Introduction of SPC to the Supply Chain

Long OP alert (probably 2 OP's)

Good morning All,

As I have mentioned in a couple of other posts we are starting to try and introduce SPC to our supply chain, and throughout our process. I am going to give an introduction to the project, and to the academic paper I am going to be writing alongside it, in the hope that when I have queries and issues, you good people will be able to help me answer them.

OK, a little about what we do as a company. We manufacture, sell and support automated bank note validation systems (note validators). We have a range of devices, which accept you bank note (bill) in any orientation, read it, check it against some stored data and then store it, telling the 'host' machine the value of the bill/note.

You would commonly see our products in casinos (top end), self-service vending machines and change machines (middle bracket), and on children’s rides at the mall (low end).

The devices themselves are electro-mechanical using various optical sensing technologies to 'read' the notes. We also do 3 devices designed to give change and/or winnings to the customer, 2 give notes, and 1 gives change. The change device is the only one in the market with a single coin bin and not multiple coin stacks. Of the 2 note return devices one is for the top end of our market, the other for the middle of our market.

The really clever bit of our machines is the software which is loaded onto them (a firmware and a dataset). The electronics anyone can do, it is controlling and using it that is hard.

I think that is enough on the product - for now.

Now for the SPC bit.

The parts in our product can be divided into 3 distinct categories (and 1 not so distinct). Plastic Injection moulded (some with a second process), Metal, Electronics, and Firmware/Dataset (not so distinct - a kind of software). I cannot think of a single product that does not have a mix of all four parts in some way. For the remainder of this post I am going to ignore the firmware/dataset element.

Physical set-up of the company:

Headquarters - UK, Sales and Customer Support Offices - World wide, Manufacture (assembly) main plant - Shenzhen, China, other assembly plants - Brazil, USA, UK - all limited capacity.

The UK HQ houses R&D, Production support engineering, IT, central admin, etc, we also have a limited production facility and conduct the final unit preparation before dispatch - tailoring it to each customer with appropriate currency data etc. The production facility in the UK is primarily a test line to set up new product released from R&D and find a 'best-fit' solution to allow future large-scale assembly.

Our suppliers for the majority of components are based in China, near our factory there. We also get a limited number of parts from the USA and UK, for certain specialist skills, not available at a reasonable cost in China.

I think that's a good place for me to look away from the screen, and allow my laptop to reset to complete installing some new software.... more shortly

:mg:

So to carry on.

We have traced the majority of probles we experiance (circa 85%) to be problems with the injection moulded plastic parts. These range from simple distortion due to cooling issues, tooling being incorrectly sizes, to incorrect temperatures and cooling changing the properties of material and causing early life failure.

It has been decided by 'da management' that SPC with our suppliers is the way to solve this problem. I started to look into this in early July, but was out of work for 2 weeks, due to other commitments. I got back from those 2 weeks away, had 5 days in work, and then flew here, to China. I have now been here for close on a week, speaking to suppliers about SPC, outlining the benefits, why we as a team need to use it, etc.

We and all our suppliers are shut down today by government edict, so I'm using the time to consolidate and prepare for Monday and Tuesday. We are spending both days complete at one of our newest suppliers who are just starting to come on stream. I am intending to go through SPC in detail with them, look at what parameters need monitoring and recording, how to analyse the data, and how many samples they will be able to take etc.

That is the physical project. The other half is the academic write up. I'm currently studying part time for my Diploma in Quality from the CQI, and my Project module needs to be complete by Janurary 2012. This module is quite large, and I am planning on using this SPC project as the basis of it.

The first stage is a literature review, I already have out of the crisis, and all the useful bits on SPC in injection moulding from the forums, can anyone suggest any other reading material which would be appropriate?

Any thoughts on the analysis I need to do or sample sizes etc. gratly appreciated.

In addition work will probably spring for some new software if i can justify it, i was going to use Excel, but there are literally thousands of parts and tens of thousands of dimensions that could be used, so think it is a little big for manual analysis.... Anyone got any recomendations?

Finally regarding the application at supplier, it is anticpated due to the way/level the chinese operators work, any inspection needed would be conducted by inspectors, and then analysed by tansmission to our computer system (web-form access). This analysis would be live. We are looking longer term to bring it even closer to the shopfloor.....

Olly

You are welcome to make use of the SPC materials here -> http://www.efcog.org/wg/esh_es/Statistical_Process_Control/index.htm

Thanks Steve, Much appreciated.

I don't suppose anyone has an .avi or .mpg or similar of the red-bead experiment? I'd love to use the one on you tube, but I'm behind the gret Firewall of China at the moment, which limits access to things like youtube.....

Try http://video.google.com/videoplay?docid=-3044393818811798956

Thanks for trying, but great firewall has stopped that one too!

It is one of the most difficult things about working out here, there are so many work related sites that are filtered out by the firewall that you have to try and work round it, without breaking the law.....

I am glad I live in the UK, and can look at anything I want. Especially as what I want to look at isn't that exciting.....

any recommendations on books to read? I'm going to put an amazon order in when i get home (I love Prime). I might even persuade work to pay for a tablet if i'm going to be travelling a lot, and then buy the ereader software for it....

Olly

pearsonow said:

.............................. any recommendations on books to read? I'm going to put an amazon order in when i get home ..............

Click to expand...

1. If you are going to Amazon, why not purchase a set of the 'red bead game' for about $200? Nothing beats a first hand experience with the game.

2. You can convert youtube contents into mp4/flv formats and email to China - but make sure you check for copyright violations.

Thanks, unfortunatly I'm IN China at the moment and due to the time difference noone will be at the office until after I'm planning to run the demo.

I have enough coloured beads at home (gaming accessories) that I can run the experiment, I'll get workshop to make up a paddle etc, that I will use for the management in the UK.

I have managed to find a nice little demo here though:

Red Bead Experiment Simulator: http://www.symphonytech.com/redbeadexp.htm

pearsonow said:

Thanks, unfortunatly I'm IN China at the moment and due to the time difference noone will be at the office until after I'm planning to run the demo.

I have enough coloured beads at home (gaming accessories) that I can run the experiment, I'll get workshop to make up a paddle etc, that I will use for the management in the UK.

I have managed to find a nice little demo here though:

Red Bead Experiment Simulator: http://www.symphonytech.com/redbeadexp.htm

Click to expand...

Symphony Technologies is a company that is owned by one of our Forum Administrators Mr. Atul Khandekar.

Thank him when the simulation is successful....

Stijloor.

I applaud your effort to train people with the Red Bead Experiment. I have a couple of observations to pass along that you may wish to use.

First, the game is primarily aimed at mid to lower management. You have willing workers pulling beads out of a tray with a pallet. The willing worker has absolutely no ability to change the outcome. So the game becomes one of scolding the willing worker. First you yell at them ‘do you understand I only want white beads.’ Management then yells at them for producing red beads. This continues several times getting louder and more vocal about the willing worker not doing their job. Then management tries a different tactic. They offer incentives to the willing worker not to produce red beads. It may be a cash bonus, recognition, or some other type of reward. Again, the willing worker does not meet the requirements even with the enticement of incentives. After this game is played several times, it becomes obvious that the willing worker has no chance of producing what management wants. The system that they have been given is forcing unwanted results.

It is here that Deming pointed out the need to improve the system by eliminating the number of red beads in the tray or, as stated in a factory setting, eliminate the causes of failure. This is what Six Sigma is about.

Another way to play the game is with mid management to upper management. You play the same game to the point where they see the futility of the job at hand. At this point you need to change the question being asked of upper management. The question is ‘Why do you allow red beads to be poured into the tray?’ It is in answering that question that the company will build a quality system, ISO9000.


I also think it is admirable that you are undertaking the effort to train people in SPC. A way to look at statistical process control is “based on past performance, what do I expect the process under investigation to produce?” SPC is executed in order to set off ‘alarms’ indicating that the process has changed. Either change for the good or change for the bad.

Many books and articles can be found regarding SPC. I would suggest much of Dr. Wheeler’s work, which is published through spcpress.com. Also, since you will be using many untrained and as yet unskilled people to perform this work, you might want to check ASQ.org and look for work done by Art Bender in the 1950’s. Specifically, ‘Look, Ma, No Arithmetic.’

The second part of your dilemma with SPC was how you define what to measure. Standing back and looking at it from an APQP methodology, you would proceed through Failure Modes and Affects and their causes. To lower DFMEA occurrence, you may be able to apply SPC in order to lower the affect of the cause. This is really a joint decision between manufacturing and engineering and if agreed to, then that DFMEA requirement for SPC is passed through the characteristics matrix to a PFMEA requirements/functions. As the PFMEA is executed, The DFMEA specific SPC item along with all the other process requirements and failure modes are analyzed for their cause of failure. At this point an action may be to lower the risk of that cause. SPC or a predictive method may or should included in the control plan.

By ranking the severity of failure mode and occurrences of your causes, you can create a list of the most important items to control to the least of the items to control. It gives you your starting point and the direction to head. (By the sound of the scale of your project, doing them all would be impossible. So just start at the top and work your way down with available resources.)

In your discussion you had noted that most (85%) of your problems lie within the production of injected molded plastic parts. You also stated that certain requirements were not being met due to distortion, incorrect dimensions (too large/too small), and changes in material properties with the effect of early life failures. My experience with injection molding tells me that the initial causes that you are already identifying are probably correct. Those causes being mostly related to molding machine set-up parameters. You identify some of the causes being the molds do not have sufficient cooling lines to take the heat out of the mold, incorrect molding temperatures, which could be the zone temps on the nozzle, back pressure on the screw, gate size and shape – many items from design, and many changes from the properties of the compound. These items are typically controlled through the use of a set-up sheet where the machine is set to previously proven parameters. These may be SPC’d to understand how well the machine controls them, but experience shows the part will reflect small changes to the process quicker than you would measure out-of-control conditions in most of these parameters.

The other issue is that typically in a molding shop environment those small process caused part changes begin to show up and the set-up technician, or molding technician, will go over and ‘tweak the process.’ In other words, he is changing the set-up parameters to another set of parameters to get a better result. SPC is about prediction. When the tech changed the settings, he changed the process and, therefore, the prediction. You cannot run SPC if the prediction is changing as that changing is really called ‘an unstable process.’ Looking for causes can be done by reviewing the changes made each time the tech changes a setting or when material lots change.

You identified several causes for these process interferences, such as, management is searching for the cheapest molds. Adding cooling lines, and heat sinks add cost. So the cheap supplier leaves them out, thereby, modifying your cooling rates and increasing hold times. Time is money. A part can be made cheaper if more of them can be produced in the same time. To do that the injection time is decreased by increasing the temperature of the molding compound. With higher temperatures, degradation of the compound and increased distortion of the part are experienced.

If upper management’s objective is to buy cheap molds and tools versus inexpensive molds and tools and produce cheap parts instead of inexpensive parts, then they are adding red beads to your tray. The willing worker will try their best to produce what management wishes for; but in many cases they will just produce cheap parts, even though they do not want to. If management wants improvement, they need to improve their ISO9000 philosophy and execute a red bead reduction process. This is why Dr. Deming stressed his Red Bead Experiment so diligently.