Writing an article about pick and place machines has been a suggested video/blog post by many people. In fact, this may be the single biggest suggested topic that I have ever received. This article will be the first part in a series on Assembly Line Basics.
A standard PCB assembly line consists of a printer, pick and place machine, and oven. Image source: ONESMT
While this is something my girlfriend would likely answer “you shouldn’t!”, at least hear me out first. My name is Kyle, and I started an electronics product design firm, MicroType Engineering, in the summer of 2017. Before that, I had been in the same line of work as a freelancer for a couple of years. Around three years ago we expanded into an in-house SMT assembly line to support some of our existing design clients. We purchased secondhand equipment: A Dek horizon 265 printer, Assembleon Opal Xii pick and place machine, and a Heller 1500 oven. We further expanded to an Ebso selective soldering machine and a Heller 1800EXL oven. Everything was bought through PFI of Florida (Not a sponsored post by any means, but I could write an entire article just on how amazing of a company they are.)
When starting on this endeavor, I had 0 experience in this line of work. Sure, I had hand assembled my share of prototypes for design jobs, but I had never touched a pick and place machine, or an industrial oven. So to say I went in to this blind would be an understatement. The people who I hired also had no exposure to any of this equipment before (most had 0 electronics experience at all!). While in hindsight this wasn’t the best way to go about things, it also means that to get going I had to learn, research, and test virtually every single aspect of running an assembly line. I’m sure I have well over 1000 hours spent just in doing research on things we had to troubleshoot.
Throughout this process I learned virtually every aspect of running an assembly line that you possibly could. While I would certainly not consider myself an expert in any one specific aspect, I do think I have a solid knowledge of nearly every single facet of the trade. I also host a popular YouTube channel and Discord server.
Our current SMT assembly line: Dek Horizon printer -> Assembleon Opal Xii -> Heller 1800EXL
With the resurgence of American manufacturing, and an ever growing hobby field in electronics, the idea of assembling PCB’s at a production scale is becoming more appealing for people and companies. There is not a lot of good information out there on the topic. Worse even, there is a ton of misinformation or flat out lies around.
It is commonplace nowadays to hear people and companies who spend thousands and thousands of dollars on a machine, and end up never being able to get them to reliably work. There is also a growing trend of home built, 3D printed pick and place machines. Unfortunately, people tend to compare pick and place machines with 3D printers. It is much easier to have a cheap and functional printer than it is a pick and place machine. The amount of technology that goes into a pick and place is incredible, and it requires excellent software support as well.
It is just not feasible to expect to buy a few thousand dollar pick and place, and expect it to assemble quality boards at scale. It is usually much faster to just assemble by hand! My goal with this article is to use my experience to help guide people and companies who may be considering doing PCB assembly in-house.
Whenever you are dealing with large amounts of money being spent, people always (understandably so) want to defend their decisions. While I will be speaking on a lot of topics as a matter of fact, these are my opinions. I base them primarily on my personal experience, and lots of first and second hand experience hearing of stories from others. There is loads of good reading to be had on SMTnet and EEVBlog all about horror stories of different pick and place machines. Just because one person had a great (or bad) experience, it doesn’t mean that is necessarily the norm. Nor does it still make it a good decision for most people/businesses.
Many experienced members in the industry would argue (successfully I might add) that a pick and place machine isn’t the most important entity in an assembly line. It is however, typically the most expensive. It is also the centerpiece of an SMT line, and the one piece of equipment that people focus on when deciding to start ramping up assembly of PCB’s. While you certainly can hand print solder paste on boards, and use a hotplate, hot air gun, or toaster oven to reflow boards, you can’t really hand place thousands of SMT parts on boards in a day!
So for this article, I will be focusing only on pick and place machines. A lot of the info, however, translates to other machines, and starting an assembly line in general. In future articles and videos, I will dive into other important aspects and machines within an assembly line.
Hobbyists
Likely the largest single group out there. They assemble boards for fun, as a side project or hobby. They may still sell products, but it will be on the order of < 100’s per month, and not nearly enough to support them full-time.
OEM’s
Original Equipment Manufacturer. These companies sell products full-time. They may build the physical products in-house, or use a third party. They rely on the income from selling their products to stay in business. Depending on the product, they are likely selling 1000’s+ units per month.
CM’s
Contract Manufacturer. OEM’s hire CM’s to build products for them. They also could be an OEM for some internal products that they sell. They are full-time, depending on this income to stay in business. Typically they will build more units than most OEM’s, as they provide services for many companies. A “small” CM would be one that builds < 10K unit’s a month, while more commonly a CM will be approaching 100K units.
Less than a few thousand dollars. Rarely has conveyors and are just setup in batch mode. Mostly plastic parts (God forbid some using 3D printed parts!). Usually use non-standard feeders, likely plastic. Machines will be quite lightweight, with a lot of vibration. Software tends to be very limited, you may have to manually program for a board instead of importing from CAD.
Examples:
Few thousand dollars, up to the $100K range new. These are typically Chinese machines, with a lot of rebranding of the same machines. They are most known for their terrible software, with a mix of Chinese and English writing. They will start to use professional feeders, the best tend to use Yamaha CL’s. Some machines at this level will have cast iron bases, but others still will be prone to vibration.
Examples:
MyCronic MY300 Pick and Place Machine. Image source: MYCRONIC
These are the big boys. They start at $100K new and can easily get up into 7 figures depending on features and feeders included. They start with CL type mechanical feeders on the low-end, with the majority using modern, advanced electric, servo driven feeders that have customizable feed rates and accelerations. All machines will take industrial 3 phase power and require a compressor with clean and dried air. They are mounted on a solid, heavy, cast iron frame. There is a large used market, with older machines easily serviceable and parts available.
Examples:
Before I dive into specifics that may make buying a machine a good idea, I’d like to mention some basics about pick and place machines in general.
Hands down, the most frustrating part of running a pick and place is dealing with feeder issues. I would guess that well over 80% of the time our machine stops running it is because of a feeder issue. Normal culprits would be:
I’m sure I could write an entire article on how much of a pain dealing with feeders really is! Our pick and place machine uses Yamaha CL feeders, which is probably the most common feeder (along with Juki’s) in the industry. They aren’t amazing by any means, but, they are a decent all around mechanical feeder. If we were to upgrade to a new machine, it would be mandatory for it to have better feeders, or it really wouldn’t be able to be considered an upgrade. They are really that important.
Nowadays the best feeders are electrically servo driven. They are much more expensive, but you can fine-tune everything about them, and for some parts they are virtually required. A lot of machines (Juki being the most well known) also allow you to use some electric and some mechanical within the same pick and place machine. This gives the flexibility to use mechanical for most parts to save money, and then electric for parts that tend to jump, or are a fine pitch. It is also a rule of thumb that on a new machine, you typically will spend as much in feeders as you will for the pick and place machine itself.
Much like the above point with feeders, your machine is only as good as its software. This is especially true for CM’s, who spend a lot of time doing changeover’s for new jobs. Software tends to be the biggest downfall for Chinese machines, as it is more of an afterthought. The best machines will supply you with an offline programming tool. This enables you to program the pick and place on a separate machine, greatly increasing efficiency. Our Opal xii has what I would consider “decent” software. We don’t often have to fight with it to get it to work, but it just lacks features that modern machines have.
Notice that I did say “Mostly”! It is a helpful metric, but it is imperative to not treat it as the golden spec. Firstly, companies love to make bold exaggerations of what their pick and place machine can actually do. You will never run a machine and meet the rated CPH value that they claim, as they did it under perfect conditions, with a perfectly tuned board. Secondly, while CPH is certainly important for how many boards you can run in a day, how reliable the machine is can be more important. If you are stopping every five minutes to fix a feeder or adjust the program, the CPH will be almost useless for the machines overall efficiency.
After all that introductory info, we are finally able to start answering the main question of this article! I will break down this section into different types of customers, and what, in my opinion, they should/shouldn’t buy, and why.
This may be the case for most hobbyists out there. They have a full-time job, and love to tinker with and design electronics on the side. This person may already sell some products, or wish to start in the future. They don’t have large aspirations of building out a company around their products, they just do it for fun. Most critically, this group also loves to tinker with, and play around with 3D printers and other gadgets, and are great at fine-tuning them.
If this sounds like you, then by all means, go ahead and buy a Hobby-level pick and place. They key is to not have goals of having the machine be cheaper than using a dedicated CM, or to pay for itself. There will be nearly constant battles with the machine. Once it is dialed in, it may be able to run fairly consistent, but once you change to a new board, the fine-tuning begins again! I have often considered building a custom pick and place machine, because it does seem like so much fun. I would never, ever, consider using it for work related jobs, it would purely be a learning project.
This person is similar to the previous, except they want to turn this into their full-time job. They have a product or two that they want to market and sell. If all goes well, they will be able to expand out and build a business around it. They need to be smart with their money, and get the most out of it. They can not afford to be limited to only being able to ship x amount of product this month because their assembly line couldn’t manage it, or broke down.
In this case, the choice is obvious: sub-contract the work out to a dedicated CM! Let them handle all the headaches, you focus on designing and selling your product! If you want to be hands on with the process, you can perform the incoming QC inspection and board testing from the CM. To save costs, using an overseas CM can be a great option. Otherwise, there are plenty of local CM‘s nearby that have the benefit of no language barrier, are able to talk over the phone, and typically much higher quality. If (hopefully when!) you get to the point to where it becomes a full-time job, you then become a “proper” OEM, and fall into one of the next two categories.
This company has been around for a while, currently uses a CM to manufacturer their products. Now they are considering bringing manufacturing in-house either to save costs, better control quality, or better IP protection. The key for this company is that they only sell a few products, likely less than 5-10.
They may be a candidate for a prosumer machine if they do not sell too many products per month/year. Since they won’t be doing a lot of changeovers, they can handle the tedious task of getting the pick and place machine setup, and dealing with the sub-par software. It is crucial to watch the cost of prosumer machines vs professional used machines.
It always upset’s me when I see a company that has spent upwards of $100K on a brand new prosumer pick and place machine, that struggles to reliably place parts efficiently. They could have instead purchased a used professional machine with feeders, for the same price or less. The professional machine will be faster, more reliable, have more spare parts available, and have a fantastic resale value. Some of these prosumer machines cost over $100K, not including feeders!! So as long as the total cost for the prosumer machine stays reasonable low (~$10K range), I think that they can be an OK purchase.
It isn’t my goal to single out any one company in this article. It’s just that there is a group of new(er) companies that try to undercut the pricing of the true professional machines out there, and claim they are comparable. They’re not even in the same ballpark. You don’t have to take my word for it either. Call any successful OEM or CM and ask what pick and place machines they use. Then, compare the used price of their Juki/Samsung/Assembleon/Mycronic/etc on the open market vs a brand new prosumer machine. It is a tough pill to swallow, how the current pricing structure of some of these brand new machines makes sense.
This company is identical the the above one, except they sell many different products, potentially 20+. They want to bring production in-house for the same reasons listed above.
I would highly advise them against going with anything other than a professional machine. The high amount of changeovers require the software and feeders to be rock solid, which you don’t get with a prosumer pick and place machine. To save money, they can purchase a used Assembleon Opal/Topaz/Emerald machine with feeders for < $30K. They will blow away any prosumer machine, even those that cost over double their price brand new.
This can be an OEM that wishes to start taking on outside orders. It could also be (like us), a design shop that wants to be able to service their design clients long-term.
No question, a professional pick and place machine is required. Further, a newer vintage machine with really good software will be ideal. Most CM’s will have multiple lines, with each line having at least two pick and place machines on them. Without a doubt, being a CM is the most difficult type of PCB assembly due to nearly constant job changeovers.
Successful US based CM’s typically will have one or two lines dedicated to lower quantity, high-mix products. Those lines use a really flexible machine (MyCronic, Europlacer, Juki KE-x020) that can place nearly all parts, albeit more slowly. They will then have a separate line(s) that are for high quantity, low mix products. This will have a standard high-speed chipshooter -> flexible machine. A CM that has at least two separate lines like this, allows them to handle virtually any type and size of job that comes their way.
If you’ve made it this far in the post, I thank you! My goal with this article was to use my experience to help explain the different pick and place machines out there, and which may make sense for you. While their certainly isn’t a “one size fits all” approach, I do think the general guidelines I’ve outlined hold true most of the time. I also hope that some of my rather strong opinions on the topic are at least backed up by sufficient data I provide. Please let me know any thoughts you may have on this topic in the comments.
MicroType Engineering is a full service electronics and mechanical product design firm. We offer turnkey support no matter how far along in the design process you are. We have full schematic capture, PCB design, firmware, mechanical design, and prototype/small-batch assembly services. Please reach out to learn more!
Securing the Future: Our Commitment to ITAR and CMMC Compliance Elevating Aerospace and Defense Technology…
Elevating Standards: Our AS9100 Certification Journey Paving the Way for Aerospace Engineering Excellence At MicroType…
New Advanced Circuit Board Assembly Equipment Introduction At MicroType Engineering, we pride ourselves on staying…
The Usage and Function of TVS Diodes in Circuit Protection Introduction Hey there! Today, we're…
This article, The DC Bias Effect in Ceramic Capacitors, is based on the above YouTube…
I recently began a livestream series on the MicroType Engineering YouTube channel. It is called…