LitePlacer.com forums

mrandt wrote:

Pixopax wrote:holder [...], consider making it shorter. Then we glue or press an acryl or polycarbonate tube on it, with M8 thread in it. Then we can detect the pressing of the nozzle optically.

At least with the Juki nozzles an available couplers / holders the design will not work - or I did not understand it correctly

It will work, take one of the available nozzle holders, cut the upper part off and replace it with a translucent part of the same size.
The you can see the pushpin, which travels up as soon as the nozzle tip goes up. A led and photodiode can then detect that.
That pin needs to be heavy, or needs a weak spring to keep it down.

Threads are not precise to make things run round. The only thing which works 100% round is press-fit or shrinking.
Everything with setscrews is not exact round, but is better as thread.
They should be able to make a 8H7 hole with a reamer which then fits a motor axis good enough to use 1-2 set screws.
If not: Let it make a bit smaller, then I can make the hole on my lathe for you.

Pixopax wrote:It will work, take one of the available nozzle holders, cut the upper part off and replace it with a translucent part of the same size.

I don't think it is possible to replace that part. It contains the steel ball lock, the outer ring and spring mechanism. All of these will get in the way of your optical sensor.

You have to extend the nozzle's pin movement by some hollow or hexagonal push rod inside the coupler so you can detect it further up - either in a clear part introduced into the nozzle holders shaft as you suggested or even above the hollow shaft of the motor.

Pixopax wrote:Threads are not precise to make things run round. The only thing which works 100% round is press-fit or shrinking Everything with setscrews is not exact round, but is better as thread.
They should be able to make a 8H7 hole with a reamer which then fits a motor axis good enough to use 1-2 set screws.

Yep, I learned this the hard way now... Next attempt will be tight fit + 3x set screws in 120° alignment. This should result in sufficient concentricity.

Pixopax wrote: Most people put these holders on hollow-shaft-steppers.

One problem with hollow-shaft steppers is you can't get them with 0.9-degree step angle. And they're direct-drive so there's no gearing between them and the pickup head. This puts some severe limits on rotational accuracy.... it's still just barely good enough for most BGAs, but only just barely.

Next time I take my machine apart I'm converting it back to off-axis rotational drive, but keeping the Juki changer. Basically what Karl is doing. Also his/Juha's off-axis-drive approach eliminates the issue of upward pressure on the motor shaft.... the nozzle holder has its own bearing which can be really beefy. You can even get 10mm-bore pillow block bearings designed specifically for this purpose.

Just had an idea for the Juki heads, based on the discussion in another thread with the inductance to digital ics.. would it be possible to touch the two ends of the spring at the end of the nozzle to measure the inductance change as it compresses? My main worry with something like this is that getting electrical contact with the spring and still be able to use a tool changer would cause more hassle than it's worth?

Mark Harris wrote:Just had an idea for the Juki heads, based on the discussion in another thread with the inductance to digital ics.. would it be possible to touch the two ends of the spring at the end of the nozzle to measure the inductance change as it compresses? My main worry with something like this is that getting electrical contact with the spring and still be able to use a tool changer would cause more hassle than it's worth?

The inductance to digital ICs posts got me thinking as well. I even went so far as to play around with a couple little coils on my oscope and then I remembered I don't know anything about analog electronics.

In any case, I was thinking that one of those ICs with the right coil embedded in the nozzle adapter above where the nozzle top is would do the trick. As the nozzle spring compresses the inner shaft of the nozzle rises up and above the nozzle body. Based on what I read in the data sheets of those ICs it seems like this would be readily detectable. The coils could probably be done on a tiny PCB that gets embedded into the nozzle holder. I was thinking a small PCB with a hole in the middle for the shaft to rise up and through, with a reference coil on the top and a sense coil on the bottom.

I'm 3D printing a nozzle adapter prototype right now to get a feel for if this is feasible.

I'll have to get some of those nozzles to play with. I'm just starting to work on building a PNP and going a bit overboard on the mechanics... so no reason not to go overboard on electronics too.

Hi,

I do have a couple of custom machined holders which support auto-change with a steel ball bearing lock. Will try to post some pictures later.

I was planning to have them machined in bulk in China and sell them in Europe as a courtesy to the community. However, I am not happy with the quality yet - especially for the price I paid.

I would also like to integrate Z-probing. So far, I had thought of
a) inserting a small contact into the coupler like Karl did
b) extending the inner nozzle shaft's movement by inserting a second tube (or hex rod) into the hollow shaft and detecting it above the motor

The inductance ICs brought another option to the table.

I think measuring the spring at the nozzle tip is not feasible, as it would be difficult to contact - making auto nozzle change more complex as well.

I like Jason's idea of embedding a hollow coil into the coupler. But in that case you still have to run at least two (maybe four) wires trough a rotating shaft and connect them via sliprings or similar

What I will would be more of a combination of the previous ideas: Insert a lightweight secondary tube into the hollow shaft so it is pushed up, whenever the nozzle's inner shaft moves up. A fixed coil mounted around a small tube above the hollow shaft motor should be able to detect that movement...

Do you think that would work?

Cheers
Malte

It might. Also, you could put a transparent tube from the hollow shaft and detect the raise of the inner tube optically.

JuKu wrote: Also, you could put a transparent tube from the hollow shaft and detect the raise of the inner tube optically.

Already tried that, did not work too well...

Reflex opto-sensor had problems with reflections from the transparent tube. A sender / receiver pair worked better but would result in different "analogue" level depending on the rotation angle of the tube.

I guess that also has to do with the material used for the transparent coupler; in my case it was clear acrylic - extruded not milled.

Long story short, I was not happy with the result.

I will order one of these TI chip and give it a try with a simple, hand wound hollow copper coil. Inductance should change as soon as another metal moves into the coils magnetic field.

Exciting

Yup, basically using the nozzle shaft as a variable inductor by changing the core material (from air to the nozzle shaft)