Open Placer A New Affordable PCB Assembly Machine

[mikeselectricstuff]

For strip holders, the super-cheap solution is just thin double sided sticky tape onto a smooth surface. You could print 4mm guidelines and an index grid on the surface ( e.g. make it from a PCB)
Not suggesting this as an ideal solution, but useful as a way to have a minimum entry cost, with strip trays as an upgrade path.

[ar__systems]

For strip holders, the super-cheap solution is just thin double sided sticky tape onto a smooth surface. You could print 4mm guidelines and an index grid on the surface ( e.g. make it from a PCB)
Not suggesting this as an ideal solution, but useful as a way to have a minimum entry cost, with strip trays as an upgrade path.

Yeah, it is super cheap, but it is also super lousy. Probably anybody who wanted a strip feeder tried this It is only marginally better than nothing.

[mikeselectricstuff]

For strip holders, the super-cheap solution is just thin double sided sticky tape onto a smooth surface. You could print 4mm guidelines and an index grid on the surface ( e.g. make it from a PCB)
Not suggesting this as an ideal solution, but useful as a way to have a minimum entry cost, with strip trays as an upgrade path.

Yeah, it is super cheap, but it is also super lousy. Probably anybody who wanted a strip feeder tried this It is only marginally better than nothing.

I do it all the time whan hand placing, the only issues I've had is the occasional plastic tape with holes in the pockets that can make parts stick. With alignment guides ( maybe use a PCB with holes, so you can pin the strips for alignment) I think it would be useable, even if only as an "overflow" for odd parts when you run out of better tape holders.
 

[girts]

Another strip feeder idea - 
http://www.nozzles4smt.com/StripFeeders_c_26.html

[ar__systems]

It is ok for larger machines. The price is outrageous though

From my experience it is not perfect. Since the side groves are fixed width, and there is no tape support from the bottom, it can't work for both plastic and paper tapes. Either those grooves are to small to accommodate the paper, or they are too large to hold plastic tapes. 

[l0wside]

Adding my 2ct. I am selling small sensor modules (PCB about 2x2in, roughly 30 components), low volume (three-digit-numbers). When I started, I thought it to be a good idea to do the PnP myself and got myself a China PnP machine.

After about two weeks of fiddling, I gave up. The machine was great at placing 0603 passives from cardboard tape. Plastic tapes did not work, as the push feeder simply ripped the feeder holes apart. With the extremely limited space in my basement, I could only barely reach the north side, where the feeders for the larger (and therefore more cumbersome) components are, so fixing a broken cover tape (not an uncommon issue) was a lesson in gymnastics. Placing .5mm components worked so-so. High components (electrolytes) are a no-go, as the feeders do not support them (and Murphy makes sure that when the head picks up the second cap, it will run over the first, already placed, cap). The Chinese SW would stop the PnP process once the bottom camera did not recognize a component.

In the meantime I have sold the machine to someone who already runs two larger PnP machines and wanted a cheap one for sample runs. He claims it works fine for him - gotta believe it.

Bottom line is: with more time, more space and more experience, I might have fared better. But after two weeks of fiddling, my patience started to wear out. I could have placed a lot of samples during these two weeks. Be ready to invest a lot of TLC before the machine saves you more time than it costs you, and expect to do constant babysitting.
If you invest 2k$ (or 2kEUR) in a student from the college next to you, he can populate a lot of boards for you...

[rx8pilot]

I had some major challenges with college students assembling PCBs. High marks in class, but still could not figure out polarity on caps or an anode on diodes. None of the 3 that where with us for a summer internship could manage to solder anything without cooking it to death.

Not saying they are all that way, but boy can they do a lot of damage in a short period of time without the same babysitting you would need with a pick and place system.

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

[IconicPCB]

Furkan,

Please check Your private messages.

[mikeselectricstuff]

I had some major challenges with college students assembling PCBs. High marks in class, but still could not figure out polarity on caps or an anode on diodes. None of the 3 that where with us for a summer internship could manage to solder anything without cooking it to death.

In every assembly subcontractor I've ever been to, almost all the assembly workers are middle-aged women.
For routine soldering work, you need someone with good manual skills, not technical knowledge or youthful enthusiasm

[rx8pilot]

In every assembly subcontractor I've ever been to, almost all the assembly workers are middle-aged women.
For routine soldering work, you need someone with good manual skills, not technical knowledge or youthful enthusiasm

The local (Los Angeles) assembly shops that I have visited can be described the same way. Dominated by females 30-50 years old.

[Furkan]

Dear all

The Indiegogo campaign is almost over. We will continue our support from Dallas, TX and TURKEY afterwards. We will keep spare parts and some devices in stock.

There were several critics about Open Placer, which we evaluated intensely. Today we discussed with Kerim and I guess we need to clarify some issues. The majority of the critics were constructive but we do not agree with some of them. And I have to explain some points about double sided assembly and max. component height.

First of all, I have to point out that we were mistaken about double sided assembly. I answered very quickly without thinking much. The distance between the nozzle and the PCB holder is a little more than 22 mm. The 6 mm component height limit was due to the feeders. You can easily place a PCB over 2x3 mm = 6 mm thick Neodmium magnets. The PCB holder is made from steel so the magnets are quite strong to hold the PCB. You can place 6 mm to the bottom side and 6 mm to the top side and still have more than 6 mm distance to bring new components. Today, I would like to apologize about my comment that Open Placer does not support double sided assembly.

Open Placer does support double sided assembly.

Secondly, we never guaranteed 402 and 201 assembly, but we received so many critics about them. During the campaign we tried to prepare a new nozzle for 402 assembly but at the end the video about 402 and 201 assembly was done with our smallest nozzle with 0.8 mm outer diameter. The inner diameter was very hard to measure, but it was around 0.5 mm. The fifth 201 component that couldn't be assembled in the video was due to the impact that removed it from the feeder while picking the fourth component. The picking speed of the nozzle can be reduced to decrease the magnitude of the impact. This can be configured by users.

Thirdly, the bottom camera views a region of 35 mm x 35 mm and its resolution is 640 x 480. So, 201 components occupies only 20 - 25 pixels. It is very hard to determine the occurrence and position with the current camera.

Next, the paper and plastic strips do not have completely straight surfaces. The small stretch on their surface enables tight assembly. The feeder structure of Open Placer works with both thick paper strips and thinner plastic ones, at most of the cases.

Next, the proposed feeder alternatives are not applicable with OpenPnp as much as we now, since the strip holes are blocked by the pieces placed to fix the strips. The setup time for that design should be quite high since the holes are missing which is required to determine the correct position in one of the axis. Using fiducials, small magnets to fix the PCB (for double sided assembly) and defining the components in feeders with OpenPnp seems to be the fastest way to initialize PCB assembly.

We have to invest more to build a machine that supports 201 assembly. At the beginning of the campaign we were expecting proper 603 assembly would be enough for majority, so we didn't make 402 tests. We were busy preparing the machines for May delivery, so we couldn't spend enough time for the critics. We are almost sure that Open Placer supports proper and reliable 402 assembly but we couldn't make enough tests about it. We will wait for the feedback from the ones who purchased it.

The prices will definitely increase after the campaign (The discounts from the list price will be lowered), since we made extremely small profit from this campaign with a lot of effort. I guess we managed to introduce Open Placer to the world.

I even don't like the poor advertising of our campaign but it was our first campaign at Indiegogo and we are engineers. Our marketing capabilities were very limited. We listened the critics and managed to announce our campaign to larger crowds every day.

I would like to thank every one of you for your support and constructive critics.

Furkan
The Open Placer Team

[rx8pilot]

Good luck with the endeavor.

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

[jmelson]

Adding my 2ct. I am selling small sensor modules (PCB about 2x2in, roughly 30 components), low volume (three-digit-numbers). When I started, I thought it to be a good idea to do the PnP myself and got myself a China PnP machine.

After about two weeks of fiddling, I gave up. The machine was great at placing 0603 passives from cardboard tape. Plastic tapes did not work, as the push feeder simply ripped the feeder holes apart. With the extremely limited space in my basement, I could only barely reach the north side, where the feeders for the larger (and therefore more cumbersome) components are, so fixing a broken cover tape (not an uncommon issue) was a lesson in gymnastics. Placing .5mm components worked so-so. High components (electrolytes) are a no-go, as the feeders do not support them (and Murphy makes sure that when the head picks up the second cap, it will run over the first, already placed, cap). The Chinese SW would stop the PnP process once the bottom camera did not recognize a component.

Very useful to hear all this.  I can compare it to my used Philips CSM84 machine.  It will handle parts up to 6.5 mm tall (if you could command the travel path of the head it could do more, but they don't give you that option.)  Of course, 6.5mm parts would logically only come on wider tape, never seen an 8 mm tape with more than about 1 - 1.5 mm part height.  (I'm thinking of SOT23 transistors and some diodes.)  I've almost never seen torn sprocket holes, but then the Yamaha feeders use the return spring to advance the tape.  If something binds up, the feeder just stops advancing the ratchet.  That does happen from time to time.  The only feeder problem that drives me nuts is 0.1uF caps in paper tapes.  These have inconsistent cover tape glue, so there are spots on the tape where it won't advance.  I put one, or even, two, C-clamps on the tail of the component tape to give it more pull.  Ugly, but it works!

Anyway, I do babysit the machine,  I am applying paste to boards, and putting them in the oven, so just constantly circulating around the area and keeping an eye on the machine.

Jon

[sokoloff]

We have to invest more to build a machine that supports 201 assembly. At the beginning of the campaign we were expecting proper 603 assembly would be enough for majority, so we didn't make 402 tests. We were busy preparing the machines for May delivery, so we couldn't spend enough time for the critics. We are almost sure that Open Placer supports proper and reliable 402 assembly but we couldn't make enough tests about it. We will wait for the feedback from the ones who purchased it.

Congrats on getting to where you are!

I would think long and hard about whether the investment to get this class of machine (and class of buyer) down to 0201.

How many people have actual money in their hand, ready to give it you, and will do so but only do so if the machine does 0201? My wager is that not many people would even say that they are in that situation and, of the people who say they're in that situation, fewer than 1/3rd of those will end up buying if you deliver 0201. The other 2/3 will forget, decide to not buy anything, balk at the higher price, find another excuse why the new machine isn't good enough for them, etc.

[IconicPCB]

Sokolof gives good advice.

Offer a solution which has been clearly identified; forget about the professional tire kickers.

They go telling you what they think, gives them a sense of self importance.

Gives You a feeling of inadequacy and drives you to offer a solution in search of a problem.

A typical trap so many engineers fall into.

I wish You good luck in Your endeavors,

[l0wside]

I could not agree more. For your target market (somwhere in the blurred zone between ambitious hobbyist and small business), 0603 is likely standard, 0402 happens occasionally, and 0201 completely out of scope. Better get it right for your customers than over-optimizing for something exotic while losing the big picture.
Double-sided PCBs, however, are a must.

[ar__systems]

I agree. Going below 0603 rarely makes sense unless you design for extremely space restricted or high frequency applications, in which case you will probably be looking for an SMT machine in a completely different price range.

[asmi]

I agree. Going below 0603 rarely makes sense unless you design for extremely space restricted or high frequency applications, in which case you will probably be looking for an SMT machine in a completely different price range.

I always use 0402 unless I can't (typical reasons include voltage rating/capacitance of caps, or power dissipation rating of resistors) because they are incredibly cheap and take up very small space. I only use 0201's under BGA as nothing else will fit in there.

[ar__systems]

Price of the passives components is of zero concern if we are talking about hobbyists.

So yes, maybe you have to use small packages but I'm betting majority of hobbyists don't work with BGAs.

[asmi]

Price of the passives components is of zero concern if we are talking about hobbyists.

Not if you want to buy them on reels. 0402 are 10-40$ for a 10k reel, 0603 are significantly more expensive on a per-unit basis. Considering that my typical PCB has 200-300 components (most of them being passives), reels get used up rather quickly.
And smaller package means PCB can be made smaller, which is a big deal when you design 6+ layer boards.

So yes, maybe you have to use small packages but I'm betting majority of hobbyists don't work with BGAs.

The reason not many hobbyists work with BGAs is because it's more expensive. I suspect that majority of those who can't afford to work with BGAs are not going to spend $1k+ on a PnP machine.
On the other hard, I totally would, as laying down 0201 is the most tedious part of assembly. Placing 0402 and above is rather quick even manually. I've already invested several thousand $$ into this hobby, so spending another $1-2k for PnP is not that big of a deal. But so far there's nothing in that price range that would meet my requirements, so I'm seriously considering building my own machine.

[ar__systems]

I don't understand what you are arguing with. First of all, the price of 0603 is about 25-50% higher than 0402, and is on the order of $0.003 per piece. So no, not considerably. I'd be surprised to see a design where cost of resistors dominates or even takes non-insignificant part of the total cost of BOM for the entire device.

Second, why do you keep talking about yourself? I think we already established that this product does not suit you

[asmi]

Second, why do you keep talking about yourself? I think we already established that this product does not suit you

This is the way forum conversations go. People express their own opinions.

[KE5FX]

Second, why do you keep talking about yourself? I think we already established that this product does not suit you

This is the way forum conversations go. People express their own opinions.

For what it's worth, I agree with you -- 0402 is the dividing line between serious industrial PnP equipment and machines that are useful for prototyping, small-scale production, and/or that are candidates for use by hardcore hobbyists. 

Unfortunately, 0402 falls on the "absolute requirement" side of that line.  0402 components must be handled reliably by any PnP machine that someone is charging money for.  I wouldn't want to spend extra money for 0201 support, though, because it will take a lot of extra money.

[IconicPCB]

"Charging money for " is a nonsensical argument.

It is more appropriate to claim : "... Ya pays ya money and ya gets ya thrills..."  as demonstrated by a variety of goods and services available

You want 0402 performance don't hope to get it for bargain basement price. Go and spend ten times more, live in a different environment.

[ar__systems]

This is not a "small scale production machine", it is barely a prototyping machine. I.e. if I want to build 10 prototypes of my boards with 160 parts that I want to machine place, this is already 1600 parts. Loading that many parts in strips is a non-starter, there is not enough space on the machine to hold the PCB and parts. Running any kind of production on this is out of question.