Hey All,
I'm finishing a PNP built from scratch, mechanical is 100% electrical is 90% software is 40%.
It was designed and built to see if I could design a PNP targeted to student groups where cost is more important than speed.
Its not amazing but i know it will work.
Prototype Specs
X travel - ~6 in (next revision will see that increased to +18in)
Y travel - ~12 in (next rev +18in)
Z travel = ~1.5 in
C Rotation about Z - 180 deg (pickup at +180 and rotate for negative rotation, giving 360 degree capability)
X and Y belt driven by stepper motor, Z belt driven by hobby servo, C driven by hobby servo
0.24mm resolution at full step for X and Y
Z resolution is ~ 0.15mm
C resolution is 0.7 deg
Bottom and top vision webcams
Aquarium pump powered suction with pressure sensor and solenoid valve to release suction
estimated cph - 600 (given the cost of the machine I'd be happy with 1 part every 6 seconds, and given test runs with stepper performance that seems reasonable)
Feeders : not sure if I should make feeders or just something to deal with lots of cut tape, as it would likely be dealing with lower volume production. Top vision will help deal with that, i think?
Again this is not me saying i'd like to build this, its built and moving and ill be making videos as software develops.
So my question to you, Internet, what is this worth? Cost is somewhere between $300-$600 so with usable software is it worth $800-1200? more?
ps, Its very hackable, arduino powered, and would come with all details on how to bend it to your will.
-Alan
What is my scratch built PNP worth?
Moderators: adafruit_support_bill, adafruit
Please be positive and constructive with your questions and comments.
-
- Posts: 164
- Joined: Fri Jun 11, 2010 9:21 pm
Re: What is my scratch built PNP worth?
I don't want to discourage you but you need to think carefully about what you're trying to achieve, otherwise you end up with something that isn't very useful for real work. 0.25mm resolution won't be good enough for QFPs, but probably OK down to 0603 This isn't necessarily a problem as even something where you manually place fine-pitch and the P&P does the grunt-work of chip parts will be fairly useful, but every compromise will make it nonviable to some part of the market. Similarly aiming too high spec will do the same on cost grounds. Getting accuracy cheaply is the biggest problem, and I can't see that anything made of wood could be useable over an 18x18" work area with enough consistency to be useable unless there are some clever tricks to correct for mechanical errors.
Are you using microstepping ? If not you may be able to improve resolution, but accuracy and repeatability, especially over an 18" bed will be difficult to maintain.
You definitely need some way to feed tapes from reels, even if an optional extra - there's no point having a P&P that needs frequent manual intervention. 100 parts is 400mm of tape, and 100 parts is not really a lot. One potentially cheap way is to use the P&P head to advance the tape.
Speed is not completely unimportant, as at some point it will just be faster to manually place.
For low-volume, fast & easy setup is extremely important - this is mostly a software issue. For example being able to auto-detect feeder positions using the camera.
If you want to get really clever, maybe apply some lateral thinking to the feeder issue - instead of making tape feeders, how about feeding loose parts from a tray, using vision to find and orient them. you could maybe even use a vibrating tray to flip them over as required. This would not only be a way to trade mechanical hardware (expensive) for clever software (cheap), it would also eliminate a perennial problem of low-volme SMT, wasted tape leaders and trailers
Are you using microstepping ? If not you may be able to improve resolution, but accuracy and repeatability, especially over an 18" bed will be difficult to maintain.
You definitely need some way to feed tapes from reels, even if an optional extra - there's no point having a P&P that needs frequent manual intervention. 100 parts is 400mm of tape, and 100 parts is not really a lot. One potentially cheap way is to use the P&P head to advance the tape.
Speed is not completely unimportant, as at some point it will just be faster to manually place.
For low-volume, fast & easy setup is extremely important - this is mostly a software issue. For example being able to auto-detect feeder positions using the camera.
If you want to get really clever, maybe apply some lateral thinking to the feeder issue - instead of making tape feeders, how about feeding loose parts from a tray, using vision to find and orient them. you could maybe even use a vibrating tray to flip them over as required. This would not only be a way to trade mechanical hardware (expensive) for clever software (cheap), it would also eliminate a perennial problem of low-volme SMT, wasted tape leaders and trailers
- asawula
- Posts: 3
- Joined: Thu Feb 10, 2011 11:30 am
Re: What is my scratch built PNP worth?
Hi mikeselectricstuff, I really appreciate the feedback.
Im experimenting with different amounts of microstepping with the cheap nema 14 motors I am using at the moment. The best i can do is double their resolution as 1/8th stepping does not give equal angular movement per step. Upgrading them is something i will likely do
As far as the problems of using wood for structure, I don't know if I understand what you mean. Not that I am set on using wood for the end product, but its great for prototyping sans-machine shop. The mechanical rigidity is provided primarily by the rails (3/4" aluminum square tubing) and bearings and there is very low play or backlash in the system..(so far)
As far as skipping steps I am going to use an optointerrupt sensor to home on-the-fly as frequently as possible to avoid accumulating any error
i definitely agree smart software will be invaluable and a quick set up time and the ability to pick up bulk parts from a tray would be a bonus to the low-volume market
Im experimenting with different amounts of microstepping with the cheap nema 14 motors I am using at the moment. The best i can do is double their resolution as 1/8th stepping does not give equal angular movement per step. Upgrading them is something i will likely do
As far as the problems of using wood for structure, I don't know if I understand what you mean. Not that I am set on using wood for the end product, but its great for prototyping sans-machine shop. The mechanical rigidity is provided primarily by the rails (3/4" aluminum square tubing) and bearings and there is very low play or backlash in the system..(so far)
As far as skipping steps I am going to use an optointerrupt sensor to home on-the-fly as frequently as possible to avoid accumulating any error
i definitely agree smart software will be invaluable and a quick set up time and the ability to pick up bulk parts from a tray would be a bonus to the low-volume market
-
- Posts: 164
- Joined: Fri Jun 11, 2010 9:21 pm
Re: What is my scratch built PNP worth?
Wood changes size and can warp with temperature and humidity, although this can be reduced to some extent with treatment.
You need to consider who might use a cheap P&P. Bear in in mind that if you're only assembling one PCB, using even the best, most cleverly designed P&P will usually take longer than hand-assembling due to setup time. This means your target market is people who do enough work to spend on a P&P, but not enough to buy a cheap used commercial one, or contract the job out.
I have no idea how big this market is, and probably the only way to find out is to build it & see if it sells.
Kickstarter may be an option but you have to have a very clear idea of where you want the cost/usefulness tradeoff point to be, and have proved that the required accuracy and repeatability is achievable.
You need to consider who might use a cheap P&P. Bear in in mind that if you're only assembling one PCB, using even the best, most cleverly designed P&P will usually take longer than hand-assembling due to setup time. This means your target market is people who do enough work to spend on a P&P, but not enough to buy a cheap used commercial one, or contract the job out.
I have no idea how big this market is, and probably the only way to find out is to build it & see if it sells.
Kickstarter may be an option but you have to have a very clear idea of where you want the cost/usefulness tradeoff point to be, and have proved that the required accuracy and repeatability is achievable.
- cyber.shifter
- Posts: 2
- Joined: Thu Feb 10, 2011 9:42 pm
Re: What is my scratch built PNP worth?
I've some experience with very small SM component assembly so I thought I would comment and give my two cents worth. When I worked with it, all of the parts we used came in small square trays that were divided individually. Each part was placed in it's own little section in the tray. The trays were notched so they could be fit into a larger P&P machine. I'm not sure how your target market is going to receive their parts, but you may want to consider a system that can retrieve parts for this sort of packaging. I used to have an old tray, but I couldn't find it as of this writing. If I do I'll post a picture.
- asawula
- Posts: 3
- Joined: Thu Feb 10, 2011 11:30 am
Re: What is my scratch built PNP worth?
Basic movement and basic picking and placing, vision coming next.
http://www.youtube.com/watch?v=BoCJFJsCXPQ
http://www.youtube.com/watch?v=BoCJFJsCXPQ
-
- Posts: 307
- Joined: Sun Jan 24, 2010 6:48 am
Re: What is my scratch built PNP worth?
It's certainly an impressive machine for a DIY build and I do hope you find a market for it!
Universities\Small businesses etc might like if IF it's easier to use than placing parts manually.
As someone else said, if it's too slow, it's going to be faster to place by hand anyway. In which case all it has going for it is accuracy, and if that's not up to scratch then there's no reason to do it by machine.
Universities\Small businesses etc might like if IF it's easier to use than placing parts manually.
As someone else said, if it's too slow, it's going to be faster to place by hand anyway. In which case all it has going for it is accuracy, and if that's not up to scratch then there's no reason to do it by machine.
Please be positive and constructive with your questions and comments.