Component Tracking - loading components etc

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Rochey

Well-known member
White Market Member
Joined
Jul 2, 2004
Messages
2,697
Location
Dallas, TX - Formerly UK
Gents (and Ladies, if there are any on this forum).

I come to you with a problem. I am fortunate enough to have a pick and place machine that picks up tiny rice sized parts and puts them on PCBs. It's has space for 48 reels of parts (things like 10KOhm resistors, 470pF Ceramic Caps, Surface mount LED's etc).

When I program the locations for components, i have to tell it which slot to pick up the part from. (e.g. pick up component on slot 10, and drop at X=1.2 Y=2.1 etc). It works great, if I don't take any parts off the machine. However, as projects change, i can see myself needing to load the machine with different parts.

So, here's the challenge: How do I track and make sure that the right parts are on the machine, in the right order?

Here's what I do today:
I have a handful of parts that live in "permanent places" on the machine - e.g. slot 1 = 0.1uF Decoupling cap in 0603 size.
Each PCB is programmed with these "Jellybean" parts in mind, parts that are already on the machine in a fixed place.


Here's what tools I have
- Basic Database creation knowledge (PHP and MySQL)
- Barcode Printer and Scanner (emulates a keyboard)


First thoughts:
Put barcodes on each slot, and barcodes on each component. For each PCB design, have a fixed list that never changes (say slot 1-24 is for Jellybeans only) and slot 24-48 are variable for each design. After loading the machine with the "new parts" scan the barcode on the slot, and the barcode on the component to make sure they are the right part, in the right place.

Am I going nuts here? Have I over complicated the problem?



 
48 unique components can cover a bunch of designs. Especially if you design for a common component set. 

Unless you are also doing rather high volume production where you might double up the same component into two reel locations to facilitate longer runs before stopping to replace reels. Machine up-time is a big consideration for commercial CM.

Speculating about the nature of your runs, just put each new part used by each additional new design into an open reel location until you fill it up, then deal with pulling little used parts. Over time your experience will tell you the right mix.
====

While it depends on what kind of file manipulation is comfortable for you, excel can handle lots of data field manipulations.

While the X-Y placement data file may only have a parts designation, you can pull another file from PCB design software to cross reference that to component values. and use "lookup" function to map values to designators.

yes, barcode sounds pretty damn useful for this. While it seems a modest expenditure of time to double check that the correct components are going into the right places...  Way cheaper than reworking mistakes. Measure twice, cut once.... If I didn't have somebody breathing down my neck to clear the machine, I might be tempted to solder up one or two and fire them up, before finishing the run.

JR
 
JohnRoberts said:
48 unique components can cover a bunch of designs. Especially if you design for a common component set. 

Unless you are also doing rather high volume production where you might double up the same component into two reel locations to facilitate longer runs before stopping to replace reels. Machine up-time is a big consideration for commercial CM.

Speculating about the nature of your runs, just put each new part used by each additional new design into an open reel location until you fill it up, then deal with pulling little used parts. Over time your experience will tell you the right mix.
====

While it depends on what kind of file manipulation is comfortable for you, excel can handle lots of data field manipulations.

While the X-Y placement data file may only have a parts designation, you can pull another file from PCB design software to cross reference that to component values. and use "lookup" function to map values to designators.

yes, barcode sounds pretty damn useful for this. While it seems a modest expenditure of time to double check that the correct components are going into the right places...  Way cheaper than reworking mistakes. Measure twice, cut once.... If I didn't have somebody breathing down my neck to clear the machine, I might be tempted to solder up one or two and fire them up, before finishing the run.

JR

I'm more or less circling around the same position.

But, i love the idea of testing the first one or two from a production run.
 
You might also do a small smd 'template' board with components 25 - 48, one of each - placed onto pads (not for exotic stuff anyway, but 0603 0805 1206 , sot23 and the like) with a pin header or test pads. run the template board first and check the values according to a spec sheet corresponding to the project. so you can make sure you have loaded the right reels in the right order. this could be  done before each batch - a small single sided pcb won't cost you all that much.

- michael
 
Michael,

that's a great idea! It took me a moment to re-read and envision it.

Wouldn't it be easier to just build one of the products on that run and test it? I already have the test rigs in place for those, and some level of automation in the testing.

 
it would be easier to measure single components not in the actual circuit (position, pin headers, possible connections between components or to ground that make it difficult to quickly check the value / type of component. of course you might do a test run with just one component per feeder to the actual board, but what to do with the board afterwards, complete? that would be an other run again.

a dedicated test board would be set up quickly, have large pads for measurements and all the usual suspects as footprints. as JR stated they wont be infinite possibilities at least in the audio world. larger components can be identified by looking at them, the smaller ones (R / C / L, sot 23) are the ones that are ambiguous I think.

- michael
 
IIRC some higher end insertion/placement equipment can confirm the value of components before they get inserted.

;D ;D  That reminds me of an automation overview meeting I was in decades ago, with supervisors from production and other parts of the company. I recall when the topic of confirming values and stopping the machine if it was determined that it was building "bad' boards and I got shouted down because it would be "stupid" to shut down the machine,,, just rework them later...  :eek: :eek: :eek:

It was a little surprising to me how some workers in the bottom of the ship didn't understand the full picture, and these were the cream of their respective areas. But I was the loud mouthed yankee who thought he knew everything about everything, so I had to pick my fights carefully.  8)

======
Confirming that parts are what they say they are before every run seems like over kill, if the reels stay loaded between runs, but it is prudent to check at least once for each new reel. After the components are in circuit, some values are difficult to accurately confirm.

The last production run I got from my CM had a red and green LED swapped... a small PIA, but he would never shut down his insertion machine long enough for me to test the board.  I once worked on a pretty expensive specialized DSP for a small company where the CM arranged to set up their machine on a friday, solder up 2 boards, and give us the weekend to check out the design... That was a bear because the design was new, rather complicated, and not very solid... That was one very short and stressful  weekend. (There was a subtle mistake in that design that was not found until months after that. A digital logic line was left floating, so boot-up of the circuitry could be somewhat random, depending on temperature and phase of the moon. )

JR
 
My current job has me working more as librarian than soldering, so I have a simple idea that may help. Label your reels A-Z etc. Make a cheat sheet for each build.
One build:         
Slot 24=A   
      25=B
      26=C
      etc.
Two build:
Slot 24=B
      25=A
      26=N
The hard part is reading a label on a reel after it's in place.
 

Latest posts

Back
Top