Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


ICT testing to improve yields

Views: 4701

We are currently checking parts on our boards with a digital... - Aug 16, 2012 by cobhambill  

#67033

ICT testing to improve yields | 16 August, 2012

We are currently checking parts on our boards with a digital multimeter to make sure values are correct. This process takes a large amount of time and is very in accurate Can someone please recommend a low capital budget system that is good and repeatable.

reply »

#67056

ICT testing to improve yields | 20 August, 2012

Are you better off with a ... * Low-end MDA ... OR * Low-end optical inspection machine

... Do you really need to check component values? ... OR Are you just checking component values to be sure that you're inspecting well?

reply »

#67058

ICT testing to improve yields | 20 August, 2012

We check the resistors to value check them. We are using resistors that are not marked in any way shape or form. We also hand build everything so it is very possible to have our operators place the wrong part in the wrong spot. From a time point of view I a mlooking at 2 to 3 hours for 10 boards of checking time. I need to reduce this if at all possible.

reply »

#67060

ICT testing to improve yields | 20 August, 2012

Try: * CheckSum LLC, 6120 195th St. NE, Arlington, WA 98223; 360 435 5510 877 CHECKSUM F360 435 5535 checksum.com * Testronics, 903 N. Bowser Road, Suite 300, Richardson, Texas 75081; 972-542-3111 testronics.com

reply »

#67064

ICT testing to improve yields | 21 August, 2012

Checksum will work if all your doing is analog/passive testing. Checksum systems do not have digital test capability. I would recommend a Teradyne z18xx system. We use 8 of them in our plant and have been for years. Used, these systems range from $10K up to $25K depending on your configuration. In your case, you don't need a VP board (digital test). These used systems are very good and dependable and won't cost as much as a Checksum machine. They are menu-driven; therefore very user friendly to program and setup. Have them spec your system with a Test Head Controller, fixed supplies, ATB II, and however many driver/receiver boards you will need (1 card/32 nodes).

Contact: Denver Test Phoenix Test Systems Rocky Mountain Test Equipment Lewis and Clark

They have other systems as well that you may be interested in. Good luck.

reply »

#67091

ICT testing to improve yields | 23 August, 2012

Hello

I would go for a Teradyne Z18xx those are kinda low budget, there are some companies that sell them refurbish right off my head I can think of Rocky Mountain Test Equipment talk to John Cowling (303 693 3322) and for the ICT fixtures so that you could test your boards I would go with Circuit Check there you can talk to Sobhi Youssef (763 567 3884) well those are the one that make all the ICT fixtures for me and they are low maintenance and they work fine.

reply »

#67098

ICT testing to improve yields | 24 August, 2012

We use CCI as well, but you might check out Q1 Test Fixtures in Ontario, CA. Their fixtures are a bit cheaper than CCI and the quality is just as good.

Contact Tim Richmond:

952-454-4474

reply »

#67103

ICT testing to improve yields | 25 August, 2012

Hello,

I already build such a system with hardware and software from National Instruments. I think is the best solution because you can buy for the start only a Multimeter (DMM), a multiplexer and a software. Is very easy to do the programming but also you can ask someone do do it for you, for few money.

reply »

#67109

ICT testing to improve yields | 27 August, 2012

An ICT solution seems to be overkill for what you've described.

What kind of volumes are you hand building? How many components are not marked? What is your current yield with the process you are using? What is/was your yield prior to implementing manual test?

We would suggest implementing assembly procedures that minimize the possibility of operators placing parts in incorrect spots, such as color coding, minimizing numbers of parts put in per operator/assembly station, etc.

If your volumes are higher, either in product or component placement, we would suggest an automated assembly process, rather than a manual assembly process, which could involve outsourcing the assembly.

Cheers, ..rob

reply »

#67110

ICT testing to improve yields | 27 August, 2012

No, in his case an ICT in not out of the question. An ICT, especially with many passive components, can give you piece of mind that the assembly was built correctly. Even with the addition of an automated insertion machine, you still run the risk of an operator putting the wrong reel on the machine. You will not get rid of the human element here. An AOI can supplement the ICT as well as give you greater coverage.

reply »

#67111

ICT testing to improve yields | 27 August, 2012

Dan_ems wrote, "I already build such a system with hardware and software from National Instruments. I think is the best solution because you can buy for the start only a Multimeter (DMM), a multiplexer and a software. Is very easy to do the programming but also you can ask someone do do it for you, for few money."

What does NI have to do with ICT testing? Absolutely nothing. This is a totally different test methodology. What you are describing is functional testing. Now this is overkill. The test bed could run in the 1000s depending on the complexity of the switching network and the number of components he would need to test. NI is great for automating your functional test, but may be of little use or help to the OPs problem. I would stick to ICT for your in-circuit testing needs.

reply »

#67112

ICT testing to improve yields | 27 August, 2012

The process is currently being performed by hand. I am trying to automate the process by switching to pick and place equipment and using AOI.

Problem number one is getting my customer to sign off on switching from hand assembly to machine assembly. Problem number two is getting my engineering team to approve parts that are labeled. Resistors are not labeled hence the testing. Problem three is quanity we are looking at 100 to 125 units a month for building. Problem four is the product was never ment for major production. When my company purchased the electronics part four years ago we did nothing to fix it we just went with it. Problem 5 yields. I wish I could tell you that but my folks just fix things and never record it. So I have no idea what the yields are.

These are just a few of the problems with this product. I am trying to improve things without braking the bank. I have been here a little over a year and I am now where near where I would like to be.

reply »

#67113

ICT testing to improve yields | 27 August, 2012

> The process is currently being performed by hand. > I am trying to automate the process by switching > to pick and place equipment and using > AOI. > > Problem number one is getting my customer > to sign off on switching from hand assembly to > machine assembly. Problem number two is getting > my engineering team to approve parts that are > labeled. Resistors are not labeled hence the > testing. Problem three is quanity we are looking > at 100 to 125 units a month for building. > Problem four is the product was never ment for > major production. When my company purchased the > electronics part four years ago we did nothing to > fix it we just went with it. Problem 5 yields. > I wish I could tell you that but my folks just > fix things and never record it. So I have no idea > what the yields are. > > These are just a few of > the problems with this product. I am trying to > improve things without braking the bank. I have > been here a little over a year and I am now where > near where I would like to be.

With the absence of any markings on your resistors, an AOI will be of little help to you except for presence/absence checking. Why are there no markings on your resistors? Why would this even be a discussion with your Engineering dept? This should be a no-brainer. Get the parts that are marked. Are these 0402 parts? Is that why they are not marked? Given this information, I would definitely suggestion the addition of an ICT in your production process. Your company can incur the cost, but knowing now that you are a contract mfg, you will have to sell your customer on the need for it; henceforth, they will have to purchase the fixture. These can run from $3500 on up, depending on the complexity, nodal count of your board. Given what you have stated about this board, you will be on the low end of cost.

You might want to look at flying probers. These are slower than ICT testers, but should give you the same coverage for your passive components. Plus, you won't need any specialized fixturing for the product. Your customer would only be responsibly for the operational cost, if you choose to charge them for it.

reply »

#67115

ICT testing to improve yields | 27 August, 2012

Reese,

We would agree about the value of an ICT system, in a general manufacturing environment. However, a couple of notes apply:

1. ICT isn't a catch-all, either, and should be used to verify the process, not validate it. Quality should be built back into the process as far as possible, to minimize failures and maximize first pass yield at ICT.

2. In this particular application, the assembly is being manually placed, leading to the assumption that it is a low component count, low volume product, at a small company with limited resources for ICT capitol equipment purchases.

2a. Given the above assumptions, a full-on ICT machine is not, necessarily, the most cost effective use of capitol.

3. I, personally, have just completed six years of work at a small EMS shop that did not have ICT capabilities. We were able to achieve 98% first pass yields at our various customer sites, including medium volume component placements that were unmarked, and impossible to identify at visual inspection. We did this by implementing different check steps at machine set-up to ensure that operators were loading the correct parts at set-up. You are correct that the human factor is always there, and that ICT can be used to help minimize it, but, it is not the only solution available.

4. I also didn't say it was out of the question...only that it might be an over-kill solution for the problem that was presented.

cheers, ..rob

reply »

#67116

ICT testing to improve yields | 27 August, 2012

We are not a contract manufacturer we are in fact a OEM. our product is used for military systems. The reason no values are on the parts is because the orginal owner of the product did not want anyone to steal the values and make their own version of this product. Stupid I know but that is one of the things I am dealing with.

Getting my engineering to sign off well to put it nicely they are afraid of the customer so they would rather do nothing to fix the process then actually take the time to fix what is wrong. Just me venting.

Getting the customer to sign off well we ent through hell because the color of the gasket inside a connector changed from pink to blue. That was 4 months of pure hell. Can you understand why switching from hand soldering to machine soldering is making my team nervous. I say we do it and move on.

The parts are 0603 parts not 0402. Reason they were never changed is again we never did anything to try and improve the product. I am trying to get engineering to look at changing to parts with values but I am being told they have no time this year to review and make changes. Welcome to my life.

reply »

#67117

ICT testing to improve yields | 27 August, 2012

> Reese, > > We would agree about the value of an > ICT system, in a general manufacturing > environment. However, a couple of notes > apply: > > 1. ICT isn't a catch-all, either, and > should be used to verify the process, not > validate it. Quality should be built back into > the process as far as possible, to minimize > failures and maximize first pass yield at > ICT.

This is true, which is why I made my suggestions. Currently, they have no way of verifying anything except manually. It sounds like they have to Quality program at all. I am a firm believer in built in quality in the process. I fully agree.

> > 2. In this particular application, the > assembly is being manually placed, leading to the > assumption that it is a low component count, low > volume product, at a small company with limited > resources for ICT capitol equipment > purchases.

Used ICT is relatively inexpensive compared to new systems. They will need to invest in capitol to grow. Your assuming they don't have the money for capitol. The OP was simply inquiring about it. I would hope he has done at least initial research into the matter and has an idea of the cost of such a system.

> > 2a. Given the above > assumptions, a full-on ICT machine is not, > necessarily, the most cost effective use of > capitol.

Again, your making assumptions. The OP clearly stated what his volume was previous. One could argue that this is low volume and they would be correct. But then that depends on the dollar amount you attach to such a product based on the man-hours you must invest into it. I did not make this suggestion for the use of one board. This is equipment that can assist him with other product and future product as well. It's a good investment for the cost.

> > 3. I, personally, have just > completed six years of work at a small EMS shop > that did not have ICT capabilities. We were able > to achieve 98% first pass yields at our various > customer sites, including medium volume component > placements that were unmarked, and impossible to > identify at visual inspection. We did this by > implementing different check steps at machine > set-up to ensure that operators were loading the > correct parts at set-up. You are correct that > the human factor is always there, and that ICT > can be used to help minimize it, but, it is not > the only solution available. >

True, it is not the only solution, and I appreciate your input from a QC perspective. But this in not fool-proof either, and as long as there is a human element, there will be mistakes, sometimes massive. That's not a risk I can take with our processes, or customers. Which is why I also suggested the need for an AOI, which it sounds like the OP already has. You are correct about the need for Quality, but it appears the OP and his company has a long way to go in this regard. One problem is going to be selling management on the idea. An ICT would be a simplier/faster solution at this point to at least verify the build and eliminate manual cost in this product. Quality has to be learned and taught. It is an institutional trait.

> 4. I also didn't > say it was out of the question...only that it > might be an over-kill solution for the problem > that was presented. >

Again, a good investment to have into the future.

> cheers, ..rob

reply »

#67118

ICT testing to improve yields | 27 August, 2012

We've seen a number of 0603 parts not being marked, more and more often in the market. Unfortunately, we never narrowed it down to whether or not it was manufacturer related, or separate product lines within a single manufacturer that were available as marked parts with a different part number. For our purposes, it didn't matter to us.

I've also dealt with customers who wanted to protect their designs. I had one customer who named his diodes as transistors on the board (used Q's for reference designators), then had me scrape off the part markings, and then, pot the area where the parts were installed...all to protect the circuit design. It was frustrating, but, it's what the customer wanted, so, we did it.

Have you considered out-sourcing the circuit board assembly to a contract house that has the core competency to build the boards for you, allowing you guys to focus on the final assembly, sales, and service? You could likely save some money on the assemblies (if you don't have to carry the overhead to build them yourselves), and you could pass that cost savings (or, better, a percentage of those savings) on to your customer. Then the CM would own any quality issues with the assembly of the boards.

I understand how things can go when your engineering department are afraid of the customer, and how easy it can be to get in the mode of just fixing the boards, delivering the product, and making the customer happy. It sounds, though, as if you understand the long-term difficulty with that mode. Eventually (if not already) your costs for producing the product will exceed the value of the product on the sale side.

For what it's worth, I'm with you when it comes to the decision to switch to machine assembly/reflow. Sometimes, it's easier to ask forgiveness than permission :) It could come down to a need to explain to the customer that you are the one that owns the product, and that you own the process, and will do what you need to do to ensure that the product line is fiscally viable, while insuring that the customer continues to receive high quality products that work for them when they're delivered.

cheers, ..rob

reply »

#67119

ICT testing to improve yields | 27 August, 2012

Which military? The US? Being a military contractor, I am surprised at your lack of quality at your facility. I suspect this is non-US military. I don't know where you are located, but being a contract mfg for the military, any military, I can imagine what you must be going through. I worked for the military for a time and I understand what changing ANYTHING on the design means, even something as simple as a resistor. Yes, this is a special case and should have been addressed when the product was purchased. This may be a hard sell and it will be an uphill battle. I can't give you any suggestions for dealing or negotiating with the military, only about the equipment to use. So, with the absence of any markings and the possibility of never acquiring such markings, an ICT/flying prober, at this point may be your best solution.

reply »

#67120

ICT testing to improve yields | 27 August, 2012

I need to clarify something for everyone. The end customer is the Military. Our customers are military suppliers who do in fact have quality systems in place. When we build the product we end up potting the unit so the customer never looks at the electronics inside they get a black box with a connector on it. With that said our customers signed off the hand solder process four years ago and they refuse to let anything change unless we test the units for a minium of 2000 hours. I am ok with that. I just want to make a better product. My current assemble time is 12 hours from hand soldering to testing. One of the reasons for trying to improve build time is to get away from paying our union people 23 dollars an hour to hand solder (they normally solder for 2 hours per unit whether its SMT or through hole 2 hours is the magic time) My people spend 3 hours every other day hand testing so you see that I am willing to put money out there if it helps reduce my touch time. I am being driven by the customer to make the product for less money again I am ok with that. Long story short I know that ICT will not solve the worlds problems but it at least removes test time by someone getting 23 dollars an hour.Yes we need to improve our quality idea but that means getting management to buy into it and when you have four million empires its hard to destroy them and make one empire. Again me venting. I think all of your ideas are great and I will use them unless the empires kill me.

reply »

#67121

ICT testing to improve yields | 27 August, 2012

Well, your on the right path. Refer to my earlier suggested make/model and well as the four contacts I provided. They can assist you with your testing needs.

reply »

#67122

ICT testing to improve yields | 27 August, 2012

Getting through the empires is always fun. And, re-qualification testing with your customer (military or not) is also fun, and time-consuming. Add a union shop on top of that, and I truly feel your pain!

ICT will solve your immediate problem, as has been described. The question will/should come down to the ROI on the machine that you purchase.

Outsourcing the production of the assembly should help you solve the problem you're having, but, comes with many other variables (not least of which could include needing union approval to outsource, as well as end-customer approval).

My best suggestion, to begin the improvement (perhaps while researching ICT machines): -Create color coded assembly drawings for each step of the assembly process. -Have no more than 8 components/drawing. -If possible, have one drawing per step of the process/per operator (as if the product were being built on a slide-line, and each drawing is an individual step of the process). -When the job is set up, have the values/reels of components verified at each assembly step...by someone other than the person that set them up. -Try to avoid similar part values at each step (ie 10 k, and 100 k). -Have assemblers use the check-do-check method (check the work before them, do their step, check their step). -Drive individual accountability for mistakes. Mistakes will be made, it's human nature...making them accountable for their mistakes helps to minimize future mistakes.

If outsourcing is an option, and you'd like to discuss further, please feel free to contact me off-list. My company would be happy to evaluate the assemblies, and provide you with cost quotes to support your needs.

Cheers, ..rob

reply »

PCB X-Ray Inspection

PCB Buffers