push is a *lot* softer, i had issues with the original version sometimes pushing the components hard into the tape during probing making the vacuum not be able to pull them out.

I measured the push force, on my machine it is almost 3N - too hard for electrolytic caps and some IC.

I however think that with properly adjusted Z-levels (backoff especially) most users should not have problems. The switch is only used to determine level for first part - all other parts of same type are placed using the determined pickup and place Z-levels.

But I agree - the JUKI springy nozzles are an elegant solution to that problem and I seriously consider upgrading to those now

[Knas]

How does your switch work? Do you check for electrical conductance from needle to the pin shown in your photos? Or is there a microswitch mounted behind the white plastic (hidden in the picture)?

So just fyi everything i do is made in the most ghetto but still reliable way i can think of, more often than not with what i have at hand at the moment because i'm impatient and poor

So the round grid sits in the tube and has a simple sewing pin that the tube from the nozzle hits, the holder of the sewing pin obviously needs to be perforated so that the vacuum can flow freely. The top side of the pin is soldered to a cable that goes through the holder assembly up to a rotating vacuum and electrical contact. The rotating contact is simply a piece of HDPE (so that it's electrically isolated from the rest of the tube which is grounded), the outside of the contact is covered in conductive aluminium tape and the wire is connected to this, again with a sewing pin going through the whole thing. So the pins head is on the outside, making contact with the aluminium tape (secured the contact with silver ink) and inside it's soldered to the wire. The tape then simply goes against a piece of steel wool - just like a brushed motor - making contact with another assembly at any rotation. This assembly is pulled to +V through a resistor, as the nozzle (which is grounded) hits something it shorts the connection to ground. Again, another little picture of my super ghetto setup.

Karl

[Spikee]

when you fully compress the spring than there should be enough pressure to trigger the switch.
If not user some other kind of detection mechanism to detect smaller changes in height such as optical, or magnetic.
You would have to redesign the head mechanism anyway. Worst case you could send the signal over RF or whatever.

You could also block that moving pin so that it works like a fixed unit as does the liteplacer now.

[dave]

Karl I think you solved me a massive problem as my Picker head is not like liteplacers where I can easy have a switch. I think I understand this.

The Juki Nozzle though looks plastic in parts. Am I correct to say the tip of the Nozzle is GND via the shaft. Probably the whole machine is GND which includes the Juki Nozzle center.

is the Nozzle is conducting to its center as standard or have you spayed it in conductive stuff?