Neoden 4 pick and place

[TheSteve]

It also desperately needs an independent speed control for the Z axis.

Seriously? you can't set this independently...? 

No, if you slow the X/Y it also slows the Z. We would often prefer a a higher speed X/Y with a slower Z for fine pitched parts.

[elmood]

These are great points / bugs... Please add them to the collaborative page.

http://collabedit.com/hfe7r

Let me know if there is any trouble editing.

[Ichan]

@elmood, thanks for the video and the list - if only i already have the machine on my hand then i surely will add some on the list.

On your video, you use drilled hole as "marking point" right? What size is it? How is the resulted placement accuracy?

The machine expects the board reference to be your first part which is an absolute value from 0,0. The fiducial points are then are then a relative value from the first part. This isn't out preferred way of doing it.

The information above is just too important to be missing on the manual.

Edit: just check the tracking, my machine just received by the forwarder company in Shenzhen - hope they ship it to me asap.

-ichan

[thommo]

Guys, don't forget that 'Fiducials' can actually appear as 'components' in a BOM / PnP file eg. FID1, FID2, FID3, and the 'First Part' can actually be FID1 with its coordinates set to 0;0 if that's what you want [albeit with a DoNotPlace status for that component/FID1].

@elmood, thanks for the video and the list - if only i already have the machine on my hand then i surely will add some on the list.
On your video, you use drilled hole as "marking point" right? What size is it? How is the resulted placement accuracy?

The machine expects the board reference to be your first part which is an absolute value from 0,0. The fiducial points are then are then a relative value from the first part. This isn't out preferred way of doing it.

The information above is just too important to be missing on the manual.

[Ichan]

When we generate pick and place files from the CAD the current origin point will become the reference point, not the first part.

"Mark points" on Neoden term are several coordinates to inform the position and rotation / angle of the board on the machine - it can be fiducials or any other precision reference mark.

-ichan

[rwb]

Exciting to see all this first hand knowledge come together.

Ichan how much do you have invested in the machine at this point?

Are you saving any money by running the boards yourself vs outsourcing the job? Just curious.

[RobK_NL]

I was seriously looking at this machine, but all of what has been said here made me think again.
Plus, this video at 2:17 and 2:28 shows the machine happily placing some TQFP's that have been very obviously used and abused before.

I'll stick with my ancient DIMA Optimat for a while longer.

[Ichan]

Ichan how much do you have invested in the machine at this point?
Are you saving any money by running the boards yourself vs outsourcing the job? Just curious.

I pay about $12K for my machine order, will have to spend some more for the shipment later.

Well, not only for my own job i take outsource job too in here - like the current job which is waiting for the machine

...at 2:17 and 2:28 shows the machine happily placing some TQFP's that have been very obviously used and abused before.

Interesting findings,  but i accept that as a positive thing

-ichan

[TheSteve]

Well later today or tomorrow we plan to build two panels that have just under 1600 parts each(10x10 matrix). So we'll see how the machine does. No tough parts, all 0805/soic.

Some other bugs to report:

-When using a tray if there is an error it will display the incorrect tray number on the screen - it is always off by one.
-When we try to use 3 fiducial points it never seems to get the parts in the proper place, using only 2 seems to work fine.

[RobK_NL]

Interesting findings,  but i accept that as a positive thing

-ichan

There's two more; at around 2:37 and 2:50

Would you accept a machine that places those "as a positive thing"?

[Jefferson]

... hm ... why bent pins in the photo

[theatrus]

Probably their firmware test unit with some damaged scrap parts run through it a ton of times


Sent from my iPhone using Tapatalk

[rwb]

I understand, they are just using the same parts over and over for demo's and testing.

It's not like they are making real boards in the training videos.

[Jefferson]

it's clear, the sight has an offset, which is then adjusted. But the loss of speed is required.
Is it always exactly?
...
I find it hard to accept that the software for XP

[wraper]

Interesting findings,  but i accept that as a positive thing

-ichan

There's two more; at around 2:37 and 2:50

Would you accept a machine that places those "as a positive thing"?

That IC actually looks like it was desoldered. Even if not, there is some residue on it's bottom, probably adhesive from the sticky tape they used instead of the solder paste. Anyway, I don't think it is the machine itself which bent those pins but it happened while the part was salvaged for reuse.

[Jefferson]

Flux is ... 100% believe my experience 30 years. It's not accurate, just laugh

[TheSteve]

Interesting findings,  but i accept that as a positive thing

-ichan

There's two more; at around 2:37 and 2:50

Would you accept a machine that places those "as a positive thing"?

All that really matters from that picture is that the vision system correctly identified the leg outline and position of the chip so it can be placed accurately - which it looks to have done just fine.

[rx8pilot]

My ancient 20+ year old machine will reject a QFP with one bent pin - which is a good thing.

[Ichan]

Would you accept a machine that places those "as a positive thing"?

So i can use refurbished chip too with the machine 

-ichan

[rwb]

@theSteve  Hey when you run those boards tomorrow will you take some video or pictures for us interested folk on here? 

[thommo]

2nd that RWB .. TheSteve!

In fact, if you had time, we'd even sit through an episode of a tripod-mounted real-time capture, from start to finish, with your narrative. Would be a great insight into what's ahead, and give Neoden a better understanding of how the machine/user interface operates.

@theSteve  Hey when you run those boards tomorrow will you take some video or pictures for us interested folk on there? 

[Ichan]

+1

Good or bad please make some videos.

-ichan

[RobK_NL]

Anyway, I don't think it is the machine itself which bent those pins but it happened while the part was salvaged for reuse.

That was not my point and neither did I even suggest that.

My point is that the vision system on this machine obviously doesn't care about the fact that several pins are bent to the point of actually touching adjacent pins! That's just completely stupid.

[rwb]

Robk_NL I think all you need to do is spend $20,000 + extra to get that software feature built in. The Neoden 4 is is a $8500 machine so you can't expect all the features of a machine that cost many times more than that.

[Ichan]

I am curious, how much is the price of Dima ATOZ PP-050 with let say 48x 8mm feeders and full vision option?

-ichan