ec-placer or other manual PnP

[electrolust]

I'm in the market for a manual PnP with video assistance. First choice would be Fritsch MP-904 but it comes at "if you have to ask" prices.

My anti-requirements are: no DIY and no feeders. No one-off manufacturers (eg SMT Caddy).

I rather like the ec-Placer. It for sure has shortcomings but at asking price they don't seem fatal ones.

The dual crosshairs on the camera is great as it saves in cost. I'm not convinced you can place BGA or QFN components reliably without split vision though? What say you? Even their own video and product page shows doing this in copy-paste mode, not freehand. Anyone have actual experience with this device and can comment? This is my #1 reason to get a PnP, as otherwise I'm quite happy just using tweezers and soldering as I go.

[wraper]

If you don't have butter fingers and your sight is OK, you fill do the job faster with fine tweezers rather than with this contraption.

[electrolust]

Oh. i just noticed a comment on the ec-placer official youtube video. copy-paste mode misplaced a TSSOP component. TSSOP! Come on!


(skip to 2:51. URL has the timestamp but forum youtube parsing seems to drop it)

I have reasonably good dexterity and a good microscope and good headband magnifier. I have no problems hand placing 0201. I'm just worried that I won't properly get 'er dun for no-lead packages. I've not built a board yet with those packages but some things I'm working on now, the parts are only available VSSOP, QFN, BGA, etc. You think it's safe enough to rely on the silkscreen?

[wraper]

On the video component placement is not that accurate and he completely misplaced TSSOP. IMHO the only advantage here is camera. If you do placement under microscope, this advantage is lost. If I worked at the speed this guy does, I would assemble 10 times less boards.

[wraper]

I'm just worried that I won't properly get 'er dun for no-lead packages.

Maybe get a vacuum pickup with foot pedal for BGA and other stuff hard to hold with tweezers.

You think it's safe enough to rely on the silkscreen?

I make alignment markers on copper layer for fine pitch components when possible. Silkscreen is not nearly accurate enough, especially from cheap PCB services.

[mairo]

Hi electrolust,

I am a bit of a "fanboy" of manual/semi auto PnP machines  . I have access and I am working with Mechatronika MM600, LPKF ProtoPrint S and the Eurocircuits ec-Placer (have done very little work on the ec-Placer). They all I would say work very similar. The camera options on these machines (later two) is a good feature, but if your eyes are OK you do not really need it as you will position the parts faster without the camera.

I would suggest to also have a look at: MAN-32, check the website, it comes with different options/accessories:
https://www.esdshop.eu/man-32-manipulator-ccd-full-hd-vision/

Another place to look at is for used machine from the above brands, or also look at Dima, Essemtec brands as they also make/made good manual machines - eBay, auction places - LPKF machines com eoften as well as Essemtec (these are almost as good as the Fritsch) and Dima.

With the MM600 (I mainly use this machine with no particular reason why as I could also use the other 2 as well, just used to it..) I can place ~100-600 parts per hour based on complexity. If I would to work lets say an 8 hour day with ~1-1.5 hours of break I could place ~1500-3500 parts and listen to the Amphour at the same time  otherwise I will get crazy  - this is by the way with zero mistakes and zero lost parts . However a components set up will need to be done, which could range from few minutes to 5-6 hours based on complexity. This is to set up all components to be ready for placement so when you sit to place, you only place. I also use 3rd party free software to help with the placement. There is a link to it in this forum somewhere and I think Kicad may be able to do similar thing now as well.

QFNs, 0402, etc., pretty much any SMD part is doable. For BGAs  I use separate  manual optical placer that is much more accurate. Not placing exactly is not that critical as you can correct later, or the reflow process will do this for you if the placement is not that much off.

[MR]

For the Mechatronika Part I would take care, the owners are almost retired and the closer the moment comes the less they care about their customers. I guess a manual PP should be pretty easy and not much could go wrong there, but still the support will completely disappear within the next years.
Also they're from Poland, they're charging full western prices on their units while the cost is significant lower (don't get me wrong there are good companies in Poland, but this is definitely no good one).

Due to my experience with fixing their automatic pick and place machine after their direct service (which also revealed the real cost of all the parts and talking with their suppliers; eg nozzles 1 piece 100 EUR+, while you can get >10 machined quickly from prototype service for that price and even they will make profit on that).
We even built another machine and are able to compare, before looking at them look at China or real western companies, it will give more bang for the buck.

If I would need a manual pick and place machine I'd rather buy a 3d printer (well I have a few already) buy some parts and look at some of those systems and design it myself within 1-2 weeks. The software is the interesting part not the hardware.
SMT caddy is using a raspberry PI you can see that in the video, it's all no rocket science.

[Jeroen3]

Relevant.
https://www.eevblog.com/forum/manufacture/getting-started-with-the-ec-placer/

[MR]

SMD Caddy = raspberry PI + camera + touch screen



https://twcn.rs-online.com/web/p/raspberry-pi-screens/8997466 (around 65 USD)

[mairo]

For the Mechatronika Part I would take care, the owners are almost retired and the closer the moment comes the less they care about their customers. I guess a manual PP should be pretty easy and not much could go wrong there, but still the support will completely disappear within the next years...

Do not worry MR, customer support is quite good from them even when I am on the other side of the world over many years - have the machine for ~ 12 years and has done over 100k parts on it, still looks like brand new with no issues, it is a work horse. With the LPKF system on the other hand, some parts (not that critical) from the feeders disintegrated over time...

With building one yourself, it may look simple, but if you have not used one before, you will need to go back and forth many times till you optimize it , so in 2 weeks? ... you will be still waiting for the 3d printer parts .... One important aspect is how comfortable the machine is and if you look at the Mechatronika machine, it is the most comfortable over long use from the 3 models I have due to the larger hand rest that it offers, simple, but extremely useful.

Long ago there were prices on the website for the MAN-32 and I think it was under 1800euros, but not 100% sure as  now the machine looks abit 'updated'.

[MR]

I wonder what kind of support do you still need for a manual pick and place machine?
Due to the age of the owners 61/65 (and the history with how they solved real technical issues) I would not count on that company. Spare parts might still be shipped of the shelve for some time I guess.
That's not discriminating, it's just like it is.

They never fixed support for plastic tapes with their feeder box design (and seems like they're using that for your MM600 too), also other users noticed that and they're either applying workarounds (squeezing multiple tapes into the slots) or put the stripe into a fixed holder inside the machine. That issue can jam up the entire feeder block. You might not have experienced that since you might be lucky and don't use plastic tapes too much - it doesn't happen with all of them but it was noticeable with many on our side. We're using those components for a half decade not a single professional SMT house ever had issues with them before.

Instead of fixing that they have said that it's our fault (while the issue was confirmed by other users).
SOT323 are packaged in such plastic tapes, a bunch of them jammed up the entire block on our side 16 positions affected, and sometimes they advance themselves due to that idiotic gravity driven stub retraction system.

After disassembling the feeder boxes the situation was very clear and we were impressed how simplistic and error prone their design is.
If their price would be divided by 10 it might be justified but they're trying to charge like over 150 EUR per lane plus overpriced services (which is good for their distributors which add like 25-50%), and that for you can good single feeders already.

http://www.dsod.pl/materials/MM600manualGBv3.pdf
page 8 - there are 2 stepper motors inside, solenoids are pushing a stub into the corresponding tape and one of the stepper motors is pushing an entire PCB plate forward to advance the tape. Very simplistic and cheap design sold for over 2K EUR (resellers go for over 3K)
On the top there's a small plate which is holding the tape down. The whole thing is made of weak/cheap aluminum parts, some screws were more like lying in the pocket instead of holding things together due hollowed out screw holes.
The other stepper motor is for taking off the plastic tape. 2 Steppers for 16 lanes (= very economical and error prone design)

[mairo]

Even their own video and product page shows doing this in copy-paste mode, not freehand. Anyone have actual experience with this device and can comment? This is my #1 reason to get a PnP, as otherwise I'm quite happy just using tweezers and soldering as I go.

electrolust, I do not have a suitable board that I can try this feature on it. I cannot promise when, but will let you know once I try it.


MR, the MM600 is a semi automatic machine not manual . For he support, they do have other machines, its not just the MM600 they offer.

[SMTech]

You really do have a bee in your bonnet about them don't you, I think they have bigger issues frankly, they are squeezed from all sides 2nd user proper machines and the Chinese ones with little to no unique selling points to choose them over the alternatives.

 The OP couldn't care less about feeders for the manual P&P they are looking for, and once you remove feeders from the equation there is nothing left to support if it arrives in good working order. Manual P&Ps have their uses, we still have one, we don't build things on it anymore, but it does get used for partial population of low volumes or putting on things that have come in inconvenient packaging. The model we have is an "MPI SM250", its old but I have seen them on eBay before, this thing even has feeders designed for those mini-reels you don't really see anymore.  I don't think a camera is all that useful except in certain situations, fine pitch QFPs etc, 0402 & below maybe although I would imagine your workrate would be quite low. Certainly nice to inspect placement after tho'

I would point out tho' that as a CEM we built all sort of things for all sorts of people and almost all of them default to 0805, there's really no need to go below that unless you really need the space, 0805 and 0603 are easy to work with manually, 0.5mm pitch isn't terrible either (trickier the bigger they get) and if you can stick to QFNs they are super forgiving and if you put them down slightly wonky you can inspect them from the side give them a gentle nudge if needed and they will sort themselves out . I would agree split vision like you find on a rework machine is the best way to reliably place devices where you cannot see pads/contacts, although as suggested above , very accurate marking in the copper probably work for devices whose dimensions you can rely on, maybe less so for those evil PCB LGA modules.

[MR]

You really do have a bee in your bonnet about them don't you, I think they have bigger issues frankly, they are squeezed from all sides 2nd user proper machines and the Chinese ones with little to no unique selling points to choose them over the alternatives.

They aren't getting squeezed they practically gave up already. Just be aware that they will shut down the business soon or just let it fade out slowly (but potential customers shouldn't be fooled, so they shouldn't be on anyone's list for new equipment)

[Styno]

We have and Essemtec Expert manual PnP that we used for 20 years and still use for prototyping, it looks like this (but without the semi-automatic stuff and monitor):


We have placed components ranging from 2.5 by 2 cm subassemblies to 0402 passives and 0.4mm pitch ic's with it.

[electrolust]

Vision assist is a hard requirement. I can place one part easily by hand without a $2-$10k vacuum pen gantry. The problem is without vision, I'm under the microscope or wearing the headband and crouched over. This is not tolerable for very long. And I'm getting older ...

I re-watched the ec-placer video and realized an absurdity of copy-paste mode, for leaded chips.

On the copy side, the stencil is quite floppy and not mounted well. You have to place leaded components in there quite precisely due to this sloppy and floppy issue. On top of that, the cutouts are for the footprint, not the leads. So there is additional slop and accuracy is overly operator dependent. In the video at 2:29 he takes great care to locate that chip on the stencil. At that point, the 'paste' step is a complete and utter waste of time! He could have just placed the part carefully, without vision assist, directly on the board! It turns out he also mis-placed it anyway! There's also no economy of scale here. It's not like computer copy-paste, which maintains the copy so you can paste-multiple. In the ec-placer it's 1:1. ec-placer has actually wasted time and effort here. And EC seemingly lack awareness of this defect.

For balled components it seems a lot better, but still operator dependent.

For no-lead packages (QFN) copy-paste seems very useless. You're left to aligning (equalizing) the extended footprint to be equal on all sides after the chip is placed. That said, even the Fritsch has this same issue, and isn't aided by a video crosshairs. In the Fritsch sample video, the operator never once looks at the video screen except at the very end to verify placement of a large, easily hand placed chip. Highlighting the uselessness of the video assist. This tells me the crosshair is very important.

At €2k the ec-placer starts to look good again. The foot-controlled X and Y locking seems useful vs other systems. I would have to fashion a rail/slot holder that drops into their grid board.

Maybe I just need to get a liteplacer.

[electrolust]

Looked at a few more. I think I've narrowed it down to ec-placer and Fritsch LM901.

@mairo, thanks for the MAN-32 reference. It's certainly interesting and the kit is complete, even with monitor. I'm guessing pricing is very reasonable. The thumb operation of the placing head, combing theta with Z, seems prone to error (theta change with Z), difficult theta correction when Z engaged, and fatigue. Have you actually used it?

Had another look at SMT Caddy. It certainly has some neat ideas, but the execution is just poor. It clearly wasn't built with a paying customer in mind. Too much money spent on the wrong aspects.

[mairo]

HI electrolust, no I have not used the MAN-32, but I do not see the issue there  with the Z/theta movement (if I understand you correctly). I always correct 'on the fly' the component, if needed, while on top and lowering toward the footprint (see also video below where the operator is doing same), but having said this, the machines I have used have slightly different mechanism to achieve this. I think just needs getting used to this type of control..

Can I ask what are you going to use the machine for and for how long/how many parts per use?

OK, if vision is what you need due to physical reasons, then you can look for a machine with a camera. My comments on the cameras on these machines are that first it slows you down alot and second I used to get dizzy while using them, altho I felt uncomfortable when I first started to use mantis microscopes as well, but then got used to it. I guess I can see why the SMT-CAdy have gone with the movable screen to probably improve this so you do not need to look away from the board which causes eye and neck pain after some time (not a big fan on this machine, but I can see its very niche use). Back to the cameras, the LPKF, Fritsch, Essemtec and few others are using cameras that you plug to monitors directly, hens they are quite fast. The ec-Placer is using USB microscope camera that is abit pain to use - lags when moving and takes time to refocus sometimes, but the cross option is good, just know that you can get this usb camera separately for $10-20. The person that mainly works on our machine I remember used to make his own footprints with a cross in them to get the full potential of the feature  . The camera option ultimately will slow you alot (same as with fully automatic machines, in general if they need to use the cameras they get slower in placement, but they have theirs workarounds to improve on this).

Watch the video at the beginning when the operator is placing solder paste, he is not looking at the camera at all and I am sure he is also not looking at it during placement of parts. By the way, for the solder dispenser, I would only get it if it comes for free (OK, make that $100 if I have to paste boards with less than 20 pads), the cheap solder stencils thesedays cannot be beaten in speed and price.
https://www.fritsch-smt.de/video-seiten/mp904

I will give you some negatives (my personal view) on the ec-Placers here:
1. The foot switch for the vacuum (I think it only works with it from memory, skip my next words if that is not the case) - I think it is the biggest downside of this machine. Having to use another body part and the brain kills the simplicity for me. If you see videos from Fritsch, Essemtec, LPKF, Dima, Atco, old OKI machines, Manncorp, Bokar/Bomir, APSNovastar, MAN-32, Mekhatronika (I will try not to use this word alot as it triggers some people here) they all use pressure switch inside the head. It works perfectly and you do not need to think about it during placing, it just works - you pick a part, the pressure engages, you drop the part in the footprint, the pressure stops and the part stays there. They are tuned (at least on the machines I have worked with - LPKF, Fritsch, OKI and M ) quite well so that as soon as it feels the resistance from the paste it turns off so you don't squish the paste under it.

2. Not comfortable on the hand/body over long use;

3. Feeders - The tape feeder slots on the left pretty much feed the empty tapes into the working area. In this case you have two options:
- option 1 is to place the board close to the feeders so that you do not travel far away (less body movement) to pick and place, but will have to often cut out the tapes as they get closer to the board/other parts on the table;
- option 2 is to place the board far away from the feeders in which case you will be cutting less often, but will be travelling further for pick and place;


If you can get the LM901 (have you check the prices of it?), then do not look at the MAN or the ec Placer  . When I got the MM600 new for ~$3-4k (air freight + some extra feeders), I also got priced for used Fritsch (priced from the manufacturer) with matching features to the MM600 for ~$14-15k. Now these machines are semi automatic machines that's why they cost more. A full blown Fritsch I think costs ~$25-35k? The MM600 is still the lowest cost semi auto machine on the market even now I think and is cheaper than some manual machines as well. A downside here is if the controller breaks in these more 'advanced' machines then it might be 'difficult' to move to the  'manual mode'.

Since you seems in USA, check APSNovastar, Atco, LPKF, Manncorp and Bokar/bomir as they are USA based, or have good presence there.

Essemtec and LPKF, both make very nice machines as well (I work often on the LPKF ProtoPlaceS) and often pop out in eBay for very good prices. I have also seen several times old OKI machines for less than $500 in working condition.
 

[electrolust]

Thanks for all the great info mairo. Clearly you have spent lots of time considering these machines.

I'm not doing too many parts per hour. CPH per se is not a concern, but ergonomics is. The camera is therefore a must.

For Z vs theta, with the small two-finger knobs you have very fine control of theta as you descend Z. As you say, while lowering you do 'on the fly' correction. But with the big thumb-press knob as with the MAN-32, it seems to me you lose fine control of theta as you descend. Perhaps the rotational friction is higher and therefore less tendency to change during descent?

On watching another ec-placer vid, I now believe the crosshairs to be fairly bad idea. First you have to use old version of .net. Why they built this machine in 2017-2018 and used already old tech is disappointing. OK, that's implementation, and can be fixed. But by design, I see that for small parts, the crosshair interferes with the visual view. The visual magnification is not large enough, so the crosshair completely or mostly obscures the target view, for small parts. For larger parts, you're still just targeting the "general area" (say center) so the crosshair doesn't help much. I prefer the laser dot of the SMT-Caddy. I think ec-placer could fix this by replacing the center of the crosshair with a hairline (like an optical scope), or translucent overlay, or an empty space and your brain will find the center.

I now also see that the single sliding rail for the ec-placer is inadequate. In multiple vids, the operator is frequently using 2 hands to get good placement accuracy. And with no elbow/arm support (wrist only is supported) the ergo seems not good. I think these 2 problems are actually inter-related. The sliding rail seems to have a jib or jib-like tension adjustment, and perhaps you will be compromising stiction.

Lastly, I didn't properly account for the exchange rate. SMT Caddy will actually be cheaper. I need to take a new look at that.

The Fritsch actually has an optional blow-off function, in addition to just releasing the vacuum. Apparently it helps for very small parts. I have indeed found that parts like 0201 want to stick to the tweezers. Yes, the full blown price is probably touching $25k. For a manual machine!! An absolute bare basic machine just sold on ebay for $1500. Too bad I missed it. They machines apparently sell like Porsches -- just over the price of the Japanese competition, but for a base model. Easily double the price after options get added.

I shied away from LPKF because I didn't like the price. I will take a fresh look. They also seem to have more of a used market.

[electrolust]

Essemtec you say?

[Clear as mud]

I don't think anyone mentioned, for manual assembly of QFN parts, you can extend the pads out from the chip edge a little bit, and this makes it easier to see when it is lined up.

[electrolust]

The visual magnification is not large enough, so the crosshair completely or mostly obscures the target view, for small parts. For larger parts, you're still just targeting the "general area" (say center) so the crosshair doesn't help much. I prefer the laser dot of the SMT-Caddy. I think ec-placer could fix this by replacing the center of the crosshair with a hairline (like an optical scope), or translucent overlay, or an empty space and your brain will find the center.

oops. I see now that the crosshairs is in fact missing the center. it seems ok enough.

[Kjelt]

With those prices I keep on wondering why there is no hybrid machine out there.
A design like the chinese 4k$ machines but then joystick camera operated for the final placing, so automatically pick the component from the trays or feeder and move aprox. where to place and let the operator rotate and place.
Something like that would safe tons of time and fatiqueness.

[SMTech]

because once that bit works, there is no point letting a human do the last bit.

[wraper]

because once that bit works, there is no point letting a human do the last bit.

Exactly. Once you have hardware which can do all of the job, what's the point of human doing the placement in real time. Set component positions and let the thing run by itself.