10 replies to this topic

[harcoskiralylany]

Good Morning

 

I just started with a new company in manufacturing and I have a question about scale calibration.

 

We weekly calibrate the scales in the lab using 4 small weight (0.10g,0.5g,1g,10g). We just tick on the sheet if its all good. 

IN one of my old job we used to write the actual weight down and put pass or fail.

We just put the week, and the 5 scales with ID and we just tick under them if they pass.

 

I feel we should do this.

 

I am trying to figure out how to put all this on an excel sheet without creating too much paperwork. We still print them of they are far away to be doing things on the computer (i am trying to move  them that way)

Does anyone has an example if you do the same 4 weight how you do it?

 

Hope this makes sense.

 

 

[MDaleDDF]

I don't use four different weights, and honestly I don't see the benefit of that, but I don't know your lab and the ins and outs thereof, so if multiple weights makes sense for what you're doing, good on you.   For our small scales, I use a 50g scale that came with the machine.   I test all four corners and the center.   If it's outside the deviation parameters in my work instruction, it's removed from service until a tech can repair it, or we replace it, whatever ends up needing to happen.  (This has happened ONCE in my 20 years.   Scales rarely go bad around here, and when they do, it has never been discovered through my weekly check, it's always something more easily noticed, like being dropped, etc.)   IF it passes, which it always does, I sign off and I'm done.   

My situation may be different than yours in that, I have a LOT of scales to test every Friday, so I keep it as simple as possible while retaining my assurance our scales are good to go.

Can you tell me why you guys do four weights?   In the history of your recording, do you find the scale good at one weight and bad at another?   I rarely work with weights in grams, our stuff in the lab is usually oz and into the pounds sometimes...

[Alex A.]

Hi Harcoskiralylany,

 

I think something simple like this would do the job

 

 scale calib example.png   11.93KB   2 downloads

[GMO]

It really depends on what you're weighing on a scale as to how many weights to use.  If for example you're sometimes weighing out 0.3g and 8g on the same set then your range is pretty good.  You want it to span the range you're doing but also to the accuracy you're weighing to.  Possibly it could be argued 3 weights would be enough.  I wouldn't use 1 as you need to check it across the range of weights.  Top pan balances don't necessarily become inaccurate across that range depending on what the issue is.

 

I agree that writing down the actual weight seen is good practice but also having space on your sheet on what the tolerance is but also what was done if out of tolerance.

 

Why not see if they would be open to having a tablet to be more portable?  But either way it's just a simple table.  Scale ID, the weights you use with boxes for actual weights seen plus a comment box if needed.

[G M]

How broadly we test instruments and how detailed the record is dependent on what the instrument is expected to do.

 

The instruments that are used to observe a wide range of measurements are the ones we record the test result for -- because I want the person performing the test to pay attention to how far off it is from the acceptable deviation.  These are also being challenged with a wide range of test pieces or values, to best simulate the range of what they will be used to measure (the expected observation range, +/- 10-20% outside that range).

 

A few instruments that are expected to measure one narrow range or specific value are challenged with the target value, and only the pass/fail of that target is recorded.

[GreyeagleA]

We use a 100g weight for our tabletop scales and a 2Kg weight for the floor scales.  We check them monthly by testing all four corners and the centre of the scale, and I have never found a scale out of calibration during my monthly checks.  Usually we find them damaged from being dropped or used improperly.

 

Our from has a list of the scales and beside each one we record the weight readings we got and either a checkmark for it being acceptable or an x for it being unacceptable.

[harcoskiralylany]

Hi Harcoskiralylany,

 

I think something simple like this would do the job

 

scale calib example.png

 Thank you for your help. This is really useful.

[harcoskiralylany]

I don't use four different weights, and honestly I don't see the benefit of that, but I don't know your lab and the ins and outs thereof, so if multiple weights makes sense for what you're doing, good on you.   For our small scales, I use a 50g scale that came with the machine.   I test all four corners and the center.   If it's outside the deviation parameters in my work instruction, it's removed from service until a tech can repair it, or we replace it, whatever ends up needing to happen.  (This has happened ONCE in my 20 years.   Scales rarely go bad around here, and when they do, it has never been discovered through my weekly check, it's always something more easily noticed, like being dropped, etc.)   IF it passes, which it always does, I sign off and I'm done.   

My situation may be different than yours in that, I have a LOT of scales to test every Friday, so I keep it as simple as possible while retaining my assurance our scales are good to go.

Can you tell me why you guys do four weights?   In the history of your recording, do you find the scale good at one weight and bad at another?   I rarely work with weights in grams, our stuff in the lab is usually oz and into the pounds sometimes...

Hi MDaleDDF 

I am not sure why they started to use 4 weight. I need to find it out. In my previous jobs we only used 1 weight. 

 

Thank you for your help.

[harcoskiralylany]

It really depends on what you're weighing on a scale as to how many weights to use.  If for example you're sometimes weighing out 0.3g and 8g on the same set then your range is pretty good.  You want it to span the range you're doing but also to the accuracy you're weighing to.  Possibly it could be argued 3 weights would be enough.  I wouldn't use 1 as you need to check it across the range of weights.  Top pan balances don't necessarily become inaccurate across that range depending on what the issue is.

 

I agree that writing down the actual weight seen is good practice but also having space on your sheet on what the tolerance is but also what was done if out of tolerance.

 

Why not see if they would be open to having a tablet to be more portable?  But either way it's just a simple table.  Scale ID, the weights you use with boxes for actual weights seen plus a comment box if needed.

Thank you for your help. I am trying to convince them to use a tablet but they are a long way away from it.

[jfrey123]

Whenever I implement a scale testing process, and maybe blame my background being wholly under SQF, but I require a lot of detail in the check form.  It's usually too easy for someone to just "tick a box" and say everything was good.  So instead of a checkmark, they have to record what weight the scale actually displayed when they test each corner plus the middle.  They also have to record the weight of the test weight, and if there's multiple of the test weights, each of them has to have some type of independent identifier that also must be recorded.  Obviously along with the scale number/identifier they're testing that day.

 

I should basically be able to trace each test on any given day to a specific weight that was used (which gets recertified as accurate at least once a year) any time I go through scale check records.

[GMO]

Hi MDaleDDF 

I am not sure why they started to use 4 weight. I need to find it out. In my previous jobs we only used 1 weight. 

 

Thank you for your help.

 

Personally I would span the weights under consideration, even if you're only using two weights as it gives you information around both accuracy and precision but if others have not been picked up for one weight I guess auditors are fine with it.