Open Placer A New Affordable PCB Assembly Machine

[Koen]

I don't feel sorry for anyone.....

Ho how fast people change. People feeling sorry for you is the only reason you won an extra scope. Think about it.

It's called constructive criticism.

No it isn't. You two are simple on repeat about how it doesn't suit *your* needs and might as well be considered useless to everyone.

[rx8pilot]

I don't feel sorry for anyone.....

Ho how fast people change. People feeling sorry for you is the only reason you won an extra scope. Think about it.

I feel sorry for people that are faced with an unfortunate problem. The OP is getting real feedback from at least two people that have gone down this path. That is a very fortunate place to be.
On the scope contest, I was faced with an unfortunate problem and many people, including Keysight clearly had some sympathy. I have been extremely grateful ever since. I have not changed at all, but rather used the result of that sympathy to rapidly grow my business.
The OP has the option to listen to a couple of real-world professionals to adjust their approach or not.

It's called constructive criticism.

No it isn't. You two are simple on repeat about how it doesn't suit *your* needs and might as well be considered useless to everyone.

Yes, it is. Neither Mike nor myself are saying this is useless for everyone, just pointing out some critical points that can be used to analyze the market size and reception. To get this feedback is hugely beneficial to someone creating a new product. If the OP thinks that the market is bigger than reality - this is the time to learn about that. If the OP is unrealistic about capability, now is the time to discover that. The more criticism the better. If the OP cannot process constructive criticism, they are not cut out to bring a new system to the market.

Lots of companies pay a lot of money to get market research.

[girts]

Yes, it is. Neither Mike nor myself are saying this is useless for everyone, just pointing out some critical points that can be used to analyze the market size and reception. To get this feedback is hugely beneficial to someone creating a new product. If the OP thinks that the market is bigger than reality - this is the time to learn about that. If the OP is unrealistic about capability, now is the time to discover that. The more criticism the better. If the OP cannot process constructive criticism, they are not cut out to bring a new system to the market.

Lots of companies pay a lot of money to get market research.

Discussion is a little "out of bounds", not only OP related.
OP is most likely a  good starting point for a little money. Just to understand what's really necessary in somebody's garage. Because main problem in this market is that most of things isn't possible to see and test in real life before they are already purchased and arrived. Of course, lot of them are useless ...
BTW - don't know any person who started his way with Lecroy 50GHz scope.

[girts]

It depends. If I have to place 100 boards, 10 parts per board, and BOM of 5 lines, sure. I'll do that manually in a couple of hours.

to set up 5 feeders and load BOM for 10 components takes 1/2 a hour with some coffee breaks. To place 1000 components with PnP machine takes another 1/2 of hour.

If I have to place 5 boards, 200 parts per board, and BOM of 50 lines, of which 40 are 0603 - no thanks. I'll load up an SMT, even if it take more time place. At least I'll know that every resistors is there it was supposed to be.

Same 1000 components, but must set up 50 feeders, most of them for 5 unique parts which are not in reels but strips instead... Ohh.. I would prefer manual placement instead. Or combined, because most likely design consists of a lot of 0.1uF capacitors and some basic resistors already loaded into PnP.  Of course, if there are only 5 boards necessary, not more in far or near future.
And... most likely this design is not for single PnP, you will need a line of 2 except if you already have TWS with more than hundred reels loaded (but missing tray exchanger and conveyor..).
 
What if you place wrong reel into feeder? 50 feeders, very simple to make mistake.
0603 is not a problem for manual PnP with camera. Smaller ones are.

[ar__systems]

What if you place wrong reel into feeder? 50 feeders, very simple to make mistake.

You need to realize that "load into the feeder" step is common for both manual and machine assembly. Yes it is an easy mistake to make, so knowing that you devise a procedure following which you will reduce the risk of having an uncorrected error to a minimum.

[ar__systems]

to set up 5 feeders and load BOM for 10 components takes 1/2 a hour with some coffee breaks. To place 1000 components with PnP machine takes another 1/2 of hour.

That's not the point. I never said I'd prefer manual over machine for that kind of job. I just said it is relatively easy to do manually, reliably.

[ar__systems]

As far as OP goes, one can live without an SMT for a long time. Or forever. Unlike a scope, which you can't live without. For small boards it is easy to make hand assembled prototype, and for any volume it is easier to outsource it. So yeah, SMT equipment falls into toy/luxury category.

It is also need to be understood that you need more equipment to run any kind of mass production: you need a printer, you need an oven, and you need an exhaust ventilation for the oven. You also need space.

[Kjelt]

grow my business.
The OP has the option to listen to a couple of real-world professionals to adjust their approach or not

that can be used to analyze the market size and reception. To get this feedback is hugely beneficial to someone creating a new product

No doubt you mean well but you are unable to see the market for this thing because all arguments are made of starting a business, pro machine this, that, stop it already. You clearly suffer from Tunnel vision.

What about grandpa radio amateur from the 60s/70s that loves to start with smt but his hands shake too much to be able to place a 0603 ?

What about the sw dev that likes to play with openpnp but the machines were too darn expensive?

How about the hobbieist that loves to play with toys but $4k or more is not fun anymore he will get arguments with his spouse but $1k would be okish?

You can burn down 60% of all the stuff on this forum with those arguments incl. all Daves products.
Probably we can even criticize your products or are they perfect? If yes then you made them  too expensive.
Let him try to make his market share , this thing has some major selling points like OpenPNP support and unbeatable price. How many millions of times have people asking the Neoden fab to do this? Not going to happen from them, do you burn them to the ground? You should with their cripple sw that is not even protected from using wrong fw for different machines.

[mikeselectricstuff]

It depends. If I have to place 100 boards, 10 parts per board, and BOM of 5 lines, sure. I'll do that manually in a couple of hours.

to set up 5 feeders and load BOM for 10 components takes 1/2 a hour with some coffee breaks. To place 1000 components with PnP machine takes another 1/2 of hour.

This is why a good workflow is important. On my machine, to set up a job with 5 BOM lines would take maybe 10 mins at most, including loading feeders. Parts that are already loaded in feeders take near-zero time.

My PCB and P&P libraries are co-ordinated in terms of component rotation relative to the taping, so that for parts I've ever used before, I don't even need to check orientation as I know it's right. I also include "fid" components in the PCB design, so these get auto-imported. The only job setup is to tell the machine where the PCB is on the placement area. I can't emphasise enough how important this is.
Another extremely useful thing I've found is to use a handheld gamepad for machine control while setting up, so you can keep your eyes 100% on what you are doing.

The placement time doesn't really matter, within reason, as long as the machine can be left to get on with it with minimal fiddling, as you can be getting on with other things. This is why it is important that the machine can deal sensibly with feeder errors by carrying on with as much as it can before stopping to get the user to fix everything in one go.
It does need to be fast enough that a larger job doesn't take all day. I'd say something like 500cph would be an absolute  minimum sensible rate.
 

[mikeselectricstuff]

What about grandpa radio amateur from the 60s/70s that loves to start with smt but his hands shake too much to be able to place a 0603 ?

He would probably have at least as many issues setting up a P&P, peeling cover tapes without parts going everywhere etc.
An auto P&P is definitely the wrong solution to this particular problem

[soubitos]

Say my modules picked up in sales and i needed not much, 100pcs a month..
If i go pcba their cost would be in the range of 1-2$ ea easy..
I am now hand assembling them right?
So, if a 1k machine could take over not all but resistors and capacitors only it would break even in a year or less no matter what... a 1k pnp handling part of the job the most intensive and repetitive part, releasing me to do other things and so on... now thing about having to assemble 10 different designs each month, 100pcs each... it could make a huge difference without breaking the bank at all.. it could pay itself in less than a year.. i call this a good investment... especially since, while it pays itself i can potentially make more money cover more orders, faster and so on... this means i get closer to having more orders, more workload for a pnp machine i can now get another similar or if it all goes so well, i can invest on a more expensive and more capable model having a first hand experience of the cheaper basic one...
If the machine/tool can pay for itself over a relatively short period, its a non-brainer when the only other alternative would be an investment 10x as much or outsourcing the job altogether loosing money in the process and demanding more capital etc...

[mikeselectricstuff]

Say my modules picked up in sales and i needed not much, 100pcs a month..
If i go pcba their cost would be in the range of 1-2$ ea easy..

That would seem a bargain compared your time to do it in-house.

[soubitos]

Say my modules picked up in sales and i needed not much, 100pcs a month..
If i go pcba their cost would be in the range of 1-2$ ea easy..

That would seem a bargain compared your time to do it in-house.

It is a fair price i recon BUT.. what i'd spend to have 300pcs assembled would pay of my own cheap pnp, i dont think one would buy a 10k machine or more to make 300pcs a year but 1k machine to help you do it? easy choice for me

[sokoloff]

It depends. If I have to place 100 boards, 10 parts per board, and BOM of 5 lines, sure. I'll do that manually in a couple of hours.

to set up 5 feeders and load BOM for 10 components takes 1/2 a hour with some coffee breaks. To place 1000 components with PnP machine takes another 1/2 of hour.

To place 100 boards in 30 minutes, you need to have a board cycle time of 18 seconds. That's aggressive without automated feeding. (Presumably anyone considering the subject PnP isn't looking to add a new placer to their existing auto-fed line...)

Even a panelized board that's 10-up, you still need to be removing the completed and placing a new stenciled board even 3 minutes.

[TassiloH]

I have been reading a lot about the possible or impossible time savings of a PnP machine for prototyping etc. For my niche, the hardware of the Openplacer would probably be just perfect, I am just not sure if the existing or future software capabilities would do the job:

I am doing a very small amount of boards, mostly one or two identical at a time, but various types. So far I haven't out-sourced production, because the involved preparation and communication time seems longer than doing hand placing. A conventional PnP machine would have to be huge and costly to save me any time, because all the various boards together have hundreds of different passives alone, and I guess if I'd have to load feeders differently for every board run that would take more time than hand placing and be just a waste of time and money.
What I would really like is a machine plus sofware which starts placing, tells me "I need five 12k1 resistors, and next I need the 27k", I take the cut tape, slide it into a holder (and peel the cover off), tell the machine which holder compartment, and the machine aligns to the cut tape position by vision and places the components, while I fish the next displayed value out of storage.
Some common components could remain in a holder (like common decoupling caps).
That would speed things up considerably (taking components out of storage and placing would happen in parallel; I wouldn't have to look for the proper pads to place things to, which gets VERY annoying over time).
I don't care about placing odd-shaped or tall components automatically, the few tall inductors, capacitors etc. I could place by hand just fine. Precision requirements wouldn't be high (mostly 0805, but some flat pack and QFN chips with 0.5mm pin pitch, but I could even place them manually).

So for me the hardware of the OpenPlacer (epspecially with the slide-in cut-tape holders and vision) would probably do fine, but I don't know if on the software side it will ever allow such an "interactive" mode - and without that there is no time (or comfort) to gain for my niche because of the many different components.

[ar__systems]

So, if a 1k machine could take over not all but resistors and capacitors only it would break even in a year or less no matter what... a 1k pnp handling part of the job the most intensive and repetitive part, releasing me to do other things and so on... now thing about having to assemble 10 different designs each month, 100pcs each... it could make a huge difference without breaking the bank at all.. it could pay itself in less than a year.. i call this a good investment... especially since, while it pays itself i can potentially make more money cover more orders, faster and so on... this means i get closer to having more orders, more workload for a pnp machine i can now get another similar or if it all goes so well, i can invest on a more expensive and more capable model having a first hand experience of the cheaper basic one...
If the machine/tool can pay for itself over a relatively short period, its a non-brainer when the only other alternative would be an investment 10x as much or outsourcing the job altogether loosing money in the process and demanding more capital etc...

No 1K machine will get make you better off making 100 boards/month, not in any meaningful way. Maybe if you have a free labour to babysit the machine, you can decide you are better off compared to outsourced PCBA.

I would order 1000 boards assembled in China, and sell them over a year. I've been doing so for years.

[mikeselectricstuff]

I have been reading a lot about the possible or impossible time savings of a PnP machine for prototyping etc. For my niche, the hardware of the Openplacer would probably be just perfect, I am just not sure if the existing or future software capabilities would do the job:

I am doing a very small amount of boards, mostly one or two identical at a time, but various types. So far I haven't out-sourced production, because the involved preparation and communication time seems longer than doing hand placing. A conventional PnP machine would have to be huge and costly to save me any time, because all the various boards together have hundreds of different passives alone, and I guess if I'd have to load feeders differently for every board run that would take more time than hand placing and be just a waste of time and money.
What I would really like is a machine plus sofware which starts placing, tells me "I need five 12k1 resistors, and next I need the 27k", I take the cut tape, slide it into a holder (and peel the cover off), tell the machine which holder compartment, and the machine aligns to the cut tape position by vision and places the components, while I fish the next displayed value out of storage.
Some common components could remain in a holder (like common decoupling caps).
That would speed things up considerably (taking components out of storage and placing would happen in parallel; I wouldn't have to look for the proper pads to place things to, which gets VERY annoying over time).

I think picking loose parts from a tray using vision is entirely doable at minimal cost - cameras are cheap. Tatmay be a good solution to the fiddliness of dealing with short bits of tape

[mikeselectricstuff]

No 1K machine will get make you better off making 100 boards/month, not in any meaningful way. Maybe if you have a free labour to babysit the machine, you can decide you are better off compared to outsourced PCBA.

Depending on the definition of "boards" and "better off" I'd disagree.
A low-end machine would certainly be useful to enough people for it to be a viable product ( especially if it's open Source and the manufacturer isn't looking to make money on the software), and I hope OpenPlacer goes some way to proving this.

 $1K is a sensible starting price, and maybe scope to reduce this once market demand is established, but trying to start out too low is a fatal mistake.
People who could really use a machine won't really care if it's $700 or $1000. If they don't they probably don't need it.
 
However it is a lost opportunity if it's a dead-end with no path for users that want to upgrade ( at purchase or afterwards) to be viable for larger requirements.
This is primarily about design, not production cost for teh entry level. e.g. decent placement area, ability to trade placement area for strip/tray feeders, PCB holding for double-sided and allowing large boards in multiple passes, multiple heads/cameras, loose part picking, and of course support for reel feeders.

[TassiloH]

I think picking loose parts from a tray using vision is entirely doable at minimal cost - cameras are cheap. Tatmay be a good solution to the fiddliness of dealing with short bits of tape

True. But it would also require a machine with a 180 degree turning station (if I put resistors on the table for picking with tweezers, invariably 60% of them land upside down  ). So I could live without picking loose 0805s.

[mikeselectricstuff]

I think picking loose parts from a tray using vision is entirely doable at minimal cost - cameras are cheap. Tatmay be a good solution to the fiddliness of dealing with short bits of tape

True. But it would also require a machine with a 180 degree turning station (if I put resistors on the table for picking with tweezers, invariably 60% of them land upside down  ). So I could live without picking loose 0805s.

Easy - Vibrator motor on tray.

[alexanderbrevig]

Replying because I find the product interesting and this is the best way I've found to keep up do date on a thread.

Nice work!

[girts]

Easy - Vibrator motor on tray.

Right.
https://www.youtube.com/watch?v=i4iUhBdVrmk&feature=youtu.be
Another youtube channel with a lot of good ideas -
https://www.youtube.com/channel/UCe7TGvrsO021fo2d22LiYoA

[Nauris]

You also need space.

Exactly! If you look in Ebay there are so many nice machines Assembleon, Juki, Siemens or whatever for just 10-20k dollars but holy shit they are BIG and HEAVY! It would be so nice to buy one but living in a single room apartment they for sure aint't going to fit thru the door or in the elevator! You have to have some workshop space or garage at least and then that needs some real money.

So something smaller especially if it is free of malfunctioning proprietary software attracts some interest.

[iwbnwif]

I really like this machine.

It would be nice to have a manual mode where I can control the head with a joystick or similar, but I guess that is something to build into openpnp.

There is a lot of negativity surrounding Indiegogo which puts me off somewhat though .

[khs]

Today I've made some prototypes. Three boards, two double sided and one single sided.

Over all 172 parts took 171 minutes. So one part takes about one minute including solder paste, some mounting work and and some manual placed parts.

If my calculation is not wrong, preparing one side to mount takes about 20 minutes - if there is no need to mount new/other parts. But there are 50 feeders, so most of the parts are mounted.

In this case the pnp was running about 18 min, so there is about 6 s / part including board handling. (My pnp runs at half speed only).

Solder paste takes about 12 min, so I need about 2.5 min per side.

So with my simple software/pnp I can be faster than manual placing if there are more than 20 parts on one side.

But even if I'm slower than placing by hand: For me it makes more fun to see the machine working rather than placing every tiny piece by hand and taking care about part  position and orientation.

So using a pnp it's not only time saving, it's energy saving too.

You also need space.

Exactly! If you look in Ebay there are so many nice machines Assembleon, Juki, Siemens or whatever for just 10-20k dollars but holy shit they are BIG and HEAVY!

This is true: Moving 600 Kg one stair down was a challenge.