NeoDen YY1 Pick And Place Machine With Under $3K Price for Hobbiest/Low vol Usag

[Smartbeedesigns]

So I've made about 50 small boards over the weekend with mine.

Some thoughts and questions for those who already have a YY1 in use.

Question:
Has anyone else had inconsistency with the placement of components? I'm placing a panel of boards, using a fiducial and component for alignment. And from one run to the next the placement of the parts varies. One run everything is more or less correct. Next run they'll all be shifted in one direction. Next run, passives will be right, but larger components will be shifted.

I've tried every combination of fiducial and component alignment with mixed results.

Thoughts:
 The tape peeling mechanism is really finicky.
 If the friction of the idler pulley isn't set just right then the tape peeling can fail and cause miss picks.
Paper tape has had almost zero issue with the film peeling. But plastic tapes is hit and miss.
Tape with film that has a sticky surface can get wrapped up into the peeling mechanism and cause issues.

Basically the paper tape components I can walk away from the machine and they'll be fine. But any of the plastic tape reels I have to baby sit.

I'm getting aboutt 1150cph real world use. That's with a tool change and picking components from both sides of the machine.

Overall I'm happy with it, it's still way faster to PnP than it is to hand assemble these boards even with the constant baby sitting the machine. And if I can get the board alignment to be consistent between runs then I'll be a happy camper.

[drgerber]

Has anyone else had inconsistency with the placement of components?

Interesting. No, haven't see such issues yet. Sometimes single 0402 resistors are a little off but nothing that really worries me so far.
0402 caps and larger passives are placed perfectly usually.

But plastic tapes is hit and miss.

For me it's almost the opposite. Haven't had an issue with 8 mm plastic tape yet. On a few feeders the paper tape was slipping through when the peeler was working. Adjusted the metal spring in the feeder a little to compensate for that.

1150cph

Roughly the same here in a run with 260 components per board / vision enabled / all 8 mm feeder used. Set the machine to 50 % speed only. Haven't tried anything faster yet.

Overall I'm happy with it, it's still way faster to PnP than it is to hand assemble

Yes, I'm also pretty happy with the machine. Setup is easy and for the low price stunningly usable. Perfect for prototypes and small runs. The early days with the NeoDen4 were a lot more troublesome as far as I remember.

[Smartbeedesigns]

Has anyone else had inconsistency with the placement of components?

Interesting. No, haven't see such issues yet. Sometimes single 0402 resistors are a little off but nothing that really worries me so far.
0402 caps and larger passives are placed perfectly usually.

Hmmm. Did you have to do anything special with the fiducials or camera threshold to get consistent board pick ups?  As it is right now, I can pick up the board and everything looks ok, then the next run it may be off slightly. I'm wondering on my setup if maybe it's not picking up the fiducial super accurately every time.

1150cph

Roughly the same here in a run with 260 components per board / vision enabled / all 8 mm feeder used. Set the machine to 50 % speed only. Haven't tried anything faster yet.

Ive been running it at the default speed of 70% and when it's picked up the board correctly it does fine. But maybe I need to slow it down a little bit.

[newto]

Something I've noticed with the peeler wheels is that they can get "stuck" if left for a few days. It's easy the break the friction by manually turning it back just to break the stickiness, and then it runs fine. Some tapes don't seem to matter, but if you've loosened the metal thing a bit, it can unpeel some tape and screw up the pin placement.

I'll probably add "turn all the wheels back before operating" to the procedure.

Question:
Has anyone else had inconsistency with the placement of components? I'm placing a panel of boards, using a fiducial and component for alignment. And from one run to the next the placement of the parts varies. One run everything is more or less correct. Next run they'll all be shifted in one direction. Next run, passives will be right, but larger components will be shifted.

Check if your belts are loose, and if any of the moves crash the head somewhere (watch the nozzle changes especially). I managed to run one of the nozzles into another nozzle in the holder doing some testing, setup a couple feeders, and then when I restarted, all my pick positions were off because the machine thought the head was slightly further to the left after missing steps during the crash.

[drgerber]

Hmmm. Did you have to do anything special with the fiducials or camera threshold to get consistent board pick ups?  As it is right now, I can pick up the board and everything looks ok, then the next run it may be off slightly. I'm wondering on my setup if maybe it's not picking up the fiducial super accurately every time.

No, not that I remember. What I do for every batch is to set the datum coordinates in the parameters menu to compensate for the tooling frame.

all my pick positions were off because the machine thought the head was slightly further to the left after missing steps during the crash

yes, this also happened once to me when a tape blocked the feeder. Unfortunately an open loop system.

[alpelectronics]

Hi,

I have also received my machine. First of all I have to say that, Im amazed for such an affordable machine. The built quality is very nice.

I havent had time to assemble any boards. But at least I tried the demo board with 0402 caps. At first I had alignment and offset issues. Then I calibrated nozzles and cameras, I also changed to a more suitable nozzle, it was perfect placement at 70% speed.

I also noticed one strange problem, there was a sunshine inside. It was exactly coming to the X-axis limit switch. When I turned on the machine, it couldnt find the home, also hit the end. Later, I realized that the sensor is affected by the light!

I believe this hardware is good, but SW needs a little tricks although it is simple. Some mods would be needed for some parts.

BTW the touch screen HMI is from DWIN. It is a dual core device I believe. I had the datasheet of their IC.

Regards,
Caner

[liviux]

How did you imported the file ? It gives me an file error was something special done to CAD csv ?

[Smartbeedesigns]

How did you imported the file ? It gives me an file error was something special done to CAD csv ?

I just made a test file on the machine itself then take out the SD card and import the information from your CAD program that spits out the positional data.

The yy1 file format has to be formatted specifically for the machine to understand it. So just reference one of the files the machine creates itself.

[liviux]

Yeah did that too. It works First ever test looks promising. I needed to adjust the origin of the machine to be able to see the parts exactly right. placement seems on point too.

[liviux]

Something I've noticed with the peeler wheels is that they can get "stuck" if left for a few days. It's easy the break the friction by manually turning it back just to break the stickiness, and then it runs fine. Some tapes don't seem to matter, but if you've loosened the metal thing a bit, it can unpeel some tape and screw up the pin placement.

I'll probably add "turn all the wheels back before operating" to the procedure.

Man, this happened to me too!!! Plus I have some 0402 resistors with thinner tape that might need 1.5nM rings on them. Fixed that on 0402 LED's too. 0402 caps seem to have no issues on 2.5nm cause their tapes are thicker.

[Smartbeedesigns]

Here is a short video review I did of my neoden yy1 after making about 400 small esp32 boards (40 panels). What I like and dislike about the machine.

https://youtu.be/7w5U1qxXqKI

[fourfathom]

Thanks for making that video!  Re. the sticky peeler tape, could you just clamp an alligator clip or clothes-pin to the tape and let gravity work for you?  You could re-position the clip after a foot of tape has peeled, or lead the tape over the edge of the table for a longer drop.  Not a perfect solution, I will admit...

[Smartbeedesigns]

Thanks for making that video!  Re. the sticky peeler tape, could you just clamp an alligator clip or clothes-pin to the tape and let gravity work for you?  You could re-position the clip after a foot of tape has peeled, or lead the tape over the edge of the table for a longer drop.  Not a perfect solution, I will admit...

Yeah I tried a few different attachment options to the tape itself to hold it down with, but didn't have a ton of luck. I did wrap a bolt into the end of the film and that seemed to work, but was less than ideal as I would have to adjust it all the time as the tape got longer.

Ideally I'd like to come up with something that is pretty much hands off and doesn't need a ton of playing around with. I was hoping the 3D printed cover would work but it still manages to stick to the top wheel and get sucked up under the cover anyways.

[eflyersteve]

Sorry for being so late getting back to anyone.  My findings on this machine:

Plastic tape parts - just forget about them.  The cover tape either pulls the carrier tape through and you dump parts or the pull needle gets stuck and when it pulls out of the hole the tape shakes and parts are lost. There are some adjustments that can help with the carrier tape pull, but I don't know any way to fix the issue with the needle getting stuck and disrupting parts.

Placement head accuracy - I have one particular component that I cannot use the camera for alignment. That's a whole other story, but suffice it to say that there's enough give in the design that I really don't have to use any optical alignment. Just pick it out of the tape and rotate it 180° and place it. Problem is that occasionally it will pick up the part and then just rotate it about 15° and then it will place it on the board with this angle error.

Up camera settings - there does not seem to be any setting that I can adjust that give us the machine the ability to place all components. For example, I'm placing a 3535 LED on a board and If I adjust the up looking camera so that is consistent with recognizing the outline of this part, it will not recognize the outline of QFPs. The camera settings need to be adjustable per component type but they aren't.

Developed a squeak in the placement head - I haven't quite figured out it if it is the stepper motor or perhaps one of the bearings. 

Neoden does seem like they are interested in the issues, but I'm really not getting much other than requests for videos and a couple of very basic suggestions.  The machine really does need some software fixes as well as some issues with a feeders resolved. 

It's a pity because the machine looks really nice, especially compared to my TVM 802. However, the TVM 802 just works.  My intent was to purchase several of these YY1 picking place machines, assigning each one to just a few circuit boards from an individual customer. Customer. If a customer needed additional boards built, I would just add an additional YY1.  This would give me redundancy in the event of a machine failure, would handle the low volume manufacturing that I have going on and also be aesthetically pleasing when a customer wanted to visit the facility. 

[Smartbeedesigns]

Plastic tape parts - just forget about them.  The cover tape either pulls the carrier tape through and you dump parts or the pull needle gets stuck and when it pulls out of the hole the tape shakes and parts are lost. There are some adjustments that can help with the carrier tape pull, but I don't know any way to fix the issue with the needle getting stuck and disrupting parts.

if its pulling through tape that it shouldn't be, like if you pick parts on feeder 1 and its slowly pulling the tape forward in feeder 2, there are metal tabs in each feeder that you can adjust to apply more pressure upward to hold the tape firmer to the ceiling of the feeder tray so it wont slide forward when other tapes are being picked. hope that makes sense.

Up camera settings - there does not seem to be any setting that I can adjust that give us the machine the ability to place all components. For example, I'm placing a 3535 LED on a board and If I adjust the up looking camera so that is consistent with recognizing the outline of this part, it will not recognize the outline of QFPs. The camera settings need to be adjustable per component type but they aren't.

im pretty sure the up facing camera threshold can be adjusted per component. its under each of the feeder settings menu.

[EEVblog]

At this price point with 40+ feeders I'm seriously tempted to get one and try it...
How hard would it be to replace just one reel?, as the mounting rod goes through all the reels.

[EEVblog]

Here is a short video review I did of my neoden yy1 after making about 400 small esp32 boards (40 panels). What I like and dislike about the machine.
https://youtu.be/7w5U1qxXqKI

One fiducial only? Why 

[nctnico]

Here is a short video review I did of my neoden yy1 after making about 400 small esp32 boards (40 panels). What I like and dislike about the machine.
https://youtu.be/7w5U1qxXqKI

One fiducial only? Why 

My assumption: If it uses optical adjustment for each component, then it would only need only one reference point to roughly know where the components are supposed to go. But it could be it doesn't. I didn't watch the video entirely so I might have skipped over that detail. What does put me off though is that the machine is quite tedious to use. If I ever wanted to setup my own assembly line, I'd probably go for a machine that needs less baby sitting / fiddling.

[EEVblog]

If I ever wanted to setup my own assembly line, I'd probably go for a machine that needs less baby sitting / fiddling.

That would rule out practically every open source DIY one and sub $5k commercial ones?

[nctnico]

If I ever wanted to setup my own assembly line, I'd probably go for a machine that needs less baby sitting / fiddling.

That would rule out practically every open source DIY one and sub $5k commercial ones?

I didn't research that far -yet-. I strongly prefer to outsource assembly (and for several designs there is no way around that due to complexity) but for some simple & cost sensitive products my go-to assembler is rather busy so lead times are unacceptably long at this moment (to put it mildly). So I'm looking at alternatives including doing assembly myself but really want to avoid turning that into a project in itself. I already have some solder paste application gear sitting on a shelve. Not sure I would invest in a pick & place as assemblers get pretty good discounts from component distributors. But I think I'm going to see what JLCPCB, Seeed, etc can do prise-wise first.

[Smartbeedesigns]

At this price point with 40+ feeders I'm seriously tempted to get one and try it...
How hard would it be to replace just one reel?, as the mounting rod goes through all the reels.

There are actually little guides one the bottom/back/top of that opening that kind of hold the reels up and in place. So you can pull that rod out without all the reels falling over/out. It's a little delicate, but I've added reels after having several already in there and haven't had any issues.

[Smartbeedesigns]

Here is a short video review I did of my neoden yy1 after making about 400 small esp32 boards (40 panels). What I like and dislike about the machine.
https://youtu.be/7w5U1qxXqKI

One fiducial only? Why 

My assumption: If it uses optical adjustment for each component, then it would only need only one reference point to roughly know where the components are supposed to go. But it could be it doesn't. I didn't watch the video entirely so I might have skipped over that detail. What does put me off though is that the machine is quite tedious to use. If I ever wanted to setup my own assembly line, I'd probably go for a machine that needs less baby sitting / fiddling.

It doesn't have any vision adjustment like you speak of.

Technically it uses a PCB origin point for X/Y (bottom right corner of your board) and then one fiducial and one component. That is how I believe they basically triangulate the board for alignment. It's not a perfect system, but it works well enough in my use cases.

[mikeselectricstuff]

What exactly fo you mean by one component? Does it try to vision pasted pads?

[EEVblog]

Technically it uses a PCB origin point for X/Y (bottom right corner of your board) and then one fiducial and one component. That is how I believe they basically triangulate the board for alignment. It's not a perfect system, but it works well enough in my use cases.

That sounds crazy. If you can do one fiducial then you can do two. Board corners and components are a very poor way to do alignment, that's what fiducials are for.

[aholtzma]

It is a very simple (mostly) open loop system. At the start of a board it looks at the fiducial and corrects for board misalignment. After that it's just plopping down parts at their corrected x/y co-ordinates. It doesn't know anything about what the board or pads look like. If the head gets jammed on anything and loses track of where it is, you need to recalibrate the upward looking camera so it knows where the nozzle is. Sometimes you also need to recalibrate the feeder positions as well, which is a pain.

That said, once it you get it in a happy place, it will happily place thousands of parts without issue. We bought it to take the drudgery out of putting together prototypes and so far it has done that well.