Automated machine downtime 4

[maxbreizh]

I am trying to determine the machine run time (10 CNC machines total).
We manufacture tool joints (drill pipes).

Currently a main server downloads programmes to each machine, corresponding ot the part to make. Each part takes approximately 10 to 15 minutes. Each time has a start cycle button that the operator must press before the cycle starts.

What i am trying to do is to get each machine to send a signal to the server (time of the day, status of the alarm) each time the operator push the start cycle button. From there, I can calculate the run time/ planned and unplanned downtimes.

Anybody has an idea of who can help me do this? There's got be some companies out there specialized in this kind of stuff...
Maxbreizh_ Houston.

 

[griffengm]

Max,
Have you checked with the machine builder? Some controllers have this function built into them.


Griffy

 

[maxbreizh]

Thanks Griffy, it's a good idea - I'll try.

I've also found these guys from a different forum:

http://www.enetdnc.com/

Exactly what I was looking for.

If you have a look, let me know if you are aware of any similar company.

Thanks again,
Maxbreizh.

 

[ofohique]

Maxbreizh:
Veryintersting question. Do you mind me asking if your company is high volume or a job shop? What are you expecting to do w/ the data?

 

[SteveMeisel]

We are doing exactly what you are asking about. Our system locks out the cycle start pb if the CNC has been down for >20 min. The setup or operator has to go to a pc on the server and enter a reason code for the downtime. The server collects data like: Incycle time, downtime, downtime reasons, Fault resets, Power on time, etc. They sent us the units, I installed them, then they came and did the software stuff.

The company is Innovative Controls Inc @

http://www.innovativecontrols.com/index.asp?sorting=0

they use Sixnet

http://www.sixnetio.com/html_files/products_and_groups/io.htm

I/O units to monitor/control the CNC. It seems to be a good setup.

 

[maxbreizh]

Ofohique,

It's high volume if you like because they could process 1000 piece orders.
I'm running KANBAN on Kaizen cells and I need the data to determine the causes for cycle failures.
Well, the data is very useful for many reasons... cycle times, unplanned maintenance ratios, process optimization, CAPEX requirements (type of machine mostly needed), training of the operators, etc.
This data is crucial if you wish implement lean manufacturing in your work environment, particularly KANBAN.

Maxbreizh.

 

[maxbreizh]

Thanks Steve for your post,

Where/ How can I get more information about your system?
Can you be my contact for that or should I use the web link on your website?

I need a quote for this system, installed in our premises.

Thanks!
Max Breizh

 

[bugnut]

GE makes a product called Cimplicity-I have installed this on machine tools it will give you all of the data you want and more-Works well with GE/Fanuc controls. If you are part probing you can also download probe data and dimensions-especially variance from zero. This is helpful if you are changing tools by usage-it will stop the machine-also will correct the part to statistical variation. Many other things are availble through the Cimplicity and best of all the operators know it is running and anyone with a web access point on the server can look at the data-how long the machine ran, good parts, parts made/shift......you get the idea.

 

[maxbreizh]

Bugnut,

Thanks.
I went to the GE website and I looked at the product offered, particularly:
"Proficy Shop Floor SPC - a complete statistical process control solution for the collection, monitoring and analysis of plant floor quality data"

I am worried that this is too powerful a system (therefore expensive) for what I am trying to do. Do you remember how much you paid for this system (give me an approximate per machine is great!).

Thanks,
MaxBreizh

 

[griffengm]

Max,
It seems to me that you have a good opportunity here if you work on this one without front end software. What I propose is that you make a daily half-hour trip to this area and, if a machine is not running, stop and ask the operator why.
This does several things.
First, it allows you to build credibility with your operators. Second, it allows you to make the call on the actual cause. Third it will improve morale. Fourth, your discussions with your management team will have more conviction from first hand knowledge as opposed to statistics. Fifth, it allows education of the operators both in operation and in analysis.
When you are satisfied that they understand what you are looking for, step back and take another look. You may not need the software at all.

Griffy

 

[bugnut]

Maxbreizh,
Sorry not much help here with cost. We had Cimplicity running so it was just a matter of paying local Fanuc whizz to get a screen made and then hooking it up to our server as the machine tool had the front end loaded when purchased. Look at this site it is a screen capture of our reporting and the homepage of the team that did the connection and screen.

http://www.controlinstallations.com/web_ci/new_page_7.htm

 

[maxbreizh]

Griffy,

What is you makes perfect sense to me.
The purpose of the software is not to determine how good the operators are at running their machines. I trust them at doing their job perfectly well, if not the issue is training and not punishment. I'm 200% behind what you say.

The main reason why this is done is to determine the ratio between planned and unplanned maintenance (curative and preventative) ; this is particularly difficult right now because the machines run according to a particular cycle time that does not necessarily require each and all machine to run at all times. For example I may thread and mill with 2 different machines, but it takes 3 times longer to thread so the milling is in standby 2/3 of the time and that's okay, but I do not have any way to measure it accurately ; so I need this software _


Bugnut,

Okay. Now I understand. I looked at the wrong website! Control Installation is indeed much simpler than Innovative Solutions. I'll call them tomorrow-

Thanks!

Max Breizh

 

[rlgauto]

maxbreizh (Industrial)
If you didn't want to purchse software right away you could collect valuable data to evaluate the process by monitoring the logic of the machine. this would provide you with exact times run, buttons pushed, change over times, up & down times plus any other combination you can think of.
Depending on the controllers in use, such as an Allen Bradley SLC series, you could program in a logic program with a few cascaded RTO functions. Also the programing of TOD clock circuit would give you all the data you need with the time of day logged without adding a lot of cost. The MSG function is an invaluable tool. These are options available at practically no cost except time to program.

rlgauto

 

[07221983]

Iam doing a feasabily study on a blister packaging machine(pharmacitical products) the problem is that the machine breaks down 6-7 times during a run of 4200 for 4.5hrs,@ a speed of 105 per min.

I have tried to dertemine some of the causes like mantainance, please give me some ideas how to solve this problem.

 

[PSE]

07221983,

What are the failure causes when the machine goes down? Take a look at the usage and failure history if available, and see if there are patterns. At the very least you could create a pareto and go after the most prevalent failure type first. Six to Seven breakdowns in a 4.5 hour run should give you quite a bit of data in just a few days.

Some Categories for failures in no particular order:

Improper run setup
Feedstock Problems
Program Problems (timing etc)
Maintenance
Facility (poor quality power, air, etc)
Package design inappropriate for machine or tooling

Regards,

 

[griffengm]

07221983,
One of the greatest analytical tool around is the graph. I suggest plotting the breakdowns over the elapsed time. Look for clusters and/or spacing. All early in the run, llok at set up. Evenly spacer would cause me to look at operator issues and late clusters might point to variation in the feedstock. Not absolutes but they may save time in the long run.

Griffy

 

[07221983]

Thank's

I have looked at the failures: I have found that it has to do with Feedstock promblems the pills get stack in the feeder now I don't know how to solve that, because I thought by changing the brushes it will be ok but it is still happening.

 

[PSE]

07221983,

How does the problem normally get "fixed" when the machine goes down? See if there is a way you can build or add this fix (or something similar) to run at the same time the machine is in operation.

Regards,