Feeders

[thatlittlefrog]

What's a good feeder type for openPNP? I'm considering making clones of the lumenPNP photon feeders, Yamaha pneumatic CL feeders, or these https://www.deltaprintr.com/product/rapid-feeder/ rapid feeders, whenever they come out. I'm also considering creating my own custom feeder with closed loop tape control using the same sort of optical flow sensor found in computer mice. If that ends up being the case, are there any motors better than those somewhat expensive N20 gearmotors? I'm considering using FOC brushless motors for their high torque and speed, but that might be overkill for a simple feeder. I would like to be able to place 0201 and maybe 01005 components reliably.

[bert shivaan]

There are lots of better motors then the N20's, but they cost much more.

Maker Marks feeders work well I believe - no motors required.

There is the 0802(?) feeder that has been around a good while, I think it is still one of the top DIY.

Rapid feeder may or may not be released. I have not heard anything since last year maybe about them.

I use used siemens feeders - pretty cheap for what you get along with an open source controller floating around here.

If you are going to make your own, I HIGHly recommend reading EVERY "I am making a feeder" thread here. Once you do if you are still planning to do it, I let us know so I can get the popcorn and watch. I am not saying you can't do it, but so far I don't recall many making it to the end - at least they did not continue to post about them.

On Mon, Apr 7, 2025 at 12:02 PM thatlittlefrog <susug...@gmail.com> wrote:

What's a good feeder type for openPNP? I'm considering making clones of the lumenPNP photon feeders, Yamaha pneumatic CL feeders, or these https://www.deltaprintr.com/product/rapid-feeder/ rapid feeders, whenever they come out. I'm also considering creating my own custom feeder with closed loop tape control using the same sort of optical flow sensor found in computer mice. If that ends up being the case, are there any motors better than those somewhat expensive N20 gearmotors? I'm considering using FOC brushless motors for their high torque and speed, but that might be overkill for a simple feeder. I would like to be able to place 0201 and maybe 01005 components reliably.

--
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 visit https://groups.google.com/d/msgid/openpnp/0f475886-368a-47f6-a200-0cc3afe75255n%40googlegroups.com.

[Jarosław Karwik]

I have these

https://www.microsmt.com.cn/blogs/microsmt-pnpv3-bing-feeder-system/microsmt-pnpv3-bing-feeder-system

Nice ones up to 0402 - robust design and long term stability, although I have still need more real live tests for them to be sure.

0201 and 01005 - I would like to see both DIY machine and  feeder which can take them....

[Marco Blancato]

Starting from Bing Luo’s MicroSMT project and the components I bought from him for just a few dollars, I decided to redesign the feeder in my own way, introducing some optimizations. But I have to say I’m experiencing poor positioning repeatability (while this issue doesn’t seem to exist in Bing’s original design). Sometimes happens an 0805 is misaligned by half its body, or a 0603 ends up completely out of place. I haven’t been able to identify the exact cause, but in any case, this problem is fully resolved by using the BambooFeeder in OpenPnP, which recognizes the tape holes.

Another critical point is the lightness of the feeder and the structure of a DIY machine itself, which makes it easier for small bumps or vibrations to cause components to pop out of the tape — especially with plastic tapes, which are thin and very elastic. I believe this issue is common to all feeders that don’t have a shutter to keep the components protected until just before pick-up.

  I also purchased a couple of those T-2 feeders to see how they perform. Sixty dollars for a 3-slot 8mm feeder is quite an interesting price, considering that printing all the parts and assembling DIY feeders like mine or the MicroSMT one at home is a time-consuming and demanding task.  

[bing l]

 this problem is fully resolved by using the BambooFeeder in OpenPnP, which recognizes the tape holes

>>>>>>>>> Bamboofeeder adds visual recognition  tape hole  on the basis of automaticfeeder.  The call commands for the two types of feeders are exactly the same, so Bing feeders can also use visual localization to improve the success rate of capturing components, especially for micro components like 0402,    Even if the assembly results of some one  cannot achieve the accuracy in my video, they can still work perfect.   The biggest drawback of it compared to Yamaha and Siemens feeders is the lack of a shutter, which cannot be solved.   But it is definitely practical, you don't need to configure a Siemens feeder with a unit price of $200 for a PNP worth less than $1000.

Unfortunately, it cannot have a more universal name, In fact, it can be valuable for all automatic feeders.    

 After openpnp provided the visual positioning function of the automatic feeder, the sprocket feeder can become a standard configuration for DIY, with a cost of only $10. The metal sprocket will not fail due to wear and tear.  nor will it fail on thin black tape (e.g. LEDs) as mentioned  by  Output. I know the reason for the failure, the tape is too thin, so the push rod often slide over the hole. I often encounter this problem on the 0816 feeder.

If anyone wants to make their own Bing feeder, please email me and I will send  the BOM, PCBA, and  the source code to your email.   

[Mike Menci]

@Marco, 

you are missing this part in your feeder - which holds and guides the tape UP & ALL THE WAY in the feeder....  
Almost all feeders have spring under the slider under the tape to keep the tape at the same hight regardless of tape thickness... 

Mike

[Marco Blancato]

@mike

Thanks for the suggestion, it's something I hadn't thought of. That's why I use different tape covers depending on the type of tape, so that the tape has no play. Of course, a spring-loaded mechanism would be more practical and functional, and I'll definitely keep it in mind for a future redesign. I can't clearly see from the photo how you implemented it—do you have a more detailed picture?

[Mike Menci]

Link to the part did not work properly - see image 15 of 36 for reference

[thatlittlefrog]

I've decided on my feeder design - I'll be going for a solenoid driven sprocket feeder. Wish me luck, I'll update on design.

[bing l]

Small soleoid does not have enough torque to drive the sprocket unless you use a large soleoid,  Your feeder will be very wide.

[thatlittlefrog]

I'll wind some thin rectangular coils and test the strength - If that doesn't work out, I might consider one of those small stepper+screw linear actuators. I have a few on hand and they're stronger than you'd expect, but maybe not enough for a feeder.

[thatlittlefrog]

I have a new feeder idea that might be interesting - You can use one of these stepper motors with an angled piece sort of like a ratchet tooth to push the tape forward when the stepper moves forward, and the friction of the tape will make sure it doesn't slide back when the stepper returns to push the next component forward. Paired with a wheel encoder to track position, this would both be more reliable than those N20 type feeders (no gears!) as well as being more precise by the nature of the stepper motor. I'm also thinking you could fit say, 10 0402 components within the stroke, which means that the stepper only has to return to it's starting position every 10 components, so a multi-head PNP machine could load up very very quickly on a single feeder, unlike the N20 feeders which take a significant time to feed a single component. The stepper motor has almost zero torque when rotating, but the screws are very fine pitch and datasheets claim up to 400 grams pulling force.

Other advantages I can think of:
Cheap - These steppers can be found for a dollar or two on Aliexpress

Software settable pitch - Unlike 0816 style feeders, you can change the pitch without hardware modification

Ultra thin - I believe with my current manufacturing and design skills I can make these feeders 9mm or even 8.5mm for an 8mm tape, as these steppers can be found in down to 6mm diameter versions, albeit those are very, very weak even with the screws.

Compact - These don't need any gears or ratchets, I believe I could design one to be 20mm tall - This means you can even fit extra feeders within the staging plate of a small pnp.

[serge a]

good luck 

--

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 visit https://groups.google.com/d/msgid/openpnp/8550e48c-e9d8-485b-be51-2a153421e2f7n%40googlegroups.com.

[bing l]

It is not usable because its movement speed is too slow, and it takes 10 seconds or even more to advance 4mm.    If its movement speed is fast, its force will be too small. 

[Jonathan Crossley]

An open forum isn't always a good place to share initial ideas, lots of negative thoughts (they are thoughts, not based on fact) and not much encouragement which is why I rarely do. So i'd say go ahead and start experimenting with the linear stepper actuator - I started looking at these a few months ago with a view to feedeer design and initial results were promising. I ordered about 8 different models from AliExpress but the one below was my preference for torque and build quality.

Features:

8mm+ stroke

0.5A running current under 'load' - can't push it much more than that due to overheating

It's a stepper so technically you shouldn't need an encoder, just a limit switch

[Mike Menci]

If use this stepper for tape strips only it might work - sooner or later you will need to turn as well the Tape Reel with Loaded tape - this stepper will not be able to take such load- job.... 

[thatlittlefrog]

Daedal - the reason I would use an encoder is for fully closed loop operation.

[Tony Huynh]

I am also doing a project to create a high precision aluminum Feeder, with the same structure as the Yamaha Feeder. After going through many types of motors, I found the N20 quite weak and it is not suitable, I am using the N050 with a pulling force of up to 0.2Nm which is quite good

[Jim Drew]

@Marco, did you ever release the CAD files for this design?

[Marco Blancato]

Hi Jim, I believe I already replied to you privately, but I'm copying and pasting it here for everyone: I can absolutely share the Fusion360 project files with you, but to be completely honest, I wouldn’t recommend replicating it.
I’ve printed and built around 20 feeders, but they all have major issues with the precise positioning of the component when the tape advances — and that causes a lot of frustration.

With cardboard tapes and components down to 0805, the errors are manageable and tolerable. But with plastic tapes or 0603s (let alone 0402s), it’s a nightmare.

So at the moment, I only use these feeders for 0805 resistors (with some cursing), and I handle everything else with simple StripFeeders or even manually.

Anyway, this is the link to download the Fusion360 project frome OneDrive,  I hope everything is visible, as there are some linked components and I have no idea if they have been embedded in the project

[Jim Drew]

Hmmm... I don't see a private response from you, but I appreciate you posting the info!  I am working on a feeder design that will provide precise positioning, but Bing has pointed out that there is some option in OpenPNP that can 'find' a part on the feeder now.  Because of this, he has eliminated the 2mm step version of his sprocket (only has 4mm now) because the vision system can find parts that are 2mm off.  Maybe you can look at Bing's info on this because apparently it was a game changer for accuracy.  I am a newbie when it comes to OpenPNP, but not for mechanical/electrical engineering.  So, I believe I can greatly improve the accuracy and simplify the entire design.

Thanks!

[Jim Drew]

@Marco, everything is there except for the peeler gear cover (rear of the feeder) and the rail pieces that hold the feeders.

Can you describe what positioning issues you had with your design?  Does the tape advance too far or not enough on occasion (or consistently)?  Since this basically uses Bing's gear setup (but with a different idler gear size) I would not expect a gearing issue to be a problem. I also do not think that rotating the beam break sensor 90 degrees from how Bing designed his unit would be a factor either unless the hub in the spur gear was out of round.