|
|
|
09/14/2010 11:10 AM
|
|
|
Hi All,
I am mentoring a project on set up reduction on brake presses and one of the things I have proposed is to track set up times on brakes by having the operators log in and out of a screen on their controllers. This data will help me track project improvements and implement SPC on set ups in future. But I will be adding an extra touch (10 seconds) and it goes against the lean principles.
Has anyone been able to track set ups without adding work to the operators?
I think in the long run collecting this information would be beneficial and I was wondering what the lean community thinks.
Thanks in advance.
Taruuna Aurora
|
|
|
|
|
09/14/2010 01:38 PM
|
|
|
Taruna -
Instead of steering your efforts towards tracking set up time, why not time it for them and then, based on the results, spend the energy to REDUCE set up times?
|
|
|
|
|
09/14/2010 01:38 PM
|
|
|
Tracking of setup times is absolutely required both before and during any setup reduction activities. You need it before you start to set your baseline. Just make sure everybody is clear on the proper definition of setup time and using it though. (Last good part to first good part - at production rate.) If everybody is using a different definition, your data will be useless. Then you need it as you move forward to see if you're actually making progress.
Another thing you may want to think about is whether or not all of your setups are equal in complexity. In a Press Brake, there are many different kinds of setups with one big variable being the number of tools. You might see your graph of setup times jumping all over the place and conclude that nobody's doing anything the same way twice - when what you are actually seeing is setups with different degrees of difficulty.
As for you actual question on the tracking being NVA and adding time, you are correct on both points. If you do this correctly though, you will add only a few seconds of NVA and you'll save far more than that down the road. I've found that it's a real kick for the operators to know how much time they've saved as they make changes. Without tracking your time, nobody will have any idea of how they're doing.
|
|
|
|
|
09/14/2010 04:27 PM
|
|
|
Aurora,
You may not have complete data, but about how long is the setup time now? You may find faster measurement methods, but 10 seconds is pretty fast. If this truly impacts the business, you must have very good setup times already--SMED, indeed!
If your setup times are very long, then I suspect this feedback about data logging time is simply an excuse to avoid change. It is unrealistic to expect tracking mechanisms to take no time. (someone has to make kanbans, after all) You end your message with the right thought: it is worth the time to measure, given the potential for improvement. Keep this in mind, and you will avoid choking the gemba with many unimportant measurements.
|
|
|
|
|
09/15/2010 11:42 AM
|
|
|
One thing you should ask yourself is How will tracking setup time effect setup performance and is it important to measure every setup?
In my experience, I will time the setup myself during a setup reduction effort only. The purpose of timing the setup then is to reduce the setup time. You will be amazed at what you will find after watching and observing only one setup. There will be more opportunities for improvement then you will think.
|
|
|
|
|
09/15/2010 03:55 PM
|
|
|
I should have said more about the project. I have time studied the set ups and they are anywhere between 7-20 mins. There are a lot of opportunities in improving the process such as reducing the programs tweaks, standardize inspection methods, reduce the variability in material and so on.
After making these improvements I want to be able to track the set up times not so much for project validation but for knowing how we are doing in set ups and if the process is in 'control'. Keep in mind these brake presses do parts with one bend to 5 bend parts so the set up will vary but as long as they are within limits I am fine. At my facility brakes and lasers see the highest volume of parts so I think control charts would be a good indicator of areas of opportunity.
Thanks everyone for your thoughts and inputs. My next challenge is going to be make sure I get reliable data.
Taruna
|
|
|
|
|
09/17/2010 02:14 PM
|
|
|
Once improvements have been made on your set up times, I suggest presenting them not only to your superiors, but also the operators themselves. Do calculations on how saving x number of minutes per setup will ultimately save x dollars per year. This not only validates the studies you've done, but encourages the operators. I would also consider creating classes of set ups depending on the number of bend parts.
|
|
|
|
|
07/19/2012 02:57 PM
|
|
|
What were your results for average setup times. I presume that you ended up with a matrix of some type, eg. 1 bend, 2 bend, etc. Automatic back stop vs manually set backstop, Etc, Etc. However, If you could use just one number for an overall average for times over a large sampling of set times, what would that be?
|
|
|
|
|
08/02/2012 11:33 AM
|
|
|
Tracking set up is a necessary NVA activity. Like you, as soon as I ask operators to track setup time after a SMED event, they say "doesnt this add time!" They are right. I counter that with the standard comment "Great job recognizing that as a waste! However, we have to know if we are meeting our standard...."
I usually only track the setup time for a period of time, 90 days or so, to see if we have made lasting improvement. The operators also like to hear that it is not an ongoing thing.
If you suspect times are slipping after the 90 days, you can always start the time study back up.
|
|
|
|
|
08/06/2012 11:18 AM
|
|
|
This thread ties in with a similar project we are conducting with our pressing operations. I would be most interested in seeing the data you have collected from your operation. The hardest part I have found in recording setup times are the huge number of variables with pressbrakes.eg: Loading 3.0M of sectional tooling vs 100mm setup, complexity of setup, type and weight of tooling....the list goes on.
Regards
Brad
|
|
|
|
|
08/17/2012 11:07 AM
|
|
|
Tracking your set-up time can deliver interesting information, especially you would want to know how your set-up time is developing over time.
Now you have two interesting questions here, first, how often do you need to measure to get the interesting information.
Clearly all handlings that need to be done to measure are waste. Therefore you would want to reduce it to a minimum concerning frequency and necessary handlings.
The other question is what actually is set-up time and what information do you really get from your measurement data. There is a widespread definition that Thomas has been regarding to (Last good part to first good part - at production rate), which, in my opinion, is clearly wrong, or at least not exactely right. You would want to know "who is busy with what and how long does this take"? As long as you are having all processteps during set-up being performed as internal steps, the definition holds. But as soon as you are doing steps for preparation or afterwork as external steps the definition doesn´t hold anymore, because then it sololey focuses on the machine itself. You need to handle a broader view telling you what the machine is occupied with and also what operators are is occupied with! During your value creation proces you want to use every resource as efficient as possible under the condition of maximum total effectiveness, therefore you need to take every resource into account for itself.
As long as you are mixing up things you would only be able to use the data as very rough sort of estimate without enabling you to do any planning calculations or things alike based on them. And then you´ll have to ask what is the gain and how much money you want to spend for that.
|
|
|
|
|
08/17/2012 12:59 PM
|
|
|
Taruna,
Another thing you might want to think about before you start measuring / graphing setup times is the definition of "setup time." If everybody is not on the same wavelength as to what the one, correct, accepted definition is, all you will have is a collection of random numbers.
Tom
|
|
|
|
>
Community >
Forums
Send to a Friend
