New Pick and Place design ideas

[mrpackethead]

and then there is just those who want to do it, becuase tey can.

[rx8pilot]

I spent some of Maker Faire on the OpenPNP booth this weekend. There was a LOT of interest in the machine being demoed, which was a pretty basic setup feeding from tape strips on the bed, but with top & bottom vision (and using vision to detect tape holes).

How does it deal with clear plastic tape?

[rx8pilot]

and then there is just those who want to do it, becuase tey can.

I wish I was in that camp. Currently struggling with business making hobbies a super low priority. All work, no play for me. 

[mikeselectricstuff]

I spent some of Maker Faire on the OpenPNP booth this weekend. There was a LOT of interest in the machine being demoed, which was a pretty basic setup feeding from tape strips on the bed, but with top & bottom vision (and using vision to detect tape holes).

How does it deal with clear plastic tape?

Not sure but there would probably still be enough contrast at the hole edge, and if the tape is clear you should be able to image the part

[Corporate666]

Personally I don't think a hobbieists needs a high cph machine or a super duper 100% reliable machine 24/7. Small starting one person businesses as yourself might as long as they don't sell >100 pcb's a week that is an ideal market for these cheap machines. If you have a real business like >200 pcb's a week, I do not understand why you do not outsource this to professionals where the whole process is certified and the chance on productfailures is kept to a minimum. Why bother yourself any longer, economicially it does not compute.
So there is this grey area between high production and manual production where these machines fit.

I would respectfully disagree with the above.

I think any PnP machine must be super reliable.  Even a relatively small PCB is going to have dozens of parts on it, and a multiple of that number worth of pads.  If you achieve 99% reliability, you're still looking at having something go wrong just about with every board.  I've placed countless thousands of boards with my PnP machines and my gut feel is that maybe once in 30-40 hours or operation I will need to do something (other than re-loading tapes, I mean).  For example... a nozzle gets clogged with solder paste and picks unreliably, or I notice parts are being placed a fraction of a millimeter off and I shut off and re-home the machine.  With some of my previous machines, I was lucky to get more than 20 minutes of reliable running before I had to futz with the machine, and that is when it became a deal breaker for me.  If the machine can't run for at least a matter of hours without problems, then it doesn't help me be more productive.  Loading boards in/out is one thing (takes literally seconds) but hearing the dreaded error beep and stopping everything else to fix the machine destroys productivity.  I realize that is a commercial environment but for home-brew stuff it's just as bad IMO.  If you can't be comfortable going away and doing something else, then you may as well put on some music and place by hand.

As for PnP'ing in-house vs outsourcing, I have done both and I am so so happy I have a PnP.  There are lots of benefits, but the main two are that I can make what I need when I need it without waiting, and second that I save HUGE on costs.  On the first one... I can run off 500 PCB's this week and another 300 next month or whenever I need them.  From start to finish it takes me a few hours to do the same work a board house would do.   The Chinese PCB assembly shop I used to use charged $0.0125 per SMD pad.  My boards are virtually all SMD.  With a couple of SOIC's, a MCU, some SOT's and 2 or 3 dozen passives, you are *easily* at 200 pads, so $2.50 per board for assembly.  So those 500 boards I can run off in a few hours would cost me $1,250.  And 200 pads is a pretty small board.  And from start to finish I am done by the afternoon, whereas with outsourced assembly, I am waiting days or 2-3 weeks if using a Chinese shop - and domestic assemblers are way, way more expensive than Chinese.  That 200 pad board would be more like $10/ea in my experience domestically, so that one run of 500 boards pays for half a good used PnP machine.

The other issue is cost.  I guess like most companies, I standardize on parts for dividers, biasing resistors, decoupling caps and even microcontrollers, regulators and switches.  I keep those parts loaded in the machine and I get economies of scale in using the same parts over multiple products.  If I needed 500pcb's that used a specific regulator, I would be buying a reel of 2,000 regulators to send to that assembler, and the other 1,500 just sit there waiting for my next board run.  Doing it in-house means I can buy partial reels from Digikey or Mouser as I need, and/or I can split that reel of 2,000 regulators across 7 different PCB designs that use the same regulator. 

Lastly (sorry for the length), it lets me do unorthodox stuff with my PCB's that assemblers might complain about.  I've pick and placed through hole stuff, silkscreened in paste to through holes, put resistors underneath other components like switches or between the pins of a connector on a high density board, violated board design rules when it suits me, had board outlines go through PTH component holes and such.  Since I know my process and my machines and I am machining the housings and all of that, I can do such "cowboy board design" stuff and get away with it when I need to solve a difficult engineering challenge.  But such stuff was always an issue when we outsourced boards and had to explain what we were doing and why and then listen to the grumbling from the PCB assembler about it.

[Kjelt]

I would respectfully disagree with the above.

Not a problem I love to be corrected by the people that do have the experience.
Still if I read your post I guess in my opinion you are not a small business or hobbieist so I would change my post with your info added to:

- Hobbieists still don't need a high cph machine or a super duper 100% reliable machine 24/7 for a few boards a week they can easily correct some mistakes the P&P made, every board is visually inspected anyway. Maybe then the only target audience for the low budget P&P machines that are discussed, but really they don't need them, it takes as much time to set the machine up as to do it by hand.

- small businesses with <100 boards a week are the target market for the medium priced PnP machines we are discussing ?

- (Semi) large businesses as yourself that do a couple of batches of 500 boards a week need to outsource or as you state own a pro machine because outsourcing is too expensive and inflexible.

So the semi-large business could best get a bankloan for $50k or more and buy a professional reliable machine with service contract and it earns itself back in a couple of years, right?
You can not buy a machine that we discuss frequently on this forum without full 24/7 service contract because if the machines is broke and you can't produce boards for a week you're loosing a lot of money.
Also it means buying more than just a P&P, you need an industrial grade reflow oven, you preferably also need a flying probe tester and a lot of small things that add up to the costs.

Pro boardhouses do more than just P&P they also do an end test of the pcb and devices with for instance a flying probe tester (also expensive machine).
They guarantee and insure their work, so for instance if the paste is bad or something else in the process goes wrong, they have to pay for it (perhaps some crappy board houses won't, but the good ones do).

[mrpackethead]

everyone needs to find there own space and figure it out for their particular situation.

Foruantly there are a lot of options starting at $1000 with Jason and open PNP and going to $$$$$$$$$$$ with state of the art 1M pph lines.

There is no universal answer.

[JuKu]

I spent some of Maker Faire on the OpenPNP booth this weekend. There was a LOT of interest in the machine being demoed, which was a pretty basic setup feeding from tape strips on the bed, but with top & bottom vision (and using vision to detect tape holes).

How does it deal with clear plastic tape?

Probably just fine. This example is not from OpenPnP, but shows how machine vision can work with clear tapes. Spoiler: clear tape is not very clear.  http://www.liteplacer.com/phpBB/viewtopic.php?f=11&t=7&p=87&hilit=Vision+transparent#p81

[Corporate666]

- Hobbieists still don't need a high cph machine or a super duper 100% reliable machine 24/7 for a few boards a week they can easily correct some mistakes the P&P made, every board is visually inspected anyway. Maybe then the only target audience for the low budget P&P machines that are discussed, but really they don't need them, it takes as much time to set the machine up as to do it by hand.

- small businesses with <100 boards a week are the target market for the medium priced PnP machines we are discussing ?

- (Semi) large businesses as yourself that do a couple of batches of 500 boards a week need to outsource or as you state own a pro machine because outsourcing is too expensive and inflexible.

I think that $10k buys a person a working and easy-to-use second-hand machine like the Quad's that I have.  Such a machine will be much more capable than anything home-made or designed for hobbyists, largely because all the bugs are worked out.  I can't imagine having to fix 10 mistakes on a board with 1,000 components on it.  It would take so long visually scanning the board and trying to find mistakes that it would drive me crazy.  And from experience, it takes 20 times as long to fix a board after reflow than to fix it before.  If a hobbyist is doing a few boards a week, I don't think any PnP is worthwhile at all.  The time required to set it up, load parts in, program it and all of that would make it not worthwhile at all, IMO.

So the semi-large business could best get a bankloan for $50k or more and buy a professional reliable machine with service contract and it earns itself back in a couple of years, right?
You can not buy a machine that we discuss frequently on this forum without full 24/7 service contract because if the machines is broke and you can't produce boards for a week you're loosing a lot of money.
Also it means buying more than just a P&P, you need an industrial grade reflow oven, you preferably also need a flying probe tester and a lot of small things that add up to the costs.

Pro boardhouses do more than just P&P they also do an end test of the pcb and devices with for instance a flying probe tester (also expensive machine).
They guarantee and insure their work, so for instance if the paste is bad or something else in the process goes wrong, they have to pay for it (perhaps some crappy board houses won't, but the good ones do).

There is a huge gap IMO between the hobbyist level you are describing above and the "get a loan, a reflow oven and a flying probe tester" level.  I have made hundreds of thousands of boards.  We paste them on a square of polypropylene plastic screwed to a desk.  I take a few old PCB's and tape them to the plastic, then tape the stencil on top on one edge.  Then just load a board in and paste it, then another.  We have pasted countless thousands of boards this way and it works great.  I can paste 500 boards in 20 minutes with zero problems (maybe once every month a pad doesn't get pasted, takes 10 seconds to put paste on with the manual dispenser).

For reflow, I use a toaster oven.  Not even with any type of controller.  Just a standard convection toaster oven.  We've reflowed countless thousands of boards in it without any problems.   We don't do any flying probe testing.  We may make a test program that's loaded into the device to test it or (99% of the time) just test it's function while still in the panel using the production program.  We program the chips with Pomona test clips.  Takes literally 3-4 seconds to program each board this way.

The failure rate is maybe 1 in 200 boards, and 99% of the time it's a component that shifted during reflow. 

Bad batches of solder paste and boards that fail test are just total non-issues for us.  As for the PnP machine itself, I've owned them for maybe 4 years and the only problems I have had was a nozzle clogged with solder paste (a maintenance issue), the occasional feeder that gets 'sticky', and the door interlock came off and caused the machine to alarm out (checked the manual, realized the problem, fixed in 5 minutes).

I spent maybe 3-4 days per month placing boards and I can do a couple thousand boards in that time.  If the machine breaks, I usually can fix it myself and if not, I don't need a service contract, just call the company and they will tell me what's likely wrong and what part I need, or I can schedule a service visit.


There's just a huge gap between the hobby guy who needs to do a few boards a week and a company that needs a full SMT line with service contracts and the like.  And that huge gap is served perfectly by used machinery.  That used machinery costs the same as what these hobbyist machines cost, but it performs a hundred times better at least.

I just honestly don't see any use for this home brew stuff.  It's sort of like building your own car.  You can buy a used car so cheap that it makes no sense to build one and while a used car can be a money pit, the amount you end up spending to build your own is well into the territory of buying a slightly used late model car that will do everything better.

Just my experience.  I am not anti-home brew/hobbyist.  I spent years chasing this dream myself.  I built my own PnP machine, I retrofitted a pro machine with bad electronics using my own motor drivers and software.  And I bought numerous used machines.  None of it makes any sense when you can get a great used machine for $10k that can support high hundreds of thousands if not millions per year in sales worth of circuit board assembly needs. 

[mrpackethead]

[quote author=Corporate666 link=topic=67734.msg948940#msg948940 date=1464217280

I think that $10k buys a person a working and easy-to-use second-hand machine like the Quad's that I have.  Such a machine will be much more capable than anything home-made or designed for hobbyists, largely because all the bugs are worked out.  I can't imagine having to fix 10 mistakes on a board with 1,000 components on it.  It would take so long visually scanning the board and trying to find mistakes that it would drive me crazy.  And from experience, it takes 20 times as long to fix a board after reflow than to fix it before.  If a hobbyist is doing a few boards a week, I don't think any PnP is worthwhile at all.  The time required to set it up, load parts in, program it and all of that would make it not worthwhile at all, IMO.

We went down a similar path, and bought some 2nd hand Yamahas.. a bit more than $10k, the entire scenerio ended up costing about $60k, but we have two machines,  and a good multizone oven,  and Lot sof feeders..      The cost recovery for me is about 6 months, but i'm able to do stuff in ways i was'nt able to before, so that also happens.. 

As for hobbiest.  I think its ok as well..   Its a hobby.. people do jigsaw puzzles for no reason other than to do them.   If they want to do PnP because its a hobby then surely thats ok..  People easily spend $10k a year chasiing a small white ball around the grass field.. What ever spins your wheels.

So the semi-large business could best get a bankloan for $50k or more and buy a professional reliable machine with service contract and it earns itself back in a couple of years, right?
You can not buy a machine that we discuss frequently on this forum without full 24/7 service contract because if the machines is broke and you can't produce boards for a week you're loosing a lot of money.
Also it means buying more than just a P&P, you need an industrial grade reflow oven, you preferably also need a flying probe tester and a lot of small things that add up to the costs.

Pro boardhouses do more than just P&P they also do an end test of the pcb and devices with for instance a flying probe tester (also expensive machine).
They guarantee and insure their work, so for instance if the paste is bad or something else in the process goes wrong, they have to pay for it (perhaps some crappy board houses won't, but the good ones do).

There is a huge gap IMO between the hobbyist level you are describing above and the "get a loan, a reflow oven and a flying probe tester" level.  I have made hundreds of thousands of boards.  We paste them on a square of polypropylene plastic screwed to a desk.  I take a few old PCB's and tape them to the plastic, then tape the stencil on top on one edge.  Then just load a board in and paste it, then another.  We have pasted countless thousands of boards this way and it works great.  I can paste 500 boards in 20 minutes with zero problems (maybe once every month a pad doesn't get pasted, takes 10 seconds to put paste on with the manual dispenser).

For reflow, I use a toaster oven.  Not even with any type of controller.  Just a standard convection toaster oven.  We've reflowed countless thousands of boards in it without any problems.   We don't do any flying probe testing.  We may make a test program that's loaded into the device to test it or (99% of the time) just test it's function while still in the panel using the production program.  We program the chips with Pomona test clips.  Takes literally 3-4 seconds to program each board this way.

The failure rate is maybe 1 in 200 boards, and 99% of the time it's a component that shifted during reflow. 

Bad batches of solder paste and boards that fail test are just total non-issues for us.  As for the PnP machine itself, I've owned them for maybe 4 years and the only problems I have had was a nozzle clogged with solder paste (a maintenance issue), the occasional feeder that gets 'sticky', and the door interlock came off and caused the machine to alarm out (checked the manual, realized the problem, fixed in 5 minutes).

I spent maybe 3-4 days per month placing boards and I can do a couple thousand boards in that time.  If the machine breaks, I usually can fix it myself and if not, I don't need a service contract, just call the company and they will tell me what's likely wrong and what part I need, or I can schedule a service visit.


There's just a huge gap between the hobby guy who needs to do a few boards a week and a company that needs a full SMT line with service contracts and the like.  And that huge gap is served perfectly by used machinery.  That used machinery costs the same as what these hobbyist machines cost, but it performs a hundred times better at least.

I just honestly don't see any use for this home brew stuff.  It's sort of like building your own car.  You can buy a used car so cheap that it makes no sense to build one and while a used car can be a money pit, the amount you end up spending to build your own is well into the territory of buying a slightly used late model car that will do everything better.

Just my experience.  I am not anti-home brew/hobbyist.  I spent years chasing this dream myself.  I built my own PnP machine, I retrofitted a pro machine with bad electronics using my own motor drivers and software.  And I bought numerous used machines.  None of it makes any sense when you can get a great used machine for $10k that can support high hundreds of thousands if not millions per year in sales worth of circuit board assembly needs. 

- small businesses with <100 boards a week are the target market for the medium priced PnP machines we are discussing ?

- (Semi) large businesses as yourself that do a couple of batches of 500 boards a week need to outsource or as you state own a pro machine because outsourcing is too expensive and inflexible.
[/quote]

I think that $10k buys a person a working and easy-to-use second-hand machine like the Quad's that I have.  Such a machine will be much more capable than anything home-made or designed for hobbyists, largely because all the bugs are worked out.  I can't imagine having to fix 10 mistakes on a board with 1,000 components on it.  It would take so long visually scanning the board and trying to find mistakes that it would drive me crazy.  And from experience, it takes 20 times as long to fix a board after reflow than to fix it before.  If a hobbyist is doing a few boards a week, I don't think any PnP is worthwhile at all.  The time required to set it up, load parts in, program it and all of that would make it not worthwhile at all, IMO.

So the semi-large business could best get a bankloan for $50k or more and buy a professional reliable machine with service contract and it earns itself back in a couple of years, right?
You can not buy a machine that we discuss frequently on this forum without full 24/7 service contract because if the machines is broke and you can't produce boards for a week you're loosing a lot of money.
Also it means buying more than just a P&P, you need an industrial grade reflow oven, you preferably also need a flying probe tester and a lot of small things that add up to the costs.

Pro boardhouses do more than just P&P they also do an end test of the pcb and devices with for instance a flying probe tester (also expensive machine).
They guarantee and insure their work, so for instance if the paste is bad or something else in the process goes wrong, they have to pay for it (perhaps some crappy board houses won't, but the good ones do).

There is a huge gap IMO between the hobbyist level you are describing above and the "get a loan, a reflow oven and a flying probe tester" level.  I have made hundreds of thousands of boards.  We paste them on a square of polypropylene plastic screwed to a desk.  I take a few old PCB's and tape them to the plastic, then tape the stencil on top on one edge.  Then just load a board in and paste it, then another.  We have pasted countless thousands of boards this way and it works great.  I can paste 500 boards in 20 minutes with zero problems (maybe once every month a pad doesn't get pasted, takes 10 seconds to put paste on with the manual dispenser).

For reflow, I use a toaster oven.  Not even with any type of controller.  Just a standard convection toaster oven.  We've reflowed countless thousands of boards in it without any problems.   We don't do any flying probe testing.  We may make a test program that's loaded into the device to test it or (99% of the time) just test it's function while still in the panel using the production program.  We program the chips with Pomona test clips.  Takes literally 3-4 seconds to program each board this way.

The failure rate is maybe 1 in 200 boards, and 99% of the time it's a component that shifted during reflow. 

Bad batches of solder paste and boards that fail test are just total non-issues for us.  As for the PnP machine itself, I've owned them for maybe 4 years and the only problems I have had was a nozzle clogged with solder paste (a maintenance issue), the occasional feeder that gets 'sticky', and the door interlock came off and caused the machine to alarm out (checked the manual, realized the problem, fixed in 5 minutes).

I spent maybe 3-4 days per month placing boards and I can do a couple thousand boards in that time.  If the machine breaks, I usually can fix it myself and if not, I don't need a service contract, just call the company and they will tell me what's likely wrong and what part I need, or I can schedule a service visit.


There's just a huge gap between the hobby guy who needs to do a few boards a week and a company that needs a full SMT line with service contracts and the like.  And that huge gap is served perfectly by used machinery.  That used machinery costs the same as what these hobbyist machines cost, but it performs a hundred times better at least.

I just honestly don't see any use for this home brew stuff.  It's sort of like building your own car.  You can buy a used car so cheap that it makes no sense to build one and while a used car can be a money pit, the amount you end up spending to build your own is well into the territory of buying a slightly used late model car that will do everything better.

Just my experience.  I am not anti-home brew/hobbyist.  I spent years chasing this dream myself.  I built my own PnP machine, I retrofitted a pro machine with bad electronics using my own motor drivers and software.  And I bought numerous used machines.  None of it makes any sense when you can get a great used machine for $10k that can support high hundreds of thousands if not millions per year in sales worth of circuit board assembly needs.
[/quote]

[rx8pilot]

How does it deal with clear plastic tape?

Probably just fine. This example is not from OpenPnP, but shows how machine vision can work with clear tapes. Spoiler: clear tape is not very clear.  http://www.liteplacer.com/phpBB/viewtopic.php?f=11&t=7&p=87&hilit=Vision+transparent#p81

Wow, amazing contrast. I know it's not super clear, but now I can see the proof. Excellent.

[Corporate666]

We went down a similar path, and bought some 2nd hand Yamahas.. a bit more than $10k, the entire scenerio ended up costing about $60k, but we have two machines,  and a good multizone oven,  and Lot sof feeders..      The cost recovery for me is about 6 months, but i'm able to do stuff in ways i was'nt able to before, so that also happens.. 

As for hobbiest.  I think its ok as well..   Its a hobby.. people do jigsaw puzzles for no reason other than to do them.   If they want to do PnP because its a hobby then surely thats ok..  People easily spend $10k a year chasiing a small white ball around the grass field.. What ever spins your wheels.

Sorry, it looks like the quotes got messed up - I apologize if I did not quote your post correctly.

I totally agree with you on the above, and I hope it didn't sound like I am poo-poo'ing someone who chooses to build a home brew PnP because it's interesting or fun for them.  Two areas of electronics I LOVE are lighting and motion, so I can totally see the appeal of doing something custom.

I am not sure why noobies gravitate to the incorrect aspects of pick and placing.  I did the same thing.  I remember when we bought our first real CNC machine.  I was enamored with how this giant hunk of iron and steel moved so fast and so accurately.  Way more accurately (I imagined) than a PnP needed to.  Yet this machine cost $50k... so I assumed a PnP would cost maybe $10k.  I looked into PnP prices and I was absolutely floored - it was a Philips (Topaz or Emerald - really nice back in the day) and IIRC it was in the $350k range (new).  I was sure it was an absolute ripoff due to low volumes of sales.  So I started looking at building my own.  The trick, I found, with PnP was not making an accurate machine... but rather that there are lots of processes a PnP needs to do that a milling machine doesn't.  Feeding components, vision/adjustment, rotation, placements against a flexing panel, and so much more.  And it has to do it really fast and with a super low failure rate.  It's all those little details that make it really hard.  I quickly realized I had been a bit (more than a bit) arrogant in thinking I could jump in and solve the problems that countless of thousands of engineers had worked on before me. 

But I definitely learned a ton tinkering with that stuff, so to each his own.

[mrpackethead]

For interest, here is some bottom images that are part of openPnp.

The camera is set up indentically for both parts, one is an 0805 ceramic cap, the other a 0.5mm pitch TQFP-64

These are using a $45 USB camera.

[forrestc]

  Etched steel stencils without frames and pasted on a desk using masking tape to hold boards in place can support multi-million per-year operations.  We can paste 500 boards (50 panels, 10 PCB/panel) in maybe 20 minutes. And we can place them in maybe 2 hours.  And reflow in about 30 minutes.  There is zero economy to be gained with paste dispensing.  The only exception is prototype boards or glue dispensing for wave soldering, but those are specialized applications.  And even then, given the time to program the machine, I've never had a case where it wasn't faster to just do it with the desktop pedal-operated dispenser.

In my specific case, programming the P&P machine also programs the paste dispenser.   The additional time necessary to set up a board for dispensing is about 30 seconds.  (Click on paste, create paste records from placement records, click on a couple of options, click done.  enter 8 numbers to tell the paste dispenser, and click on done).   

BUT, I agree... I'd rather stencil.   We very quickly move from dispense to stencil just because of defect reduction.... unfortunately, some boards just don't have high enough volume for a framed stencil.  Which is why I want to ask you about the process above.  Do you have some pictures and/or more details to describe?  Did you re-tape at each changeover?  Did you just use 'painters tape' or have you tried something else (aluminum foil or kapton tape, etc)?

I'm thinking that for my lower volume products I might be able to do something similar, but using a piece of 3/4 MDF as a base so we can avoid re-taping and re-aligning each time.   Make one for each board type, and keep them between uses.

[Kjelt]

There is a huge gap IMO between the hobbyist level you are describing above and the "get a loan, a reflow oven and a flying probe tester" level.

Yes I put three steps, it could be four or more, but I do agree.

I think that $10k buys a person a working and easy-to-use second-hand machine like the Quad's that I have.  Such a machine will be much more capable than anything home-made or designed for hobbyists, largely because all the bugs are worked out.

Again I agree, if you get it working it is really nice. There is just a risk that some parts will not be available in the near future anymore and you have to be innovative to come up with a solution. Or other problems arise since the machine was build a long time ago, such as the 0204 feeder problems (half step needed, only full step available) that rx8pilot now encounters. If there are new parts coming out of the factories you need to come up with new solutions, nothing insurmountable probably but still takes time so you ARE at that moment busy with building and adapting your own custom solutions for your machine 

I have made hundreds of thousands of boards.

Ok then I really really do not understand the following:

For reflow, I use a toaster oven.  Not even with any type of controller.  Just a standard convection toaster oven.

You must be kidding right? Have you at least temp. profiled the oven to see if it matches the reflow profile of the parts and paste?
The risk of microcracking, over or underheating or uneven temp zoning causing bad contacts is huge with a standard oven. Most can not even handle the temperature curve needed (too steep).
Rohs leadfree paste is even more difficult.
Even I , a hobbieist that does at most 50 boards a year have a 2nd hand semi professional reflow oven bought cheaply from a business that outsourced (new price $2k to $3k) and I do not even make money of this or a business it is just hobby.
I can't believe you run a business with this kind of equipment, I don't know what to say, for a few k$ you can have a really nice good controlled reflow oven, is the profit of the business that bad or do you just not care if your products could fail in a few years due to microcracking or overheating? Maybe I am misinformed or overexaggerating, but for what I have read this is really a big no go concerning reliability and product lifetime.

If the machine breaks, I usually can fix it myself and if not, I don't need a service contract, just call the company and they will tell me what's likely wrong and what part I need, or I can schedule a service visit.

If you have stock to compensate production outage that is not a problem. If you have a customer that paid you to deliver the order at a penalty in the contract of each day delay that is different ballgame.

I just honestly don't see any use for this home brew stuff.  It's sort of like building your own car. 

I agree, a hobbieist does not need it, nice to have, they don't run production numbers that ratify such an investment, still there are hobbieists that own equipment better than some pro's do, think about photography and such. So if there is a market why not serve it, hence the chinese machines we see so many on this forum.

[rx8pilot]

I think that $10k buys a person a working and easy-to-use second-hand machine like the Quad's that I have.  Such a machine will be much more capable than anything home-made or designed for hobbyists, largely because all the bugs are worked out.

Again I agree, if you get it working it is really nice. There is just a risk that some parts will not be available in the near future anymore and you have to be innovative to come up with a solution. Or other problems arise since the machine was build a long time ago, such as the 0204 feeder problems (half step needed, only full step available) that rx8pilot now encounters. If there are new parts coming out of the factories you need to come up with new solutions, nothing insurmountable probably but still takes time so you ARE at that moment busy with building and adapting your own custom solutions for your machine 

Some clarification on this point.....Every single part of my machine is still supported and available from PPM and a few other sources. Nozzles, belts, rollers, PCB's, etc. That was one of the most critical considerations when I chose to buy it - I had no time to buy an off grid system where I was totally on my own. I have repaired a ton of things on my own, machined special fixtures and wrote some interface scripts - but PPM has been CRITICAL to my success, both in parts supply and troubleshooting setup problems.

Case in point: The PPM owner called me after seeing a post I made on another forum about the 2mm indexing problem. He showed me how the software supports a 'double pickup' with 4mm index feeders to accommodate 2mm pitch tapes - DONE! I had found a bunch of feeders on eBay that are precise enough for 0402/0201 but only index 4mm. The solution was there and I didn't even know it. PPM unlocked the mystery.

So far, I am just a pinch under $10k, and a few weeks of labor and learning.

[Corporate666]

BUT, I agree... I'd rather stencil.   We very quickly move from dispense to stencil just because of defect reduction.... unfortunately, some boards just don't have high enough volume for a framed stencil.  Which is why I want to ask you about the process above.  Do you have some pictures and/or more details to describe?  Did you re-tape at each changeover?  Did you just use 'painters tape' or have you tried something else (aluminum foil or kapton tape, etc)?

I'm thinking that for my lower volume products I might be able to do something similar, but using a piece of 3/4 MDF as a base so we can avoid re-taping and re-aligning each time.   Make one for each board type, and keep them between uses.

I don't have any pics handy but I can describe.

We have a "pasting station" at work.  It's just a standard 4 foot desk.  I buy 24" x 24" x 0.5" pieces of polyproplyene.  I use PP because it's just about the cheapest plastic... polyethylene is cheaper but too slippery for the tape to stick.  Anything less than 0.5" thickness means the plastic will likely be quite warped before you get it.

I have tons of old PCB's lying around of various sizes/shapes/thicknesses.  I just put the panel to be pasted on the plastic sheet, then place some PCB's around it, making sure I am tight against the corners so the board can't move.  Then I use blue painters tape to stick them down.  My stencils usually have 4-5" of free space around the actual openings, so I make sure to build a big enough 'platform' to support the whole stencil.

I secure the stencil in place with a strip of painters along the top side.  Then scoop a blob of paste from the container onto the top of the stencil and silkscreen the board.  I have perfected a "flick" maneuver at the end of each pasting stroke which keeps the paste on the blade after pasting.  Then I have a little scrap piece of steel and I scrape the paste off the pasting blade back onto the top of stencil in a blob.  Then lift the stencil up a little (with blob of paste at top - but it's close to the pivot point so no problem), grab the board using a pair of pliers and set it aside, put another one in and drop the stencil back down - I also check the panel I just pasted to make sure it's all good.  I always get my boards gold plated since the contrast between gold and solder paste is easy to see, making any un-pasted pads stand out.

Sometimes, maybe after every 20-30 panels, the piece of tape at the top of the stencil gets worn out from using it as a hinge and registration between stencil and board gets off.  I just peel off the tape, clean that area of the stencil with acetone and re-apply.

When I am done, the leftover paste goes back in the tub, and I immediately clean off the stencils with acetone from a spray bottle, a toothbrush and cloths. 

I have several of these 24x24 plastic panels and I tend to have 2 setups on each one - one for the front and one for the rear of the PCB.  The only trick for doing the second side of the PCB's is the height of components from the first side, but I can stack PCB's or buy plastic shim stock from Mcmaster to get the height perfect for 2nd sides of PCB's.

Maybe once every 1 or 2 thousand panels pasted, I will take my fixture/holder off the plastic panel, clean it off with acetone and re-build the fixture.  But that's just because I am a neat freak and hate seeing dried blobs of paste on the plastic panels.

I don't do much pasting myself anymore, but I have shown maybe 6 or 7 people here how to do it, and within a couple of boards, they get it down perfect.

Nowadays I get my stencils from my PCB fab in China, but I have used Stencils Unlimited - $150 for the frameless ones I use.  I have also used OSHstencils kapton prototype stencils on this same setup.  Works fine too.

[Corporate666]

You must be kidding right? Have you at least temp. profiled the oven to see if it matches the reflow profile of the parts and paste?

I haven't temp profiled the oven.  I bought one that was recommended years ago by someone who was selling some sort of controller and tested some ovens to see which worked best.  But if you buy the right kind of oven, most of them are fine.

The risk of microcracking, over or underheating or uneven temp zoning causing bad contacts is huge with a standard oven. Most can not even handle the temperature curve needed (too steep).

Based on what data?  The boards reflow evenly with very little time difference between the first and last part of the board to reflow.  The trick is to use an oven with multiple elements and it has to be a convection oven.  I don't know of anyone who has done any serious testing or analysis on using toaster ovens, so the above concerns are greatly overstated.

I can't believe you run a business with this kind of equipment, I don't know what to say, for a few k$ you can have a really nice good controlled reflow oven, is the profit of the business that bad or do you just not care if your products could fail in a few years due to microcracking or overheating? Maybe I am misinformed or overexaggerating, but for what I have read this is really a big no go concerning reliability and product lifetime.

Yes, you are misinformed and exaggerating.  I have a bench top reflow oven.  It sucks.  It heats unevenly, it isn't consistent and it produces inferior results to the toaster oven.  I also had a 10 zone Heller 480V reflow oven that I got just about for free that was a year old and with less than 100 hours on it.  The amount of time it takes to set up, tweak, adjust and mess with it wasn't worth it.  We get overall superior results using a toaster oven.  We've reflowed hundreds of thousands of boards spanning 15 years, offer lifetime warranties and out of the maybe 30 failed units I get back each year, in just about every case the problem is physical damage or water damage.  If you're not using new paste that you store in a fridge and throw away the unused amount after you open it... and if you don't store all your components in temperature and humidity controlled environments... then focusing on reflowing with a toaster oven is being chicken little.

If you have stock to compensate production outage that is not a problem. If you have a customer that paid you to deliver the order at a penalty in the contract of each day delay that is different ballgame.

You're not listening to what I am saying...  there is a huge huge area between a guy doing hobby work and a large company with a full SMT line(s).  The people and companies that occupy that gap are not running their PnP machines every day.  They do not do contract manufacturing.  They are not doing JIT manufacturing.  Most companies aren't doing JIT manufacturing.  Service contracts are not needed.  All that is needed is the ability to get service when it is needed.  I have that.  I do have customers that we have to deliver to with built-in penalties, and service contracts or the lack thereof don't add anything or harm anything in that regard.  I am challenging your worldview that there are 3 categories of people - hobbyists, big companies running SMT lines, and "everyone else" who should be outsourcing.  I know numerous companies in my area and elsewhere who are making from hundreds to maybe several thousand PCB's per year and they are doing it just like I am doing it and making a good living at it. 

I agree, a hobbieist does not need it, nice to have, they don't run production numbers that ratify such an investment, still there are hobbieists that own equipment better than some pro's do, think about photography and such. So if there is a market why not serve it, hence the chinese machines we see so many on this forum.

I am not suggesting that hobbyists shouldn't buy PnP machines because they don't need them.  My position is that all of these hobby-level PnP machines are vastly inferior to buying a used machine that is still supported and was built as a pro level machine. 

This is a Quad 4C placing parts.



There is one on eBay now for $10k.  A more realistic price is probably $5-7k for the machine - that's what they sell for in working condition with feeders.  Not everyone can fit a machine like that in their basement, but other than for people who just don't have the space or who just like to tinker with stuff because they enjoy it, I don't see the value in the home-brew PnP machines. 

[mrpackethead]

My $90 oven + Sparkfun Controller ( no longer avaialable ) works brillantly, and yes, i've profiled it and its every bit as good ( result wise ) as the expensive multizone oven we have in teh factory.. And it takes about 10 seconds to start, versus 45 minutes.   It uses just 4kW, vs 12kw. 

It just does't work for doing 10000's of boards.. 

[mrpackethead]

Yamaha YX100's with no feeders are going for about $10-15k, in various states of repair..    Just be prepared for some work to get them back to to "sweet"

[rx8pilot]

There is one on eBay now for $10k.  A more realistic price is probably $5-7k for the machine - that's what they sell for in working condition with feeders.  Not everyone can fit a machine like that in their basement, but other than for people who just don't have the space or who just like to tinker with stuff because they enjoy it, I don't see the value in the home-brew PnP machines.

That machine does not have feeders. Even if you get a great deal on feeders, it would be at least double. That also assumes the machine is in great shape and works out of the box - mine was advertised as READY TO GO which was not even close to being true. Since it was only $5k with about 55 feeders, I did not care. On top of that, it was local so I only spent $50 renting a trailer for shipping.


I use a JEM310 batch oven that is very good for what I do. Easy programming and very consistent. Personally, I would be scared of the toaster oven setup if only as a precaution. My products are in a critical role and operate all over the world in remote locations. A huge amount of my work is reliability and longevity. I have read a lot about the soldering process and it scared me enough to get a tightly controlled oven. No way for me to know without lab testing, which I have not done. The profile I use very closely matches the recommendations.

[forrestc]

You must be kidding right? Have you at least temp. profiled the oven to see if it matches the reflow profile of the parts and paste?
The risk of microcracking, over or underheating or uneven temp zoning causing bad contacts is huge with a standard oven. Most can not even handle the temperature curve needed (too steep).
Rohs leadfree paste is even more difficult.
Even I , a hobbieist that does at most 50 boards a year have a 2nd hand semi professional reflow oven bought cheaply from a business that outsourced (new price $2k to $3k) and I do not even make money of this or a business it is just hobby.
I can't believe you run a business with this kind of equipment, I don't know what to say, for a few k$ you can have a really nice good controlled reflow oven, is the profit of the business that bad or do you just not care if your products could fail in a few years due to microcracking or overheating? Maybe I am misinformed or overexaggerating, but for what I have read this is really a big no go concerning reliability and product lifetime.

I agree with Corporate666.... but I'm going to vote misinformed.

Microcracking generally is only an issue with moisture sensitive parts which haven't been stored correctly.   

As far as reflow profiles go, most boards just aren't that picky.   Most components just aren't that picky.   SAC305 is a bit of a pain, but still, fairly wide window with the newer paste chemistries.  As long as you are able to move from active paste to reflow within a reasonable amount of time - and then cooling at the proper time, you're golden.   Too slow and things don't want to melt because the solder has re-oxidized (after being cleaned) - too fast and the paste doesn't have time to work.   But most paste handles a wide range of slopes.   Of bigger concern is cooling time for grain formation.  But even that has a wide range of working slopes.

[forrestc]

I use a JEM310 batch oven that is very good for what I do. Easy programming and very consistent. Personally, I would be scared of the toaster oven setup if only as a precaution. My products are in a critical role and operate all over the world in remote locations. A huge amount of my work is reliability and longevity. I have read a lot about the soldering process and it scared me enough to get a tightly controlled oven. No way for me to know without lab testing, which I have not done. The profile I use very closely matches the recommendations.

I use a batch oven myself - a toaster is a bit ghetto even for me...  for similar reasons you mentioned.    If it failed I'd probably replace it with something similar.  But not a high end conveyored oven (I have a low end one here collecting dust - anyone interested?)

However, having rebuilt this oven like 3 times over the years I've had it, and reengineered the profile countless times,  I've discovered that things are not as critical as one would be led to believe for most things.   I have built a relationship with a fair number of people in the solder and paste industry, and they have all generally told me the same thing:  As long as you get everything hot enough roughly at the same time, and don't take too long getting there (and pick the right paste), everything should be good.   Cooling is FAR MORE CRITICAL than the heating.   Which is sad since everyone focuses on the heating side, and it's not the critical part.

The nightmares people talk about in the industry are usually designs which have temperature sensitive small parts on a board right next to something with big thermal mass.   Some of the giant ceramic BGA packages are the worst.   Then the problem is that the ceramic package is still getting up to temperature while the poor little temperature sensitive part in a SOT23 is melting.

-forrest

[JuKu]

BUT, I agree... I'd rather stencil.   We very quickly move from dispense to stencil just because of defect reduction.... unfortunately, some boards just don't have high enough volume for a framed stencil.  Which is why I want to ask you about the process above.  Do you have some pictures and/or more details to describe?  Did you re-tape at each changeover?  Did you just use 'painters tape' or have you tried something else (aluminum foil or kapton tape, etc)?

I'm thinking that for my lower volume products I might be able to do something similar, but using a piece of 3/4 MDF as a base so we can avoid re-taping and re-aligning each time.   Make one for each board type, and keep them between uses.

I don't have any pics handy but I can describe.
...

And on this page, you can find the video, from which I learned the same technique: http://www.beta-estore.com/rkuk/order_product_details.html?p=13.
It really is as easy as Corporate666 says and the video makes it look.

[Kjelt]

I started out that way by placing other pcbs around the target pcb.
Since I don't make all my pcbs right away but only when and how many I need, I wanted a more flexible approach.
So i stick them to a piece of cut laminate floortile (see picture old).
It works but there was room for improvement which I found by milling the piece of laminate floortile with the correct depth and dimensions of the pcb and in one or two corners
some extra space to easily remove the pcb after (see picture new). \
It takes a bit more setup time (around half hour or so) if you have a normal decent mill but you have it available whenever you want/need it.