Christian's PNP

[Christian...@gmx.at]

Hello all you dear PNP enthusiasts,

A few months ago I came up with the idea of ​​building a PNP machine myself so that I can be quick and flexible in the production of small series. In addition, my customers always have spontaneous changes, which is why I cannot keep huge amounts of finished circuit boards in stock in order to keep the assembly costs low with third-party providers.

Of course I could have bought a PNP machine, but where's the fun?

Since I didn't have the time to pursue the project in the meantime, it just had to wait that long ... until now !!

Because I am entering uncharted territory by building such a machine, I occasionally have to rely on third-party and qualified help.

I hope that you will help me with my project, because without you I probably wouldn't make it.

Now to the machine:

I had planned to build a desktop machine.
Not even close!! I guess I overdid it. In the current expansion stage, the machine already weighs 80 kg ... without feeder, without control, without placement head, ... without almost anything.
The machine is designed for the Siplace / Schultz feeder and offers space for a total of 34 pieces, e.g. 2x8mm feeder. The active assembly area is 300x500 millimeters. Of course, PickUp positions can also be set up in this area. I designed a 4-way head as the placement head.

Here is the construction of the machine:

And here is the construction of the placement head:

And here is a photo of the machine showing the current production progress:

I still have to manufacture all machine parts that are still missing, which unfortunately will still take some time. Mainly because my CNC milling machine was unfortunately struck by lightning and I need to repair it first.

I am always open to discuss suggestions for improvement.

AND YES, if anyone wants them, I can provide the construction files ;)

By the way, if someone can produce 8mm aluminum milling parts for little money and get it shipped cheaply to Austria, please contact me :-) (I still need the three aluminum plates from the assembly head ... but - as you already know - my milling machine is dead at the moment)

So, hope you like it.

Best regards,

Christian

[Nick]

Head looks interesting. Can you share some details about the way Head is designed.

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ce82f761-f292-4c59-a8dd-ba3e7f211692n%40googlegroups.com.

[chende...@gmail.com]

Nice build, good luck with it. Your customers looks awfully similar like mine :)

I would recommend you to install a laser displacement sensor on the head, it will drastically reduce setup times by automatic measuring Z pick locations. My machine has one. And it's a bit heavier than yours, around 120kg so far, no feeders, no heads.

Regarding your request for milling parts: in my build, for testing purposes, I am using a 22 places feeder plate support (yamaha CL feeders) from robotdigg. But those guys does not have anymore those plates, nor have them in the future. No one else has them as well, as far as I was able to look. So I will have to build them, in the end I will need support for 80 feeders. I obtained some local contacts in my country with businesses which are capable of manufacture them . If you want, I can share them with you, just leave me an email address. But be advised, I did not worked with them before, I do not know them.

[eklun...@gmail.com]

Hi Cristian,

I might be able to help you.

I don't promise any thing, but I can at least take a look at it.

--Per

[Christian...@gmx.at]

A laser displacement sensor, hmm, not a bad idea. What kind do you use I will do some research on the sensors and see how it can be implemented in the controller.

Hey Per,

Do you mean that you can help me with the milled parts

If so, which file format do you need? I design with solidwork but i can save in different formats

[chende...@gmail.com]

I am using an HG-C series from panasonic.

There are several models in this series, 5 to be precise, each of them with their specific precision and measurement range. Better precision comes with smaller measurement range and vice-versa. You have to pick one suited to your case. Obviously better precision is preferred but you have to check that all things you want to measure will be in the sensor range. Sensors require anything between 12-24 volts to run and they have an analog output, between 0-5volts. You have to ADC this and feed it into openpnp. You can find them at around $200 on aliexpress.

[eklun...@gmail.com]

Sorry, I missed your answer yesterday.

Yes, I can probably help you to mill some parts.

I use Fusion 360 and I think step-file works fine, but I also think it's possible to import solidworks-files.

--Per

[Christian...@gmx.at]

Hey Per

No Problem ;)

That pleases me!!
Am I correct if I assume that you are from France?
Shall I send you the raw material?

What would your help cost me?

Christian

[Christian...@gmx.at]

I also bought fusion 360 a few months ago because Solid works is way too expensive for me in the long run.

And I also work with eagle, and the premium version of eagle is only available together with Fusion.

Unfortunately, I am not very familiar with Fusion 360 and I have already drawn the PNP machine completely in Solid Works.

I will send you a selection of different files. I'm sure a few of them will work :D

[eklun...@gmail.com]

Hi,

I have imported from Solid Works to Fusion before, so it's no problem.

I actually live in Sweden.

Let me have a look at your parts first. If they are not too big I probably have aluminum.

Regarding cost.After looking at the parts I hope I can give you a good price. Shipping could be an issue thou. 

--Per

[Christian...@gmx.at]

Hi Per,

Ohhhh Sweden ... thats a little bit different :D

Here are the three parts i need.

It is not essential that the round holes are drilled to the correct diameter.

It would be enough for me if only centering holes were made instead of the normal holes.

I can then drill them myself on the pillar drill without any problems.

The threaded holes on the lower face of the side plates are also not absolutely necessary.
The parts are 8mm thick.

[eklun...@gmail.com]

Hi Christian,

I had a look at the files and I'm afraid it will take me too long time to mill those parts.

Two problems, Work holding and narrow slotts.

So unfortunately I will not be able to give you a fixed price that is ok.

--Per

[Christian...@gmx.at]

Hello Per,

Thank you for looking at the files.

The bevel and countersink on the underside of the "Quadro-Head_Rahmen" part is not absolutely necessary.

The 1mm groove left and right on the top is more important.

Both parts could therefore be milled with just one clamping.

The two narrow elongated holes can also be 1-2 mm wider.

With these changes, it may be possible to manufacture the parts more cheaply.

If it doesn't work, it's of course not a problem;)

Greetings from Austria,

Christian

[Christian...@gmx.at]

Hey Nick,

I am sorry for not answering until now.

Look here for more information about the head:

https://groups.google.com/g/openpnp/c/BOmK72AsypY/m/OzyIPzKGAwAJ

[Christian...@gmx.at]

Hi Guys,

I wrote to several companies and asked for an offer for the aluminum milled parts. The cheapest offer I received was this:

Quadro Head Ground plate: € 156.00 (one needed)
Quadro Head Side Plate: € 192.00 (two needed)

Shipping Europe-wide € 12.90

Anyone interested in the parts ?? I would then order several parts at once.

[Clemens Koller]

Hi!

You might want to check if you can get along with laser-cut plates. These could possibly lower your costs because of much lower tooling cost in comparison to milling.

Just my 5 Cents,

Clemens


--

> PNP construction.JPG
>
> And here is the construction of the placement head:
>

> Quadrohead.JPG
>
> And here is a photo of the machine showing the current production progress:
>

> IMG_2227.jpg
>
>
> I still have to manufacture all machine parts that are still missing, which unfortunately will still take some time. Mainly because my CNC milling machine was unfortunately struck by lightning and I need to repair it first.
>
> I am always open to discuss suggestions for improvement.
> AND YES, if anyone wants them, I can provide the construction files ;)
>
> By the way, if someone can produce 8mm aluminum milling parts for little money and get it shipped cheaply to Austria, please contact me :-) (I still need the three aluminum plates from the assembly head ... but - as you already know - my milling machine is dead at the moment)
>
> So, hope you like it.
>
> Best regards,
> Christian
>
> --
> 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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ce82f761-f292-4c59-a8dd-ba3e7f211692n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/ce82f761-f292-4c59-a8dd-ba3e7f211692n%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> 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 <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/4c92d2e9-9600-4eb3-8443-ad5e9f8a9891n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/4c92d2e9-9600-4eb3-8443-ad5e9f8a9891n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Christian...@gmx.at]

Laser-cut plates - a great idea!!

why haven't I come up with the idea myself?!?!
I try to get a quote for the laser cut parts
I will then tell you whether and how much it is cheaper.

[ja...@loveelectronics.co.uk]

FYI If your getting this laser cut, 8mm is on the very thick side for Aluminium, it doesn't laser very efficiently. Changing to steel would be cheaper, but increase your cost, if you reduce the thickness of the parts down to maybe 4mm this would be more effective, but then if you've reduced it to 4mm then the milling would be cheaper, as your 3mm slots are only 1.5D instead of 3xD.

Just some thoughts from a machininst.

[ja...@loveelectronics.co.uk]

I meant to say changing to steel would increase your weight.

[Yevhenii Shcherbakov]

Steel - bad idea. 

It will weigh 3 times more. This is extra weight that will affect top speed and acceleration. 

For this product, in terms of strength, 3 mm of aluminum is enough. 4KW laser cuts 6mm thick aluminum. This is maximum. One can find waterjet cutting and cut on it. There, the thickness is noticeably greater. But in all these methods there is a drawback: the poor quality of the cut end. If there are precise fit of the parts, you will have to additionally process the surfaces.

пятница, 30 октября 2020 г. в 23:44:13 UTC+2, ja...@loveelectronics.co.uk:

[Christian...@gmx.at]

I will not use Steel, it is way to heavy.

Just if it is still to expensive i will reduce the thickness of the aluminium plates to 5 mm.

I have a company nearby that can laser-cut aluminum plates up to a thickness of 10mm with an accuracy of +/- 0.1mm without problems.

Since I don't need an exact fit for the parts, that's perfectly adequate for me. I drew the threaded holes all with a 0.5 mm smaller diameter.

I then drill this on the column drill to the core diameter and cut the thread into it myself.

I can make the four fitting surfaces on the parts with the manual milling without any problems.

So now I hope that laser cutting is significantly cheaper than milling.

regards

[Dave Park]

Plastic is cheaper and lighter than both of those. A well thought out carbon/PLA part or similar would do really well.

On Sat, Oct 31, 2020 at 3:16 PM Christian...@gmx.at <Christian...@gmx.at> wrote:

I will not use Steel, it is way to heavy.

Just if it is still to expensive i will reduce the thickness of the aluminium plates to 5 mm.

I have a company nearby that can laser-cut aluminum plates up to a thickness of 10mm with an accuracy of +/- 0.1mm without problems.

Since I don't need an exact fit for the parts, that's perfectly adequate for me. I drew the threaded holes all with a 0.5 mm smaller diameter.

I then drill this on the column drill to the core diameter and cut the thread into it myself.

I can make the four fitting surfaces on the parts with the manual milling without any problems.

So now I hope that laser cutting is significantly cheaper than milling. 

Dave 

[Christian...@gmx.at]

Hey Dave,

thank you for your suggestion.

At the moment I am waiting for an offer for the laser cutted parts.

If those are still too expensive i will make it out of plastic.

Your are absolutely right, plastic is totally sufficiant.

[Clemens Koller]

Hi!

On 31/10/2020 07.46, Yevhenii Shcherbakov wrote:
> Steel - bad idea.
> It will weigh 3 times more. This is extra weight that will affect top speed and acceleration.

It depends... once it's on the laser cutter, you can take all material away, you don't need to make it very light again.

> For this product, in terms of strength, 3 mm of aluminum is enough.

Usually, it depends if you want to tap your screwholes or not. Otherwise, I would even consider going down to 2mm or 1.5mm aluminium sheets or
add some construction (bents, additional rails) to make your design more stiff.
IMO, a lot of (open source) constructions suffer in oversized everything. A single M4 screw 8.8 steel should be able to hold up to 7kN (~700kg).

> 4KW laser cuts 6mm thick aluminum. This is maximum. One can find waterjet cutting and cut on it. There, the thickness is noticeably greater. But in all these methods there is a drawback: the poor quality of the cut end. If there are precise fit of the parts, you will have to additionally process the surfaces.

A friend of mine (I believe he is a really good mechanical engineer with many years of experience) was stacking several similar thin metal sheets he got lasered together to achieve different functions. The reason was that in his case, the programming costs for the laser cutting were more pricy than the plates themself, once the machine was running. So, he reduced the number of different plates to an absolute minimum and just got many more of them puzzled together in a modular fashion.


Regards,

Clemens
--

On 31/10/2020 07.46, Yevhenii Shcherbakov wrote:
> Steel - bad idea. 
> It will weigh 3 times more. This is extra weight that will affect top speed and acceleration. 
>
> For this product, in terms of strength, 3 mm of aluminum is enough. 4KW laser cuts 6mm thick aluminum. This is maximum. One can find waterjet cutting and cut on it. There, the thickness is noticeably greater. But in all these methods there is a drawback: the poor quality of the cut end. If there are precise fit of the parts, you will have to additionally process the surfaces.
>
> пятница, 30 октября 2020 г. в 23:44:13 UTC+2, ja...@loveelectronics.co.uk:
>
> I meant to say changing to steel would increase your weight.
>
> On Friday, October 30, 2020 at 9:43:48 PM UTC ja...@loveelectronics.co.uk wrote:
>
> FYI If your getting this laser cut, 8mm is on the very thick side for Aluminium, it doesn't laser very efficiently. Changing to steel would be cheaper, but increase your cost, if you reduce the thickness of the parts down to maybe 4mm this would be more effective, but then if you've reduced it to 4mm then the milling would be cheaper, as your 3mm slots are only 1.5D instead of 3xD.
>
> Just some thoughts from a machininst.
>
> On Friday, October 30, 2020 at 8:41:54 PM UTC Christian...@gmx.at wrote:
>

> Laser-cut plates - agreat idea!!

> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/080b04d1-402b-4209-974f-16668407abb6n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/080b04d1-402b-4209-974f-16668407abb6n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Yevhenii Shcherbakov]

> It depends... once it's on the laser cutter, you can take all material away, you don't need to make it very light again.  

The more cuts in a small area, the more the part heats up, the higher the stresses (hardening) arise at the edges. This leads to deformation of the parts. Then it will need to be straightened out.  

>Usually, it depends if you want to tap your screwholes or not. Otherwise, I would even consider going down to 2mm or 1.5mm aluminium sheets or

In the thread, the first 4 turns take up to 80% of the entire load. Therefore, when using M2 screws, a thickness of 2 mm is sufficient, for M3 - 3 mm.  

>add some construction (bents, additional rails) to make your design more stiff.

The bent edge, in this case, can only serve as a stiffener. Bending accuracy may not be sufficient to position the elements.  

>IMO, a lot of (open source) constructions suffer in oversized everything. A single M4 screw 8.8 steel should be able to hold up to 7kN (~700kg). 

The screw itself can withstand such a shear load. At break, the thread in aluminum will break first. 

P/S/  In this discussion as a participant, not a topic starter. Just sharing my experience. 

 My head structure is already assembled and installed on the machine. I'm just waiting for NEMA 8 to run the machine.

понедельник, 2 ноября 2020 г. в 02:40:51 UTC+2, Clemens Koller:

[Christian...@gmx.at]

Next week I will get my laser cuted plates.

Meanwihle I am busy making the other parts:

Here are some pics of some parts I have finshed during the last days:

[ja...@loveelectronics.co.uk]

Looking good Christian - I bet those 90 degree air connectors were a challenge to machine!

[Christian...@gmx.at]

I made the 90 degree connectors on a 35-year-old Thoma-Tiger lathe and an equally old milling machine with a dividing head.

Manufacturing such extremely small parts on such large and worn machines is a bit tricky, but not a big deal if you are trained and can use the machines well. ;-)

[mike....@gmail.com]

Christian you will need a strnger Base Plate for Feeders - see here: https://groups.google.com/g/openpnp/c/hBWwKGdBlgE/m/zuAC2HfTBAAJ

Mike

[betzt...@gmail.com]

Wow, I am impressed with those air connectors as well!! 

Regards,

Peter.

[Christian...@gmx.at]

Hi Mike,

The base plate for my feeders is made of 20mm birch plywood covered with a 1.5mm stainless steel plate.

17 pieces of 20x5mm neodymium magnets were glued into the plywood panel at a distance of 31mm.

These pull the feeders down so that they don't wiggle.

On both sides of the base plate there is a profile rail with which it is attached to the machine frame.

At first I wanted to support the plate in the middle, but that doesn't seem to be necessary

because the plate does not bend measurably when fully equipped with feeders.

The entire construction is electrically conductive and quite stable.

Is there some other reason I should build the plate stronger? Have I missed something important?

[Christian...@gmx.at]

Hey Peter, thank you for the flowers :-)

Regards

[Christian...@gmx.at]

Hey guys,

the day before yesterday I made a vacuum tank and today I tested everything.

The Juki nozzles are unfortunately quite leaking but the rest of the system is completely tight.

The pump creates - 0.75 bar negative pressure. When the tank is "full" I can use it to hold a part for a minute until it is empty.

But it depends on how airtight the individual Juki nozzles are. Maybe I can optimize the Juki nozzles a bit.

Maybe I'll build another tank for vacuum that is much bigger. I will then use this tank for air pressure (blow off is needed) instead.

Now here are a few pictures of the vacuum system and the tank:

[mike....@gmail.com]

Christian, this way might work, as long as you pull the feeders down so that they don't wiggle.

Mike

[Christian...@gmx.at]

Mike,
In the beginning I was looking for a way to hold down all feeders in the same way.
I use 2x8mm, 12 / 16mm and 24 / 32mm. All three types of feeders use a different form of assembly.

A continuous rail like the one you made only works with the 2x8mm and 12 / 16mm.

A variant with M5 screws, where the screw head holds the feeder down, only works with the 12/16 and 24/32.

Since all feeders have a ferromagnetic metal plate in the same place, I decided on the variant with strong magnets.

As long as you don't hit the side of the feeder bank with full force, the feeders won't wiggle.

If the magnets are not strong anough i can take stronger one. At this time i just use N35 - 20mm diameter.

I can increase to N52 - 25mm diameter - They are MUCH stronger.

[mike....@gmail.com]

My front U channel is used to hold the feeder down and horizontal, Bolt head is keeping the proper start line position and distance of 31mm...

[Christian...@gmx.at]

I have the bolts too for alligning the feeders in a 31mm distance.

But the wedge is not useable for all kind of feeders.

you can see it on the attached pictures.

12/16mm - Feeder:

24/32mm - Feeder:

[chende...@gmail.com]

Hi Christian,

I do not get one thing. Please tell me if I understand it right. Your machine have several nozzles and you use one vacuum pump and a 1:4 distributor for all?

Do you use vacuum sensors? How do you detect if a part is present in a particular nozzle or not if the vacuum circuit is common for all nozzles?

My machine is not ready yet but it has separate vacuum circuits, each with its sensor.

I made some tests with different juki nozzles and I found that for large nozzles, vacuum difference is large and vacuum level with open circuit is quite close to 0, but with small nozzles there is significant vacuum level present even with the part off, that's because of the flow restriction induced by the small channel of the nozzle. Iirc, around 400mbar with a 501 nozzle and no part attached.

[Christian...@gmx.at]

Hello,

Sorry if I caused confusion.

The picture above does not show the entire system.

It's just the test system.

All right, one by one:

1.) The vacuum is generated by two of the pumps shown.

      These two pumps run at the same time and are only controlled by a vacuum sensor

      - like a conventional air compressor.

2.) The vacuum is stored in a vacuum tank.

      The sensor described above is mounted on this tank and controls the two pumps, as I said.

3.) The outlet on the pressure side of one of these pumps fills an air pressure tank for the blow-off function.

      This tank is also controlled by an air pressure sensor, which opens the tank when the pressure exceeds,

      for example, 200 mbar in order to avoid a loss of performance of the pump

4.) The vacuum is then distributed to the separate nozzle units by a solenoid valve block

      also shown in the picture. This block is controlled by openpnp.

5.) The air pressure is divided and controlled in the same way.

6.) Each nozzle is monitored by a vacuum sensor.

      The sensor is not installed directly on the nozzle unit to prevent high air pressure on the sensor

      when blowing the parts - SMC vacuum sensors do not like high air pressure.

      Perhaps I will mount each sensor via a 3-way distributor directly after the valve outputs.

So in total there are 2 pumps, a vacuum and an air pressure tank, 5 vacuum sensors, a pressure sensor,

2 pieces of 4-way solenoid valve blocks, a single solenoid valve and many hoses and distributors.

I hope I have cleared all the clarity :D

Regards

[ma...@makr.zone]

Hi Christian

> this tank is also controlled by an air pressure sensor, which opens the tank when the pressure exceeds,  for example, 200 mbar in order to avoid a loss of performance of the pump

So you run the pumps all the time?

Why not run them on a hysteresis to keep the tank pressurized to a certain level?

https://makr.zone/vacuum-sensor/192/

_Mark

--

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5932dca5-8d69-4f46-97b1-a9946dff970bn%40googlegroups.com.

[ma...@makr.zone]

> I made some tests with different juki nozzles and I found that for large nozzles, vacuum difference is large and vacuum level with open circuit is quite close to 0, but with small nozzles there is significant vacuum level present even with the part off, that's because of the flow restriction induced by the small channel of the nozzle. Iirc, around 400mbar with a 501 nozzle and no part attached.

Have you found a solution for that problem?

I tried to improve the OpenPnP vacuum system with relative measurements etc. but while this greatly improved sensing for the other nozzle tips, even these enhancements did not fix the problem for the finest nozzle tip (0402). There is no conclusive threshold that has neither false positives nor false negatives:

https://makr.zone/openpnp-advanced-vacuum-sensing-part-on-part-off-detect/421/

I guess for that nozzle tip one would have to build a switchable flow restriction behind the vacuum sensor.

_Mark

--

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a4c0ab09-298a-4059-9fea-6313ffd2a47dn%40googlegroups.com.

[chende...@gmail.com]

Hi Mark,

You improvement with relative measurement is tremendously helpful precisely for this reason, that depending oh the nozzle type, the vacuum without part does vary because of the restriction.

I did not performed real tests yet, currently I am working on installing a quad head on my table and there is much to work on the controller side.

The tests I am talking about above were made off machine. I had a compressed air source (I usually use cylinders at 200 bar, I am scuba diver and have plenty of them). Those are connected to a reduction to 10 bar, then to an air purifier unit. From there, I go in a relay and a vacuum ejector. From vacuum I go to nozzle and a vacuum gauge. I like to keep the vacuum section of the circuit as short as possible to minimize time required for vacuum to build to the required level. This is the configuration I used and by placing small parts on the various nozzles, I was able to get the numbers. I will keep the idea with the restriction and try to test it after finishing the installation process of the heads.

And to avoid being off topic, Christian, I (think) I understood your configuration, and hopefully it works. In the initial stages of my design I also considered the idea to use a general vacuum source, I also have vacuum pumps available, but I discarded it for two reasons: the way I designed it, I could not differentiate between vacuum circuits of different nozzles, and because the pneumatic circuits where the sensor is connected gets bigger. I am not saying I am right, my experience with pneumatic circuits is very limited, but those were the best reasons I could get at that time.

[Christian...@gmx.at]

Hey Mark,

> So you run the pumps all the time?

No, I will not do that. The two pumps run with a hysteresis that is controlled by the vacuum level.
One of the two pumps fills an air pressure tank at the same time.

This tank will only be filled when the vacuum level is below the threshold.

Since a lot more vacuum than compressed air is required, I vent the compressed air tank

with a second hysteresis regulator so that I have a continuous air pressure of approx.

To make the Juki nozzles more airtight, I lubricate the spring-loaded lienar guide

of the nozzles with a highly viscous lubricant. Any ideas if this could work?

[Harjit Singh]

Christian, when you get a chance, can you post closeups of the head your designed and built? I checked you other posts and didn't see them and might have missed them.

With the pictures, it would be great to have you comment about interesting aspects and/or learnings.

Thanks.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/af96f04b-0768-4b93-8524-67e0ce674883n%40googlegroups.com.

[Christian...@gmx.at]

Hi Harjit,

Once the head is built, I am happy to do that.
That should be ready in a couple of days. I still have to wait for the laser parts.
I also need to do some tests as a basis for the comments

[Christian...@gmx.at]

Today I've made a bigger vaccum tank.

Next to it is the old tank for size comparison :D

[Harjit Singh]

Is love to know more about the tank making process, please.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/4b17ecc9-aaa4-471f-b909-e3e04e18c39an%40googlegroups.com.

[Christian...@gmx.at]

Hello Harjit,

I made the tank from a stainless steel pipe with a diameter of 130mm and a length of 200mm.

I turned both ends flat on a lathe and then attached a 2mm stainless steel sheet.

I then sanded this roughly as round as the pipe and then welded it to the pipe.

Finally, I re-sanded the edges. I bent the fastening straps on a folding machine and the rounding by hand.

I spot-welded the tabs before I welded the second plate to the tank.

After the tank was welded shut, I drilled the holes in it and welded the fittings on.
That’s it. Not a big deal

Unfortunately I didn't take many pictures, I only have two more:

[Christian...@gmx.at]

Hi Guys,

Due to too much work and even more lack of time, I couldn't come here to share the progress of my PnP Machine.
For the same reason, my machine is taking a lot longer to complete than I had hoped.

In the meantime I was only able to assemble a base for my machine and attach a few attachments.

The base is extremely stiff! The whole structure with its, now already 150 kg, does not give in a bit, not even if you bump against it properly.

[Christian...@gmx.at]

Hey Guys,

here are some new pictures.

best regards.

[Jesse Lackey]

TOTALLY Awesome! I wish you were in the Bay Area, I'd love to help
fiddle and make.

J

> --
> 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

> <mailto:openpnp+u...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/7097d9d1-5f34-464e-99e8-e3bd8e14ca5dn%40googlegroups.com
> <https://groups.google.com/d/msgid/openpnp/7097d9d1-5f34-464e-99e8-e3bd8e14ca5dn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Christian...@gmx.at]

Thank you Jesse for your nice words.

To be honest, I'm a little proud of this machine, but now it has to work.

And even if it looks so good, that doesn't mean that it will populate my circuit boards.

When everything is done and the machine works, I'll publish all the plans.

Of course, I'll give them away beforehand, but there are still a few improvements to be made.

best regards

[Ian Arkver]

That's some serious engineering. It looks very solid and neat. Good luck with getting it finished and working. Looking forward to the videos. :-)

All the best,
Ian

[Dennis Deyen]

I also plan to use Panasonic HG-C laser displacement sensors on my Signking rebuild. There are a lot of features I would like to add to this machine like calibrating the Z height of all of the feeders in one shot.  Currently there is no way of telling what the height is of each tape and there are differences due to different feeder manufacturers and tape thickness whether paper, plastic, etc.  The laser sensor can also scan a PCB after parts are placed and compare part placement to a gold standard board to check for differences.  I will probably build a separate machine for this function so the PNP can stay busy placing parts.

There are 2 Firewire cameras on the machine connected to an older Pentium 4 industrial PC.  The PC is fast enough for what it is but I will be upgrading the PC to either support USB3 or 100mb/s ethernet global shutter cameras.  The Firewire cameras are global shutter cameras and need to be global shutter because the machine snaps a bottom picture of the 6 nozzles on the fly.  It doesn't stop to take a pic of how the parts were actually picked and I wouldn't want it to stop because any unnecessary deceleration or acceleration may shift parts on the nozzles after the bottom pic was taken.  It also kills throughput.

This will be quite the rebuild for me.  I have a full size Bridgeport CNC mill so making parts is not an issue.  I am toying with making the head a quick bolt on assembly with a built-in pneumatic interface block for the 6 vacuum valves and 6 positive pressure placement valves.  Currently it is a wiring and pneumatic hose mess and very difficult to work on the head.

The main reason to rebuild the head is to use inexpensive Juki 500 series nozzles and automatically pick nozzles from a storage rack in the corner of the machine.  Signking is a Chinese company that went out of business maybe 10 years ago.  They had proprietary nozzles and I'm down to my last few.  They easily break off if a feeder doesn't respond fast enough.  The Yamaha style feeders use a pneumatic cylinder to advance the tape and have a spring return.  

I have a friend in Nanjing CN who tracked down an engineer who used to work for Signking to get the maintenance/calibration password but the engr didn't want to give up the pwd because he was afraid someone would duplicate and sell these machines.  If the company is out of business what does it matter?  Without the maintenance pwd my hands are really tied so it makes sense to switch to OpenPNP.

The motion system has 400W Sanmotion brushless servo motors and steppers for everything else.  There are 12 small steppers on the head for up/dn and rotate for each nozzle.  I want to simplify the stepper motor drivers and use current production driver ICs.  In the past the 4-channel Sanyo-Denki stepper drivers have overheated during use and toasted drivers and steppers.  Used parts are available on ebay but not a great solution.  I added 2 fans to cool the drivers and steppers and have not had a failure since.  The 2nd pic is a top view of the head assy.

The X-Y axes are pretty fast. Maybe 1 second from corner to corner.  The current software does allow slowing motion but sometimes will drop a larger inductor because the inductor has some momentum when accelerating or decelerating. 

Sorry for the long msg.  I plan to interact with the OpenPNP forum quite a bit in the next year or so during the rebuild.  I am located in SW Iowa USA if anyone is interested.

This is my introduction. Have a nice day.

Dennis Deyen

And here is the construction of the placement head:

And here is a photo of the machine showing the current production progress:

I still have to manufacture all machine parts that are still missing, which unfortunately will still take some time. Mainly because my CNC milling machine was unfortunately struck by lightning and I need to repair it first.

I am always open to discuss suggestions for improvement.

AND YES, if anyone wants them, I can provide the construction files ;)

By the way, if someone can produce 8mm aluminum milling parts for little money and get it shipped cheaply to Austria, please contact me :-) (I still need the three aluminum plates from the assembly head ... but - as you already know - my milling machine is dead at the moment)

So, hope you like it.

Best regards,

Christian

--

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ce82f761-f292-4c59-a8dd-ba3e7f211692n%40googlegroups.com.

[Christian...@gmx.at]

Hello, my dears!

Once again I found the time to continue working on my machine.

Not least with your help, I have also managed to get the many small and large problems under control.

In the next few days I will then take care of other problems:

1.) The 12 Nm closed loop stepper for the Y-axis causes a positioning error of approx. 0.5 mm.

      With the backlash offset I haven't been able to get the problem under control until now.

      If that doesn't work, or I can't find the cause of the error, I use a servo with a planetary gear

      instead of the stepper motor. (This already works perfectly for the X-axis)

2.) The serial communication with the Siplace feeders does not seem to run completely stable.

      I also have to connect a USB isolator between the nano every controllers and the PC to protect all other

      components. The 12/16 mm feeders cause nasty transients - that's why I've already lost a camera.

3.) I still have to design reliable board holders.

4.) I still have to think about the best way to connect the emergency stop switch - any ideas??

Etc...

I took a few pictures of my machine again. This is what it looks like now:

[Mike Menci]

Hello Christian,

Nice looking machine ! 

Ad.1) Do you use one or two motors for Y axis ? Is the error of 0.5mm only when you move in forward (or as well beckward?)

Ad.2) Yes I have similar problems with Serial to Siplace feeders - latly I upgraded to Win11 and I tried without the USB isolater/ filter ... but than my new HP computer died and its at servise office now...

Ad.3) PCB holder - have a look at my low profile design: https://grabcad.com/library/pnp-openpnp-low-level-pcb-fixing-1

Ad.4) If you use Smoothieboard : see http://smoothieware.org/emergencystop

Cheers

Mike

[Christian...@gmx.at]

Hey Mike,

thank U!

I Use one motor for Y axis.

To be honest, I don't know in which direction the error occurs. I only noticed it in one direction, but that could also be because I always reference the machine from the same direction.

I noticed the mistake like this:

First movement:
Camera over pick up position of any feeder -> move to the nozzle over the pick up position and then back to the camera over the pick up position.

In this case, the last movement of the Y-axis takes place in the + direction. A position error of approximately 0.5 mm (Y) occurs.

Second movement:
Camera over pick up position of the same feeder -> move the head to the maintenance position (Y max, X center) and then back to camera over pick up position.

The X and Y positions are reached exactly.

I know it really looks like a backlash problem, but there is no measurable mechanical backlash in the system.

All pulleys are connected with feather keys on their shafts and the whole construction is extremly stiff.

The transmission ratio is 1:1 and the pulley that drives the linear belt is about 50mm in diameter

- that's about 157 mm per round. I use 5Mx25mm HTD belts and they are strong tensioned.

Maybe the rresolution of the closed loop stepper is not sufficient or it's a chinese driver of bad quality.

Ad.2) I ust have to make a lot of experiments and i hope i find the reason for those communication issues.

Ad.3) Very smart design!! I like it, but for my rail supportet system i need a different solution.

Ad.4) I use marlin runnig on a BTT Octopus, so it's completly different. But maybe there is an idea for a similar solution

all the best

Christian

[Mike Menci]

Hi,

Ad.1)  I wanted to move on my PnP  Y servo to have Shapeoko type of Y drive -

but it did NOT work for some reason in one direction I was loosing some 0.4mm.

see:  https://youtu.be/F-9-kBaeEek

I was never able to find mistake why!

Ad.2) I still face communication issues - when setting up Schultz feeders in OpenPnP,  but after I start a Job / PnP and Feeders work (???!!)

Good Luck Christian!

Mike

[Mike Menci]

To make it more Clear - error happens when I change direction - from home  +10 ......+10, +10, +10, to -10, (-10 -10) this was my observation!

Perhaps you want to test this Christian as well.

Mike

[Christian...@gmx.at]

Thanks Mike I will test it.

But when I remember right I only loose those 0,5mm one time when changing direction.

regards

[Christian...@gmx.at]

Hey Mike,

Hi Mike

I have now done some tests and have come to the conclusion that the backlash already begins on the engine.

I measure a gradual deviation of about 0.5 degrees directly on the engine's toothed belt pulley.

It gets a little more on each additional pulley. I've used keys on all shaft connections, but they probably have some play.

I will now replace the motor (closed loop stepper) with a servo.

The servo is not as loud as the stepper and hopefully a bit more accurate.

I will compress all feather keys in a press so that they become a bit wider and then glue them with screw lock.

I hope this solves the problem with my backlash. Finally, I also have backlash on the X-axis, but this is only 0.009mm

Best regards

Christian

[Mike Menci]

Hi

I would use glue on motor shaft only after all tests are completed and I am sure all is OK ! 

Best would be a pulley with conical bushing that fixes pulley to shaft and no glue! 

BR, 

Mike

***************************Disclaimer***************************

The contents of this e-mail and any file transmitted with it are confidential and intended solely for the individual or entity to whom they are addressed.  The content may also contain legal, professional or other privileged information. If you received this e-mail in error, please destroy it immediately.  You should not copy or use it for any purpose nor disclose its contents to any other person. The views stated herein do not necessarily represent the view of the Company.

Please ensure you have adequate virus protection before you open or detach any documents from this transmission. We do not accept any liability for  viruses!

On 16 Jan 2022, at 12.22, Christian...@gmx.at <Christian...@gmx.at> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2a9e99a3-935c-4ade-b182-99899db51f28n%40googlegroups.com.

[Christian...@gmx.at]

Hi Mike,

I think you mean this:

https://antriebstechnik24.eu/Zahnriemenscheibe-HTD-TB-36-5M-15-mit-Taperbuchse-1108-inkl-Bohrung

Sure, I wanted to use these pulleys with taper bushes, but couldn't find them anywhere for a 25mm belt width with 36 teeth.

[Mike Menci]

Yes this type would be perfect for you

BR, 

Mike

***************************Disclaimer***************************

The contents of this e-mail and any file transmitted with it are confidential and intended solely for the individual or entity to whom they are addressed.  The content may also contain legal, professional or other privileged information. If you received this e-mail in error, please destroy it immediately.  You should not copy or use it for any purpose nor disclose its contents to any other person. The views stated herein do not necessarily represent the view of the Company.

Please ensure you have adequate virus protection before you open or detach any documents from this transmission. We do not accept any liability for  viruses!

On 16 Jan 2022, at 14.13, Christian...@gmx.at <Christian...@gmx.at> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/91538871-d082-46cc-bfce-1857c537541fn%40googlegroups.com.

[Christian...@gmx.at]

Hey guys,

here are some new modifications:

1.) BOARD CLAMPING:

Beveling the 2x10mm brass strips on the edges was quite a challenge, but it was definitely worth it.

The PCBs are held extremely tight and very reliably. But you can get them in and out easily and without effort.

With a lower

     

I will add a second rail below the circuit board holder so that JEDEC trays can also be clamped reliably and always in the same place like the next picture shows.

2.) REEL TRAY:

I also found a simple and inexpensive solution for the reel trays. And it still looks good to me.

Where I only have belt sections, I inserted a sheet metal strip so that the wound belt section does not fall through.