Cheapest reliable feeders.

[TheCunningFellow]

What is the cheapest reliable feeder for 8mm tapes?

What are the numbers to beat?

Are any of the homebrew laser cut or 3D printed feeders acceptable or are knock of Yamaha CL8 at $70 a pop the most economical way to go?

If you could make a reliable feeder for $160 for 10 lanes ($16 per lane) would that be an achievement or is the bar already higher than that?

[Trampas Stern]

The cheapest and most reliable are the used CL8 Yamaha feeders off ebay, I can often get them for $20 in lots.  I got ~20 feeders this way about a year ago, you just have to wait to find the deal. I also found a vendor in China that was selling used feeders for around $20 as well, I think it was the same guy selling the used plates, so contact him and see:

 

https://www.aliexpress.com/item/YAMAHA-mounting-block-for-feeders-40-port-unit/2055212235.html?spm=2114.search0104.3.1.77335cacwXn7ld&ws_ab_test=searchweb0_0,searchweb201602_4_10152_5722917_10151_10065_10344_10068_10130_10324_10342_10547_10325_10343_5722817_10546_10340_10548_10341_5722617_10545_10696_10084_10083_10618_10307_5722717_10059_100031_10103_5722517_10624_10623_10622_10621_10620,searchweb201603_11,ppcSwitch_5&algo_expid=3f54b655-36af-458b-b839-417e7bc6170b-0&algo_pvid=3f54b655-36af-458b-b839-417e7bc6170b&transAbTest=ae803_1&priceBeautifyAB=0

They do require a plate and air pressure to operate. Robotdigg sells a mounting plate: https://www.robotdigg.com/product/1190/Pick-and-place-machine-Yamaha-Feeder-mounting-block.  

The air solenoids can be gotten of aliexpress for around $5-$10 each. 

Note the used feeder plates are great as they come with solenoids, but I find they will need new o-rings. 

If you want I can provide a bare PCB for 16 channel solenoid driver, http://misfittech.net/16-channel-fet-pwm-driver/  for $5. 

This puts the total initial cost per feeder around $40 per feeder.  However for this you get very reliable feeders that just work. 

I have looked at designing and building feeders, there is nothing that I could design and build that could compete with Yamaha feeders, even at $100 each. That is the Yamaha feeders just work, and my time is not free so the cost of good feeders verse my time makes the Yamaha feeders the best deal in my opinion.  It is much cheaper than one miss feed on a panel... 

Trampas

[TheCunningFellow]

Trampas,

Thank you.  I have been keeping an eye on eBay for cheap CL8 feeders for some time, but have not found any cheaper (who will ship too me) than the all day every day $60ish price for knock offs on ali.

I am not in a super hurry yet and am getting by with cut tapes.

So I may try design my own feeder and see if I can't get the price down to <$10 per lane and still be reliable.

[TheCunningFellow]

Oh - and I guess the other thing about self made designs of feeders is they don't need compressed air like the CL8

[Marek T.]

Don't think you can build self made <10 with the same speed and accuracy as CL8 have. To get similar parameters you need 2 or at least one flat stepper (which are nor cheap) if you don't want have it air pressured.
Sure you can make something below 10 but it's not to compare with CL8. Btw you can try find Chinese clones of CL8 for near $30.

[Jason Rush]

Is a CL style feeder a standard? I see different brands for sale using the CL term. Does this mean that an 8x4mm CL from Yamaha is compatible with an 8x4mm CL from Philips? Also, what is the meaning of a black handle feeder vs a green handle?

[Trampas Stern]

I did some calculations and designs on feeders and to get below $30 for a good feeder I found that I would need to build in large volumes.  That is I would need to build 1000s of feeders, and spend lots of time designing testing and debugging.  .  I did consider doing this but found that the market would be the home built PNP folks who would rather spend their time trying to build something than buy something that just works.  

The only cheaper alternative to the CL8 feeders was simple drag feeders with weights for the cover tape.  For small runs this is the cheapest/easiest thing you can do, however it is slow and can be error prone. It also requires lots of time to move weights and adjust things.  However many people seem to do this and it works for them.  

Personally I would do the drag feeder of cut strips as that once the cover tape is removed from parts, you should write off the parts. So if you remove the cover on 100 parts and use 10, most likely by the time you try to use the other 90 you will have bumped machine and parts will be everywhere.  

[Trampas Stern]

I think they are standard, the different handles are different design generations.  The ones I got are old (metal handles) and they work fine.  I got a bunch of 8x4mm and cursed as that I went to 0402 and found that I need 8x2mm for 0402 (this is one of the reasons the 8x4mm are cheap used). If I was doing again I would buy 8x2mm and just pulse twice for the 0603 and larger parts. 

[Cri S]

Philips GEM use the same CL feeders as Yamaha, (it's a rebranded yamaha).
green handle is normally 0201, blue is 0402 and black is 4mm pitch.
green and blue is 2mm pitch.
white handle is >16mm tape width. Hovewer there are electrical feeders
> 16mm width that
have black handle,
Tape depht differ, it is not granted that on blue/green handle you can
place 0603 and
push twice.

Here in europe, the cheapest electrical feeders are 19€ in quantity 8
(bulk feeder of
8x8mm).

> --
> You received this message because you are subscribed to the Google Groups
> "OpenPnP" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/bf42a36d-8d6a-42b0-be45-384254933b81%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

[Trampas Stern]

Thanks that is good to know... 

Do you have a link to the electric feeders? 

Trampas

[Cri S]

Here a picture.

[Marek T.]

Cri Is it electric or drag?

[Cri S]

Electric, using electric motor and solenoids.

2018-06-13 15:20 GMT+02:00, Marek T. <marek.tw...@gmail.com>:
> Cri Is it electric or drag?
>

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

> https://groups.google.com/d/msgid/openpnp/39e88d88-d402-4961-8ed9-5aad43e1e8d9%40googlegroups.com.

[Mike Menci]

Cri -is there a web page - where to get more info?
Thanks
Mike

[Marek T.]

So it costs 8x€19 and only 8mm tapes?
Is there a shutter and cover tape roller?

[phon...@gmail.com]

Not only 8mm, but price increase if you count per lane.1x32mm + 2x16mm+1x24mm means, cost is not 19 but 39€ / lane.

No shutter, cover taper roller..

[TheCunningFellow]

I don't think my design is going to compete with the pneumatic feeders on speed.

Just as long as they can get to the next part by the time my slow PNP machine needs it.  It is not like I have a gatling gun like chip shooter head and can place 100,000 parts per hour.

Reliability is my main concern I think from what I am reading here.

Plan was to use regular wide (nema14) steppers.  Saw how expensive thin steppers are and decided that I would not make single lane feeders but make a gang of 10 feeders.  The stepper motor then does not have to live in 12mm wide space.

I realise this makes changing parts on lanes much harder - but that does not bother me.  My 10 fixed lanes are going to be for common parts every board uses many of

100nF

1uF

4K7

10K

The things that have 10 or 20 of on each board.

When I need a 9K81 resistor for feed back path on a variable regulator there is only one of them on the board - so a cut tape that can hold 80 parts will be plenty.

My PNP is not doing commercial quantity and that might be a different ball game to what others are playing.

[Marek T.]

So Cri proposal sounds good for you. Or your adventure of self-made play :-).
However consider the thickness of the feeder. You never know how many lanes may need in the future. Wasting place without necessary has no sense.
Btw: if you want make cover tape roller, relatively easier is to use the second motor for this. Then costs go high.

[TheCunningFellow]

Not sure I understand what you mean.  What is Cris proposal.  I only saw he mentioned feeders for 19 Euro but didn't have a link.

If I decide to build my own design then the plan is to have a gang of 10 feeders that is 120mm wide.  12mm per tape is not very wide.

They will not be separate units so that normal thick nema 14 motors can be used instead of expensive thin ones.

I have not decided if I will go with buying CL8x or design my own yet.  Buying CL8x will certainly be easier.

However if I can come up with an elegant design that is cheap to make then that will benefit the OpenPNP community a lot in the long run.

[Trampas Stern]

For an electric design you can move parts forward slowly, else you will need a gate to cover part while moving. Faster feeders have to have gates that cover parts like the CL8 do.  That is f=ma and with fast acceleration even the low mass of a resistor gets lots of forces applied. 

Another thing is that the cover tape should be removed before the pick location. When the cover is removed at the same opening as the pick location it often causes tape to jerk and parts fly off. Hence most commercial designs remove the cover tape before pick location and between cover removal and pick is a shield over the parts.  This way sticky cover tapes do not cause parts to jump out. 

N20 gear motors and SG90 servos have been popular for feeder designs. The DC motors require some encoder, and by the time this is done the cost is over the $20-$100 budget. For stepper designs by the time they stepper driver design is done the cost has exceeded the budget (and people quickly find that they still need encoder on the stepper designs to know where toothed wheel (tape) is). 

No matter what motor design you use the key to all the designs come down to the advancement system. A motor driven system usually uses a toothed wheel to advance tape. The toothed wheel costs is where most fail.  That is the cost to make the toothed wheel/gear in low volumes (less than 100 units) is high (more than a CL8) and that is where most people who have tried give up.  As far as I know no one has found a way to get the toothed gear cheaper, some have even though of making as a FR4 PCB.  The toothed wheel has to be thick to be rigid, but then the teeth need to be tapered on sides to fit in the tape.  Thus simple laser cutting or water jet cutting parts is not good enough as afterwards they need to be post processed to add the taper on the sides. Again in really large volumes you can get costs down it but not in small volumes. So you end up doing the post processing by hand... 

The cover tape system is the next thing people fail on. Here if you have to pull the cover tape off with constant force not constant retraction distance.  So the designs have to have some force based slip system. Some use belts which slip when the force is reached and then have manual belt tension, some use a switch on the cover tape and spring, when the tension is correct the switch cuts off the take up motor.  Some use rubber bands get tension, CL8 uses springs,  the drag feeders often use slip clutch on the take up reel, other people using drag feeders use fishing (lead) weights and gravity.   This cover tape take up is the second hurdle you will face and it can be a pain to get working. Here is a video on the CL8 system: https://www.youtube.com/watch?v=ArWzMyBTSK0

After the advance and cover tape is done the next hurdle is the varying tape thickness. Here some type of spring system is usually needed to accommodate varying part heights. 

Again for slow reliable feeders the cheapest and easiest is the simple head drag feeder.  Here is a video of a head driven drag feeder: https://www.youtube.com/watch?v=vfukCmMRR5s

These are simple and easy and I think OpenPnP supports this.  For slow machines this is by far the easiest and cheapest way to go, I seem to recall that robotdigg sold the plates for this but I might be wrong.  Additionally with a little vision help in OpenPnP this could be super reliable. 

Again I know so much as I have done several designs myself and learned that the head drag feeders are cheapest, and the CL8 is second.  

 

 

 

[Cri S]

You are kidding, the sprocket (toothed wheel) cost is cheap.
The cost is near 30cent but single run is 20€ for 40 Wheels.
Add 5€ shipping cost and then it is 0.625€ / wheel. If you need 10,
and trash the rest or keep it as spare, the cost is 2.5€.
For one single wheel you'r statement is true and no postprocessing is needed.
If you want get it for free, go the the photoshop and ask for the
trasheds oneway use cameras.
It works only for 4mm or larger spaced parts, w8p2 and w8p1 or w4p?
don't work, but probably
you don't use this tape carrier formats.

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

> https://groups.google.com/d/msgid/openpnp/95c8a0f6-bef7-4b70-98d2-77cf7949bd3c%40googlegroups.com.

[Cri S]

Check out the fast hover devis feeders, no shutter at all.
shutter is a sign of cheaper not so precise feeder that have rought motion.

[Marek T.]

Never thought about it this way that advanced feeders don't need shutters...
Maybe you are right but if the feeder need to be really quick then seems to me, maybe I'm wrong, that hard to get it and don't be rough.

[Trampas Stern]

F=ma not matter how you look at it. So to move fast you have to accelerate and decelerate rapidly, if you do a S-curve acceleration and deceleration then you keep part more stable, but it takes longer to move using S-curve than instantaneous acceleration (ie a step in acceleration). 

The CL8 uses air and the air pressure could vary the gate is needed, especially with SOT23 and larger mass parts, that is the acceleration and deceleration is not very controllable and more of step. Hence a gate is required to have the fast change in acceleration. Note the rate of change of acceleration is called jerk. 

Cri, if you have a link to how to a supplier for the toothed wheel please send it. I know I have not found a supplier that can do them as cheap as you have quoted in the US. I know I would love to change the wheel on the CL8 to be 2mm for some of my feeders... 

Thanks

Trampas

 

 

[Marek T.]

I'm not experienced with CL8, have one piece but never used it yet. But I've a lot of earlier feeders of Yamaha with mechanical shifter. And in their case it's opposite than you say: if you dismount the shutter there is not any problems with larger parts. But small parts like 0603 or 0402 you can forget - the fly to every directions around the feeder.

If you have dwg file for the wheel you need - where the problem to go to any company with laser cutting? I've made that wheels in company which cuts for me the laser stencils and it works good. Little problem that they could do it with 0.5mm steel max instead of original 1mm - but can't see problems with this.

[TheCunningFellow]

Thanks again Trampas, Marek and Cri,

These are all the kinds of things I am wishing to hear.

With respect to the cost of steppers.  A small Nema 14 is less than $15 including the cost of driver.

My design which you could think of as an upside down drag feeder only needs 2 stepper motors for advancement of the whole 10 lanes.

I was thinking about a shutter but didn't know if it would be needed or not.  I will try first without, seeing as I can control the acceleration of the stepper very easily.

I think that mechanical parts to hold 10 tapes, two stepper motors and a solenoid to advance those tapes another stepper or a DC motor to pull the cover strips sounds like it should be doable for the $160 budget (excluding labour obv)

Of course this idea would get cheaper per lane the longer the gang of lanes is, but 10 seems to be about what I want.

Its not an idea that is viable for a commercial machine because changing tapes is hard and the speed is not the same as a CL8x, but then a normal drag feeder has that problem also.

[Trampas Stern]

Your labor is not free, even at minimal wage. 

If you did 10 lanes at $160 it would be $16 per lane and would be amazing, if you include labor at $5/hour you will find the CL8 are cheaper. 

Trampas

[TheCunningFellow]

>> if you include labor at $5/hour you will find the CL8 are cheaper. 

I understand this, but the spirit of open source is to make things and give back.

If I do make it and release the design - someone else making it is not going to waste 100s of hours like I have too.  The 2nd person to make one may be more like $160 and 4 hours.

Again - I have not decided CL8x or self design.  I am still in the thinking stage.

[Cri S]

I think everyone here know this.
https://www.youtube.com/watch?v=PpcUNTBp_SA

The strip is feed on wrong direction, feeder lane pitch is 23mm for
8mm carrier tape.
16 lanes need 25cm. Peeling tape work with fish line weigth,
Material cost is 70 euro/$ and time needed is 2 hour and 4-5 hour the
first time.
But you need access to cnc mill.
If doing 42cm as example instead of 25 (40cm moving area).
16x8 + 4x12 `+ 2x16 for a total of 22 lanes is used.
Doing half of that, 11 lanes (8x8mm + 2x12mm + 1x16mm) 22cm with arduino usb.

Do you think that it is worth sell such kit for 50 euro where the peel
tape is removed
using fish lines, maybe adding some tray strip feeder in order to
increase the user value
as add-on.. I was always thinking that there is too less demand for such thing.

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

> https://groups.google.com/d/msgid/openpnp/042827fc-6f11-439c-898f-75b2d25ae225%40googlegroups.com.

[TheCunningFellow]

Yes - I have seen that one.  It is very wide and needs a servo for every lane.

Maybe one stepper motor for tape advance (X axis) and one stepper motor to select lane (Y Axis) can be skinnier and cheaper.

If drag feeder that uses the main PNP head to advance is OK, then surely an independent X/Y+solenoid must also be OK/Reliable.

Anything I make is going to need a CNC machine to produce.  It is 2018.  Almost everyone has either their own CNC machine or access to one.

[ma...@makr.zone]

Here is a video on the CL8 system: https://www.youtube.com/watch?v=ArWzMyBTSK0

Nice video.

I wonder if it were possible to drive the air using a dedicated electric piston instead of having to buy a compressor, valves, plates, o-rings, etc. pp.

_Mark

[Trampas Stern]

I thought about the same thing....  However the plates are $10 per feeder, the valves are around $5 each, a compressor is $40...  So the time and cost to retro fit to electric would end up costing more. 

Back in the 90's I worked as an auto mechanic. I got a Carvan with a blown 4 cylinder engine for a $300.  A friend of mine owned a junk yard and so I pulled a V6 engine and transmission for the van. Required wiring in computer for engine, computer for transmission, new axles, etc.  I got it all working...  Then I went to junk yard and got power windows and seats and put in the van....  I then added power mirrors....  However no matter how much I worked on that van it still was a $300 POS caravan.   I thought about it and realized if I had purchased a Mercedes with problems for $1000 and then put the same time and money into the Mercedes I would have something more valuable when done.   Worse still was because it was a Frankenstein monster, no one else in the world could work on it and I could never sell it, which meant I had to drive my time and money out of the thing, which was great as it kept the pain going longer to enforce the lesson.  

The feeders are much the same way, if you convert to electric then one day in the future when you need a 12mm feeder and buy one, then you will have to spend time converting it.  If you just take the time to use air and set it up the way they intended, you would just need to plug in the new feeder. 

I am a cheap person at heart, and when I get the feeling that I could save some money, I think about my caravan story and ask myself would I really save time and money or would I be just trying to turn a caravan into a Mercedes. 

I spent more money on my home built PnP than I did buying a used Manncorp, not including the time... So I still fall into that same trap.... 

[ma...@makr.zone]

Thank you Trampas

A very nice story and lesson. Yes, yes, yes.

Has anyone ever really built a working CL feeder bank for OpenPNP?

BOM?

Fotos?

Videos?

I've seen CL feeders mentioned many time, but I have never actually seen a finished DIY CL feeder bank working with OpenPNP.

I've seen some parts on RobotDigg but I can't glean enough information from it to find the rest of the parts needed (or if those two even match...)

https://www.robotdigg.com/product/1190/Pick-and-place-machine-Yamaha-Feeder-mounting-block

https://www.robotdigg.com/product/1330/Pick-and-place-operating-valve

...and????

The threads here and elsewhere are full of unsolved problems right down to obscure imperial tube diameters etc. Sadly, people often don't take the time to report back once they've solved their problem. Or they never actually solve it...?

Also if I buy a compressor, I would like to create the vacuum with the compressed air. Apparently this can be done with less noise. Anybody know how?

_Mark

[Marek T.]

Search "micro ejector ME05-E1". It creates -100kPa vacuum from 5bar of pressure.
Have this on my machine, one pc for each head (nozzle).

CL8 plate build Berndt here (if finished) using robotdigg plate. I've started build the plate on cnc but not finished yet, have plate already but no valves. Some problem is the kind of collector to mount all the valves and don't use meters of pipes and connectors (and to save the place).

[ma...@makr.zone]

Thank you Marek for the tips.

Using your info I found this:

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

Bernd Walter thankfully even added a BOM in the description (reproduced here):

Parts are 8mm apart, so the feeder has to be triggered twice.
The feeder can be configured to feed 4, 8 and 12mm in one action by screwing the pneumatic actor into different locations by using the alignment plate - I didn't know that when I took the video.

As people keep asking:
The mounting plate is using the following parts and valves from Robotdigg:
TG23-06 in 24V
Y-ring6921  6*9*2.1mm
PC4-A2  PC4-01
PUTube-4*2.5mm_1.8
FMB30

I also needed some short M3x6 flat head countersink screws on the mounting plate as it needs 31 screws, but only came with 2 or 3.

Some valves are similar types from another supplier, which turned out less reliable.
The pneumatic tube adapters can be bought from other suppliers, the ones from robotdigg may break when screwing in with the loctite, which others didn't.
You also need those adapters on the valves.
The pneumatic adapters need to be screwed in with LOCTITE 55 or something similar.
The sealing applied won't work in the threads of the mounting plate.
The feeders need 6-8 bar pressure, low pressure airbrush compressors don't work.
A high air volume isn't required, but a silent running compressor with a big air tank is a nice thing to have.

[TheCunningFellow]

So mostly complete cost of parts for a 10 lane Yamaha CL setup from robotdigg is

x10 $5.40 TG23-06 in 24V

x10 $0.20 Y-ring6921 6*9*2.1mm

x30 $0.15 PC4-A2 PC4-01

x1 $1.80 PUTube-4*2.5mm_1.8

x1 $89.00 FMB10

x10 $8.90 CL8FC

x10 $69.00 CL84

$930 or $93 per lane

Plus a couple of hundred for a compressor

I imagine you could better this by 20% on aliexpress by hunting around, but as people keep making the argument here - what is your time worth trawling aliexpress to save $10 here and there.

And we hear the stories of people getting used Yamaha feeders on eBay for $20 a pop. Again your time to find them (if you ever do) and in my case the cheap ones you do find won't ship to you.

If you are going to accept the possibility of finding cheap ones on eBay as the yardstick to measure what something new could be worth then we all should be driving Lotus Esprit because they come up on ebay for $1

http://news.bbc.co.uk/2/hi/uk_news/england/west_midlands/4122842.stm

I Still think there is merit in trying to make something usable for the openPNP community that is cheaper, easy to make and an adequate match for the most common slow 20x20+Hiwin+Smoothie kind of setup.

Now if I decide to go it alone and try design something then

A, I will be having fun designing something

B, I Won't care if doing the first unit actually costs me more than $930

As long as the design I come up with is useful and subsequent units can be built for around the target price.

I'm not saying Yamaha engineers are stupid and I can design something better and cheaper.

What I want to design has a different goal.

Yamaha feeders are modular and easy to swap out lanes - Mine wont be modular

Yamaha feeders can feed a part very quickly - Mine wont be fast

Yamaha feeders can run 24/7 for years before wearing out - Mine will wear out faster

My design goals are

Cheap

Reliable feeding

Neatly handle cover tape some way

Electrical (not pneumatic)

Low width (12 to 14mm per lane)

Low profile (40 to 50mm high would be nice)

Speed only needs to be faster than the slow PNP head

What I don't care about

Modularity / Ease of changing tapes

Feed speed of 100mS

Service life and 24/7 duty cycle

I THINK I can achieve this for $16 per lane with a setup that has

Two stepper motors in a Cartesian X/Y format under a drag feeder plate.

A Solenoid for the drag pin

A 3rd motor (stepper or DC) for tape take up