Looking for an economical XY (well and Z) solution that is not for the normal use

I'm looking into the possibility of making a machine to aid me in circuit board assembly. I don't want to make it too complicated or I end up designing a machine that I can get off the shelf anyway or I end up reinventing the wheel and spending all of my time on the new wheel design instead of the projects I want to make with the thing.

I need to move a table in X and Y with an accuracy of 100µm (50µm would be nice but like I say I don't want to reinvent the wheel). The table will move around under a gantry that will have the part picker/placer on it that will travel in the Z. The reason I am not planning the table to move in one axis and the picker on the gantry in the other is that I want to avoid the risk of having parts fall off my picker that is pneumatic but it's an option if I can sort the picker out to be as robust as possible and if it solves a load of other problems.

200x300mm movement should be enough but given space that could be a bit bigger. I've been thinking something along the lines of a couple of those sliders with bearings in them on each axis with a stepper motor controlling them.

I don't know if what I want exists off the shelf or if I should just settle on cobbling my own together.

https://www.amazon.co.uk/Roeam-Upgra...0YDMXV00VP4A59

This sort of thing might do. I can take the motor off and replace with my own mechanism, but I am not sure what the T8 screw means in terms of movement. They say they have 42 stepper motors, is that 42 steps per revolution?

T8.= Trapezoidal screw - that's likely to bugger your positional accuracy requirement unless you introduce some anti-backlash mechanism. Stepper "42" - can't see that in the link, but could be related to the frame size that translates to a NEMA17 motor - it's just an indication of size and expected power range. Most common steppers are 200 steps/1.8 degree/step resolution. On a T8 that's 25 steps per mm or 40 micron - but you could pretend that micro-stepping offers higher resolution. Of course that assumes everything about the screw thread linearity / accuracy, as well as the frame geometry. Do you really need 100 micron accuracy? I don't think so (I'm pretty sure that working with 0402 components I'm not that accurate by hand).

How are you planning to feed your P&P - what's the component delivery mechanism? How are you going to achieve component alignment (and necessary re-alignment...particularly with vibration moving through the frame). Have you not thought of solder stencilling or solder-paste deposition as a "glue" for the components before placement?

I spent a few months in a clean room pick&place PCB assembly room when I was "a lad" - okay, quite a few years ago, but working incredibly compact boards (and high layer-count PCs) and the early generations of SMD devices. P&P with a less than heavy investment can go spectacularly wrong quickly.

I'd be interested to hear how you get on; unfortunately none of my projects get anywhere near complicated enough to make me desire such a machine myself.

Edit: reference to 0402 - I tried to prove I could... I much prefer 0603 for sanity, or 0805 to rent out as student digs :).
Surface tension of molten solder will draw components into alignment over distances substantially greater than your 100 micro accuracy.

What sort of backlash can I expect? 40µm is plenty. These days IC's have pin pitches down to 0.5mm. I'm not looking for an automated solution. I tend to use a lot of the same parts so it would save a lot of time just to load one BOM line at a time so that the machine can automatically pick each part up once told where the first one on the tape is and be driven to the location where it is set down. My plan would be to feed it the coordinates of each part on the board but that I give it the OK before placing a part so that I can make a manual adjustment if required. So for the average passive that there are a lot of and that do not require too much precision it will probably work automatically and quickly but the IC's that I won't be able to pick up reliably anyway I can move precisely into position by hand after it goes to about the right location.

Boards are stencil pasted before part placement. the paste indeed will hold parts in place which is why I'd rather move the board around than the PnP head.

Large items that there are not many of I would still put on by hand. I'm trying to speed up the fiddly small parts that there are lots of and not try to do the easy big ones that would be near impossible to pick up and move like large electrolytic's as I can put them on faster by hand anyway

Actually, thinking about it - you can design out backlash - you'll be moving from the component pick-up in one direction across both X/Y axis - so backlash doesn't really play as much a part as e.g. a router. You'll still suffer from screw accuracy, but I think you overestimate your accuracy requirement in any case. You can't eliminate backlash, but you can probably get it down to a reasonably tolerable level. You might find few people with experience in the DIY P&P domain - you might have to have a play here and come back with specific questions. Clearly you're going to be developing a lot of the system and software yourself - but I can't see anything that would particularly worry me if you're playing man-in-the-loop with the placement.

We want vids :)

If the backlash is known I suppose I can factor it in when I reverse direction just like I have to when using a manual miller and the Vernier's on the handles.

Not sure what to do with the software yet as windows programming is not my thing but have been told VB is dead easy with lots of tutorials. It will likely be a work in progress starting with the mechanical build and a simple stand alone controller I can punch coordinates into followed by talking to it from a PC with a USB/serial converter.

Yes it will be very much man in the loop, it's really pointless trying to over automate, more of a helping hand that saves me being bent over a table for 3 hours. I just built a board with 219 parts on it at an average rate of one per minute. A lot of time was wasted trying to pick up the 0603's that then just stuck to the tweezers. I just bought one of those hand vacuum pickup tools to make it easier but then started getting ideas.

To stop backlash can't I put a threaded item that is rotation locked to the moving body with a spring between the two to keep the table or head on the gantry always pushed one way?

Budget?

cheap but without making it useless! basically you can buy pathetic hand tools for £2k that you just put your arm on and slide around from part pickup to placement but I see that as problematic as you need fine motor control and in the demo of how fast you can go parts are not dropped right and of course they neglect the time taken to hunt out the parts locations so they are just matching footprints for effect. If I can buy a good base with motors and drivers for under £300 then that's a really good start but a bit bigger than 300x160mm would be good as if I can put multiple boards on (they can easily be 200x100mm each) this further greatly speeds up the processing even as a man in the loop.

https://www.amazon.co.uk/VEVOR-Machi...12981&sr=8-100

This one actually states 50µm repeatable accuracy, I guess as it's a bit more heavy duty but not sure how I hack into the motor controls.

https://www.amazon.co.uk/VEVOR-Engra...&sr=8-277&th=1

Finally one that admits to 80-100µm accuracy not bad, that's within 20% correct for a 0.5mm pitched part but for just over twice the money the other one is twice as accurate with twice the working space.

Quote:

---

Originally Posted by SparkyLabs

I guess as it's a bit more heavy duty but not sure how I hack into the motor controls.

---

If you're in the electronics field then 'hacking' the motors is the least of your worries (though carries cost) - standard steeper motor drives and some bespoke control system. Concentrate on the mechanical as that presents the greatest challenge to deliver against cost. You seem to be trying to adapt a machine to your use-case, might be easier designing a machine to actually match your requirement,

well I'm not the greatest wiz with software. The cheaper machines look like they have a simple protocol that I could talk with a micro controller or if it comes to it cannibalise the manual controller to simulate button presses. The more expensive machine looks like it is much more solid. Ultimately I need an XYZ table which is what any router/engraver/miller is once you toss off the head. If I try to make the same mechanism it will quickly becomes at least as expensive. With these things I just need to toss off whatever tool there is on them and put my own on the already convenient mount holes.

The more expensive one talks about MACH3 control. I thought that was a CNC program not the communication protocol. So the MACH3 controller has a USB socket so has to be driven by a computer unless there is a USB/serial driver in there so that I bypass the USB and go straight to serial with a micro controller. But as ultimately I want to computer control this as that saves me making the keypad and the display not to mention bridging with a computer anyway if I can control that USB from a VB program that is practically my solution out of the box minus the PnP head.

Quote:

---

Originally Posted by SparkyLabs

The cheaper machines look like they have a simple protocol that I could talk with a micro controller or if it comes to it cannibalise the manual controller to simulate button presses.

---

In the interests of transparency: I'm not convinced that anything you're proposing will be easily viable - but not for the reasons that you've identified - the bits that scare me is the whole component management and handling. XYZ is straightforward, and to some extent the machines that you've identified would be broadly usable (but think of the ancillaries that you'd need to add - you need something that you can build upon). But it could be an interesting experiment, and you'll learn from it.

You mention "simple protocols" - for the machines linked they either come with a little Arduino / combined stepper driver module running the GRBL software, or with some form of BoB controller intended to connect to a PC to be driven from Mach3. The latter could be USB (popular), parallel (old-hat) or ethernet (not in your price range). Just to explain what these options are doing: The typical tool chain involved with 2.5d routing/engraving takes a design from modelling software into a CAM process (software) that generates a series of machine motion commands (e.g. goto X,Y... goto Z(down).... gotoZ(up), goto X...) - most typically in a format called 'g-code'. What the GRBL and Mach3 software does is trajectory planning - translating these "simple" instructions into a complex series of motion events to the steppers (including acceleration/deceleration, linearisation of arcs, etc) as well as affording some manual control. That, you *could* use, but it would be clunky - and it wouldn't integrate well with the component feed system (you'd have to play around with some form of hacked A/B/C axis) or a more complex macro system driving external hardware, or with the man-in-loop verification before a component is released to the board (e.g. component rotation prior to placement). It'd result in spaghetti code of the worst form. A g-code solution is not appropriate, in my mind, for your problem.

Far easier, in my mind, considering the very simple, and very different requirement that you have, to disregard (or re-appropriate) the control hardware to move away from GRBL/Mach3 and simply generate the motor stepping signalling. All you need for each axis is a "step" (clock) signal and a coincident direction (boolean) signal to move the stepper one step in the necessary direction. You can/should consider some simple mechanical sympathy with gentle acceleration but the basic behaviour is to simply Seek-home-position, move-to-component-tray, pick-up component, move to target XYZ, verify with operator / machine-vision/control for orientation - rotate as necessary, micro-step to adjust for component holding offset on tool, move to board (Z0 minus component height), release component, recover to safe-Z, then traverse-to-home and rinse-repeat. In describing this general placement strategy it should be clear why P&P needs a very different approach than the usual g-code interpreters - there's too much closed-loop control required - and in my view needs a different solution. That's why I say to cut your losses at the steppers (or stepper drivers) and throw another control system in its place. You may find there's open-source software out these with the usual hobbyist/hackspace type of sites that can help with that, although the software shouldn't be difficult to derive a basic capability.

BTW, the USB/Ethernet controllers mentioned - you probably want to avoid these (or replace them) - these migrate parts of the trajectory solution from the Mach3 software into a local microcontroller - Mach3 would packetise elements of the trajectory solution to the controller for localised signal generation... that's the worst place to try to "hack" into the protocol. The simpler parallel-port BoBs are much easier to interface to, or bin all of that and use local signal generation.

My view: You could cobble something together, but unless there's already an open source hardware/software solution then this forms a substantial project in itself. Don't expect that you can find a COTS/turnkey solution.

Thanks Doddy

So the USB controller is indeed complex and a bit specialized. That is why I originally saw the cheap 300x180 ones as they have the simple controller that you describe that I could then put my own electronics on that could be connected to a computer. I'm not really set up to build good XY tables and all of the other mechanical stuff. I work for a mechanical engineering company so have access to people that can do some milling and turning of parts for me which I can draw up in 3D CAD.

My original thought was two sheets one on top of each other with sliders and to drive them with stepper motors or linear actuators with encoders. But the motors are about £100 each before I have done anything else. So I wondered if there were any existing things out there that would have a lot of what I need.

I will still need to design a picker that yes will need to rotate. I'm thinking something along the lines of a shaft with O-rings that can swivel in a body so that the air can come in between the two seals and go down the centre, then it can be turned from above.

PnP machines typically are complicated with camera's and all which is why I don't want to go there and make this a helping hand. So I have been through a few ways of doing it.

1) manually drive it around with direction buttons like the controller on that simple engraver thing, but that will take forever and longer than by hand
2) Have a keypad and punch in the coordinates of the part to pick up and the location
3) have a program drive the machine, this does away with the interface and HMI are a project in themselves. So I learn to do some VB that will provide an interface and commands to electronics that I need to make.

I think the third option is as much as I can do before it becomes easier to buy an off the shelf system. So camera's and software to inspect the part picked up is not an option, this is where the entry level fully automated systems fall down. Part feeders are also out at these are complex, bulky and are projects in themselves. So I have done away for the need for a camera with MK1 eyeball putting me in the loop. I will design holders for the standard tapes that the passive parts come on and the IC's. The tapes will be slid into these holders on the table with the film removed. Passives are always on a 4mm pitch so once I have learnt the location of the first part the locations of the rest are known and it is for me to confirm that the picker is in the correct location above the part to be picked up. If it is not I can make corrections that my program can remember so that if say after 10 parts it has drifted when I correct for the 11th the next ten are most likely to be fine. The parts are then taken to the location of the part is to be placed in and lowered just above the board where I again confirm correct location.

Now I won't be bothering with parts taller than 3-4mm, basically just small IC's and the small passives up to 1206/1812, no large capacitors or larger chips like TO stuff that can be hand placed it's just worth making a machine capable, this is about speed not full automation, the automation is to take out time I waste not be clever. My expectation of the accuracy is that the passives are in fairly tight pockets in the tapes so their initial pickup is likely to be accurate enough and these are the parts that there are most of so the least error is on the parts I have most of. The chips tend to be more loosely packed so their pick is not guaranteed and I may have to correct the placement of each one, that is fine as it is the chips that I find don't always self align on reflow and they need to be placed more accurately than I can by hand. So I don't mind these taking a bit more time but the idea is that as I can more them with a machine more accurately than I can by hand it saves the time of trying to get the solder bridge out later or a ruined board.

The picker will have a camera so this will save the strain on my eyes. So really what I m trying to do here is add the magnification that on my eyes is tiring but as a picture on a screen would not be. I avoid bending over a bench and getting back ache and I can place parts more accurately when required than I can by hand. Obviously doing such a system does mean that I have to be careful to not misjudge how good a human is at doing certain things that it is harder to get a machine to do. It will take some time to position parts that need manual adjustments but if I can feed the machine the coordinates and it broadly gets the parts there that is faster that me looking them up by hand.

this is by no means meant to be a pick and place machine but a helping hand. I am just trying to divy the work up between me and the machine. I will do what I do best which is visually check all is well and give the OK and the machine will do what i waste time doing which is trying to get parts out of tapes and find where an the board they go.

I don't know if this helps but https://www.youtube.com/channel/UC5j...yZGhhokC5j49tg

Quote:

---

Originally Posted by SparkyLabs

this is by no means meant to be a pick and place machine but a helping hand. I am just trying to divy the work up between me and the machine. I will do what I do best which is visually check all is well and give the OK and the machine will do what i waste time doing which is trying to get parts out of tapes and find where an the board they go.

---

Okay, I have more questions (and more challenges) with what you say - but I think you know your own mind. An hour soldering under a stereo mic is enough for me and my back - so I appreciate your intent. Despite this, progressing this will be a project no matter what.

If I was looking to do this within the use/cost constraints you say, I'd pick up the most physically robust of the GRBL-based machines and look to adapt the firmware or integrate with the control software a joystick/joypad controller. The mechanical builds are... appropriate for your use-case (and it pains me to say that). You're not going to get fantastic placement speeds - imagine 5 seconds traverse speeds in the working area. but that gives you time to think. I've since read that the thread pitch on at least one of those machines is 5mm, so that's higher resolution (but lower speed) than I mentioned earlier.

It's a project no matter how you look at it. It's do-able, I wish you luck. Personally - mic, a pot of of tweezers, soldering pencil and frequent breaks. But I rarely go north of a 100 components on the occasional board.

The basic professional P&P machines position the placement head above the pads on the board, using the info from the manufacturing files. That's one of the most useful functions you need here. The next stage is presenting the correct component for placement, either by operating a carousel and/or compartmented tray or tape and reel dispenser and moving the head there. Then the ability to pick up and orientate the part, using a sucker and a (manually) rotating head. Even if you have loose parts in a tray and have to turn them the right way up, this can be a big bonus.

If you can get those elements covered, you have a pretty useful system. It will help you to select the right component and place it in the right location in the right orientation. Determining the correct component and its required position from the manufacturing files would likely be one of your main challenges.

Incidentally, I have a rather nice (professional) solder mask printing frame I want to sell on (don't recall the correct description). I keep planning to get it on ebay but it never seems to happen. If you are interested, I could send some pics.

Quote:

---

Originally Posted by Doddy

Okay, I have more questions (and more challenges) with what you say - but I think you know your own mind. An hour soldering under a stereo mic is enough for me and my back - so I appreciate your intent. Despite this, progressing this will be a project no matter what.

If I was looking to do this within the use/cost constraints you say, I'd pick up the most physically robust of the GRBL-based machines and look to adapt the firmware or integrate with the control software a joystick/joypad controller. The mechanical builds are... appropriate for your use-case (and it pains me to say that). You're not going to get fantastic placement speeds - imagine 5 seconds traverse speeds in the working area. but that gives you time to think. I've since read that the thread pitch on at least one of those machines is 5mm, so that's higher resolution (but lower speed) than I mentioned earlier.

It's a project no matter how you look at it. It's do-able, I wish you luck. Personally - mic, a pot of of tweezers, soldering pencil and frequent breaks. But I rarely go north of a 100 components on the occasional board.

---

Yes this will be quite the project as it is which is why I am being careful about not trying to design anything I can already buy and to avoid features that are too complex for what the intent is and just keep reminding my self that this is to be a relatively simple aid not full automation.

At the moment it's taking over a minute per passive, some of that I can reduce with a hand held sucker tool if it works as a lot of time is spent trying to get parts up the right way having tipped them out of their tight tape pocket. So a 5s movement time is fine really. The few 300x180mm machines that state an accuracy say 80-100µm which is not bad and the controllers seem easy to work with. I spend several seconds as it is entering the part designator into the PCB software to locate where the part goes so in less than that time a machine can just go to the location and in less time than I spend carefully rotating the PCB around and "getting into position" the Z axis can lower and await my confirmation after any minor adjustment. At the moment to deal with rotation I rotate the entire board as I have to go in from the side that the location is closest to and be very careful about not knocking anything else. My last board was 175x90mm so most parts had to go in from two sides rather that all 4.

The PCB design software will output a part position file. I use KiCad so I can add my own fields to parts (and I am sure other programs do it too) as I create them which I am doing a lot of at the moment. So I can put into the bill of materials columns for things like the height of the part in the tape and on the PCB so that the Z axis knows how far to travel and stop at pickup before confirmation and on confirmation. I can put in the pitch of the tape. So if I load the BOM and position file into a program I will have all of the information it needs about the part to pick up and where to get more after the location of the first is known and where on the board to move to.

So yes if I use one of those off the shelf systems I can then concentrate on the picker head and software as the one thing I am not equipped to do is start making CNC beds. I can do a controller that over a serial to USB adapter will talk to a PC so that apart from the 3 axis's that the basic bed comes with I can also issue from the same USB port commands to control the rotation stepper motor, pump, pump solenoid and anything else I need. The only other USB port I will need is for the camera that I will have to look into and should be available off the shelf that sits on the picker mechanism to give me a nice big picture on screen of what my picker is doing. I think that is as simple and functional as I can make it without trying to do the impossible.

For example rather than relocate to an origin on each operation if I have to manually correct I can decide to tell it that it made a permanent positioning error and to use the corrected location as the new true location so all of the position offsets can be recalculated on the fly if it drifts.

Quote:

---

Originally Posted by Muzzer

The basic professional P&P machines position the placement head above the pads on the board, using the info from the manufacturing files. That's one of the most useful functions you need here. The next stage is presenting the correct component for placement, either by operating a carousel and/or compartmented tray or tape and reel dispenser and moving the head there. Then the ability to pick up and orientate the part, using a sucker and a (manually) rotating head. Even if you have loose parts in a tray and have to turn them the right way up, this can be a big bonus.

If you can get those elements covered, you have a pretty useful system. It will help you to select the right component and place it in the right location in the right orientation. Determining the correct component and its required position from the manufacturing files would likely be one of your main challenges.

Incidentally, I have a rather nice (professional) solder mask printing frame I want to sell on (don't recall the correct description). I keep planning to get it on ebay but it never seems to happen. If you are interested, I could send some pics.

---

Hi Muzzer

Yes you are correct and I should also be able to rotate part placement. So long as the parts are always the same way around and the XY of the PCB relative to the parts or the bed is known then the rotation information in the position file will allow that to be automated but again with manual confirmation I can have a button that rotates parts 90 degrees if the are wrong before placement confirmation because as they are all passives with no polarity that is very quick to correct for and of course I could just do +90, -90 and 180.

Yes by all means send me some pictures of your "stencil printer". I have just bought an SD-240 for frameless stencils that makes life easier although the 175*90 PCB was a struggle and did not come out so well. I am always trying to get my cheapskate employer to buy me more gear so maybe they will be interested and hopefully it's compatible with what I have. I am currently working at home which makes assembling boards a lot easier as I have my own equipment which is more than they have and more importantly my lab is a large bedroom not a shaky shipping container (I have already binned an item worth £300 just because someone walked across the floor and everything shook so badly that the culmination of a difficult attempt to solder a missing part to board that was very difficult to get up to temperature ended in failure and i could not be arsed to have another go or risk damaging it anyway with a second reheating attempt).

Quote:

---

Originally Posted by Muzzer

Incidentally, I have a rather nice (professional) solder mask printing frame I want to sell on (don't recall the correct description). I keep planning to get it on ebay but it never seems to happen. If you are interested, I could send some pics.

---

Not a Blundell unit, by any chance?

I was in dialogue with someone in you neck of the woods a couple of months back about (me) getting rid of a reflow oven (then I talked myself out of it)... not you by any chance?

Quote:

---

Originally Posted by Doddy

Not a Blundell unit, by any chance?

I was in dialogue with someone in you neck of the woods a couple of months back about (me) getting rid of a reflow oven (then I talked myself out of it)... not you by any chance?

---

My SD-240 was from Blundell yes although they are made in the Netherlands and shipped through Blundell, people sell the same unit under different names. No it was not me you were talking to about the oven, I'm sorted there for now.

If you've not seen it already - google smoothieware and pick and place. You're not alone.

Oh I know there are loads of attempts out there which is why I am wary of not falling into the same trap of trying to develop a solution that competes with £15k entry level machines that cost that much due to the development and then spend my life developing the machine and not my work. This is why I want to get as much off the shelf stuff as possible and just do the work that ties the bits together which will teach me software I need to learn anyway. Also for just one or a few boards it is probably not worth setting all the feeders up and once you try to do that you have made a limitation for yourself in that you won't fit all the parts so still have to keep stopping and reloading. This is why I would load one part at a time, take the head to the first and tell it to start from there in this direction and that the parts are at this pitch and avoid trying to be too clever.

The other problem I have with these open source projects is that they rise and fall and you have to be able to keep up with them and know what they are on about and if you don't your out of the gang. I tire of being told that if I don't like the way their open source project does something then go and write a fix, open source seems to mean that only the nerds that can design one from scratch anyway can make it work, I design PCB's I don't want to become a co developer on a project that sort of does what I want but not quite but just fails in the end. I am literally looking to solve my limitations which are:

1) not bending over a table
2) not having to wear magnifiers
3) not wasting time finding parts on the board myself when a machine can be told as the data is already available but the coordinates are not that useful to a human
4) not waste time trying to get parts out of carriers and picked up the right way around
5) be able to make precise parts placements that I can't by hand even if that takes longer but saves the build.

Wasn't me - I only have the screen printer. Mine's a German SEF brand. Professional stuff, originally at professional prices(!). Needs an appreciative home. Uses std frame and stencil sizes. You can get the stencil made at most PCB houses.

I posted an ad in the For Sale area so I don't clog up this thread. Sorry, have to dash now, so rather brief post / ad. BTW, I'm near Blackpool.

Does anyone have any information on the GRBL controllers/motor drivers? I see the machines on Amazon have the manual controller but there are lots on ebay without the manual controller ("joystick") bust still the same motor driver with the socket for one. All have USB connectors.

My recommendation.

For little money, get polabs pokeys ethernet, and machx.

For trivial money, get screws that are tight with small rise.
You can easily get 0.01 mm or better repeatability for not much money, around 300€.

My own 120k€+ machines run machx, and I was 2011-2012 the country sales manager for Haas, and sold 6M€+ of cnc hw to industrial customers.
Pokeys is the best cheap option.
Cslabs -IP-S is the best there is, I use it on my lathe, but it´s quite expensive. 2000€ all-in.