I originally had a USB microscope mounted, but ended up taking it off. Because the microscope was at an angle, it was difficult to figure out where the device would end up being placed. I tried using a Processing program that displayed crosshairs, and this helped. However, both my friend and I found that using our eyes worked better.
you would need to mount it parallel to the needle, or perpendicular with a mirror
Anotehr idea: two cameras (arbitrary angles for depth reconstruction) with some image processing to augment the picture like Steve Manns welding helmet
why? Image processing is cheap (laptop), cameras are cheap, this would turn manual P&P into monkey on a tricycle/video game type of affair, where you can put totally clueless person behind the desk and ask them to do what the pretty green lines on the screen tell them to.
btw cheap vacuum pumps are in hot air stations
http://www.ebay.com/sch/i.html?_trksid=p5197.m570.l1313&_nkw=Aoyue+Replacement+/+Repair+Pump&_sacat=0this is the pump used for hot air / desoldering station combo
why? Image processing is cheap (laptop), cameras are cheap, this would turn manual P&P into monkey on a tricycle/video game type of affair, where you can put totally clueless person behind the desk and ask them to do what the pretty green lines on the screen tell them to.
a few thoughts of why not:
- a manual PnP machine will become quite large with a laptop , the screen of the laptop has to be in view with the attached keyboard infront not very ergonomical, an external large monitor will take even more space.
- you have to boot the laptop for 30 seconds, run the correct program, be annoyed all the time with all kinds of OS bullshit popup messages like updates/virusscans etc.
- there are a lot of problems reported to get two identical USB camera's on one PC working
- you probably have to write your own software which will take weeks/months to get correct, only interesting if you go to a DIY fully automated PnP
why? Image processing is cheap (laptop), cameras are cheap, this would turn manual P&P into monkey on a tricycle/video game type of affair, where you can put totally clueless person behind the desk and ask them to do what the pretty green lines on the screen tell them to.
a few thoughts of why not:
- a manual PnP machine will become quite large with a laptop , the screen of the laptop has to be in view with the attached keyboard infront not very ergonomical, an external large monitor will take even more space.
- you have to boot the laptop for 30 seconds, run the correct program, be annoyed all the time with all kinds of OS bullshit popup messages like updates/virusscans etc.
- there are a lot of problems reported to get two identical USB camera's on one PC working
- you probably have to write your own software which will take weeks/months to get correct, only interesting if you go to a DIY fully automated PnP
A $300 android tablet and be done with it.
http://opencv.org/platforms/android.htmlhttp://www.hardwarecanucks.com/forum/hardware-canucks-reviews/67076-nvidia-shield-tablet-review.htmlHas one camera too many but in any event, in a year this probably goes for $200 or even less
why? Image processing is cheap (laptop), cameras are cheap, this would turn manual P&P into monkey on a tricycle/video game type of affair, where you can put totally clueless person behind the desk and ask them to do what the pretty green lines on the screen tell them to.
a few thoughts of why not:
- a manual PnP machine will become quite large with a laptop , the screen of the laptop has to be in view with the attached keyboard infront not very ergonomical, an external large monitor will take even more space.
- you have to boot the laptop for 30 seconds, run the correct program, be annoyed all the time with all kinds of OS bullshit popup messages like updates/virusscans etc.
- there are a lot of problems reported to get two identical USB camera's on one PC working
- you probably have to write your own software which will take weeks/months to get correct, only interesting if you go to a DIY fully automated PnP
Why are you mentioning manual ? I was referring to a vision correction system for auto P&P.
For manual all you need is camera+monitor
That's much like I run on my manual PnP - the 272(H) pixels is a bit crappy. Sure, you can go closer to get 0603 to look massive, but then you can't see enough of, say, QFNs or QFPs.
Hence my rambling earlier in this thread about building something better. And 3D, with 2 cameras.
Why are you mentioning manual ? I was referring to a vision correction system for auto P&P.
I was not responding to you Mike, it was another subdiscussion mentioning manual PnP, at least how I read it.
why? Image processing is cheap (laptop), cameras are cheap, this would turn manual P&P into monkey on a tricycle/video game type of affair, where you can put totally clueless person behind the desk and ask them to do what the pretty green lines on the screen tell them to.
a few thoughts of why not:
- a manual PnP machine will become quite large with a laptop , the screen of the laptop has to be in view with the attached keyboard infront not very ergonomical, an external large monitor will take even more space.
- you have to boot the laptop for 30 seconds, run the correct program, be annoyed all the time with all kinds of OS bullshit popup messages like updates/virusscans etc.
- there are a lot of problems reported to get two identical USB camera's on one PC working
- you probably have to write your own software which will take weeks/months to get correct, only interesting if you go to a DIY fully automated PnP
you are effing with me, right? complaining about how difficult and unwieldy laptop is to operate?
especially compared to setting up full blown P&P
Consider the benefits of manual, but vision augmented p&p - you put absolute MONKEY in front of your manual pick and place machine, and all that untrained monkey has to do is follow pretty pictures on the screen. This means
you dont have to do the boards, intern/your kid/cheap manual labour.
Most expensive part of P&P machine is ... machine itself
precise repeatable quick mechanical bits. Software is cheap in comparison.
AR program wouldnt even need to know how to read board files, it could learn on the first pass - make first board yourself, and on the next pcb program knows where all the components go and overlays accurate positions/orientation/directions.
hmm those guys could give it a go (granted it should work better than their slow clunky demos)
Why are you mentioning manual ? I was referring to a vision correction system for auto P&P.
I was not responding to you Mike, it was another subdiscussion mentioning manual PnP, at least how I read it.
oops - sorry I speed-read the "...are cheap" and thought it wad a reply to my comment about not trying to re-use vision systems from old pick/place machines!
I'd certainly agree that using a laptop to just display an image from a camera is a very far-from-optimal solution!
you are effing with me, right? Consider the benefits of manual, but vision augmented p&p
No I am not effing with you, I am dead serious. But reflecting on your last post I think I just have different requirements for a manual PnP machine than you do.
I just want a simple design but accurate manual PnP where the human still has to do some labour and handling and can judge on two small monitors if the position of the part is correct.
Most important for me it should be finished in < 100 hours of work to be able to complete it at all in my spare time off work.
For me the manual PnP is a tool I want to use to easier do my hobbie projects, not a goal on it own.
If you want your described finished product for the end of the year, it should become an open source software group community project with a lot of very enthousiastic people, or you must be a student/unemployed with 80 hours a week to spend on it and a brilliant programmer. The database of all possible physical packages and their possible orientations alone would scare me away from starting
IMHO it will be a very ambitious project with a high risk of never ever being completed (working 100% properly), but if that is your ambition and you can and will do it, my hat goes off to you sir and I wish you succes.
Well, I'm building 007. The 7th Firepick Delta. Why.. Because I can. Will it work. Ask me sometime later. What will I loose. Some time. about $350 of materials ( incl freight ). What will i learn.. Heaps probably. And i'll be be able to apply this to some real world problems.
Hey, i already have a might fine Manual Pick and Place machine ( Dima FP-600, bought 2nd hand from Ebay ) and i can place about 400 parts per hour on that. ( the ladies at the factory get more like 800 but they are awesome ). We use a big Juki based system for doing production.
bit like the First guys who went to everest, they did it because it was there.
Cut these guys some slack. If they fail, it won't hurt you.. if they succeed, well, that will be great won't it.
hmm those guys could give it a go (granted it should work better than their slow clunky demos)
I was almost really excited, then I realized they do work for corporations and charge $250 a month minimum plus a $10000 starter fee package.
*bump
I finally have some spare time to finish the manual pick and place machine mentioned in this post -->
https://www.eevblog.com/forum/reviews/o-m-g-cheap-pick-and-place-machines-coming!/msg489146/#msg489146. Regarding this pen, if someone has it, please precisely measure the diameter and post it here. I think it's 10mm but I have to be 100% sure. The pen has to slide through this linear bearing. There are bearings for 10mm and 12mm shafts. I don't want to order this stuff and get a wobbling disaster. Thanks guys.
The pen.
Linear bearing.
12.02 mm but it varies slightly (it is a tube). The internal diameter is 10.03mm.
pickle9000,
Thank you very much mate for your fast response!
Ok, I'll order the 12mm bearing and the pen and hope it perfectly slides in.
... and hope it perfectly slides in.
a 0.02mm bigger journal than the bearing is almost the standard spec for what they call an interference fit where you have to heat the bearing to get it onto the shaft. Even then the shaft is usually solid
AlfBaz,
that is worrying me. If the pen slides in and comes out that it's too stiff, especially those difficulties while rotating, pushing down etc, then I'm afraid the pen will be useless. Anyways, the pen is, like you aussies would say, dirt cheap. I'll use the plastic tip and find an appropriate aluminium tube.
That ball bearing is sloppy as hell, it will accommodate 0.02 mm without any problems.
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
AlfBaz,
that is worrying me. If the pen slides in and comes out that it's too stiff, especially those difficulties while rotating, pushing down etc, then I'm afraid the pen will be useless. Anyways, the pen is, like you aussies would say, dirt cheap. I'll use the plastic tip and find an appropriate aluminium tube.
If it's to tight you may be able to fit the pen onto a drill to spin it and use some very fine grain emery cloth to reduce it's diameter. Since the pen's body is probably made from aluminium, be very careful not to take too much off. If you get it to the point where it is to loose, you can apply some glue. Normally one of various grades of Loctite will do but I can't imagine a great deal of mechanical strain that couldn't be overcome with whatever glue you have lying around
If you want to get it to slide go to the bearing supplier and look for needle roller inner races, which come in assorted diameters and lengths. Those have a very controlled outer diameter, so will fit in the linear bearing perfectly, though you will probably need one that is 15mm outer diameter so you have the 12mm inner diameter, and a linear bearing to suit that. You can simply fit 2 or 3 of the inner races on the shaft to handle the length, the ends will simply leave one ball in the linear slide loose as it rolls past, which will do nothing.
This is unbelievable. I can't find any of those linear bearings over here in local shops. So, of course, I ordered it over ebay and it might take 2 weeks to get them. Anyways, the base is finished and the trolley. Also bought a cheap arse usb microscope and the AOYUE 932 vacuum unit. The vacuum station needs some hacking. Need to add some solenoid valves and add an option for the foot switch pedal. Will post some pics soon.