Advice on a Desktop Pick and Place Machine under $2500

[mohala]

Hello All,

Hope you all are well, I've been racking my brain on which pnp would be best which can also fit in my office, I’m a bit tight in space and if I throw away my only sofa I can some how manage to get at least an area of 1200x1400mm of working space. I was thinking I would have the pick and place machine upstairs and the reflow oven downstairs in the garage, away from me due to the fumes (as I spend most of my life in my office sometimes sleep as well!)

I have been eyeing the OpenPNP options but there have been quite a lot of versions being offered and would like some honest opinions on this matter. Some of the work I do can get me to be forced to work with 0201's and some 0.35 pitch BGAs. But speed is not a focus for me accuracy and ease of use would be very important.

I do like the idea of the SmallSMT Panda, but its looks to be too out of my price range

Currently I have seen the following under the OpenPNP:
Pandaplace A1 (https://pandaplacer.com/)
TornadoSMT (https://www.alandesign.cn/en/home)
T2SMT (Aliexpresshttps://www.aliexpress.com/i/1005005772454749.html)
LumenPNP (https://www.opulo.io/products/lumenpnp)
PSE-3000 (https://www.picksoeasy.com/PSE-3000.html)

I know these pnp have advertised to do 0402 and 0.5 BGAs, but wondering if anyone has had experience with them enough to know you can do certain modification to achieve better performance, or getting the right nozzles and settings?

Lastly I have been eyeing the more expensive options like the SmallSMT Panda Advance and the Tronstol E1 or V3. But unsure if its worth me aiming too high rather than focusing on the OpenPNP route for now?

Please let me know your opinions, I'm very curious what you all think?

Mo

[Kean]

If you have lots of spare time to waste, then maybe... else you're dreaming.

[vespaman]

Hi Mohala,

What is the purpose of the machine, few boards here or there, or a decent amount of boards? I am using OpenPnP, but I'm a open source guy, haven't really used any closed source software the last 15-20 years. The learning curve is steep, but if you are into these things, good and useful in the end. I don't think I would like to buy a closed source machine (then I would probably convert it to OpenPnP) for my small production.

I'd say if you only need to do  a few prototypes, it is not worth it. Especially, if you have a lot of "new components" that needs to be entered into the PnP. So, I'd say - if you design the boards yourself, and have control over the selection of components it gets easier, also a little bit easier if using KiCad.

Regarding hardware, it I doubt you can find someone that can answer and give you a "best in test". It depends on how much you swap components, type of components and how many different components you have in total, and on each board and e.g. how many components you purchase (only cut tape, full reels, 12" reels et c et c. Maybe read up a bit on the OpenPnP mailing list to see what other users of the ones you are looking have experienced.

 

[ebastler]

There's a narrow space between the realm of building a prototype now and then (which you'd be better off placing by hand), and building a small series of baords now and then (which you'd better outsource to a PCBA house). In that narrow space running your own small pick & place might be beneficial -- but is that where you operate?

And I am not sure whether there is any space at all between the technical requirements for placing 0201s and 0.35 mm pitch BGAs reliably, and your budget cap of $2500.

I'd recommend that you forget about the pick & place and keep your sofa instead. Much better for sleeping in the office too! There's just not enough room for a sleeping bag under those small desktop PnP's. 

[trevwhite]

There are benefits to having your own machine. I think a belt driven machine will struggle to place 0201. The cheaper options seem to be that style.

The cheapest machine I would consider so far to be up to the task you require is a SMT220 also known as a HW-T2-40F. It’s more money and uses CL feeders which you would also need to buy.

I can’t state anything about it’s build quality, etc but it looks to be very accurate as it runs with servo drives.

[jmelson]

You mention having the oven on a different floor than the P&P.  After placing, any heavy joggling of the boards may cause components to move or even fall off.
I have my oven about 4 meters away on the same floor, and I would not want it much farther than that.
Jon

[SMTech]

You mention having the oven on a different floor than the P&P.  After placing, any heavy joggling of the boards may cause components to move or even fall off.
I have my oven about 4 meters away on the same floor, and I would not want it much farther than that.
Jon

This + @ebastler's comment previously are key issues even ignoring the elephant in the room - your requirements.

There are lots of these threads (normally with a bigger budget) there are also plenty of people with this type of machine here and posting away on Youtube. With a lot of tweaking, some of them are getting 0201 out of some of the better machines, but we don't know their definition of success, how stable that tweaking is or or how much post placement tweezer nudging they aren't quite admitting too. uBGA however is to me a laughable ambition; you will struggle to paste that let alone place it. A uBGA stencil (&PCB layout for that matter) requires a decent bit of knowledge to specify a decent one.

This year we had an enquiry that involved placing a few thousand uBGA devices, this isn't something we do with our SMT line as we lack both the experience and the inspection equipment to validate our process for BGA. So we shopped around for the client with other CEMs . A significant number declined for the same reasons we did despite being much bigger and better equipped, the ones that said yes were 50:50 split on the need for X-ray on every board and a much laxer attitude to this being a step they needed at all once tooled up. The no's all didn't have x-ray, despite what you might read in other threads, in the UK at least it is still quite common not to have this capability.

[jmelson]

I have a ~$150,000 commercial machine (Quad QSA30A) made by Samsung, similar to their CP30 but with Quad electric feeders and Quad alignment cameras.
I got this one from an auction for $500, but had to spend $7500+ getting it repaired and obtaining feeders and accessories.  I do passives down to 0603 and 0.5mm LQFP parts and it is fantastic for those.  Once I have everything dialed in, I never have to mess with part alignment.  Sometimes the pickup height or placement height is not set right and I have to find the ideal settings for that.  I have a home made stencil printing fixture that is just repeatable enough for this size of parts.  But, I sure would not want to do anything smaller than 0603 parts.

Jon

[mohala]

Hiya All,

It really was for myself to step into the field of of PnP originally while supporting my work, but as work deadlines are being tighter, and getting a PO to ok to assemble a board would end up being a mammoth task with constant hassle with finance and assembly manufacturer to send free issue parts. After seeing this happen on a few projects where it should have taken a week is now a month long process, of shipping things across and getting PO's to be ok'ed. I started on doing a lot of them myself, with hand soldering. It was great to do on a few boards when you are set to do one or two project but now I'm put on multiple projects which require smaller components and in a smaller area to work with (and double sided).

At first I was thinking to maybe invest on Fortex MPP1. But Now the idea that some of the projects are looking to have from 10 boards to 50 boards at a time.
If the projects progress to testing phase they may need more boards to assemble.

Secondly, I also want to equip myself in the future in the event I change jobs or possibly start working for myself that I can ready equip myself from designing all the way to prototyping and be efficient in doing so.

Its a bit of a bold claim especially in this economy, its is good to hear your opinions, and advice, to do it or not.

Also, the mention of the oven, I need to rethink that maybe best I put it near to the pick and place and chuck the hose out of the window. Its not a great look but its an idea, maybe sticking it in the toilet I have an air vent there and it would warm up the toilet during the winter months!

But going back to the point, I want to try to be a bit more self sufficient and not get bogged down even with everyday work but same time. Hopefully do my own thing during and in the future.
 
P.S I do use Altium all the time, but I think also learning to use KiCAD would be good for me, too.

[ebastler]

[...] constant hassle with finance
[...] now I'm put on multiple projects
[...] I also want to equip myself in the future in the event I change jobs or possibly start working for myself

Hang on -- so you are not on your own but working as an employee? In that case you should really do a cost/benefit analysis, weighing the investment into equipment and your time to get it up and running vs. future time and cost savings. And if that comes out positive, convince your employer to spend money on a PnP that's up to the task. (Probably spending well above $2500 if "the task" really includes placing 0201s and 0.35mm pitch BGAs.)

And maybe you should also rethink the "sleeping in the office at night" policy. That's too much dedication to a dependent job.

[vespaman]

>Secondly, I also want to equip myself in the future in the event I change jobs or possibly start working for myself that I can ready equip myself from designing all the way to prototyping and be efficient in doing so.

This is why you should do it. If you are curious about it, and can have this financed. What you learn (regardless of machine) will be yours, and could be a useful thing going forward.

Regarding the 0201 and 0.35 BGA - I think actually many machines can do this, there's even that guy that modifies 3d printers and with the "lego" build setup that does 0201. Granted, slow and so on, but still.. But no, for all sorts of reasons, I would not consider doing boards like this myself, mostly because of the paste printing, as has been mentioned. Maybe an odd 0201 here or there, but I doubt I'd want to do it in too many boards.
Large (long) higher density connectors are worse for the pnp machines, in my experience.
Expect to nudge few components here and there, especially with such small small series there will be few iterations to tweak. But this is manageable, and beats hand placing in every way, anyway.

I can say that for me, the learning curve was steep, when I bought a used machine, and converted it to opnenpnp. At times, I thought about giving up. But now, I fully enjoy making boards, and like the flexibility it gives me,
I have (for now at least) set a limit of max 1000 boards/yearly on my machine - if it goes higher, I'll subcontract. Mostly because of purchasing and planning time, that I totally underestimated, coming for larger companies, starting my own business (as a recovering workaholic), but also, because I don't want to spend more than a few days a month doing this type of work.
 

[mohala]

The thing is we presented the idea to the company if we can have one and did the full presentation.
But the problem, we got is not the money situation (which they have a lot of it).

The response we get is why can't our external partners do it (which we already argued the point, but very redundant as the problem is that the external partners are too slow to do it, because of our very slow internal approval process)
second response is use our China division to do the prototyping, Which make me think my managers aren't bothering to listen as its the same problem too many internal approval process. (Also, that process has an extra, step be friendly with the head of China prototyping team then have your boss speak to thier boss and ask permission which needs to be a morning teams meet. Then send out a email to approve that its done. Ten send the copy of the email to Finance which then will assign a cost code which then you can provide to the prototyping team. Once that is done you can start to tell them what needs to be brought for the BoM) All this takes a few weeks because the bosses have no time to talk they are speaking with senior leadership and when they do they don't wanna hear obstacles)

I will try however to see to push it again on the next financial year to push the idea, but I'm thinking maybe to make them to push one which is more attainable in my price point so I can learn about how they work and then buy one for myself. The nice thing is if it does get approved its easier to hide the cost of a second machine, its just hard to find where to put it 

As for sleeping you are right. My wife hates it. But hey its the only room in the house which has an air conditioner so her mind changes when its summer   

[ebastler]

[...] the external partners are too slow to do it, because of our very slow internal approval process
[...] our China division [...] same problem too many internal approval process.

Sounds to me like that is where the company should really invest some effort (and maybe money into suitable software tools). If there are external or internal partners which would otherwise be perfectly qualified and efficient, but the internal ordering process is the bottleneck, your bosses need to fix that process. Buying a piece of equipment to let you (inefficiently) work around the problem by doing things yourself is the wrong answer.

As for sleeping you are right. My wife hates it. But hey its the only room in the house which has an air conditioner so her mind changes when its summer 

Maybe suggest to her that you want to move the soldering oven to the kitchen, so you can connect it to the extractor hood over the stove. 

[thm_w]

Sounds to me like that is where the company should really invest some effort (and maybe money into suitable software tools). If there are external or internal partners which would otherwise be perfectly qualified and efficient, but the internal ordering process is the bottleneck, your bosses need to fix that process. Buying a piece of equipment to let you (inefficiently) work around the problem by doing things yourself is the wrong answer.

Exactly, either they deal with the problem, or if you want try to find an efficient workaround: get a blanket PO setup for dev builds, a company card you can easily expense a build to, etc.

[mohala]

Sounds to me like that is where the company should really invest some effort (and maybe money into suitable software tools). If there are external or internal partners which would otherwise be perfectly qualified and efficient, but the internal ordering process is the bottleneck, your bosses need to fix that process. Buying a piece of equipment to let you (inefficiently) work around the problem by doing things yourself is the wrong answer.

Exactly, either they deal with the problem, or if you want try to find an efficient workaround: get a blanket PO setup for dev builds, a company card you can easily expense a build to, etc.

I totally Agree we have pointed this out not from my team but a few teams have now. Instead, these issues caused more issues, instead of fixing and streamlining they put more processes and good god they implemented Teamcenter PLM into the mix of it.

One Thing I'm learning about huge corporate companies they love to make things more complicated to complain that they need more people, so they can be managers to boss those people around.

It was a pain.  But I'll try both ways mentioned. I can see I might be able to get one for a learning tool, and call the equipment to expand our internal capabilities and see if that spins a PnP. But we have been pushing the whole streamlining thing even given and answer how to do it. (Which made matters worse, because I think it embarrassed them, and now knuckled down on making it more complicated with more people involved)

P.S I think this is becoming more of a therapy session  which Really is helping, come up with ways i can navigate this company.

As for sleeping you are right. My wife hates it. But hey its the only room in the house which has an air conditioner so her mind changes when its summer 

Maybe suggest to her that you want to move the soldering oven to the kitchen, so you can connect it to the extractor hood over the stove. 

Oh she probably kill me for that , The house is her Domain and the garage is my Gaming/Man Cave/Office/Sleeping for late work or getting told off

[prutser]

Hi Mohala,

The learning curve is steep, but if you are into these things, good and useful in the end. production.
I'd say if you only need to do  a few prototypes, it is not worth it. ones you are looking have experienced.

@vespaman:
It seems you have some experience. I am currently also looking into something for hobby purpose. This is for occasional prototypes.

The main reason is not the amount of boards/parts, but the fact that parts are getting smaller. This makes it more time consuming and risky to place by hand.
Recently did a small PCB with some 0402 parts and 0.5mm pitch QFN. Hand PNP took me almost 4 hours (and you still have a risk accidentally moving the QFN or place it wrongly. If that happens, cleaning and stencil would have been  the only option).

I have currently no idea how difficult  openPNP is to setup (just downloaded it), but for my purpose I consider an "hybrid" approach. IC's, decoupling caps, 0402 and 0603 parts by PNP (If possible I mostly use 0805) and do the larger parts by hand (easy and no machine setup)

Any experience/advice on this ?

 

[mikeselectricstuff]

Forget 0201's - this is as much about feeders as anything else- even multi $100k machines sometimes need special feeders for these.
Fine-pitch BGA is also very dependent on the paste print process - stepped stencils etc.

[vespaman]

Hi @prutser

It is hard for me to say how hard it is to set it up, because it depends so much from where you are starting knowledge-wise.  I jumped into it, seeing a used Charmhigh 48VB not too far from where I live.
From the start, I knew I wanted to convert it to OpenPnP, so I did not even try the original FW (something I realized later, I should have, just to get to know how the machine was supposed to work).

I think it depends on your perspective and willingness to spend time to learn, because, yes, it is a lot to learn. Not just OpenPnP, but the whole process.
If you think this can be a hobby, then go for it. Why not? But don't expect to pick new populated boards out of your machine next month. 

0402 is no problem at all in my setup, this is my preferred component size, I absolutely avoid 0201 and smaller. Here's  a small clip https://youtube.com/shorts/ie137bqQuyg of the typical jobs I do.
The build-your-own kits are typically a little slower, since they don't have the mass, but that shouldn't be a problem.