SMD Assembly line within 30k budget

[MR]

Don't even think to get a Mechatronika (used) I went through that (there are a few machines available in Europe every now and then ridiculously overpriced and spares are also sold as expensive as gold), even new I would not buy anything from them.
I end up studying the machine, writing my own software for it and after moving to another country (due to family reasons) building another machine myself rather than transferring the Mechatronika machine because it's just not worth it, the shipping cost would exceed the effective value of the machine, while the machine was actually very expensive.

If I could go back in time I'd rather go for a used well known brand (the guy who advised it to me just went as silent as Mechatronika itself when confronting them with all the bad things of that particular machine; they even serviced and changed the motors which just end up adding more problems to the machine rather than improving anything, eg. oscillating Y axis on the machine due to a loose pulley - those are no user errors).
A panel of 1200 components randomly stopped to pick and place due to pick up issues, rarely a panel completed without pick up error (due to poor feeder design), it required the user to babysit the machine.

Also I would insist on testing the machine, if something is not documented and the company wants you to fly to them forget about them. As mentioned I built my own machine in the meanwhile there's absolutely nothing that cannot be documented or explained via phone.
I admit designing good feeders is not easy (I did that and it took quite some time), but if a company is selling that commercially and after a service cannot get things right - and even after receiving another complaint then they should be avoided.

Also be careful with recommendations - you have your products - your requirements and those might not necessarily map to other "small scale" smt-line owners.

[jayx]

For used machines, have a look on https://smtbroker.pl/, https://tools.adoptsmt.com/equipment/search/, but likely the best way would be to find a company which is upgrading their machine (or even better the entire line) so you can see it working.
I can advise about Autotronik, likely it's better than Neoden, at least seems that it has better vision system for detecting components, also can change nozzles automatically. Used machines are not expensive, but feeders are and used ones are rare.

[andrea.longobardi85]

To everybody who advised against Neoden PnP...you were right!!!

Yesterday I have seen the Neoden K1830 in action at one of my client facility.
I cannot judge the expertise of the PnP operator but the experience was a total mess.

Every single board coming out of the PnP had to be reworked manually before going into reflow making it unusable into a production line. The board was a mix of 0603, SOT23, and QFPs.
First I though it was a setup issue, but the problems where always different on different components.
In the end it was a repeatability issue, I have seen a QFP which was placed on spot for couple of times at a certain point being paced shifted by one pin and the PnP program was the same!
Bottom line I think this is a no go for me on Neoden and the local sale representative told me that Neoden 9/10 are not much better than K1830...

[MR]

Not taking actions for Neoden but from my own experience during my DIY bringup experience.

There can be several reasons for this.
* Vacuum pressure too high / low
* Wrong nozzle (too small, too big might be an issue)
* pickup distance not correct, correlating with vacuum pressure that the component was flying through the air for some time and sit on the needle the wrong way (there's only a corridor for correction +/- certain degrees, if the component is too misaligned the correction might mistake the sot323 legs and cause completely wrong rotation)
* the cherry of it all - certainly not good optical recognition,
I made my optical recognition pretty good, some SOT323 are different and make it hard to determine the actual location of the part.
That said I'm taking care that components are picked up +/- 30° (max, usually it's just within 5-10°) the optical recognition will do the rest and I have absolutely no issue with SOT323 anymore after fixing my software and HW issues on my custom machine.

I have some variations for component detection passives, QFN, SOT232, SMA.. practically I an use QFN for SMAs but it won't take care about the shape of the part, SMA would knock out components if they're wrong (I never needed that feature actually)

The most important part during the entire pick and place process is to have a good stencil and stencil printer. (I probably should underline this part and write it bold)

On the Mechatronika I also had different issues per board, that was always related to their shit feeders, eg. couldn't advance random full reels sometimes. Those unreliable owners Leslaw Gajda, Dariusz Pieczerak from Poland never cared either (and I always gave them a window to fix their issues). Sometimes 2-3 stops during the pick and place process telling the user couldn't pick up part it made it absolutely necessary to babysit their MX80 because it might have stopped after 10 - 20 - 70% or whatever pure randomly (unless someone walks through all their feeders and unwinds them regularly). Also some hidden calibration settings in their firmware made it hard for customers to fix issues themselves (basically they made customers nicely depend on them just to get some extra cash from them). There's an array of key/value number pairs exposed in their application that's what I'm writing about.

[SMTech]

The more you understand the process & the machines, the more you will appreciate certain features in both hardware and software. The low end is very low end.

MR is quite correct, there are a host of things that could be causing really shitty placement, they can include some modicum of setup/tuning by the user as well as all the quality of build/design issues on the machine itself and its algorithms. Its possible the machine they sure could do better with more effort, but that effort is not easy to do and not easy to determine  how best to apply.

Higher end machines have encoders on axes, sensors on vacuum (not to mention high end servo control boards/modules) and in some but not all cases measure component height &/or placement force this lets them compensate for things like board warp or components that are taller or shorter than programmed (e.g a different manufacturer of the same value capacitor).

It is certainly possible to get a used line for your budget it just needs you to be in the right place at the right time with the funds ready. The best deals will be found buying direct from a user, lots of people scrap old machines when they upgrade. If you want support or slightly refurbished, then you might need a bugger budget.

[48X24X48X]

Usually it's pretty much a configuration and setup problem when you have many parts not placed properly. This includes many component height mismatch (even thin and thick film resistor can be different let alone others), vision not set correctly like the light intensity which in return affects the shape of the component under inspection, pick up location not accurately set, inaccurate nozzle size usage, and plenty more.

There're few more and much seasoned company in China like HWGC, Bovi and Boreytech and has been in the business much longer than Neoden has been in making larger machine (don't count their earlier days on those smaller machine but count from their K180 and later machine). Take a trip to China and visit them yourself, you'll see it yourself. There's also many 2nd hand machine available especially in Shenzhen if Japanese or European machine is what you are looking for.

[Mangozac]

It is certainly possible to get a used line for your budget it just needs you to be in the right place at the right time with the funds ready. The best deals will be found buying direct from a user, lots of people scrap old machines when they upgrade. If you want support or slightly refurbished, then you might need a bugger budget.

This is the best advice. OP sounds like they don't need this urgently, so I would get in touch with the local distributors of high end machines and let them know that you're looking for a secondhand line. They will gladly put you in touch with their customers who are upgrading to new machines.

My recommendation is to look for Yamaha/Philips/Assembleon placement machines <20 years old. They are built like a tank from only high quality parts and you can buy brand new clone feeders for them from China. I get the impression these are less common in Europe though so you're probably best to look for what the equivalent is. Probably something like a MyData/Mycronic?

Buying a used machine from China is high risk. I wouldn't do it unless you really know what you're doing and definitely not for your first machine.

[loki42]

I've seen a bunch of used Philips / Yamaha gear used in Europe.  I was going to buy a Philips line from Poland but ended up getting the uic because it had better odd form handling but the complete line with spare parts etc would probably have been a better deal.  There's a good dealer in Austria too that had good prices.

[jmelson]

Not taking actions for Neoden but from my own experience during my DIY bringup experience.

Yes, I had really frustrating problems with my used Quad QSA30A (made by Samsung, similar to their CP30).
It was spitting SOME components onto the board with bad placement.  It all happened so fast I couldn't see what was going on.
Finally, after having this issue for a few months, I saw the pattern!  It was ONLY on parts that I had defined into the parts library as entirely new profiles.
There was a menu button that I was afraid to touch, that said something like "change placement parameters". Well, when I touched that button, a screen came up with all the vacuum and Z stroke timings!  I compared the default settings with those of parts that were successfully placed, and they were WAY off.  I changed those settings and all was good.
Jon

[andrea.longobardi85]

Don't even think to get a Mechatronika (used) I went through that (there are a few machines available in Europe every now and then ridiculously overpriced and spares are also sold as expensive as gold), even new I would not buy anything from them.
t's just not worth it, the shipping cost would exceed the effective value of the machine, while the machine was actually very expensive.

I have got a quote for a Mechatronika M10V at 30k+ with feeders. On the spec side seems to fulfill my need but I haven't seen it in action...
I think after your comment I will be cerefull on this machine...

[MR]

Don't go for that Mechatronika I built my own machine in the meanwhile - it's easy to build a machine like that with one single pick and place head forget about that.

Take your money and run to another vendor with it.

30.000$ is absolutely overpriced, resellers only list it because they earn a good amount of commission from that junk.
It will pick and place and if you're lucky with your parts you will not run into problems, but there are problems with certain parts which are on the market.
The owners are retired - there's no motivation to fix anything, but it's still nice to count money coming in at that age.

If you can make them fix our machine and all the problems we had with it with *their* software (the machine is now decommissioned and idling in Austria because I moved to Taiwan (I built my second machine in Taiwan with the knowledge I gained from the Mechatronika machine, but with multihead) - the MX80 now works okay with our own software and can handle the problematic parts but that's not our job - neither your job as a potential customer).

They probably tell you that we have no idea about hardware, software and mechanical engineering and that we damaged the machine ourself, it's scam.
Their website has no valid SSL certificate, no new product updates for over 8 years either.
Just look at the picture of the top camera (which is supposed to be sold as new and judge yourself, some other customers also confirmed that their cameras are modified like that - even on the M10V (picture attached)).
Also after their scam upgrade the speed did not improve, it was just as slow as before due to wrong settings (1700cph max) their top of the line product.


Overall this has happened quite a few years ago, we are over that - I see the MX80 which we own as our product now we can make anything out of that since we have full control over everything now. But before when they just cut the support it's a shock you invest that money and get no help to successfully build your products. And by time we figured out that the problems are on their side - misconfigured machine and handicapped software.

I offered them even to work for free to fix their shit software (even under NDA). It was not possible to pick and place a single panel (1700 components without babysitting). Sometimes 1 component made the process stuck, sometimes 3-4 times and all nicely randomly.

I have 2 days left before going back to Taiwan, maybe I'll disassemble their feeder box again and take some photos .. it's just a disgrace what they have built.. well for 600$ a box would be okay but 4000EUR/20 lanes (there are only 2 cheap stepper motors + solenoids inside!). It's just a push feeder under the hood, no sprockets and it will cause issues with some plastic tapes.
From a business point of view well done... they're in the range of something customers are willing to put on the table but the value is just not there.

my time is more important than this Mechatronika pick and place project, especially since our DIY works better (luckily). I consider the Mechatronika junk as a hobby for the future again.

[andrea.longobardi85]

Has anyone experience with small Vapor phase oven like this:
https://www.vaporflow.eu/

I have read good feedbacks compared to bench convection reflow. Price is reasonable and there is a new model coming out which can reflow 440x235 cm PCB.

[Chad.Wagner]

My Neoden 4 has allowed me to stay in business.