Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

SMTnet
»
Electronics Forum
»
Cpk Variables Control for Paste Printing

Cpk Variables Control for Paste Printing

Subscribe to Thread

Hi Guys, Can anyone help us out here, please? We need ... - Mar 19, 2002 by ianchan

It might just be me, I have been known to loose my mind, but... - Mar 19, 2002 by JAX

FORMULA OTHER THAN LxBxZ? - Paste is X% [X=~50] metal by vol... - Mar 19, 2002 by davef

Heyz...at this rate i'am losing my mind too! We have been... - Mar 20, 2002 by ianchan

Are you saying L x W x H didn't answer his question? Unle... - Mar 20, 2002 by Steve Thomas

Your customer's question is reasonable. Before you can ... - Mar 20, 2002 by davef

You based the aperture size on LxBxH. Where did you get t... - Mar 20, 2002 by davef

thanks, everyone for your valuable feedback and shared opini... - Mar 31, 2002 by ianchan

Just one minor point to bear in mind, most stencils have tap... - Apr 03, 2002 by

To date, much of this discussion has focused on determining ... - Apr 03, 2002 by davef

Good points... Thanks everyone. ref back to Mr.Bob's/Dave... - Apr 04, 2002 by ianchan

What are the normal control limits in the industry, for a we... - Apr 04, 2002 by Ioan Tempea

Equipment companies like to talk about �accuracy� and �repea... - Apr 04, 2002 by davef

You got me dander up now. Anything gettin' out the door the... - Apr 05, 2002 by

I have found a consistency in certain package types which ho... - Apr 18, 2002 by

ianchan

#19172

Cpk Variables Control for Paste Printing | 19 March, 2002

Hi Guys,

Can anyone help us out here, please?

We need to know what formulae (other than Length(L)xBreath(B)xheight(Z)), is used to determine what volume of paste is required for 20mils pitch leadless chip carrier (LCC28), for the paste printing process using a 6mils stencil?

How to achieve a constant process capability (measured in Cpk), for this 20mils pitch LCC28, for the reflow soldering process?

any experiences or info to share, is really appreciated!

Thanks, to all :)

reply »

JAX

#19174

Cpk Variables Control for Paste Printing | 19 March, 2002

It might just be me, I have been known to loose my mind, but I don't believe I understand the question. If your asking for an idea of how much paste volume you should have for this part, I think its a little late. You already have the stencil so your apeture sizes and stencil thickness is set. As far as calculating what you have, the three measurements you listed is all there is. As far as Cpk is concerned in the printing process, a baseline starting point is usually between +- 25-35% although I think I'll just stick to the " Looks good. ; Looks bad." test. Let me know what you mean and I'll try to help more. That's just my opinion, I could be wrong!

reply »

davef

#19179

Cpk Variables Control for Paste Printing | 19 March, 2002

FORMULA OTHER THAN LxBxZ? - Paste is X% [X=~50] metal by volume. Your paste supplier knows the value of 'X'.

HOW TO ACHIEVE A CONSTANT REFLOW PROCESS CAPABILITY? - Buy a capable oven, operate and calibrate it properly.

reply »

ianchan

#19186

Cpk Variables Control for Paste Printing | 20 March, 2002

Heyz...at this rate i'am losing my mind too!

We have been doing the "looks good; look bad" approach, and backed it all up with Z-height paste measurements. Only thing there's this customer QA asking how we established those USL & LSL utilized in our Cpk calculations?

We have leadless chip carrier (LCC) packages on the board, and all paste falls into our UCL and LCL, its juz the said QA is asking us on what formulae was all this specs derived from??? We based it all on LxBxH! Shheeeezzz.....

Anyhow, thanks for everyone's help!

reply »

Steve Thomas

#19194

Cpk Variables Control for Paste Printing | 20 March, 2002

Are you saying L x W x H didn't answer his question?

Unless he wants to know how you're verifying that the post reflow fillet is adequate (which is determined by his pad size as well as your print process and the % metal of your paste), I don't get it.

reply »

davef

#19210

Cpk Variables Control for Paste Printing | 20 March, 2002

Your customer's question is reasonable.

Before you can establish your control limits, you need to take 25 to 30 data samples of 4 to 5 measurements. This says you need a 100 or more measurements before you can start charting. And even this much data does not assure good determination of control limits for x-bar and s charts. The scary part, while these data are begining gathered, the process is running without any attempt to exercise control.

reply »

davef

#19211

Cpk Variables Control for Paste Printing | 20 March, 2002

You based the aperture size on LxBxH.

Where did you get the LxBxH? Obviously the thickness of the stencil is kinda fixed. Who gave you the other numbers?

Consider searching the fine SMTnet archives for background on aperture design.

reply »

ianchan

#19342

Cpk Variables Control for Paste Printing | 31 March, 2002

thanks, everyone for your valuable feedback and shared opinions. Appreciate your efforts!

been following up with this "nitwit" issue on volume calculations (sortof a forumulae to confirm a formulae???) for paste printing process. What we did was to buy a new oven (hopefully our purchase review on machine capability was hey'O'k) and start afresh on the production/process run controls... we will know in a cuppa weeks how r we fair thus far...

reply »

Bob

#19382

Cpk Variables Control for Paste Printing | 3 April, 2002

Just one minor point to bear in mind, most stencils have tapered apertures. If you want to get really accurate volumes you will need to work put the volume based on the taper of the walls.

Probably going to far but at least you can sound knowledgeable to the QA guy.

Bob...

reply »

davef

#19384

Cpk Variables Control for Paste Printing | 3 April, 2002

To date, much of this discussion has focused on determining the volume of solder paste. That is but one piece of the measurement of process capability. The other aspect that needs to be evaluated is: an assessment of the accuracy of the positioning of that 'brick' of paste.

reply »

ianchan

#19391

Cpk Variables Control for Paste Printing | 4 April, 2002

Good points... Thanks everyone.

ref back to Mr.Bob's/Dave's observation (not in order) on brick accuracy, position & placement, Stencil apperture geometrical opening, will feedback to my team to brain-storm further.

reply »

Ioan Tempea

#19393

Cpk Variables Control for Paste Printing | 4 April, 2002

What are the normal control limits in the industry, for a well controlled process, using a 6 mils stencil?

Should they be different according to the surface finish (higher UCL for ENIG, Im SN, etc, lower UCL for HASL)?

Thanks, Ioan

reply »

davef

#19394

Cpk Variables Control for Paste Printing | 4 April, 2002

Equipment companies like to talk about �accuracy� and �repeatability�. These measures provide limited information. Flaws of the measure aside, some printer companies have begun to report Cp. * DEK ELA: Cp = 1.33 @ �0.025mm * DEK 265: Cp = 1.6 @ �0.025mm * DEK Infinity: Cp = 2 @ �0.025mm * Ekra E4: Cp = 1.6 @ �0.025mm * Ekra E5: Cp = 1.6 @ �0.025mm * Minami 878: Cp = 2 @ �0.025mm * Minami 880: Cp = 2 @ �0.025mm * MPM 1500: no capability reported * MPM 2000: no capability reported * MPM 3000: no capability reported

Similar to placement machines, Cpk for printers should be differentiated by pitch [eg, ultra fine, fine, normal, etc.].

Read �Developing A Low PPM Defect Level Ultra-Fine Pitch Printing Process� Steve Hall 'Circuits Assembly' 9/92

reply »

darby

#19398

Cpk Variables Control for Paste Printing | 5 April, 2002

You got me dander up now. Anything gettin' out the door there?

reply »

Eric Wolf

#19572

Cpk Variables Control for Paste Printing | 18 April, 2002

I have found a consistency in certain package types which hold a very close tolerance. We use a six mil laser with electropolish. BGA's are tight. Some IC's drift. I think the range and Cp is more important than the Cpk for today. The reason being is the nominal expected mean drifts. We use Cyberoptic with 100%BGA site inspection. I have distributions, but they are a manually created report. In a few weeks I will have it automatically downloading. I will share.

reply »