Open Placer A New Affordable PCB Assembly Machine

[mikeselectricstuff]

It also needs thermal hardening glue or your parts from the first pass could fall off during the second reflow stage.

No it doesn't, unless you have large or heavy parts, in which case, where possible, you plan the process to put those on the second side. Solder surface tension will hold parts on just fine.
 

In practice I see very few two layer smt used by hobbieists and small startups.
Most use one layer smt combined with TH parts.

Such generalisations are meaningless. Everyone's needs are different. I rarely do double-sided but wouldn't consider buying a machine that didn't support it as i don't know what future jobs may need.
 
My point is that it is trivially easy to support double-sided assembly, it's only about how the PCB supports work.

Not supporting DS is an unncesseary limitatation on a product that already has a limited market.

I do see it in high density boards having 4 or more layers and that is a different ballgame,

PCB layer count is irrelevant - I've done more double-sided assemblies on 2L than 4L, and 4L PCBs are cheap these days

[ManCave]

It also needs thermal hardening glue or your parts from the first pass could fall off during the second reflow stage.

If you are super paranoid you can always use lower temp paste for the second side. No glue required.

[mikeselectricstuff]

It also needs thermal hardening glue or your parts from the first pass could fall off during the second reflow stage.

If you are super paranoid you can always use lower temp paste for the second side. No glue required.

Glue would be easier than messing about with multiple reflow profiles and different pastes, which would be a bit of a nightmare.
Just planning the layout to put all heavy parts on one side is the easiest.

[Fire Doger]

It also needs thermal hardening glue or your parts from the first pass could fall off during the second reflow stage.

If you are super paranoid you can always use lower temp paste for the second side. No glue required.

If you own the PNP before ordering big or multilayer PCBs (which may be expensive) you can just make a dummy PCB 2L with heaviest parts, put on the components and check if the fall down. If they do change their side...

*And make a topic here with a list of the components which fell so everyone else can avoid it and contribute their own lists

[IconicPCB]

and yes.. you do get wet when having sex

[Furkan]

Dear Mike

OpenPlacer is an open source hardware. We had chosen the open source option to improve the current design and end up with a better device. All critics are appreciated. But we have an ongoing sale at Indiegogo and there are only 12 days left. I have to tell the current status of Openplacer. If people will spend money they need to know the limitations of the device. So, I said that

"For the moment OpenPlacer does not support double sided assembly" and then I added that

"It can be adapted in the future easily because it is very modular. Users can also design their own holders and replace the standard holder if they want."

I guess I was honest to the buyers and appreciating your critics and saying that we can implement that property in the future easily.

We are open to all critics and trying to improve our device day by day.

Regards
Furkan

[rx8pilot]

To Mikes point.....

Of the last 15 or so PCBs that I have done, all but one are double sided. It's not for fun, just the only way to fit the PCB into mechanical restraints.

I have very rarely needed smd adhesive for double sided assembly. Like others have pointed out, just for the larger parts. IC's and passives never need adhesive.



Short and misplld from my mobile......

[Kjelt]

To Mikes point.....

Of the last 15 or so PCBs that I have done, all but one are double sided. It's not for fun, just the only way to fit the PCB into mechanical restraints.

To my point you spent $80k on a cnc machine another $10k++ on a refurbished p&p machine, you have an $60k scope so you are far from the target audience for a cheap p&p machine as this.
People that buy these are starters , hobbieists and the like. As Furkan said if you do a bit of DIY you can easily mod the setup for two sided if really needed, or you place the few components on the backside manually, faster than setting up again the machine.

[rx8pilot]

True....to some extent.

My P&P was the first 'expensive thing I purchased when I decided to go professional. I had been looking at all the available ultra low cost options from the Lite Placer, to Maddell, etc, etc...

In the end, I learned that the limitations of the ultra-cheap machines were not worth it. It was better to keep hand assembling as I already was. So I struggled along and had my eye on the Quad, but all of them were out of financial reach. The one I got was DOA from eBay for $5k and it took me a year and another $5k to make it useful. All the while I was hand assembling everything. That choice, however, got me a system that I can make a living off of now. That P&P paid to get the CNC machine and all the other pieces of gear I now have. Because my assembly system has very few limitations - I can design and assemble just about anything which allows me great freedom. My business would have suffered if I only had a very limited assembly system - I would not have grown and purchased all the fun stuff I have now.

If I was not a pro, I would never have purchased ANY of them regardless of cost. A true hobbyist does not assemble enough to mess with it. A hobbyist that sells stuff - is a professional. I only buy things that I can make money from and have nearly zero budget in my personal life for a hobby.

[soubitos]

True....to some extent.

My P&P was the first 'expensive thing I purchased when I decided to go professional. I had been looking at all the available ultra low cost options from the Lite Placer, to Maddell, etc, etc...

In the end, I learned that the limitations of the ultra-cheap machines were not worth it. It was better to keep hand assembling as I already was. So I struggled along and had my eye on the Quad, but all of them were out of financial reach. The one I got was DOA from eBay for $5k and it took me a year and another $5k to make it useful. All the while I was hand assembling everything. That choice, however, got me a system that I can make a living off of now. That P&P paid to get the CNC machine and all the other pieces of gear I now have. Because my assembly system has very few limitations - I can design and assemble just about anything which allows me great freedom. My business would have suffered if I only had a very limited assembly system - I would not have grown and purchased all the fun stuff I have now.

If I was not a pro, I would never have purchased ANY of them regardless of cost. A true hobbyist does not assemble enough to mess with it. A hobbyist that sells stuff - is a professional. I only buy things that I can make money from and have nearly zero budget in my personal life for a hobby.

So, according to your mindset, you would not spend 1k for a machine that would help you do even 50% of your work better and faster NOW even while spending 5K + another 5K over a 1 year period during which you had NO PnP machine to help you out.... hmmmmm i am indeed missing something here..

Furkan, nice job, i wish i could afford it!

[rx8pilot]

So, according to your mindset, you would not spend 1k for a machine that would help you do even 50% of your work better and faster NOW even while spending 5K + another 5K over a 1 year period during which you had NO PnP machine to help you out.... hmmmmm i am indeed missing something here..

It is more than $1k - it is your time. You have to learn it, learn the quirks, deal with a likely variety of problems and limitations that you did not originally know about. Then for each job you have to set it up, run a board, check it, then run a couple more and check them to make sure you did not do something silly. In the P&P world - there is always something going wrong. A fiddly feeder, a loose cable, a software glitch, a cut tape that got bumped and your diodes fell on the floor. At the end of all that - you only have a half-assembled PCB so you then have to setup your manual assembly anyway. The sole reason to have a P&P is to assemble faster and that would hardly be faster. I know this first hand because my system was initially very limited. I had to hand place a lot of parts so I did quite a few projects that were 50% P&P and 50% hand assembly. It was at least the same amount of work - only harder to predict.

The overall throughput was about the same as just doing it all by hand and not messing with the machine even after I learned how it worked. The first dozen or so setups on the P&P were 1/10 the speed of hand assembly. It took a long time just to get the P&P system to match hand assembly speed. It was not until I got all the parts in the machine, all the nozzles needed for all the parts and the ability to do panels that it really started to look good in terms of time. By the time I got to that point, I had huge amounts of time invested in the whole system and process and all of that time needed to be made up. So, now that I am running 'full speed' I am thinking about all the time I need to make up. In reality, I will never really know for sure, but I do know that a machine running 3600cph and has 100 parts loaded at a time is still hard to make up that time.

A super cheap machine most likely would have been a distraction that could never make up the time needed to justify it's direct cost and time costs. After going from zero to full system......I believe I made the right choices.

[soubitos]

A super cheap machine most likely would have been a distraction that could never make up the time needed to justify it's direct cost and time costs. After going from zero to full system......I believe I made the right choices.

What you say may have some short of logic to it but it is like telling me, who cannot even afford a HAKKO praised by most, i shouldnt buy a soldering iron because its a waste of time etc, i should get a HAKKO anyway... this is how i understand what you say... may work for you but you cannot expect everybody else to have the money in the first place. There are more people who would benefit from the experience etc buying a cheaper machine with which they will definitely gain in productivity and earn more while saving money and making real life decisions about the machine more suitable for them...
Again, please understand what i am saying, what i get from what you did is that you spend 5k now, 5k within a year to get to the point to have a working machine and you didnt spend 1k now to get a working machine... so your productivity in 1 year remained the same... of course perhaps when you made these choices you couldnt find something really cheaper but if this PnP we are talking about here can work today as promised, you cannot keep thinking like you did when you made your original choice. I hope my english is good enough to explain all this as its not my mother tongue and i am not taken wrong....

[mikeselectricstuff]

A super cheap machine most likely would have been a distraction that could never make up the time needed to justify it's direct cost and time costs. After going from zero to full system......I believe I made the right choices.

What you say may have some short of logic to it but it is like telling me, who cannot even afford a HAKKO praised by most, i shouldnt buy a soldering iron because its a waste of time etc, i should get a HAKKO anyway...

The key thing is it is important to have the right tool for the job. 
A limited-functionality P&P machine is only the right tool for a small subset of jobs, as fiddly setup and feeder limitations mean that the process of setting up and running it takes longer than doing it manually.
Any sort of P&P makes no sense for most true hobbyists financially, though of course it can be a fun "project" and learning tool.
There is nothing like trying to get a pick/place system running if you want to learn about all the issues.
This is why I would not take any potential machine seriously if its creator doesn't have any experience of running a proper machine to do real work.


 

[soubitos]

A super cheap machine most likely would have been a distraction that could never make up the time needed to justify it's direct cost and time costs. After going from zero to full system......I believe I made the right choices.

What you say may have some short of logic to it but it is like telling me, who cannot even afford a HAKKO praised by most, i shouldnt buy a soldering iron because its a waste of time etc, i should get a HAKKO anyway...

The key thing is it is important to have the right tool for the job. 
A limited-functionality P&P machine is only the right tool for a small subset of jobs, as fiddly setup and feeder limitations mean that the process of setting up and running it takes longer than doing it manually.
Any sort of P&P makes no sense for most true hobbyists financially, though of course it can be a fun "project" and learning tool.
There is nothing like trying to get a pick/place system running if you want to learn about all the issues.
This is why I would not take any potential machine seriously if its creator doesn't have any experience of running a proper machine to do real work.

All i am saying is, i'd get ANY PnP machine i could easily afford before spending 10x and wait one whole year without a pnp at all... same as a soldering iron... or any tool... if i need a soldering iron but cannot afford a top notch brand name but need to get a job done, i'd get a decent cheap one to keep me going and help me make more money to buy the "good" one...

[Smallsmt]

Here, good ideas are being discussed as long until the idea got broken.
The idea of building a cheap machine for hobby purposes is great.
The price is also no problem.
That this system cannot be perfect is logical!
But think how many people have built or bought useless 3d printers for several hudred dollars but good machines starting above 3000usd.
This is a machine for hobby!.
Professional equipment costs money and also takes time for commissioning.
This machine is certainly a good start to learn automatic SMD assembly works.

[jmelson]

It also needs thermal hardening glue or your parts from the first pass could fall off during the second reflow stage.

If you can tolerate two passes through the reflow oven, there's no need for glue.  I do almost all my boards with parts on both sides.  I have one that has a big regulator chip (5-lead TO-220) on the back, and have never had one fall off!  I've done hundreds of boards with big Oxi-cap (Niobium Oxide, similar to tantalum caps) on the back, never had a problem with them.

The glue is used in cases where the back of the board is printed and placed, then the front of the board is printed and placed, and then ONE reflow pass is done.

Jon

[Fire Doger]

Comparing soldering iron with PnP is wrong.
The only thing that matters is to get the boards done. Without soldering iron you can't, without PnP you can.
Does PnP worth buying to assemble boards? ONLY If it is more efficient than your hands. Efficient includes personal time, problems, broken components, etc...

If you need a PnP to increase production you don't touch cheap (yet). It just doesn't worth even the 1k, by keeping it you need 1k less to buy a proper one.

A manual PnP with with x-y stepper controlled from a joystick with velocity and a simple pickup nozzle has better value for money. You get great stability on hand and with a usb microscope near nozzle its a killer for small productions and it may cost less than 100$ with Ebay parts without fancy machined component holders...

[rx8pilot]

Pick and place is not a machine - it is a process. It is a process that is intended to automate a manual process that can be done by a person. The only reason to have it is to save time.

The only reason to have it is to save time.
The only reason to have it is to save time.
The only reason to have it is to save time.

The decision is not like a soldering iron, oscilloscope or CNC machine. For 15 years, I used a Weller WES51 soldering iron which is entry level and limited performance. Still, it was vastly better than using a match to solder wires and components. A cheap CNC machine, with all of its limitations, is vastly better than a chisel or Dremel tool. An entry-level oscilloscope is vastly better than just guessing what signals are in a wire.

The point is that those processes cannot be done at all without a minimum tool present designed for the job. In PCB assembly - you can do it by hand with very basic tools. You can speed up the process considerably with cheap additions like a stereo microscope, vacuum pen, parts trays, etc. You can speed it up further with organization and process planning. The next jump is to add a pick and place machine which, on the surface seems like it would amplify you speed dramatically. The problem is that the process is complicated and fiddly - it takes it's own time to accomplish. In general, it takes a lot of time to setup and verify. If you skimp on this process to save time - you will end up with wrong values and backward diodes that have to be re-worked by hand. Every minute that you spend goofing off with the P&P system is a minute that you could have been manually placing boards.

In my world, even with experience, it can take 4-5 hours to get a PCB setup and ready to roll. It may take 30min per PCB to hand assemble them with an hour to set up. So, in 5 hours I could hand assemble 8 PCB's or I could setup the machine and have 1 PCB. If I only needed 4 PCBs the P&P is CONSUMING time, not saving it. The time starts coming back after making 10-12 PCB's for most of what I do. If you have a slow machine with other limitations that only allow half your PCB to be assembled - you may never save any time at all. There are, the human factors that can be helpful. It is tedious work making 10 PCB's where the machine will offload the tedium.

A person is a pick and place machine, if you are going to buy a mechanized version just make sure it is actually better than you are.

[IconicPCB]

Can a human pick and place for four hours straight?

Adhesive is a MUST  in wave soldering of SMD assembly.

Wave soldering is an order of magnitude faster than reflow process.

[girts]

The only reason to have it is to save time.
The only reason to have it is to save time.
The only reason to have it is to save time.

Time is money, money is the only reason why to go to job, no reason to waste time if nobody pays for it, only reason why we are making electronic stuff is money...

Definitely not right! (may be except United States). May be... some of us are doing all this just for fun?

Difference between child and adult is only one - cost of toys.
About OP promoted in 1st post - it's a toy. Like railway models, RC drones etc.
For educational purpose, to understand how steppers and machine vision works, what is SMT, how PCB's are assembled etc.
Nothing more.

Can a human pick and place for four hours straight?

Really no problems. With average speed of 500cph using manual placers mentioned some posts before.
Automatic PnP machine always has some limitations, manual PnP is much more flexible and covers all remaining components and demands. It is much better for prototypes or small batches.

Anyway, if somebody is going to load strip feeders into "automatic" PnP machine, reason is very questionable. Because it is much faster to place these components directly on PCB using manual placer.

[mikeselectricstuff]

Can a human pick and place for four hours straight?

Yes, but they probably wouldn't want to
 

Adhesive is a MUST  in wave soldering of SMD assembly.

Yes, but who does that apart from consumer stuff with lots of through-hole ?
Can you wave solder BGAs, or DFNs, or even 0.5mm QFPs ?

Wave soldering is an order of magnitude faster than reflow process.

Soldering speed is rarely important as production rate is usually limited by pick & place.
 And reflow ovens are pipelined.

[Koen]

I feel sorry for the bad reception this is receiving. It's nice, it's cheap, it's convenient and it's a solution to some. At the very least, it would be a great starting point for the OpenPNP beginners.

Double-sided would be easily solved by replacing the flat tray on which the PCBs currently sit by two rails.

Hang in there Furkan, it's nice and it has a market. A few clients sharing their experience with it would probably help most.

[rx8pilot]

Time is money, money is the only reason why to go to job, no reason to waste time if nobody pays for it, only reason why we are making electronic stuff is money...

Definitely not right! (may be except United States). May be... some of us are doing all this just for fun?

I think you are missing the point. If you pretend that you have an ideal pick and place system you will quickly learn that to save time, you have to make a lot of boards. If your machine is slow or unpredictable - you have to make a ton of boards to get your time back (if ever). If you are hoping for a fun project just to see if you can sort out a P&P system - great, this is perhaps a good option. I am not being difficult I am being realistic based on actual experience building a small business out of thin air and no money. I have been down this road with the blood, sweat, and tears to prove it. If I did it all over again and did not have the rare opportunity to get a commercial system - I would be looking at the manual assist systems or building my own manual assist system. That way I could earn enough money to buy a more or less professional system and then still have a nice manual setup as well.

How many hobbyists out there build so many PCB's that they would justify the complexity of introducing automation? It really has to be a LOT of PCB's before it makes sense from a time perspective (which is the only thing a P&P system does - it buys time). The only people I have met that can justify the time and money needed are the ones that have turned their hobby into a business. Even a small business like mine can justify it, but once you are in business - your perspective changes rapidly. Time becomes money.

All this is to say, that the market is tiny overall in my opinion. If I wound the clock back, I would probably buy something like this just for the challenge. I like messing with automation ideas for fun without regard for practicality. Not sure there are a lot of folks that have automation hobbies though, lol.

I feel sorry for the bad reception this is receiving. It's nice, it's cheap, it's convenient and it's a solution to some. At the very least, it would be a great starting point for the OpenPNP beginners.

Double-sided would be easily solved by replacing the flat tray on which the PCBs currently sit by two rails.

Hang in there Furkan, it's nice and it has a market. A few clients sharing their experience with it would probably help most.

I don't feel sorry for anyone.....the OP is getting a TON of feedback for free. That is hugely valuable.

[mikeselectricstuff]

I feel sorry for the bad reception this is receiving. It's nice, it's cheap, it's convenient and it's a solution to some. At the very least, it would be a great starting point for the OpenPNP beginners.

Double-sided would be easily solved by replacing the flat tray on which the PCBs currently sit by two rails.

Hang in there Furkan, it's nice and it has a market. A few clients sharing their experience with it would probably help most.

It's called constructive criticism. I at least am just contributing comments based on my experience of running a low-volume pick & place setup for several years.

IME most people who have not used a P&P system have unrealistic expectations of how useful it will be in practice, and little understanding of what factors make it more or less useful.

For the record I'd love to see a viable system with low entry cost and good software ( in particular that streamlines the setup process) ,  realistic specs ( i.e. forget 0201s) that can be upgraded as required to be more capable  (proper feeders, decent assembly area size), and not a dead-end that only serves a small sector of a small niche market.





 

[ar__systems]

Really no problems. With average speed of 500cph using manual placers mentioned some posts before.

It depends. If I have to place 100 boards, 10 parts per board, and BOM of 5 lines, sure. I'll do that manually in a couple of hours.
If I have to place 5 boards, 200 parts per board, and BOM of 50 lines, of which 40 are 0603 - no thanks. I'll load up an SMT, even if it take more time place. At least I'll know that every resistors is there it was supposed to be.