Rail/Slide Based PNP Machine

[Anthony Webb]

Wanted to start a new thread to report on my findings.  Up until now I have been posting stuff into the thread discussing inexpensive PNP options.  For those who havent followed there I will recap here.  I'm on the hunt to build a cartesian pnp.  I'm not looking to try and build the cheapest machine out there, but I am out to build a highly reliable and accurate machine for the least cost possible.

I've been through countless 3D printers, CNC machines, and other various motion projects and have a pretty good feel for what I can live with.  For that reason I have chosen to go the rail/slide and belt route as opposed the other various options out there.  I think I will be able to find a motion platform for this project that checks all the boxes (cost/quality/simplicity).

To this end I feel the need to know how the various price points of rail/slide options compare to each other.  You can pay $20 for the inexpensive stuff, and $1000 for the real pricey stuff.  Question is, when you pay more, do you get more?  I'm about to find out.  But I am looking to find out EXACTLY how much more you get.  I'm going to measure each set of rails against each other in a torture test of sorts using a tools called firenodejs that Karl Lew has put together (https://github.com/firepick1/firenodejs) It will be a ruler capable of drill into just how well each of the rail/slide combos perform compared to each other.

In order to do this analysis I needed a common platform to test them with.  Task 1 was designing a test rig.  I started with an actuator.  Small problem is that I didnt actually know any CAD tools, but I knew what I wanted.  I wanted to design something that was easy to build, modular enough to be expanded/shrunk to meet varying budgets, and I wanted the actuator to be able to mount onto other actuators to give a nice x/y platform that would ultimately be the bones for my PNP.  After days in front of Fusion360, Here is what I came up with

The nice part about extrusions is that they are relatively easy to work with, you can cut them easily, bolt bits and pieces on without much trouble, etc.  The beauty of what I ended up with is that I could very easily bolt on various rail/slide combinations for testing (WIN).  I have a CNC and 3D printer, but I wanted the brackets to be something that a guy could make with a drill press and some delrin, so I kept them very simple. With the actuator design in place I set out to join them together in an x/y test rig.  This would require some custom brackets to join together the actuators.  Here is the next rev

As you can see above, I really wanted to see if a passive rail would fly. For simplicity and cost I am going to first see if this works, if not I have lots of workarounds I'll try too.  I initially targeted 300x300, but after looking around I found that 500mm rails would be a better size to target.  I felt ready to go ahead and start buying parts.  The last 3 weeks have been spent bringing in the hardware to build what you see above.  Of most importance was the rail/slide combos. I bought 3 of everything, the cheap chinese hiwin knockoffs, the run of the mill authentic hiwins, the uber expensive THK option, 12mm, 15mm, as well as some of the igus drylin stuff.  Companies found out what I was doing and sent me samples of their product to test too.  Basically, I have enough rail/slide combo hardware to provide a really good picture of what works and what doesnt, but more importantly to know where is the point of diminishing returns, ie at what point does spending more money not necessarily buy you more quality.  From there I figure I can pretty accurately decide how much quality I want to afford :)

I'm not going to make any predictions at this point, but I can tell you that I was not expecting the good stuff to be that good, nor was I prepared for the bad stuff to be that bad. I nearly didnt even bother trying to test the chinese set, sure it was 1/15th the price of some of the other stuff, but man, I couldnt even move the slide on the rail at all, I'm not joking.  But after a dose of WD40 and white lithium grease it at least slides along the rail now. Definitely not the butter smooth action of the THK's out of the box.

Off we go, I cut brackets and cleaned off a spot on the bench and began assembly today.

Chinese rails are up first, and as for as much as I had written them off initially after getting them all locked in I must say that they seems to move great.  No slop at all.  The passive rail I was worried about?  Well for now I am not seeing a problem, but the proof will certainly be in the result.

Thats where I left off today.  Mechanical is complete, electrical is next.  I have a tinyg laying around that I have tasked with driving the motors I installed.  These are off an old 3D printer, nema17 400 step (.9) per rev.  I am using 32 tooth 2mm GT2 belt, not sure what kind of speed I will drive with that, but excited to give it a whirl.  I am a little nervous about the steppers, I have had such great luck with my clearpath closed loop servos I almost hate to even try anything else, but if I can make steppers work I could save a bundle of money along the way so I am starting with them.

I'l be sure to post my findings as I go, the good/bad/ugly.  As always if there are any questions or feedback please feel free to call me crazy, offer insight, etc.  I'm very new to all of this and know many of you have been through, at the end of the day I hope to be picking and placing with the rest of you!

A short video of the rolling chassis.

https://youtu.be/vqTw7lJbhqU

[Michael Anton]

I'll look forward to seeing your results!!!

Mike

[Jason von Nieda]

Awesome work Anthony! Can't wait to see the results! 

One quick note: One thing that isn't always obvious is that those rails (and linear rod / bearings, too) all *require* lubrication before use. You can really quickly destroy a carriage / rail without it. And WD-40 is a no-no. It actually strips grease from the bearings. The white lithium is good but it does tend to get gunky. I really like "Super Lube". I have this bottle and it has lasted me years: http://www.amazon.com/Super-Lube-51004-Synthetic-Viscosity/dp/B000UKUHXK/ref=sr_1_2?ie=UTF8&qid=1450505017&sr=8-2&keywords=super+lube

I've found that the Chinese bearings tend to come from the manufacturer pretty dry and if you try to run them without lube it's like chewing marbles, but as soon as you lube them up they are pretty smooth. Very interested to hear about your much more scientific results!

Jason

On Fri, Dec 18, 2015 at 9:40 PM Michael Anton <3d.m...@gmail.com> wrote:

I'll look forward to seeing your results!!!

Mike

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e4df3a36-2a33-4664-a9ca-7b5febb787ab%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Graeme Bridge]

Nice work, I haven't had a chance to do much more with my design as i have been messing around on my other project5 a large format 3D printer design.

What i have been doing though is an epic amount of work on researching lead screws.

The printer is designed around these now and will be the test bed for the PnP.

Come the new year when christmas and tax season are finally gone i will look to start ordering parts to test the screws.

One thing i have come across is this ball screw spline shaft, I'm trying to hunt this down at the moment but i think it could be the key to the head 

[Robert Walter]

Anthony,

I am currently converting a Chinese machine to openPNP but will be following your progress quite closely, as I will likely start a ground up machine design in the spring. I too am considering ball rails and belts, but I may go the lead screw route on the Y-axis to eliminate the inherent stretch and syncing requirements of having two belts. This axis would have the highest load and largest span, so it only makes sense to consider a ball screw. The cost of belts, bearings, pulleys and shafts to link everything together starts pushing the price up towards a lower end ball screw, so it is a plausible solution.

With regards to rails, given the placement accuracy requirements of a typical pnp, you will likely have satisfactory results with just about any ball / rail guide type system. Pay close attention to build quality, flatness, etc  of your frame / table and if you find a little slop here or there, some preload on a axis rail can work wonders. You aren't likely to be exerting any forces that are even close to the limits of a ball rail (smallest rails aside), so pay closer attention to the axial play. That is the amount of play in the bearing in each axis. Using multiple block on each rail helps, and some blocks have preload adjustment built in, otherwise, pre-loading between one static rail, and one rail that you can adjust a few thou will suffice.

Also, if you hare planning on just building a one-off machine, ebay is a great way to pick up surplus / overstock high quality rails at really cheap prices. Still likely a little more money than cheap offshore stuff, but a ground THK rail and bearing block set is a work of art that will run well for years. But of course, the higher quality your rails, the better your build quality needs to be, as the tolerances have to be spot on in order for the rails not to bind. Essentially, you build tolerance has to match or be less than the axial play of your rail, otherwise it will bind. That is why most low cost machines use shafts, as they typically deflect enough to reduce binding, at least in one direction.

I find that getting rail bearing blocks with grease nipples / zerks is best if you can get them. It makes the whole lubing procedure much simpler as the lubricant mostly stays with the block internally and kept in place by the wipers.  If you are going to lube the rails, lithium grease works well, but is a pain if you are in a somewhat dusty environment. Picking small smd components out of the grease is not fun either!

I tend to go the grease nipple route, and wipe the rails with light machine oil / sewing machine oil when necessary to keep clean and lightly lubricate / add corrosion protection.

Good luck with your build.

[Arthur Wolf]

On Sat, Dec 19, 2015 at 8:29 PM, Robert Walter <ttsther...@gmail.com> wrote:

Anthony,

I am currently converting a Chinese machine to openPNP but will be following your progress quite closely,

If you document how to do that on the Smoothie wiki, there's a free Smoothieboard for you :)
 

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/b9c803d4-c7a2-45a8-b3f7-669061594a94%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Courage et bonne humeur.

[Robert Walter]

By the way, beautiful work so far. Also, you may want to consider linking the two bottom rails with a shaft so that you can have a belt on each side. Driving from one side will inherently cause binding. It has been done before, and will almost certainly cause you grief one you start adding weight to the axis' on top of it, and adding acceleration / deceleration into the mix.

Rob.

On Friday, December 18, 2015 at 9:09:18 PM UTC-8, Anthony Webb wrote:

[Robert Walter]

Arthur,

Thanks for the offer. I will definitely keep notes on everything. I already designed up a buffer board that will plug into the smoothie step / direction ports in order to work with my external stepper drives. They need about 15mA drive capability. Those boards should be here late in the week (hopefully before x-mas). Pretty simple board, but a necessity.

Rob.

[Anthony Webb]

Wanted to post a brief update on progress.  After  some of the feedback I got I felt it was important to lug around as accurate of weight as possible to the system I am trying to build so I took the time to make a couple real simple brackets to do 2 things: 1) hold the camera and provide mounting holes for the PNP head and 2) mount to a chunk of extrusion to act as the mounting point for the new brackets.

So, without modification the brackets I have been using on the slides are now able to hold the head/camera secure and snug, slick.  I'll be testing with a full load so there are no surprises when I decide which rails are the best for my application.  I realize that the head is too low right now, the final machine will be on legs that are long enough to give the proper drop.  Right now I am only worried about getting some data back from the rails/slides I have on hand.  Getting closer everyday.

[mojalovaa1]

You are buy or make head?
I m attach my head design that you can see , if you look I m add limit switch for boot nozzle ( head ) that can know when is on home position.

[Anthony Webb]

As this is my first try at PNP and never having seen detailed head plans before I simply bought a head used on other machines so I could see how they work.  Now that I have seen how they work I think I could design my own and likely will at some point.  For now I'll use this one http://www.robotdigg.com/product/559/Headset+for+PNP+Machine

[Anthony Webb]

Some new developments.  The motors I had wired up to the tinyg were 0.9 400 step per revolution steppers.  They were moving pretty well, but I wanted to ensure that they were always powered during my tests.  Somewhere along the way in tinkering with them they overheated and melted almost entirely through the 1/4" delrin mount.  So, I had to build new mounts, haha.

At any rate, the 400 step steppers I had just didnt have enough torque to hit the target rapids I was looking to achieve (250mm per second)  So, after building new mounts I installed 1.8 (200 step) steppers from my old shapeoko which seem to pair very nicely with the tinyg and hit the target speed no sweat.

I am in the process of formalizing a test suite to use, but I have done some preliminary tests and was pretty surprised at what appeared to be very favorable results.  I even moved all the weight of the gantry out near the passive rail and still had near perfect results.  If that is the case, these tests may be fairly boring as I am only able to test to about 50 microns, and if the cheap chinese rails are testing consistently within 100 microns I dont know how much more perfect I would expect to find.  I still intend on working my way through all the rails with actual repeatable test cases.  Not too far off.

[Anthony Webb]

Made some more progressing in finalizing the bed.  I am very happy with the rigidity of the machine, it is pretty bombproof in this config.  Getting some 3/4MDF for the bed which will then have 1/8" galvanized steel on top for a firm surface to magnetically attach board holders, strip feeders, trays, etc in whatever config I need.  Left plenty of extrusion exposed all the way around to attach feeders we well.

[Jason von Nieda]

Looking good Anthony! 

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d549996c-d476-4fd7-a741-c62bbdc97724%40googlegroups.com.

[Anthony Webb]

I realized that I didnt get a great shot of the down looking camera yet.  It kind of hides out behind the Z head.  At any rate the Z head attaches to a couple delrin caps that cover a small chunk of extrusion which mounts to the slide.  There is a set screw  (which is probably not needed, the delrin is a very tight fit)  The camera appears to be supperb and has a nice focus ring up top.  Only downside was that when focusing, the wire would want to turn inside the housing, which in turn rotated the camera too.  So I took off the cap and put a small dab of silicon in there, problem solved. Time will tell if this camera is a great solution or not, but for now openpnp seems to like it and I am happy with how rigid it is, the minimal offset from the pick tips, easy focus that is very accessible, and how well it nestles in there without needing separate mounting bracket.

[Jason von Nieda]

I've seen a lot more people using that camera, or at least one similar to it, recently. Sorry if you've posted it before, but do you have a link to that one? There seems to be a lot of cameras that look just like it and I don't know if they are all the same.

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/974dcd3f-233c-42ce-a720-2e0b4df2ac99%40googlegroups.com.

[Rich Obermeyer]

Nice PICs Anthony.

Thx for sharing your assume progress.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jxqTZqkBV_TMY%3D2ZCPFrrOSNmJpnuOUMn%2Bt-P2rw4b84w%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Rich

[Anthony Webb]

Yes, I am anticipating doing a full BOM for my project and may even make available the brackets for those who want to jump start their own PNP.  Until then, here is the camera I am using: http://www.ebay.com/itm/121292898829

I also am keen to try this one too:

http://www.amazon.com/gp/product/B0066HA1PU

The ELP I bought is pretty darned nice, and I will likely use it for upward looking, But I do like the form factor for this one for the downward looking if it ends up working well once I am picking and placing.  I have a little ways to get to that point, still am testing all these rails!

[Jason von Nieda]

Thanks Anthony, That Andonstar camera is the one I've seen referenced. I think a lot of the Liteplacer folks are using that one with quite a bit of success. I think I will pick one up for testing. 

I've been leaning towards using something like this for the head and the ELP (or a Logitech C270, also on order) for the bottom vision. 

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/855cb263-baa0-44f9-90ee-d07c3cbd1ac0%40googlegroups.com.

[Anthony Webb]

A little note on some of the test results, and a movie of the process.

As I mentioned earlier, I built a tinyg driver for firenodejs and am using that to test the precision of my design using various rail options.  I first mounted the cheap chinese rails I sourced from here: http://www.robotdigg.com/product/493/MGN12-1H-L600+Linear+Rail+and+Carriage+for+Kossel+XL

I have had these rails on for quite a while.  I can start by saying that these things are cheap in every single way, and I dont mean "kind of cheap".  The literally clank their way down the track, even after cleaning and lubing them up real good.  Of the 3 I got 1 was VERY bad.  I put that one on the X becuase I figured as long as it was directly driven by a motor it might be OK.  I took the best of the 3 and put it on the passive rail.  Hoping to make some lemonade.

Based on my tests the machine would move maybe 8-9 times nearly perfect. Below is a video of the cheap rails in motion, as you can see, pixel perfect precision (each pixel is about 50 microns) As you will see, motion is perfect during this leg of the test.  BUT, inevitably, and eventually, that catching, binding comes back to bite you, and when you bind in a movement, it might throw you off as much as 1mm. The binding only occurs in the Y axis, that is the one with that passive rail that I thought might bind on me, it does.  The good news is that is the rail with the worst rail/slide combo I had, and believe me, it is TERRIBLE.  Yet in all the test it NEVER was off more than 50 microns, pretty impressive.

This leads me to my conclusion on the cheap chinese rails... they will work, IF you actively drive them.  If you actively drive them they will be accurate too.  Be aware that you will need to crank up the pots on your stepper drivers to deal with the friction, but they can work.  That said, after I began the rail swap to the second set I will test, I found the rails/slides to be very dirty, a few metal shavings here and there.  I'm not convinced that these will last, so even if I could make them work, I'm probably not inclined to every build anything with them again. $70 bucks down the drain all in the name of science :)

https://www.youtube.com/watch?v=pz1ZOkmSzG0

Moving along to the HIWIN's

[reade...@gmail.com]

Hey Anthony,

Do the cheap rails have the 'seals' at the ends of the carriages? I noticed on the HIWIN carriages moving things around by hand seemed to have a bit of drag and catch, I took a razor to the rubber and removed the part rubbing and things became incredibly smooth. Granted I need to be more mindful about dust but I figured since its not on a mill/router type machine it should be pretty manageable debris wise.

[Anthony Webb]

Good point on the seals, I will take a closer look in the morning.  Odd you are seeing drag and catch on your authentic hiwins, those are usually pretty flawless.  I have 5 of them and 4 of the 5 are absolutely perfect.  I do have 1 of the 5 that is wanting to catch, which is the first I have ever seen.  I have never seen a THK catch or drag at all.

[reade...@gmail.com]

It may not have been bad enough to effect the machine, I also did not lube the rails with the 'nipple' after cleaning things off so that could have played a part. It just felt strange moving it around by hand with the seals on.

[Anthony Webb]

I was asked how heavy my Z head is.  While switching rails tonight I weighed it fully loaded and it is 1.7lbs (or 27 oz)

[Anthony Webb]

Got the final bed installed. size is about 22 inches square.  It is 3/4 inch mdf with a layer of 12 gauge (~1/8 in) galvanized steel laminated to it.  It is probably a bit overkill but gives us a nice heavy base to stick magnetic holders and boards to. 

[Jason von Nieda]

Looks great! Love seeing the progress - keep it coming!

Jason

On Thu, Jan 7, 2016 at 6:01 PM Anthony Webb <anthon...@gmail.com> wrote:

Got the final bed installed. size is about 22 inches square.  It is 3/4 inch mdf with a layer of 12 gauge (~1/8 in) galvanized steel laminated to it.  It is probably a bit overkill but gives us a nice heavy base to stick magnetic holders and boards to. 

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7f133f01-1eb1-4ca7-879a-0110c9169283%40googlegroups.com.

[Anthony Webb]

Spent some time with the rails from HIWIN over the last couple days.  And I can tell you straight up, these things are in a totally different class than the cheap chinese rails found all over the web.  Are they on par with the pricey THK's?  Time will tell, but I can tell you these things are butter smooth.  

I've yet to have any binding on a passive rail that is over 20 inches (500mm) away despite only running a single slide on the passive slide,  I bought 2 slides for the passive side thinking perhaps I might need to explore have a wider base on that side for stability, not needed.  I've done some pretty high speed jogging and analysis and the drift over the course of them has been less than 50 microns (near perfect).  In short, these things are great, definitely keepers.  

These are the first rails I have tried that I would feel complete confidence putting into production today.  The only knock?  They don't have a grease nipple like the THK's, I'm not sure if it is an option on these small 12mm rails either.  But for the amount of use that a typical maker PNP system might see it might not be needed.

[Michael Anton]

That looks great!!!  Here are a couple of things that spring to mind...

Is that the final location of the bed?  It looks like you don't have much room between the nozzles and the bed.  I can't imagine how you could get board supports, and component height in that short a distance, unless you are planning on having the PCB resting on the bed.  Which would mean that you can't do double sided assembly.

The mass that the steel plate adds to the machine may be a benefit for adding some stability to it.

Mike

[Anthony Webb]

Thanks for the watchful eye, I've never built a machine like this much less been around them so the likelihood of me painting myself into a corner is pretty high.  Funny thing is I have been worried about the Z height, only I have been concerned that I have too much room, not too little.  The good news is the bed can easily slide up or down the legs and be attached right in the sweet spot.

Here is how it measures out today...

There is 31mm (~1.25in) of space between the lowest tip and the top of the steel.  The max travel of the fully extended tip is 21mm, which leaves me 10mm (just under 1/2in) that my board mounts will need to lift up the board (at a minimum) so the tip can reach the board.  Same goes for parts.  Since I will probably be designing/milling my own mounting system I can really make them however thick they need to be, but I didnt really want them to be much over 1/2 inch.  Does this sound reasonable?

[Jason von Nieda]

Anthony,

That sounds pretty good to me. My build is quite similar and I use 10mm height for all of my magnetic accessories such as board holders and feeders.

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/15afd2bb-d272-41c0-8b1e-5658ee0a55a4%40googlegroups.com.

[Cri S]

As example Magjack components (ethernet port) have 17.5mm height. 

Now pcb can be 3.2mm height + 0.2 solder paste + 2.5mm populated components =

23.5mm . 31mm-23.5mm =  7.5mm . If smt tape passes below, like tm220 machine,

then it's 6mm max height of second side with little play. Sounds good.

If you raise the pcb to 20mm, then not can populate magjack, or more critical, cannot

populate the 10.2-10.4mm height caps and passing over with other components.

[Michael Anton]

If there were multiple magjacks on the board, it would not be possible to place them with only 31mm of clearance, and 21mm of Z travel.  One needs to be able to lift a jack up high enough to clear previously placed ones, so you would need 2*17.5mm=35mm of Z travel.  But that is perhaps a special case.

Mike

[Jason von Nieda]

Except for very specific circumstances that is a problem we could solve with software by calculating a movement route that would avoid the component.

Pull requests accepted! :)

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/fef40ec6-5d52-48b7-9153-adaf08104e1f%40googlegroups.com.

[Michael Anton]

The 21mm of travel will limit your max component height to about 10mm.  With a short travel like that, you will likely want the pick height of your components to be at the same height as the PCB surface (or slightly higher than), to make maximum use of the travel you have.

Mike

[Cri S]

It is not true, as for high components, the job planner places the
most far component first.
The point of reference is the unique camera used for checking
components that are highter
then 1/2 of free space. The only requirements is that there is at
least 1.5 mm clearance,
better if 2.5mm . I have calulated that in.

2016-01-09 0:38 GMT, Michael Anton <3d.m...@gmail.com>:

>>>> <https://groups.google.com/d/msgid/openpnp/15afd2bb-d272-41c0-8b1e-5658ee0a55a4%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>> .

>>>> For more options, visit https://groups.google.com/d/optout.
>>>>
>>>
>

> --
> You received this message because you are subscribed to a topic in the
> Google Groups "OpenPnP" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to

> openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/fef40ec6-5d52-48b7-9153-adaf08104e1f%40googlegroups.com.

[Anthony Webb]

I've come to the realization that my machine wont do everything perfectly, but I'm fine with the trade off for now, in exchange for the simplicity and cost savings.  Most of the huge components I am fine putting on by hand, its those tiny ones that are most tedious and require more of a steady hand that I have :)

--
You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/fef40ec6-5d52-48b7-9153-adaf08104e1f%40googlegroups.com.

[Anthony Webb]

I've been thinking a lot about lighting.  There are some spall LED's on the end of my USB microscope camera, which can get fairly bright, but some of the realy nice CNC machines seem to have lights that are beyond powerful on their up and down looking cameras.  Can some one chime in on how important lighting is, and if there are some ideas around lighting that I need to take a look at?

[Cri S]

Lightning is very important and there are tree independent things.
First forget buildin light. It works but its really worst.
Sure it depend what you want and need to do.
Light intensity. Without enough light fps get down and you need more then 15 or 33ms or 66.5ms
To get the image. If the CMOS sensor have high noise you need average several images. Time adds. Light uniformity and sunshine interference especially it is another issue. Depending what you want to do you need more direct or more indirect diffused light. You need to find out a compromise. As example having decent lightning its possible to know if inside pockets are components only making fixed threshold and blob detecting based on the shadows the component makes. This is less the 10ms operation. If you have time it don't really matters.
For up looking cam on modern PNP, there check coplanarity too and for this light is sequenced and in combination with plastic sheet that act as mirror there have side view of pins too.
Otherwise up looking is less critical except more diffuses light is needed. First thing is you need enough light for 30fps or higher. Second thing light uniformity. Third thing you need diffused light too. Mostly because if uniformity requirement you see a lot of LEDs

Il 14/gen/2016 04:26 "Anthony Webb" <anthon...@gmail.com> ha scritto:

I've been thinking a lot about lighting.  There are some spall LED's on the end of my USB microscope camera, which can get fairly bright, but some of the realy nice CNC machines seem to have lights that are beyond powerful on their up and down looking cameras.  Can some one chime in on how important lighting is, and if there are some ideas around lighting that I need to take a look at?

--

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/6b7f5364-af23-4347-a729-560c5232059c%40googlegroups.com.

[mojalovaa1]

Anthony , you wont share with as  yours files for pnp machine  make on FUSION360 ?
Where you found bearing and rails for fusion360 , you are make it or have some where for download it ?

[Graeme Bridge]

try here http://www.automotioncomponents.co.uk/en I've pulled most of my fusion360 parts from this website

[Anthony Webb]

Actually, fusion360 makes this SOOO easy!  Under the "insert" menu you can insert parts direct from McMaster-Carr, or (my favorite) parts4cad.  You may also import other types as well.  Modeling your stuff couldn't be any easier.  Give it a try!

[Anthony Webb]

Cleaning up some of the 3D designs and bringing them up to speed with the machine we actually built.  We opted to go with a far more heavy duty frame than we originally planned, cost increase was minimal.  As compared to our previous plan for the bed, this new bed can be adjusted up and down very easily, which came in handy as we fine tuned for the position and travel of our head unit.  Attached are shots of both with and without the bed. On the unit we built we bolted aluminum guides onto the bed in order to easily put parts and board into the same place again and again.  You'll see those details in the upcoming videos.

Another note (spoiler) we were able to actually do our first pick and place operations!  More to follow.

[Anthony Webb]

First pick and place in the books! Note that in our dry run, we enabled both nozzles and it was a lot faster to grab and place both parts in a single pass, but for this video Jason disabled the 2nd nozzle because we have a HIWIN rail/slide on the head that keeps sticking.  Not sure if we should get a more powerful spring or what.  Robotdigg was kind enough to send us a new rail/slide to replace it with, but it is just a rough as the one it replaces.

Here is the successful pick/place

https://www.youtube.com/watch?v=SCdOUyYe_1Q

Here is a dry run with 2 nozzles enabled

https://www.youtube.com/watch?v=DnyG4xOWCoo

[Jason von Nieda]

Looking great Anthony! Isn't that first placement just the best thrill? I was so psyched for the first one!

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/b51831bd-24db-47b4-9544-73aae7ae4536%40googlegroups.com.

[Anthony Webb]

It is very magical to see it come from a set of drawings all the way into the physical world and actually perform the task you set out to accomplish.  It's absolutely amazing to watch it do its thing, especially on its maiden voyage!  Thanks for your brilliant software that's powering it!

[mojalovaa1]

Anthony ,can you put video when place more parts ?

[Peter Betz]

Wow I am very impressed with this!!!

As you might know, I am over on the FirePick side, and was throwing the idea of switching to cartesian around, like you did. I' don't have tons of money to throw around, like abandoning the Firepick completely. I really just need the PnP functionality, I already have a 3D printer (and CNC G0704).

May I ask what you are into this machine for? Also, do you have an assembly you are willing to share? I am also using 360 (Ultimate).

Thanks for posting all of this for the community, and congratulations on the first pick and place!

Peter.

[Anthony Webb]

Hi Peter, yes I do recognize you from the FirePick side. I ended up taking a bath on my firepick and went back to square one to do this right.  We've made great progress so far and are nearing a point where I think we can start to confidently say that this works, and works well.  We're in the process of compiling a full BOM and models which we will make freely available to the community, might be another week or so before I know the exact answer to how much this machine costs to build.  I can say the big expenses on the machine are the head unit (we bought that for $300 from robotdigg, but are exploring engineering our own), and the rails. Unfortunately the cheap chinese rails did not get the job done for our design (we have a passive rail, so real smooth and accurate motion is a must),  so you are stuck sourcing a set of 3 rail/slide from THK (SRS series) / HIWIN (MGN Series) or similar, our brackets will work with most 12mm or 15mm rails we have tested.  Ebay can be a good place to find a pretty reasonable deal on these.

If you have your own way to print/mill material then this should be a slam dunk build for you.  For those who dont have the right tools to make the brackets needed we will likely mill some extras and make them available for people who want to build one.  We may even go so far as to have pre-cut extrusions and hardware kits to minimize the hunt for all the parts needed to get to your first PNP operation.  I think the more people we can get to the point of basic PNP on a similar hardware platform, the more great and open collaboration that can happen which ultimately benefits us all.  We hope to help the maker community arrive there.

We're working on getting the homing switches in place and finalizing our cable management strategy, little things that we want to make sure are in place before we finalize things as a V1.

@mojalovva1 more videos are on the way next week.

[Joao Matos]

Hi Anthony, I've been following this thread for the past few weeks and its been great to see all the progress.

I'm also interested in building a PnP machine and your design seems like a pretty good start. I'd be potentially interested in purchasing a kit from you in the future.

What are the main advantages of your design vs. say Jason's reference machine based on OpenBuilds parts?

Looking forward to more details of your builds (like the BOM). Keep up the good work!

[Anthony Webb]

Thanks for the kind words Joao.  The designs Jason and I have posted are very similar.  Let me explain the main difference and why I went the way I did.  The main difference is  his design uses wheels that ride in a slot on the extrusion, and mine use steel rail and slides.  The reasons I chose rail/slide as opposed to openbuilds:

1. I had experience with other x/y platforms like the shapeoko and quickly grew tired of tinkering with the eccentric nuts dialing in the tension.  Too tight and it wouldnt move (or skip steps), too loose and there was slop.  And even worse (perhaps because it was CNC) it changed every time you used it so you had to go over the machine regularly tinkering  to keep things operational.  I have heard from several people that this wasnt their experience, but as I evaluated my options, and give my experience, I knew that if I designed around wheels on extrusion, and it didnt give me the accuracy or precision I needed for small pitch PNP parts, then I would have effectively painted myself into a corner.

2. The rail/slide option gave me a variety of options in quality/cost which would be easily swapped if an option was not able to deliver the speed/accuracy/precision I was after.  Had the cheap chinese rails works (they didn't) they would have been a more cost effective solution than the openbuilds option.

3. Slide/Rail were already proven to deliver great speed/accuracy/precision over a long lifespan with minimal maintenance, I wasnt sure the openbuilds solution could deliver that.

4. I wanted to have a passive rail and avoid the complication/expense of having a second stepper to drive the Y axis.  Only rail/slide could deliver the kind of butter smooth motion that I would need to do that.

5. I love clearpath closed loop servos, and wanted to leave the door open to jump from cheap steppers over to the more capable servos which come in a nema23, the openbuilds brackets dont come in a nema23, so building my own gave me that flexibility.

It obvious that either route will give you the ability to pick and place parts. I would bet that the openbuilds route may come in a little cheaper, but I think long term we will be able to eek out more performance, with lower maintenance, from a rail/slide based solution.

-Anthony

[Juha M]

Hi Anthony, 

Did you have any issues with alignment of Y rails?

I have also linear rail based machine in the works.

- 20x40mm aluminum profile

- standard 15mm linear rails

- using nema17 steppers also as a belt idler to get rigid belt drive (just have to remove rotors from shaft)

- GT2 9mm belt

- dual y drive

- for x-axis thinking about some weight reduction if possible: cable carrier, aluminum profile to 20x20mm

Regards

Juha

[Anthony Webb]

Looks real great Juha!  I did not have any alignment issues with the Y, left stuff loose and let the X gantry tell me where the Y rails wanted to be.  The 15mm rails will be great but if you are buying rails or looking to save weight you may consider 12mm which appear to be plenty in my estimation.  Keep us posted on how your progress goes.

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a880b432-465c-4080-8ef8-82027903e798%40googlegroups.com.

[Anthony Webb]

We ran into an issue where our NEMA8 steppers for the nozzle tips were not powerful enough to twist the rubber tube attached to it, even with the 180degree limit enabled.  Looks like a couple people have gone with NEMA11 and special (expensive) adapters.  We opted to print a little boot for the top of the motor that leaves the shaft to freely rotate inside.  Problem solved!  The part is up on thingiverse here http://www.thingiverse.com/thing:1280038

[Jason von Nieda]

Nicely done! I use a similar setup on the OpenBuilds head but I really like how you integrated the tube nipple on there. Looks great!

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/196eea6d-a00a-4196-8983-c8118ef9976b%40googlegroups.com.

[Anthony Webb]

Its starting to feel more and more like we will be able to pick and place board flawlessly.  There's very little that is lacking.  We placed and entire board flawlessly a couple times now.  Just a few little things to iron out, mostly just learning how things work.  Here is a shot of us placing parts in a pretty deliberate and steady manner: https://www.youtube.com/watch?v=pLlzr1Asz9E

[Michael Anton]

I would be worried that this approach will might pull dust into the bearings of the motor over time, but perhaps I am just being pedantic.  Other than that it looks to be a good solution.

You might be able to come up with a similar design that uses a greased o-ring to seal the shaft instead, and would not present too much load to the motor.

Mike

[Anthony Webb]

I'm not following how dust would enter the motor, the motor is still sealed, this isnt a replacement cap, it merely sits on top of the cap.  It is also sealed off by the rubber hose attached to the top so I am not sure how dust comes into play any more than attaching the hose direct to the shaft would?

[Rich Obermeyer]

@anthony, curious why you and a lot of others only put .stl files on thingiverse?

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/99fbe591-9c23-4fea-a566-82c035c31615%40googlegroups.com.

[Anthony Webb]

I think it is the most common format you can drop into a 3D printer and hit print with.  If you want to download the original or play with it in your browser you can hit this link http://a360.co/23dJGMF (I will post that link to thingiverse for good measure as well)

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/33D168FC-C113-4909-AF33-87CABBF0CCF6%40gmail.com.

[Michael Anton]

The bearings are not likely sealed, or they would present too much drag, so dust will be pulled in from the front of the motor (due to the applied vacuum, when a part is on the nozzle), into the front bearing, and eventually into the rear bearing, through the motor.  Now, the dust may not make it into the inside of the motor, as it will get caught by the grease in the front bearing, but this will likely cause problems over time regardless.

Mike

[Michael Anton]

This is the common format that everyone can usually use, and is what is used by most 3D printers.  CAD files are a pain, as you really need the appropriate software to make good use of them.  Importing a non-native file format into a CAD package, doesn't work very well most of the time.

Mike

[Graeme Bridge]

As long as the files are Step or Iges most cad systems will operate just fine

[Rich Obermeyer]

That's what I was thinking.  .STL while 3D printable can not be modified, measured or much of anything universally.

A step file and the native format it was created with would give more value if the user wants to tweak the design for there machine use.

Most files show up on Thingiverse as .STL and while they may be great designs as is, they are limited because they can not be changed.

Everybody wants to tweak the design to make it their own.  That's just the way us engineers are :-)

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f5af66ab-ddd4-4c03-977a-d04453f3dee0%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Rich

[Peter Betz]

Sites like grabcad are more for .step sharing. Thingiverse is more for people sharing 3D printable designs. 

Peter Betz

BETZtechnik.ca

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAOvy4mZ1-vJq0B4BCEg-4RubHps5CfbpXVBQZvwJVN_%2B-Yr2dw%40mail.gmail.com.

[ja...@tetontechnology.com]

I checked the download formats in the file I shared here http://a360.co/23dJGMF

Looks like the download formats are plentiful

[Michael Anton]

Cubify usually pulls these in as a solid blob, without the creation tree.  That makes them of limited use, though step tends to be the most useful.  Sometimes they can be modified ok, sometimes not.  Other packages perhaps handle this better, but I am not familiar with anything else.

Mike

On Wednesday, January 20, 2016 at 11:09:24 AM UTC-7, Graeme Bridge wrote:

[Anthony Webb]

People have asked which model of HIWIN rails I am evaluating.  I found mine on Ebay here http://www.ebay.com/itm/HIWIN-MGNR12-linear-guide-rail-600mm-23-6-with-MGN12H-blocks-3d-printer-CNC-kit-/262229961840?

Overall they have been very impressive.  A huge jump up from the cheap chinese rails.  However, one of them was SLIGHTLY sticky (orders of magnitude less sticky that the el cheapo rails) every once and a while.  The other 2 were perfect.  I picked the best rail and put that one on the passive side, the sticky one is on the driven side.  So far, flawless performance.

I do get nervous about buying these so called "authentic" products out of China.  Of the 9mm "authentic" rails I have gotten (3 of them) 2 of them are pretty bad, I mean, they work, but they are not perfect.  I would expect authentic rails to be pretty perfect out of the box.  The THK's which were still in the factory box are absolutely perfect, all of them.  I imagine they do build some fairly convincing knock offs, if these HIWIN's are not authentic, then they did a mighty fine job of a knockoff.

So, YMMV, not sure if I should be nervous about these HIWIN's from China or not, they are getting the job done today, and the price ($275 for 3 rails) is fantastic.  Quality is pretty spotless with a couple exceptions.  Retail from HIWIN directly I am being quoted $200 per rail, $600 total.  Retail on the THK's is in the same ballpark.

[Jason von Nieda]

Honestly, I would be pretty surprised if the Hiwins that aren't directly from Hiwin are real. During my time in China I saw rails that looked just like the Hiwins (green and red color scheme) almost everywhere. With as much as everything is knocked off, I can't imagine they would be immune.

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/0994d6a6-c680-427b-b778-e8968830baad%40googlegroups.com.

[Anthony Webb]

Spent the day rewiring all the electric in cable chain and generally tying up loose ends.  Of course there were some hiccups.  First, the USB microscope we were using would not work once we extended the USB cable.  This was odd because I think we were overall only like 8-10 feet which should be within spec.  I had one of the ELP cameras laying around, and was able to extend the cable on that just fine, so I am thinking we're going to make a new bracket to accommodate that camera instead, the quality of the ELP was strikingly better in every way, it's probably just a superior solution.  Designing the mount and will install tomorrow, I'll add a pic once it is done.  Milled part holders for both 12mm and 16mm strips.  Other than that things will look really nice with all the wires tucked away in the chain.

[Anthony Webb]

As I mentioned yesterday, we were enable to extend the USB on our microscope camera we had been using for downward vision.  This ended up being a huge benefit for us because after looking at the images from the ELP camera (same used in Jasons reference machine) we were blown away at how bad the USB microscope was in comparison to the ELP.  So, we went to work designing a bracket to mount the ELP, We recessed in the PCB so it would nice and flat, turned out great.

With the new camera in place we were able to return to getting the cable chain wrapped up.

We're now getting a mount together for all the plumbing and we can return to more testing in a more final type configuration.  Wishing there were more hours in the day.

[mojalovaa1]

 

Anthony , whom camera you are buy ?

look nice , you are use that machine for some job ? 

[Anthony Webb]

The camera I bought is this one:

http://www.ebay.com/itm/221448995232?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

They have them all over ebay and amazon.  I've seen the 5MP versions as well, mine is only 1MP but I almost think the lower resolution will mean easier/quicker image processing?  The picture is beautiful, even with only average lighting conditions.

We'll be using the machine to assemble small runs of test boards in house once it is complete.

[mojalovaa1]

Why you not use more endoscope or microscope camera ?

[Anthony Webb]

The image quality and focus of the endoscope/microscope is not even comparable to the ELP.  In addition, the endoscope/microscope did not have the same low light characteristics.  Finally we were not able to extend the microscope USB cable enough to route through our cable chain.  I had originally planned on only using the ELP on the upward looking camera, but now I may use it on both.  I still need to run some PNP jobs with the new camera, but it looks pretty good so far.

[FredG]

And you use the original 6mm lens or the 8mm lens from Jason's BOM? What about distortion and field of view with the 6mm lens? Which lens would you recommend?

[DAniel Dumitru]

Good question about lens.
I am interested about resolution :
http://www.elpcctv.com/industrial-usb-720p-camera-module-for-monochrome-stereo-p-79.html
says : Max. Resolution 640X480 VGA
is is right ?

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/b08baf4f-a3f9-4d29-b7da-05c6dc672ef5%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

This email has been sent from a virus-free computer protected by Avast. www.avast.com

[Anthony Webb]

I am using the 6mm lens that came with my device.  I dont have an 8mm lens to compare to.  It does appear in the camera info that the resolution is 640x480.  The lens of my camera is exactly 28mm from the surface of my PCB.  The calibration wizard in openpnp calculated values of .23 for both x and y.  Distortion does not appear to be a problem. 

I'm unclear on the amount of settle time that is ideal for these ELP cameras as compared to other models, I think we are using 250ms now.  What I am shocked at is how well the image looks in simple ambient lighting, the ELP is more of a "microscope" than my USB microscope ever was.

[cf]

ELP has so many different models to choose from, what is the ideal properties ?

http://www.elpcctv.com/

Resolution ? Frame rate ? Light sensitivity ? etc ...

I see they have some low light cameras, black and white only cameras, some high framerate cameras 120fps etc ... andy advice ?

[Anthony Webb]

I would like to know this as well. Jason would probably know best what the real limitations are of the current vision system. I'd also like to know how quickly I should get another one coming as an upward looking camera :)

[Jason von Nieda]

Unfortunately this is not something I have really quantified. It's something I'd like to be able to just say "You need this resolution, this view size, this mm per pixel" but I haven't taken the time to figure out those numbers.

Jason

On Sun, Jan 24, 2016 at 4:41 PM Anthony Webb <anthon...@gmail.com> wrote:

I would like to know this as well. Jason would probably know best what the real limitations are of the current vision system. I'd also like to know how quickly I should get another one coming as an upward looking camera :)

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ee8a79e1-2795-462b-bf68-9e0b119f1c4f%40googlegroups.com.

[Anthony Webb]

Finally got around to putting together a hackaday.io project page.  As we post BOM and design files it will be much easier to share and organize them there instead of buried in an email on this list.  Note this page isnt meant to replace this thread, we'll be posting here in tandem, join us in either place.  https://hackaday.io/project/9319-diy-pick-and-place

[FredG]

As you mentioned earlier you were using a Tinyg board for testing. Which board are you using finally?

Am Samstag, 19. Dezember 2015 06:09:18 UTC+1 schrieb Anthony Webb:

Wanted to start a new thread to report on my findings.  Up until now I have been posting stuff into the thread discussing inexpensive PNP options.  For those who havent followed there I will recap here.  I'm on the hunt to build a cartesian pnp.  I'm not looking to try and build the cheapest machine out there, but I am out to build a highly reliable and accurate machine for the least cost possible.

I've been through countless 3D printers, CNC machines, and other various motion projects and have a pretty good feel for what I can live with.  For that reason I have chosen to go the rail/slide and belt route as opposed the other various options out there.  I think I will be able to find a motion platform for this project that checks all the boxes (cost/quality/simplicity).

To this end I feel the need to know how the various price points of rail/slide options compare to each other.  You can pay $20 for the inexpensive stuff, and $1000 for the real pricey stuff.  Question is, when you pay more, do you get more?  I'm about to find out.  But I am looking to find out EXACTLY how much more you get.  I'm going to measure each set of rails against each other in a torture test of sorts using a tools called firenodejs that Karl Lew has put together (https://github.com/firepick1/firenodejs) It will be a ruler capable of drill into just how well each of the rail/slide combos perform compared to each other.

In order to do this analysis I needed a common platform to test them with.  Task 1 was designing a test rig.  I started with an actuator.  Small problem is that I didnt actually know any CAD tools, but I knew what I wanted.  I wanted to design something that was easy to build, modular enough to be expanded/shrunk to meet varying budgets, and I wanted the actuator to be able to mount onto other actuators to give a nice x/y platform that would ultimately be the bones for my PNP.  After days in front of Fusion360, Here is what I came up with

The nice part about extrusions is that they are relatively easy to work with, you can cut them easily, bolt bits and pieces on without much trouble, etc.  The beauty of what I ended up with is that I could very easily bolt on various rail/slide combinations for testing (WIN).  I have a CNC and 3D printer, but I wanted the brackets to be something that a guy could make with a drill press and some delrin, so I kept them very simple. With the actuator design in place I set out to join them together in an x/y test rig.  This would require some custom brackets to join together the actuators.  Here is the next rev

As you can see above, I really wanted to see if a passive rail would fly. For simplicity and cost I am going to first see if this works, if not I have lots of workarounds I'll try too.  I initially targeted 300x300, but after looking around I found that 500mm rails would be a better size to target.  I felt ready to go ahead and start buying parts.  The last 3 weeks have been spent bringing in the hardware to build what you see above.  Of most importance was the rail/slide combos. I bought 3 of everything, the cheap chinese hiwin knockoffs, the run of the mill authentic hiwins, the uber expensive THK option, 12mm, 15mm, as well as some of the igus drylin stuff.  Companies found out what I was doing and sent me samples of their product to test too.  Basically, I have enough rail/slide combo hardware to provide a really good picture of what works and what doesnt, but more importantly to know where is the point of diminishing returns, ie at what point does spending more money not necessarily buy you more quality.  From there I figure I can pretty accurately decide how much quality I want to afford :)

I'm not going to make any predictions at this point, but I can tell you that I was not expecting the good stuff to be that good, nor was I prepared for the bad stuff to be that bad. I nearly didnt even bother trying to test the chinese set, sure it was 1/15th the price of some of the other stuff, but man, I couldnt even move the slide on the rail at all, I'm not joking.  But after a dose of WD40 and white lithium grease it at least slides along the rail now. Definitely not the butter smooth action of the THK's out of the box.

Off we go, I cut brackets and cleaned off a spot on the bench and began assembly today.

Chinese rails are up first, and as for as much as I had written them off initially after getting them all locked in I must say that they seems to move great.  No slop at all.  The passive rail I was worried about?  Well for now I am not seeing a problem, but the proof will certainly be in the result.

Thats where I left off today.  Mechanical is complete, electrical is next.  I have a tinyg laying around that I have tasked with driving the motors I installed.  These are off an old 3D printer, nema17 400 step (.9) per rev.  I am using 32 tooth 2mm GT2 belt, not sure what kind of speed I will drive with that, but excited to give it a whirl.  I am a little nervous about the steppers, I have had such great luck with my clearpath closed loop servos I almost hate to even try anything else, but if I can make steppers work I could save a bundle of money along the way so I am starting with them.

I'l be sure to post my findings as I go, the good/bad/ugly.  As always if there are any questions or feedback please feel free to call me crazy, offer insight, etc.  I'm very new to all of this and know many of you have been through, at the end of the day I hope to be picking and placing with the rest of you!

A short video of the rolling chassis.

https://youtu.be/vqTw7lJbhqU

[Anthony Webb]

The TinyG doesnt have enough motor drivers for my PNP (I need 5) so while it was great for testing, I needed to use (or build) something else for PNP.  I began design on a TinyG2 board that would allow for 6 stepper drivers, but an held up waiting on some IO tweaks that will enable the IO I need to control the pump solenoids.

In the meantime (and perhaps longer?) the SmoothieBoardX5 is working great.  It required minimal configuration too since my design is pretty similar to Jasons reference design.

--
You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2e659587-cc47-4853-9e18-3df86c695647%40googlegroups.com.

[Anthony Webb]

Today the endstops got installed which means I can finally home the machine.  In addition the plumbing and electrical got cleaned up.  Photos are posted here: https://hackaday.io/project/9319-diy-pick-and-place/log/30826-endstops-plumbing-and-electrical

[Anthony Webb]

Very reliably picking and placing parts. Not having any failures at all. Jobs that took us around 10 minutes by hand are now under a minute in the machine.

I still think we have some optimizations left. I need to quiz Jason on why the dwell time is so long when placing the part on the board? In the machine xml for the nozzle there is a pick and place dwell that is in the xml but not in the UI (oddly the only params not in the UI) so I am not sure if mess with them? Seems to me that even though in the xml the are both set to 250, the placing appears to take longer than the pick?

I think we will be even more productive once we can use bottom layer fiducials, and of course the upward looking camera will make it that much more bombproof, but there isn't anything stopping us at all now from creating production boards.

This board is all 0603 and 1206. Waiting in some 0402 parts and test board, but I can't foresee a problem with those. Everything is spot on right now for us using the tray feeder.

https://youtu.be/Ril6AWOdqfg

[Jason von Nieda]

Hey Anthony,

There is a hardcoded 500ms delay in the place routine in OpenBuildsDriver that is there to let the vacuum leak off after opening the exhaust valve. It can probably be MUCH shorter. I set it to that and kinda forgot about it. It should, at the very least, be made configurable.

As for the pick and place delays, feel free to edit those in machine.xml to suit your setup. I just haven't gotten around to adding them to the UI.

And aside from that, awesome video! Makes me really happy and proud to see OpenPnP being used in the wild :)

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/6eb0f2a7-4bc4-4fa2-9f20-8bb836e49622%40googlegroups.com.

[Peter Betz]

Amazing work Anthony and Jason. You guys should be proud.

I am 100% building this machine, it suits my needs better than the firePICK, I need lots of parts mounted on the deck.

Anyone want a good deal on an assembled and tested firePICK? ;) 

Peter.

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/w46THEk0Mjc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jz4S8unFuqv5wU1yydp4pnBhNpGHCydN84Rs22ZCd2ycA%40mail.gmail.com.

[FredG]

It's great to watch your machine growing step by step, good work! I'm a member of the apertus open source cinema camera team and at the moment we are assembling our first boards by hand. There are a lot of 0402 components to place and I'm also thinking of building our own PNP! 

Jason and Anthony thumbs up! You do a great job! 

Manfred

[mojalovaa1]

Anthony , you are use vacuum generator or vacuum pump ?

[Michael Anton]

There is a fellow on the FPD list looking for a machine.

Mike

[Anthony Webb]

Finally was able to pull together the full BOM and 3D rendering to the point it is ready for you to take and build one yourself.  We've tried to consolidate the pieces in order to minimize the number of unique parts on the BOM.  Go to the project page on hackaday and look for the links on the left.  https://hackaday.io/project/9319-diy-pick-and-place

We have heard from several of you who either dont have a mill, dont have a 3D printer, dont have the time, or just would prefer a kit instead we are considering putting together some quick build kits.  Still not sure what the kits would include, but likely would be a kit that includes all the custom brackets needed for the machine, and perhaps even a bundle of pre-cut extrusions as well.  As for a FULL kit, with all the hardware and electronics, we'd be open to that if there was enough demand.  

For now, we'd just like to get as many people enjoying rock solid PNP using OpenPNP, and we feel that this machine accomplishes that and more.  To that end, we're happy to contribute the design back to the community.  We'd love your feedback, collaboration, and know that collectively we can make this thing even better.

[Jason von Nieda]

This is really amazing work Anthony. Congratulations! I'm pretty seriously considering building one myself as it solves many of the problems I've been wanting to solve in my own design.

Are all of the parts that require manufacturing designed to be 3D printable, or do they have features that must be milled?

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e6568f93-c9ed-44f1-afb8-24b02d55c94e%40googlegroups.com.

[Anthony Webb]

Thanks for the kind words Jason.  Everything was designed such that it can be printed well.  In fact, most of the high stress brackets (the motor mounts) can be built with a jigsaw and a drill press.  All of our stuff was milled from HDPE for test fitting, and the final pieces in delrin for max accuracy/strength.  Most of them could be cut on a waterjet out of aluminum too.

Regarding printed parts, I dont have much feel for how accurate and strong those can be, the motor mount and the idler pulley bracket might have a bit of stress once you really tighten that belt.  I'd love some feedback on how well that may or may not work out.

What I do know is that with the delrin we are using today there is no flex or play at all, and with our passive rail, I would expect that the rigidity it pretty key.  As you look over the parts, if you see anything that looks like it could be improved let me know.  Thanks!

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzmQ0wkUDqaQmNozcRwDzvwwmjKALcojQfV1vZejftyMg%40mail.gmail.com.

[Peter Betz]

Great work!!

 I am going to slowly start acquiring the materials. Going to price out some derlin stock tomorrow and start machining ASAP.

One question, I imagine a Arduino Mega and Ramps 1.4 will work in leu of the smoothie? Seems to have 5 axis and the Mega combined with FIREstep on the FPD did not seem to be an issue for smooth motion.

Peter.

[Anthony Webb]

On Friday, January 29, 2016 at 8:41:44 PM UTC-7, Peter Betz wrote:

Great work!!

 I am going to slowly start acquiring the materials. Going to price out some derlin stock tomorrow and start machining ASAP.

The delrin stock you will want to source is the 1/4" stock for all the motor/idler brackets, and 1/2" for all the carriage brackets that join the actuators.

 

One question, I imagine a Arduino Mega and Ramps 1.4 will work in leu of the smoothie? Seems to have 5 axis and the Mega combined with FIREstep on the FPD did not seem to be an issue for smooth motion.

Jason has several sample drivers for open PNP, so long as you could connect over serial and your controller could consume GCode I imagine it would work.

[Malte R.]

Hey Anthony,

congrats - your machine is looking great!

When it comes to manufacturing kits, I think laser (or waterjet) cut steel / aluminium plates would make it much easier and faster to make larger quantitities.

That is very easy to outsource, many shops will cut plates based on your DXF or SVG files - a fact that helped Juha a lot with the LitePlacer kits I suppose.

Many of the custom parts could probably be redesigned for that; you may want to consider that if you are really going to sell kits.

For all those people wanting to build a one-off instance of your machine, CNC milling + 3D printing is probably fair enough - but if you make more than a few that extra effort might pay off.

Just my $0.02 ;-)

Thanks for all your hard work and sharing the results by the way!

[Peter Betz]

Hi Anthony,

 For me, I wanted to only have to buy one size of Acetal stock. So I modified your carriage blocks to split the bodies into 2, 0.25" pieces. I was tempted to raise the recessed slots to the top of the new stock line but I wasn't sure if there was enough clearance for the head of the bolt.

The bolts for the T nuts should clamp everything together just fine.

A360 link: http://a360.co/1m5L8yQ

Peter.

[Anthony Webb]

You and I were thinking about the same thing today :)  I was trying to make that bracket fit in the 1/4" stock, didnt consider stacking it, interesting.  If we changed the tensioner bolts over to M3, it may fit, but it would be very close.  The M3 attaching the bracket to the carriage would have to be a low profile head like https://fci.thomasnet-navigator.com/cache/Geomstore/20160130-193030-97884962.pdf

After looking at your changes I realized that I left something pretty critical off the original file I shared, that was the screw holes for the tensioning belts.  Since the holes were not part of the milling process (drilled them on the press after milling) they werent in the model.  At any rate, I added them in, here is an image.  You could probably glue together your halves then punch the holes for the tensioning belt.  I'd have to double check but I dont think raising the recessed slot would be a big deal.

[Anthony Webb]

Peter, here is a version of the brackets that use M3 for tensioning, this way you can get away with 1/4" for the bracket without having to use multiple layers.  I have not had a chance to build one of these yet to see how it compares, but theoretically it should work:

http://a360.co/1RSZvEE

http://a360.co/1SQV4dI

[Joao Matos]

I'm also starting to source the parts needed to build this machine, and I have a few questions.

About the vacuum source, you're using a seemingly much more powerful (and expensive) pump compared to the cheap 12V ones. I'd like to understand the tradeoffs here. Is this for performance reasons, like being able to pick parts better?

I haven't been able to find your particular pump model in Europe, would something like this work? I have a feeling it might be overkill but I really have no clear idea.

http://www.ebay.co.uk/itm/3CFM-Single-Stage-Refrigerant-Air-Conditioning-Vane-Vacuum-Pump-HVAC-AC-Parts-/231826381886?hash=item35f9eddc3e

About the head design, you're using the pre-built one from Robotdigg, this is not ideal for me since I'd have to pay extra shipping and import charges to get it in Europe. How hard would it be to adept the design from Jason's reference machine? If you think it's doable without a lot of extra work, I could try to figure out a way to adapt it once I build the machine core.

Also, and this is directed at you guys from Europe, any idea of some good websites for sourcing most of the hardware parts?

Cheers,

Joao