3D printed kits with OpenPnp - Pandaplacer versus Bing V3

[Jarosław Karwik]

Hi,

When trying to update CHMT36VA with automatic feeders I have concluded that it would be  actually faster and cheaper for me to buy whole machine kit ( expecially that I own Prusa XL and Voron 3D printers )

 

This was not really possible few years ago, but now it seems that the progress in open source feeders makes possible working with  0402 components

Bascially there are two contenders:

- Pandaplacer - https://pandaplacer.com/

- Bing Luo - https://www.microsmt.com.cn/

Full machine setup ( without printed parts) would be in 1000....1500$

Both machines have rather good comments, although I have not seen full independent reviews by people using them for small productions - right now mainly first impressions are known.

Pandaplacer:

- Seems more people bought it, US distributor with documentation

- Cheaper

- 0402 possible but tricky ( according to peoples comments)

- Better overall machine arrangement ( mainly cabling)

Bing Luo:

- Potentially better feeders (special 0402 version). more complete set ( up to 24mm)

- Documentation mainly as youtube videos

- I have tried to contact them with some feeders questions - no answer

Is there anybody who seen them both ? Or used any of these extensively ?

[PP.ca]

Disclaimer: I'm in charge of the North American store for Pandaplacer

Both machines are good and do what they are supposed to do quite well. 

I seriously doubt there is anyone who has both, it's usually one or the other. 

There is someone on this group that had the LumenPnP and did a side-by-side comparison with Pandaplacer. (then sold the Lumen)

Let me add some more to the list for Pandaplacer:

* all metal construction, nothing needs to be 3D printed to get the machine running

* 0402 - super-reliable due to the OpenPnP BambooFeeder what has vision calibration and due to the BambooFeeder v2.0 (I'll publish this weekend)

* 0201 - video recorded, will be uploaded soon - proves further precision and reliability

* 5 feeder controllers included for a total of 65 feeder ports

* significantly cheaper: $729 for the machine + 7 x $25 feeder kits = $904 USD for a machine with 65 feeder ports and 70 total feeders available (buy more kits to keep feeders with the reels like I do)

* well documented

* EST timezone first level support (sometimes even video meetings)

* available in stock near you

* shipped within 24 hours from the Toronto area, usually takes 3-5 days to deliver via USPS almost anywhere in US

* active development, both hardware & software, including contributions to OpenPnP

[bing luo]

Actually, I want to sell at a higher price, like Lumeng. It's not because I want to make more money, but because I'm afraid of harming designers in the United States or the European Union, making it impossible for many original designed accessories to get profits. Without their start, there would be no existing . Because the production cost of these accessories in China is much lower than in Western countries.

And I have always been solving the problem of automatic feeders, making them closer to industrial feeders at the lowest cost.I never thought of reducing a $1 N20 motor on top of the 0816 push-pull feeder as an innovation, because I think it's meaningless.

In addition, I think some rules of PNP cannot be violated, such as the heavier the machine, the higher the accuracy it can achieve; And you can only use aluminum alloy, and absolutely cannot use materials with poor rigidity like iron to make it.

At the current price, I already feel that it's too difficult. It takes me too much time to package a set of PNP accessories. I hope to sell it more expensive so that I don't have to sell so much. People buy my machines just because they like my original feeder. It's hard for me to imagine how selling a PNP for $700 can get profits and even develop agents, where almost everyone is working for free.

I prefer to see some special designs like https://www.youtube.com/ @AFARCO PnP.     Instead of being cheaper and cheaper.

Can we not compare them together here? 

[bing luo]

Based on most of the feedback I have received, lumenPNP works well, but because it is not a finished product and requires assembly and debugging, some individuals may give up due to difficulties. LumenPNP is produced in the United States and its labor and material costs are clearly high. But it appeared very early and solved many people's problems, constantly upgrading. So it's best not to simply compare these machines, such comparisons are meaningless. Currently in China, you only need to spend half of Tesla's price to get a BYD electric car with the same performance. Without tariff barriers, Tesla will no longer exist.

[bing luo]

It's hard to imagine using 2020 lightweight aluminum profiles to make PNP .I feel that the aluminum profile of 4020 is not heavy enough.

 

In my current PNP V5 project, I will use 10mm thick aluminum plates as a platform.   And  Taiwan HWin 15mm linear guide rail and 57 stepper motor   will be used. I am willing to spend some extra cost on weight, PNP requires stability brought by weight.

 

[Jarosław Karwik]

Well,

@Bing Luo,

-  if V5 is current design, where do you have V4 ? ( I have seen only V3 in sale)

@PP.ca

- I will wait with my decision for you new feeder design release

- I have seen only 8/12mm feeders, but some comments suggest you have version up to 24mm ?

- Your machine is actually more expensive for me then Bing Luo. I am located in EU, so I cannot buy from US/Canada shop ( no delivery).  Buying from Aliexpress  costs me extra 23% of VAT - as a company I could claim it, but Aliexpress does not allow such option ( it collects it at checkout and I cannot recover it)

- There is a mess in Aliexpress kit options - there is  Pandaplacer-A1 (full option plus) in comments but in in product selection ( only 'Customized' )

About the profile - regardless on the machine bought I have planned to order 10mm waterjet cut metal base plate to stabilize the machine.

[bing luo]

There is no V4, so only V5. V5 is a finished machine because some people tell me that they like OpenPNP software, but they don't want to spend time DIY. They think many finished PNP software is not as good as OpenPNP. And as a finished product, I can make it very sturdy.

[Flynn Dreilinger]

Hi Bing, are you planning to offer the V5 design soon?

[Marshall S. (Alakuu)]

Jarosla,

I've got a PNP US - Ohio. Had it since I think the start of the year.
I've got the bamboo feeders (v2 or something like that).
I had originally expanded the old design to make up to 24mm feeders. They worked, but there was tape tensioning issues.
Having messaged Leo and sending along my designs as well as some details about the 'problem points' he actually a couple of days later send me new updated files!
These latest designs seem to 100% work better. I've been placing with 0402 all the way up to components on 12, 16, and 24mm tape with basically zero issue.
The kits clock in at ~ 2.5$ a feeder and then add about 2$ per feeder in plastic. Assembly is maybe 15 minutes.
The slotting system once you get a feeling for how to assemble it is down right fantastic. I can walk up, grab a feeder and pull it out. Insert a new component and have functional placement in a couple of minutes.

My problems with the PandaPlacer are as follows:
1. Aluminum frame. Tramming everything is difficult. Surfaces aren't perfect. But honestly the machine placed 0402 components 'okay' with my first assembly where I did basically nothing to try and square the frame. But it does move around and really likely needs something to deaden vibrations.
2. The ribbon cables for data and power to the axis' worry the heck out of me. I've got quite alot of hours accumulated on my machine without any notable issues. But they're not proper strain relief cable chains. It might be that this will work better. The ribbon cable is probably quite cheap to replace too.
3. PCB holding system. It works, but I keep coming back to it as a point I'd like to improve. Been trying to figure out maybe a way to put linear rods and a screw system on the underside of the 2020 to create some kind of solution that clamps squarely. Also I don't think a conveyer system is possible but that's really really not something that should even be considered for this class of machine.
4. Cameras / lighting. This is a subject that I completely lack the understanding of, but vision does require the nozzle to go 'down' ever component, so there's a fair bit of time loss. I also have some miss identification issues with SOD323 diodes that I just can't seem to tweak out of. 
5. Stepper drivers get hot. I added a fan for both the control board driver and the one on the axis. The head mounted steppers are already cooled. Leo ages ago mentioned they should be upgradeable I just haven't chased that yet.

I haven't tried Bing's design, but it's a very solid looking machine. His feeders are well liked too!
I do agree that the PandaPlacer feels almost too cheap for the work it does. I also completely agree that these machines very likely SHOULD be more expensive for what they are. I do worry that the floor for 'okay' professional machines being around 3k means there isn't much room for machines to increase in price without bumping into the impossible to compete with professional machines. 

TLDR
PandaPlacer and the bamboofeeder work great. I've pushed ~6k SOD323 diodes without a bad pick. 0402 work. Up to 24mm work get the latest files tho!
Bing's machine and feeders are very well tested by the community.

You probably can't go wrong either way.

[Jarosław Karwik]

It is tough choice - both good products. Two different feeder designs - let me show you some pictures.

Bamboo feeder versus 20+ years old mechanical industrial feeder from DIMA

Both use same principle. Dima was not motorized, but gantry was used to push the lever. Here 0603 Dima version - there are also 0402 with dual levers

Siemens (well known here - 3 lane version on the picture) versus Bing

 

Also basically same principle.

There is actually one important difference ( except that industrial feeders could take millions of components) - both have "shutters"

Here ( Dima) you can see metal part which moves along tape and pulls back with tape to expose the component.

This allows rapid tape progres ( component is covered and will not jump out). Then it retracts - when component is  steady in its pickup place.  

While not essential ( feeders do work nicely without it), it allows faster tape advancement and as end result faster operation.

It is really nice to see that very expensive industrial solutions can be optimized so much and still retain most of its precision and functionality :-)

I will likely go with Bing Luo - his machine is a bit bigger and heavier ( I will reinforce it with some metal plates - this would be trickier with Pandaplacer). This is unfortunately more expensive option  ( as I would like to build feeders for all components I use).

Hmm, would it be out of place to mix feeders - and use cheaper Bamboo Feeders for some of them (these less often used) ?

[Jarosław Karwik]

This I really love :-)

https://www.youtube.com/watch?v=rSL4-6i3cXM

[PP.ca]

Big shoutout to Nick that was kind enough to be one of our beta testers for BambooFeeder AS2!!!

He also had to suffer through some interim versions with tremendous patience.

Add to this system the visual calibration from BambooFeeder-AutoVision and reliability for 0402s and 0201s is at a whole new level for about $3-$5/feeder.

[Danh Trinh]

@pandaplacer - I would like to see a video of your V2 feeders placing 0402 parts.

--
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/a5216721-38ae-4863-8b9f-27645ad20d00n%40googlegroups.com.

[bing luo]

If you have used the 0816feeder, you will know that it is impossible to work for 0402 through push-pull operation, because you need to push it forward through the holes on the tape. The minimum spacing between the holes on the tape is 4mm, so you will push forward at least 4mm each time, while the spacing between 0402 is 2mm.

The solution to this problem is to modify the software of OpenPNP so that the coordinates of the components picked up by OpenPNP for the second time are offset by 2mm from the first time, and then repeat the work.

If this direction is correct, then just use the classic 0816 feeder that appeared 7 years ago, because the 0816 feeder only uses an additional $1 N20 motor ,Its total cost has not exceeded $4 either. But you cannot see any   push-pull feeder  today in any SMT factory.

[bing luo]

https://youtu.be/6aExyo1NG2s    

In this video, I made improvements to mgrl's 0816 feeder, which is very cheap and costs no more than $4. The price of the servo and N20 motor is $1 respectively, and the price of the spring and PCBA is $1. You may need another $0.5 PLA to print it, so the total cost is approximately $4.   

Due to the presence of N20 motor, you don't have to worry about the Tape film being full. 

Bamboofeeder should be compared with 0816feeder, and their principles, structure, cost, advantages, and disadvantages should be similar.

I would prefer someone to design a more creative sprocket feeder for comparison.

[Jarosław Karwik]

Well, push-pull feeder become obsolete for several reasons:

- they are bigger then modern motorized solutions ( width matters - 2,3 lane electric feeders save space )

- slower then modern solutions

But they are smart design and basically timeless .... ( 2mm DIMA Optimat with 2mm feed advanced using 2 levers)

https://www.youtube.com/watch?v=VyLRNqCEI5U

Also video of shutter operation ( for 4mm feed version):

https://www.youtube.com/watch?v=FOffg7-mMO0

BTW. Bing -  I thought I would go for your machine, but AS2 ( and as result - Pandaplacer) is what I really need ( or rather want)

It was always my dream to have closet full of feeders with components which I can easily install in my machine.

And their management system - feeder with magazine and embedded connector ... blows my mind.

[bing luo]

1:  But they are smart design and basically timeless .... ( 2mm DIMA Optimat with 2mm feed advanced using 2 levers).

3D printing cannot achieve this level,  If made of metal, its cost will be unacceptable.

2:  Many feeder developers, such as Rapid and Lumen, use Rs485 or I2C for communication between feeders and control boards, so they do not need any cables

https://groups.google.com/g/openpnp/c/tHkY0duicdI/m/71xyRri3EQAJ.     

I am currently testing infrared communication, and if there are no issues, it can be more concise. 

[Jarosław Karwik]

Well, I know that metal production is much more expensive.

For me the price compromise would be having laser cut flat metal sheets supplemented by 3D print - it would save a lot of time on printing ( even if assembly would be more tricky). This maybe not an option if you do it only for yourself - but when you sell kits it sounds much more viable.

About your IR - you still need supply. I know you mentioned placing batteries inside, but then maintenance of such feeders would be troublesome.

I think AS2 has  better design choice. If I were you I would go that way - a bit better connector (integrated with PCB - there are plenty cheap and reasonable) with CAN/RS485 . Maybe even multi pin connector with address coding on matching side - there are plenty options.

[bing luo]

Mgrl has already achieved a push-pull feeder level of 0816. Why should we focus our efforts on reducing one N20 motor.  It is meaningless to compare the cost between the push-pull feeder and the sprocket feeder. 

在2024年7月8日星期一 UTC+8 16:43:05<jarosla...@gmail.com> 写道：

[bing luo]

I don't put lithium batteries. Maybe I will make a sample to attract attention. You know, with lithium batteries, it's not possible to ship them by air.

Metal is impossible. When its price reaches $50 each, you can consider using a second-hand Siemens electric feeder.

[bing luo]

A NEC infrared controller can directly control 99 feeders, which is more attractive to me. If it has no other issues. Has anyone ever made such an attempt? Tell me the result so I don't have to take a detour.  
Connecting the feeder directly through a PCB may be my final consideration.

在2024年7月8日星期一 UTC+8 17:22:05<jarosla...@gmail.com> 写道：

[Jarosław Karwik]

Well,

I have not seen anything like that - there were some using IR leds to detect nozzle ( as trigger for feed). But such approach is not very attractive.

I can only guess this was never tried because:

- You cannot use TV in the same room or another machine  ( well, there might be some workarounds )

- Where would you put the led ?  Might be tricky as the machine controller would need more then one transmitter to cover whole machine - you cannot depend on reflections nor the led transmitting beam cone might not be sufficient to cover whole machine .

While you can make it running - I would use simple 3 wire solution . One extra communication cable - for unidirectional communication. Would work even without CAN/RS485 transceiver as it is easy to make one wire half duplex uart.

https://ww1.microchip.com/downloads/en/AppNotes/USART-in-One-Wire-Mode-ApplicationNote-DS00002658.pdf

it is possible to make even uart over supply, but such design is more troublesome and likely more expensive.

[bing luo]

The bing feeder does not require communication, it only requires a high-level signal to trigger. So even if a PCB is used for connection, its wiring is still very simple. One wire is used to transmit high-level signals, and two wires are used to provide 5V voltage.   

[Jarosław Karwik]

Then it is only matter of neat arrangements of slots with cable plugs.

Nobody really counts cables/pins - here the end result matters.

And simple solution to just plug without thinking - wins for me.  I could plug cables (like you have now) and it is working solution, just less appealing.

[mark maker]

> If you have used the 0816feeder, you will know that it is impossible to work for 0402 through push-pull operation, because you need to push it forward through the holes on the tape. The minimum spacing between the holes on the tape is 4mm, so you will push forward at least 4mm each time, while the spacing between 0402 is 2mm.

> The solution to this problem is to modify the software of OpenPNP so that the coordinates of the components picked up by OpenPNP for the second time are offset by 2mm from the first time, and then repeat the work.

I've only followed part of the discussion, so maybe I misunderstand what this is about. Just to make it clear: 

OpenPnP already supports that kind of 2mm pick location arbitration for 0402 (and smaller) with some feeders including the ReferencePushPullFeeder. Recently the BambooFeeder was included in the OpenPnP test version, ReferencePushPullFeeder and BambooFeeder now share some common code and both can do 0402. In fact if one day there are 1mm pitch tapes, they will be supported too.

The ReferencePushPullFeeder (despite its name) can support other types of feeders just as well. It also supports electric actuation via two configurable actuators.  It certainly does support a 0816 feeder and can provide it with feature such as OCR/QR part detection, one-click auto-setup etc.

This video is about the all-3D printed push-pull feeder, and demonstrates the 0402 picking in this scene:
https://youtu.be/cNCjjvCT4Fc?t=377

The same here in the video about the OpenPnP setup:
https://youtu.be/5QcJ2ziIJ14?t=255

Yes, I know it is slower than an electrical feeder, but also much cheaper, and quicker to make. The idea is to have your tapes permanently loaded in these cheap feeders (like a cartridge), which saves a lot of time on setting up jobs:

https://makr.zone/new-all-3d-printed-tapereel-feeder/399/

https://github.com/openpnp/openpnp/wiki/ReferencePushPullFeeder

_m

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a180ef6a-3b23-4e21-8d12-87beb19d5692n%40googlegroups.com.

[bing luo]

So, if  need a cheap push-pull feeder, just choose the 0816 that costs $4. It has been around for many years, and for DIY, there is no problem with this，Openpnp used another method to pick up 0402. 

But  the purpose of developing   Sprocket  wheel feeders is simply to approach the speed and stability of industrial feeders as closely as possible. Because even on the JUKI PNP produced 30 years ago, only pneumatic sprocket feeders could be found, and push-pull feeders could not be found.    People will have different choices based on their own needs, and simply comparing prices is meaningless.

[Jarosław Karwik]

Well, unless you develop sprocket based feeder with shutter - it will not be much better/faster then push/pull. Here physics is the limitation - you need to keep component when moving it fast.

[PP.ca]

One would be incorrect to assume AS2 is just another PushPullFeeder or 0816 servo-based AutoFeeder in a different shell.

Don't be deceived by its simple looks, there is more than meets the eye.

The AS2 is the true result of hardware, firmware and software coming together and without any of these 3 it wouldn't work as well as it does.

All sorts of things were considered, including the impact of vibrations over the tendency of small parts to jump out of their pockets...

We are happy it got to a point where the new hardware could be released, but there's more work in progress as we speak.

[mark maker]

> So, if  need a cheap push-pull feeder, just choose the 0816 that costs $4. It has been around for many years, and for DIY, there is no problem with this，Openpnp used another method to pick up 0402.

Sorry, I don't understand the point of this answer. Isn't it simply a matter of preferences?

If you're more concerned about production speed (you have larger series), then yes, a 0816 might be a better choice. But if you're more inclined towards faster job setup (you have smaller series/prototypes), then there are still valid reasons to favor a mechanical push-pull feeder, e.g. no need for cabling and connectors (which cost too), no need for more controller(s), faster 3D-print and assembly of a new feeder.

Plus I don't think I can get all the needed 0816 hardware for 4$ here in Switzerland 😉.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7f57ef54-0e4a-40fc-9d28-305fabe2fec4n%40googlegroups.com.

[bing luo]

Yes, if affordability is an important point, then a mechanical push-pull feeder like the one in your video is the best solution, with a cost of only about $0.5，and  slower speed is actually an advantage, as slower speeds prevent components from jumping out.   

Comparing these completely different plans together is meaningless. If the push-pull feeder is the end point, and there is no need for MGRL to develop 0816, then there will be no rapid, lunmen and other people constantly launching various sprocket solutions. The advantage of OpenPNP is that there are various solutions for people to choose the one that best suits them.   

[Jarosław Karwik]

I can speak only for myself - I do not care if it is sprocket or push-pull. I do care about easy and fast setup and reasonable operation. And as Mark said - having "cartridges" with components is excellent idea.

I prefer easy setup over speed - I do only prototypes and small series. I do not care even if it it 600cph instead of 1200cph - as long as I do not have to do it manually with microscope.

And I would say that anybody with even semi-serious production will never print his feeders and likely buy second hand professional feeders ( larger volumes of dual lane Siemens Siplace could be bought for ~100..150$ ). I guess the second hand market of these feeders makes almost impossible to manufacture own feeders in this price range.

Please remember that it is not the best technological solution which wins people's hearts. It has to fulfills their needs/wants - even if it looks sub optimal for you.

[mark maker]

> The advantage of OpenPNP is that there are various solutions for people to choose the one that best suits them.  

 👍👍👍

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2f78738e-63cb-4071-bf8b-107a418cbca3n%40googlegroups.com.

[bing luo]

These are different requirements, and Mark thinks his mechanical push-pull feeder is enough for him to use. MGRL felt that this was not enough and upgraded it to an automated 0816 push-pull feeder. I have seen many people using this PNP for production and selling their PCBA, which is also why the price of Lumen's Sprocket  wheel feeder is as high as $90 per unit, and there are still many users. Openpnp has been advancing and will not stop at designs from 6, 7, or even 10 years ago.

[bing luo]

I hope to compare the advantages and disadvantages of Bing   sprocket   feeder, Lumeng sprocket feeder, and Rapid sprocket feeder, so that people can develop better sprocket feeders.  It should not be compared here with the 0816 push-pull feeder or its variants. The software performance of OpenPNP is actually stronger than most desktop PNPs you can see in the market, such as Charmhigh or Zhenbang, QIHe, and so on.  

So it should have the opportunity to fit  a more professional sprocket  feeder instead of a cheaper push-pul  one.

[mark maker]

> MGRL felt that this was not enough and upgraded it to an automated 0816 push-pull feeder

That's not how it happened, historically. The 0816 feeder was first. 🙂

I specifically developed the push-pull feeder (the all-3D printed one) because I found all the electrical feeders at the time to be way too complicated for my use case of prototypes and very small runs. Back then the 0816 feeders were not really "mountable" in a practical sense, frankly, I consider such cabling "permanent".

https://docs.mgrl.de/_detail/maschine:pickandplace:feeder:0816-feeder-machine-photo-rear.jpg

This means loading tapes in and out of feeders all the time, which looks difficult to "impossible" and sometimes involves loosing parts. It also means getting the right feeder assigned to the right cable, actuator (index) and/or part with many opportunities for error. All fine if you then produce thousands of PCBs, but rather inconvenient for prototypes and small runs. 

Designs with practical connectors or even proper "plug-in" bus rails only came later. But those either are very expensive, or they don't look like surviving more than a few connections, some don't even mention any ESD protection. And for the bus connectors, I sometimes wonder about signal quality in the presence of harsh machine motion. In most demo videos I saw, the machine is usually moving so turtle slow, even a mechanical feeder is faster.

The ReferencePushPullFeeder also specifically uses OCR/QR to auto-discover where each feeder is loaded, addressing the error prone feeder mapping.  Simply being "there" mechanically, makes this a no-brainer. This is immediately much more complicated with electronic feeders, but solutions with proximity sensors or feeders with QR-code that correspond to built-in MCU UUIDs on bus systems have been discussed.

Note, the ReferencePushPullFeeder combines the OCR/QR vision op with (sprocket hole) calibration, which most practical slot-in electronic feeders require as well (or you have to do it manually), so the head must move there anyways. With all the setup time honestly accounted for, I don't think the mechanical actuation actually costs as much extra time as one might think.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d511f228-25ce-4f22-85bf-24103e084829n%40googlegroups.com.

[bing luo]

0816 has two main drawbacks. 

The one  is that N20 motors interfere with each other. The solution to this problem is not to remove them, but to stand them upright and use a worm gear for transmission. This way you can take out the feeder without touching other feeders
The second issue is that after multiple reciprocating movements of the push rod, due to wear and tear, the tip of the push rod cannot accurately align with each hole, which is why people abandoned it and instead developed  sprocket   wheel feeders. The sprocket has not failed due to wear.

This is an AFARCO feeder made of aluminum alloy  with a push-pull design.   It solves the problem of wear and failure.  

https://youtu.be/PHoDNB5aD00?si=XR-sQENdnZdPQ3El

[Danh Trinh]

A+++ for design cleverness and simplicity. Kudos for showing 0402 picking while others only claim.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/16b2c1ee-20ee-4aa3-88ef-3faf127f802e%40makr.zone.

[Jarosław Karwik]

3D printed parts - side by side

Both printed without any issues  - even if AS2 has smaller parts.

Bing feeder is much more stiff . I know it will be better once fully assembled , also AS2 stiffness  depends on its machine mounting to keep it stable - and this is something I do not like as it feels a bit malleable.

I myself would add a bit more internal  reinforcements, (even if current solution works)  no matter  that it  means more plastic and printing time.

And the concern about wear and tear is quite valid - these parts are quite tiny. I think I will print some of them from PETG with CF ( carbon fibers) for extended life time.

While AS2 for me is ok ( I have really low volume) I would say that if you want to do more - Bing feeders are superior and will live much longer without maintenance.

[bing luo]

This structure is no different from 0816, except that the N20 motor is removed to save costs. It can even use the same control board and firmware as 0816 to drive the servo. 0816 has been used for many years, but it has inherent flaws. The accessory marked with the red arrow will not work many times and will wear out, making it difficult to accurately push the circular hole on the tape forward. This situation may occur at any time, requiring your eyes to constantly focus on PNP work. In this discussion group, you can see that many people are working hard to develop cheap and reliable   Sprocket  wheel feeders, which is why.  

 At present, people who produce finished OpenPNP machines are not surprised to choice  the sprocket feeder as the only option, and you cannot find a push-pull feeder to use.

Of course, if   people who want the feeder to be very cheap, don't care about its speed, and don't care about the need to frequently replace the servo push rod, then  can continue to improve the push-pull method.

They are two different levels of feeders, so there is no need to discuss them together. As for the communication stability and ESD issues that Mark mentioned, I hope someone can improve them.

AFARCO used 3D printing to make push-pull feeders before, but now it has launched an aluminum alloy version, and I believe he has also encountered the same problem.

[bing luo]

The sprocket is made of stainless steel with a thickness of 1mm, and it will never fail due to wear and tear in its lifetime. 

This is imitating the sprocket of the Philips  feeder in the following picture.  The Siemens feeder approach is impossible to achieve. Because it has no feedback signal, its accuracy depends on precise mechanical transmission.

Philips

Siemens

在2024年7月9日星期二 UTC+8 13:49:57<jarosla...@gmail.com> 写道：

[Jarosław Karwik]

Well,

These day I use at work plastic bearings (https://www.igus.com/) which stregth is somewhere between bronze and steel - nothing else so strong which also works directly in sea water.

So not always metal means ultimate solution.

I plan to print the fragile parts with nylon reinforced with carbon fibers - something like this:  https://bigrep.com/filaments/pa12-cf/

Even though I admit that you are right - your way is much better. Just that mine is barely sufficient :-)

These are interesting reads:

https://markforged.com/resources/blog/replacing-aluminum-parts-with-3d-printed-carbon-fiber

[bing luo]

Our direction is different, Our ultimate goal is to  for the finished PNP,  We need it to be more advanced than CHM-T36VA   , rather than becoming more primitive .   So we can only choose the sprocket feeder  as the direction of development.   

We cannot tell the user to look for materials similar to metal to print the feeder. If he are not satisfied with its effect, he can  throw it away and switch to a different one.

[bing luo]

I am very unwilling to compare it with kits like Panda  so  here because everyone has their own ideas. If you like Panda, you can create a separate conversation. If you like PNPV3, you can also create a separate conversation. I am different from PP.ca, I have no intention of promoting my PNP here. I hope this is a discussion group full of novel ideas, not a big store. 

在2024年7月9日星期二 UTC+8 15:55:08<jarosla...@gmail.com> 写道：

[Jarosław Karwik]

Well, we are discussing different ideas - and materials.

Please understand that I appreciate you work.   And I also  think we all benefit from the discussion.   

If your solution had reel holders and different cabling - then I would take your machine ( even with higher price). I just do not have enough time to make the modifications myself.

Once I get Panda kit, I will create new thread.

[huynht...@gmail.com]

Hello, I am the founder of AFARCO, 0816 has one drawback that I had to give up: the plastic push-pull bar, and the top-down push-pull mechanism is not stable, sometimes it will slip on plastic rolls.  , and of course when there is constant friction from top to bottom it will wear out.  So I changed the push-pull mechanism, the push-pull head will push and pull from the bottom up.  Everything works quite well, however there is only a problem with the 2mm feed.  I'm looking into adjusting the software to handle this problem

Vào lúc 13:14:55 UTC+7 ngày Thứ Ba, 9 tháng 7, 2024, tianl...@gmail.com đã viết:

[Bingluo]

I don't care if you buy my kit, I said you just choose what you like. What I dislike is that pp.ca is promoting his pnp in any thread that mentions my suite. This is definitely not a normal person's approach, and no one else would do it. I never promote my PNP here because I hope it is a discussion group without commercial purposes.

---- Replied Message ----

From	Jarosław Karwik<jarosla...@gmail.com>
Date	07/09/2024 17:55
To	OpenPnP<ope...@googlegroups.com>
Cc
Subject	Re: [OpenPnP] Re: 3D printed kits with OpenPnp - Pandaplacer versus Bing V3

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/4ba2997b-c0d8-410a-9770-509da022661an%40googlegroups.com.

[bing luo]

I want to say to PP.ca, friend, making money is not everything in your life. 

If you like, you can open 10 threads here to promote your PNP, but I ask you not to promote your PNP in my threads.

[bing luo]

If pp.ca's behavior is not stopped, this discussion group will eventually become lifeless and become a discussion group promoted by many people who hope to sell kits.

[bing luo]

Disclaimer: I'm in charge of the North American store for Pandaplacer

Both machines are good and do what they are supposed to do quite well. 

I seriously doubt there is anyone who has both, it's usually one or the other. 

There is someone on this group that had the LumenPnP and did a side-by-side comparison with Pandaplacer. (then sold the Lumen)

<<<<Lumen developed three years ago and has the most users in Europe and the United States. Their team has also developed a great automatic feeder, which is definitely not as bad as you said.  

[Jarosław Karwik]

Well,  please do not be upset.  I would say that (at least me) agree that your design is better and Pandaplacer, while works, has some limitations which your design does not have.

I have created separate thread for Pandaplacer and will not continue this thread anymore.

For good or bad, even the thread owner cannot close it.

[bing luo]

To be honest, I don't know if the owner of this discussion group supports PP.ca's behavior, at least I wouldn't do such a thing. Because it turned this discussion group into a Pandaplacer sales conference. 

[bing luo]

It's a good idea to create an independent thread, as there are many people who make the package, lumen ,liteplacer， It is available in the United States, Canada, Europe, Canada, and China. I don't think you want to see everyone come out and say that my kit is the cheapest, my kit is the best, and Lumen's kit has been thrown away and replaced with Panda.

在2024年7月9日星期二 UTC+8 21:48:30<jarosla...@gmail.com> 写道：

[mark maker]

bing, I don't know where your frustration is coming from. Watching from my side, there was no such "sales conference". There was a lot of valid and sometimes hard technical discussion. That missing video of the machines placing 0402 for real is still lingering there 😉

The first post by PP.ca contained a disclaimer, so everybody knew where this was coming from, and that it is potentially biased. A disclaimer is a good thing, in my book.

IMHO, we should be allowed to compare commercial products, and for members of the community to develop products and offer them to the rest. As long as the group is not overly spammed, and as long as the spirit of Open Source is strictly upheld, I believe this is a good thing. The most important thing is for you guys to engage with the community. To offer your products, yes, but also to give back, for the next guy to develop an even better product. Both you and PP.ca have both profited from the guy before you and then also given back, great! 

Ultimately, Jason has the say.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/64cc14a6-a915-45da-9248-0f285e8296b4n%40googlegroups.com.

[bing luo]

This is just my personal opinion, it may be different from your philosophy. But I respect you.

If I don't like it, I can skip the discussion.

[SM]

What strikes me about all these commercial OpenPnP machine kits from the far east is that they all have one thing in common:
It is primarily about business interests and not about the "Sprit of Open Source"

[bing luo]

From 2022 until now, regardless of whether we sell PNP kits or not, we have each donated a fixed amount of $100 to the OpenPNP project.

But we never promote our PNP kits in this discussion group, so we won't go to all the threads about other people's PNP kits to promote like PP.ca. 

When the user fails to debug the PNP of Lumen, PP.ca are not providing meaningful suggestions to  the customer to complete Lumen's PNP , but rather asking the customer to replace Lumen's PNP with Panda's  at  once.

Here  has indeed become a boring  PP.ca hypermarket.

[PP.ca]

This is enough Bing. Stop the lies. You're into defamatory hate speech territory.

[Danh Trinh]

Not hard to imagine that a post titled ".... Panda Placer vs Bing 3" would result in some heated arguments. I have a second hand Lite Placer and welcome other affordable alternatives and comparisons.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/23e0116c-07dc-43f0-bf1b-8455ea30f4a2n%40googlegroups.com.

[bing luo]

I don't oppose any user reviews, both good and bad are normal. I just don't want this place to become a PP.ca hypermarket.

[Jason von Nieda]

Let’s keep the thread on topic please. Comparing commercial offerings is fine, bickering between companies is not. No more messages on that, please. If someone believes there is an issue that requires moderation, please email me directly at ja...@vonnieda.org. 

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ea7db56e-5c08-46b3-be73-ed929d972946n%40googlegroups.com.