O-M-G cheap pick and place machines coming!

[AlfBaz]

If you want to get it to slide...

Oh, I thought he wanted the pen to just turn and not slide up and down

[AlfBaz]

That ball bearing is sloppy as hell, it will accommodate 0.02 mm without any problems.

Any bearing that accommodates a shaft 0.02mm larger than its bore diameter needs to be thrown through the manufacturers window, hopefully hitting them in the head

Sorry ataradov, you may be right. I looked at the picture and immediately thought it was a pedestal bearing, with a normal ball bearing in it. I see now it's a linear bearing

[SeanB]

If you want a hollow long shaft to replace the aluminium go to a bike shop and look at gudgeon pins, as there you get a long precision ground outer with a reasonable inner diameter hole. Used ones are either free or scrap steel price, as there when the piston is worn the pin is good enough to still work in a linear race, even if it has been in the engine and was picked out of the exhaust after it tore itself to pieces.

[M4trix]

@Sean,

thanks mate for the tip. I already found a candidate for the pen --> http://www.tme.eu/en/details/gn480.1-d12-200-ni/profiles-and-holders/elesa_ganter/gn4801-d12-200-ni-os/

The tube length is 200mm but I'll cut it down to 150mm. Seems enough for the overall height (linear bearing+spring+tip). Also, there is enough thickness to thread it on both sides in order to mount the plastic tip and the air hose plug. I still don't have the linear bearing. 

[pickle9000]

This looks like a massive commitment.



https://www.vbesmens.de/en/pick-and-place.html

[Kjelt]

This looks like a massive commitment. 

The paste dispensing should be left out. A stencil is not that expensive and if you have the numbers (why else use a P&P) its peanuts to the total BOM.
As seen too many bridges and under the Fets/transistors he removes at the end you clearly see only 3 dots of paste while it should be almost completely covered for good heat exchange.
But overall looks like a very nice machine, if it is 100% DIY, I take my hat off to the designer.

[Psi]

Why have one pick-n-place head when you can have 20

[kayvee]

This looks like a massive commitment.

A lot of flex/movement on the PCB when the head comes down onto to it.  That can't be helping placement accuracy much.

[george graves]

A lot of flex/movement on the PCB when the head comes down onto to it.  That can't be helping placement accuracy much.

I saw that as well.  I'm sure it's something that can be tweaked.  Really not an issue worth mentioning IMHO.

[Rasz]

This looks like a massive commitment.

A lot of flex/movement on the PCB when the head comes down onto to it.  That can't be helping placement accuracy much.

dont the big boys suck pcb to the table? I seem to remember watching YT clip of one of expensive P&P boxes using vacuum to keep board in place.

[M4trix]

*bump

Here is an update. Got the linear bearing and the vacuum pen. The pen perfectly slides through the linear bearing but for the project it's too short! So I dismantled it, kept the plastic tip, the cap, needle and the suction cups. Here is a preview photo made in Solidworks. The base and the trolley is finished. Oh, that's not a wireless camera. I only showed a bit of the usb cable. 

[JohnSL]

I have a similar design I've been using for a little while now:

http://www.thingiverse.com/thing:385567

http://trains.socha.com/2014/07/building-fsr-circuit-boards.html

I originally had the camera like in your design, but discovered it didn't help very much. The problem I had was a result of the angle, so I couldn't use the camera very easily for alignment. I ended up removing the camera and just using my eyes. It's amazing how much a difference stereo vision makes for accurate alignment.

I've thought about having two cameras, 90 degrees apart, but I haven't tried that.

  -- John

[M4trix]

Hi JohnSL,

Yea, that's a well known problem with these kinds of microscopes.  I tried to place it as much as possible over the working spot. I don't expect problems with small packages like 0805, 1206, SOT23 etc but certainly there will be problems with bigger chips. Well, like you said, more cameras might do the job but that makes the project more complicated. Anyways, I can't work with SMDs without a camera anymore. When you step into your 40's your eyesight isn't perfect anymore. Btw, nice job you did there John. 

[pickle9000]

I've been raiding my junkbox and putting a manual machine together. It's falls more under the catagory of a "junkbox special". My eyes are poor so that's the reasoning. I'll grab a few picks to show how it's going.

Don't laugh too hard this is mainly so I can see how I want a system to work before I do a finished unit. Knowing me if it works out to be "good enough" I'll just use it as is.

I'm using a security camera with a telephoto lens on it, 4.5 inches off the board at 10 deg off 90. It's on the back of my tool so I the board is not obscured.     

[slburris]

If you happen to be at Maker Faire this weekend, there's a working FirePick being shown there.  I saw at least one other PNP machine there as well.  While I wasn't being distracted by the 3D printers, CNC machines, plasma cutters, and laser cutters!

Scott

[ProtoVoltaics]

I've been working on a Pick and Place Machine now for a few months. Maybe you guys can look at it a pick it apart. To really see all the info it might be easier to look at our Hackaday page hackaday.io/protovoltaics when you're there just click on our pick and place project.

Here is the latest video on our machine we took yesterday. The video is from a phone camera, because I forgot my camera at home 

https://youtu.be/C7Pi4akL5og

Thanks!

[alexandru]

Although I didn't know when I started my own Pick and Place machine, I am quite amazed there is an entire community and a lot of Pick and Place machines solutions on the internet. Unfortunately, most of these afforable P&P machines are just for prototyping - for intance most of them don't work with real reels. I actually started 8 months ago as builing a home-made P&P Machine .

Right now, I am working to design VisionBot which is an affordable Surface Mount Technology assembly shooter that can be used to place about 2,000 SMD chips onto Printed Circuit Boards for PCBA. Practically the VisionBot machine will enable the dream of millions electrical engineers, hackers and makers just like me to manufacture electronic devices in their own garage.  One week ago we changed VisionBot design from using a needle to advance the cut tape with automatically feeders.

VisionBot  is still in the beta testing and in a few months it will be available to the makers through our website and maybe a crowd-funding platform. Using advanced Computer Vision VisionBot is avoiding electronical waste and is professionally assembling Printed Circuit Boards. It is estimated that VisionBot will be able to assembly about 75,000 electronic boards per year.

Although, VisionBot is compatibile with standardized files like RS-274X Gerber files, the machine works also with the classical CSV files. VisionBot will enable the dream of many makers who want to turn their homemade prototypes into industrial products by starting an electronic production.

VisionBot Main software showing the Gerber Editor

A picture of the VisionBot machine


How it looked like two months ago:

VisionBot very old

[NickE]

I am planning to make a PnP machine. Can someone recommend me a good bottom camera for vision system? I have tried a few cam but I was not satisfied with the results.

[EEVblog]

The VisionBot is looking quite promising!

[mrpackethead]

I've found a cheap-ish chinese made machine, that has vision, and takes up to 40 reels, plus tray pick ups..   About $10k USD, for a 4 head vacumn pick up and 40 reels.

Made by a crowd called BoreyTech.  They have smaller ones listed on alibaba.

http://www.aliexpress.com/item/Desktop-Automatic-Visional-SMT-Chip-Mounter-LED-Pick-and-Place-Machine/32416534844.html?ws_ab_test=201407_5%2C201444_6%2C201409_2&spm=2114.01020208.3.57.eOCqr3

Its certainly an interesting option..  Next time I'm in China, i'm going to go and check this out.. 

NB. I have no interest in this company other than i hope its the real deal.

[mikeselectricstuff]

At least that looks like a real P&P machine - warning light  tower & all.
But why the big leadfree sign on it?
 
 

[Kjelt]

Hardware is only 50% of the package. I have seen beautifull chinese hardware with desastrous software ruining the whole experience. So check out the software.

[mrpackethead]

At least that looks like a real P&P machine - warning light  tower & all.
But why the big leadfree sign on it?

I guess they think it will sell? Maybe you get a PB-free and a PB sticker in the box?

[BFX]

VisionBot software UI looks horrible.   

[KL27x]

I have a large batch of boards arriving in the next week, and I tuned up some of my tools in anticipation. Then I came across this thread, and have been blown away at discovering the "manual PnP machine."

e.g. http://www.thingiverse.com/thing:385567 from John

I think this bears being repeated:

How about a 45 degree angled mirror, with a small hole through the middle for the needle?

Absolutely brilliant idea, credit to Mike, and I can't imagine why this comment seems to have disappeared without a trace in the discussion. I suppose you would need to reverse the image, though, and I don't know if there's an off the shelf solution for that.

Also, as for rotation without twisting the tubing, the tube-in-a-tube idea is really all it takes, which Mike has also pointed out. I have used simple brass tubes from the hardware store to make linear bearings, and for an application with zero side loading, like this, it would be more than adequate. There are a couple pairs of standard K&M Engineering brass tube sizes at your local hardware store that slide within each other with relatively high precision. Even without grease, you probably have a good enough vacuum seal. Heck, this could be your linear bearing, pen body, and your free X degree rotation without kinking the tubing, all in one.

All that said, I am baffled at the engineering and time being misapplied to this device from the start. It seems to me this design completely hits its mark, but it's the wrong mark. I believe the valuable features can be done in possibly more efficient ways without having to make the compromises that come with this design. I think it might be a bit of a Rube-Goldberg machine.