Drawing number the same as part number? Why not? 1

[agthad]

I know this topic has come up before but I can't seem to find definitive reasons of why the drawing number should not be the same as the part number. I'm working on a new system for my company and I believe I've read just about every thread and external reference on the topic of numbering systems. Here's a summary I've put together of all I've found.


Goal

[tab]The goal of any numbering system is to uniquely identify an item.


Best Practices

[tab]• Non-significant, sequential number
[tab][tab]• The part or document number has no meaning itself but simply points to the product information
[tab][tab]• Encoded information in significant part numbers usually only serves one specific group (engineering, purchasing, production, sales, etc.)
[tab][tab]• Time is wasted determining part classifications
[tab][tab]• Mistakes are made classifying parts making them hard to find, often incurring the cost of recreating the part
[tab][tab]• A significant part number is a descriptor (rather than simply a unique identifier) which can be misinterpreted
[tab][tab]• Incorrect initial assumptions and evolving technology render original classifications obsolete
[tab][tab]• Most significant part classification systems fail due to improper classification and overuse of “Miscellaneous” categories due to changing business needs
[tab][tab]• When using PDM software (such as SAP), all item information is already tracked so there is no reason to place that information in the part number
[tab]• Numeric only
[tab][tab]• Alpha characters are on a different part of the keyboard increasing entry time
[tab][tab]• Restricting numbers to numeric characters provides the fastest data entry for heavy users
[tab]• Short
[tab][tab]• 5-7 characters
[tab][tab][tab]• 5 characters = 100,000 numbers
[tab][tab][tab]• 6 characters = 1,000,000 numbers
[tab][tab][tab]• 7 characters = 10,000,000 numbers
[tab][tab]• Data entry errors increase as the number of characters increase
[tab][tab]• 7 is a widely accepted “magic limit”
[tab]• Part numbers and drawing numbers are unique
[tab][tab]• Practice of tying part numbers to drawing numbers was established when item searching was more labor-intensive
[tab][tab]• Each are a unique item/product
[tab][tab]• Documents and parts have different life cycles
[tab][tab]• Drawings are a specification of a part (just like a written description)
[tab][tab]• Multiple drawings may be created of a part or assembly to convey different information
[tab][tab]• Multiple part configurations (i.e. different part numbers) can be represented on a single print
[tab][tab]• Documents get revised for reasons having nothing to do with a part’s definition
[tab]• Parts do not have revisions
[tab][tab]• Follow the principles of interchangeability – Form, Fit, Function
[tab][tab]• If parts can be stocked in the same inventory bin without any negative effects then they are interchangeable
[tab][tab]• If parts are not interchangeable then they should have distinct part numbers


This is the concept I've come up with for my company

[tab]• Non-significant, sequential number
[tab]• 5 or 6 digits
[tab]• Numeric only
[tab]• Unique part, drawing and document numbers
[tab]• No part revisions
[tab]• Create a new internal part number for purchased parts used in assemblies (source control drawings)
[tab]• Can be tied to the SAP database if necessary
[tab]• Engineering department can maintain the Master List



I'm still straddling the fence, though, on drawing numbers not being the same as the part that is on the drawing. (We use SolidWorks so parts and drawings have different file extensions making this easy to do.) Are there any problems this causes that out-weigh the benefit of the numbers being the same? Examples would be appreciated.

Thad, PE (Texas)

 

[KENAT]

One simple reason I know of why part & drawing number aren't always identical is...

Some drawings define more than one part, typically with "-" numbers but not always.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[agthad]

KENAT, that still means the part and drawing have the same base number, correct? I assume you're talking about different configurations of a part.

Thad, PE (Texas)

 

[swertel]

Yes, same base number.

An example:
Drawing number is 12345
Part numbers are 12345-1, 12345-3, 12345-4 (for opposite hand of -3), etc.

Each of the dash numbers is shown on the same drawing 12345.

--Scott

http://wertel.eng.pro

 

[looslib]

I question your choice of "No part revisions".
How do you control the revision of a part or drawing if you don't track them? Using a new number for each change is not effective, and can cause problems with BOMs and other product structure methods.
Does your company use a CAD system to design parts and assemblies and create drawings? Look at how they handle part revisions and incorporate their scheme into what you ant for your company. I know UG/NX uses revisions and will load the latest revision of a component in an assembly.


"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[PeterStock]

I believe he means that there should be no need to track different versions of the same part. If the part is changed such that it is not interchangable for form, fit or function, then it would need a new part number.

Peter Stockhausen
Senior Design Analyst (Checker)
Infotech Aerospace Services

 

[KENAT]

I said with "-" typically not always. I've seen drawings that had at least 2 parts defined on them with different part numbers, not same base # with dashes. They were consecutive numbers, but not quite the same situation as with a base number and “-“. This was in systems that didn’t accommodate “-“ numbers.

When you are talking part numbers, do you mean part numbers for the real thing, as you'd seen in an ERP system or parts catalog etc. Or are you talking Model file V drawing file in CAD?

In CAD, we have revisions of models in the file properties, but not in the file names to simplify data management in our non PDM world.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[swertel]

Configuration control is not maintained by revision. The OP does have the concept correct and has explained it clear enough.

Drawings, though, do have revisions. The revisions are for things that do not effect form/fit/function (the interchangeability rule) such as typos, new specs, etc. If anything changes to a part that is not just superficial to the drawing, a new part number is required.

And your reference to UG/NX loading the latest revision, that is an issue with File Management (with or without PDM/PLM) and not one of part numbering or configuration management. I don't believe the OP has intended to bring this topic into the realm of CAD and file management, but I could be wrong.

If that is within scope of the OP, then there is good reason to not include a revision deliminator on the solid model file and only put it on the drawing file. But then you get into things like what is the "master"? The part or the drawing? If you are truly paperless and don't use the drawing for anything, then the solid model is the "master" and it should have revision control because things will change in the solid that do not affect interchangeability. But, if drawing notes are used to clarify requirements, then the solid can not be the master because those notes are not transferred via the solid model (file) and thus the drawing becomes the "master" by default.

Determining the "master" also matters because of the issue of "faking" dimensions. If the drawing and/or solid don't match - via fake dimensions, non-incorporated deviations, notice-of-revisions, etc. - then which is used to create and inspect the end-item?

--Scott

http://wertel.eng.pro

 

[KENAT]

looslib, while I'm aware of places that use 'significant' revisions, the classic way is as Pete & swertel say that real life 'parts' don't have revisions, just part numbers.

Parts made to drawing rev A should be fully interchangeable with parts made to drawing rev B etc. or else they should have been new part numbers. This is effectively the definition given in ASME Y14.100 of when a new part number is required. When a new part number is introduced, on of the aspects of this is chasing it up through the documentation until you reach a level that is interchangeable - sometimes introducing new assy numbers.

This can, as you point out, be a pain for some types of ancilliarly documentation but relying on significant revs has its own issues.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[agthad]

Thanks everyone for the input so far. In the interest of keeping us all on the same topic, yes, I am referring to actual part and drawing numbers, not file names.

At present, I'm still leaning toward the part and drawing number being the same. I can't yet justify the added confusion of them being unique for the occasional exception where it would be beneficial. Maybe I'll see it differently in the future.

I'm still looking forward to more discussion of the matter.

Thad, PE (Texas)

 

[KENAT]

I think your plan of keeping them the same will probably work (untill it doesn't at least;-))

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[shoelace]

Yes, keep the part (or assembly) and drawing numbers the same. Only one part detailed per drawing (multiple parts detailed on one drawing is old-school that was of particular value in the days of manual drafting on a drafting board ... and Autocad emulated this old approach).

Having the part numbers and drawing numbers different would be a big mess on modern Cad platforms IMO (eg Pro/E, SolidWorks)

 

[RTTaylor]

One of your bullets states:
• Multiple drawings may be created of a part or assembly to convey different information

How would your document numbers differ when multiple docs are needed? (I am assuming the part modifier is the same "######-01" and you have multiple docs...

I decide on this approach:
#####@@@-##
For a board design: 12345ASY-00 =Assembly(dwg), -01, -02 =BOMs(part), 12345PWB-00 =PWB Fab(dwg), 12345SCH-00 =schematic (dwg)

 

[agthad]

RTTaylor, that's a good approach. I've really decided to stay away from alpha characters, though. The bullet points in my original post are what I had found while researching the topic. I think if I need multiple drawings of a part or assembly I'll just make it on a different sheet of the main part/assembly drawing.

Thad, PE (Texas)

 

[tloutzen]

I'll just list what we currently do:
Drawing number: 123456
Part number: 123456-001

Revision for Class 1 change (revision will affect form, fit, or function) results in need for new part number: 123456-002 (-001, -002 not interchangeable).

Class 2 change does not cause interchangeability issue; this change can be to correct drawing error, add alternate material, loosen tolerance, and so on.

Any change to drawing by Engineering Change Request results in review of interchangeability issues and classification of type of change. Any change to drawing results in bump of revision letter (rev A to B).

 

[ptruitt]

There are two good books that defend two different approaches. Both approaches are fine but, after learning the nuances of each one, you can determine which is best for you:
1- "Engineering Documentation Control Handbook", by Frank Watts
2 - Bills of Material for a Lean Enterprise" by Dave Garwood

I am sure that there are other good books, but these are 'deep' and current. I have been trying to convince folks of my own personal opinions on this subject but I have come to the opinion that this is a deep subject that requires serious study instead of a quick discussion.

 

[pejaer]

Thad-

I am curious if during your research that you found anything other than "data entry time" as a reason not to go with alph characters?

For what it is worth, I created a system recently. I am using A123A as my format; since I am a small custom design firm, I do not need more than 5 digits and data entry is not an issue either. Furthermore, the first 4 of the 5 are my root and the last is a suffix for design variations.

IMHO:
Part numbers and drawing numbers should be the same.
Adding a table for design variations are not old school, but essential for R&D development/pre-production.
PNs should be dumb (opposed to smart numbers)
Models can be revised; it is a good check for the "changing form fit and function?" question; that is if you are changing the model, the answer is likely "yes".

The suffix in my system is the "revision" to the design, hence it automatically changed the PN while maintaining history. I use a REV number for updating drawings.

I am curious about limitations to the apha numbers---I don't think it will be a problem, but I am worried about other standards and bar coding---it is another reason I stayed away from dashes and other strange characters.

Paul

 

[ctopher]

Part/drawing numbers - same.
Assy/drawing numbers - same.

Chris
SolidWorks 09, CATIA V5
ctopher's home
SolidWorks Legion

 

[naratama1980]

Out of the box,
As project eng, i always find typical project here (Around south east asia) if project doc number (Whatever purpose i.e drawing, pfd, p&id, requisition doc, datasheet) always different with part number. I usually in charge into onshore, offshore oil&gas facility.

Let say we'll adjust same number, do you think will not make confuse management when we put into server database? Please note if every project have own purpose and number, include field development, unless if your manufacture.

Best regards,
Naratama
at south jakarta - Indonesia

 

[ptruitt]

I don't think any of the best practices I have read would have room for "A123A". The problem is not with the first "A", but with the (significant) last "A". It 'could' work, if the document number was the first four characters and all five characters represented the full, hybrid part number. In other words, the last 'A' would be the 'tab'. But you would still need a revision designation for the document so you would end up with A123 Rev.A for the document and A123A for the part number.

I set up a slightly similar numbering system. Consider A12A34-01 Rev.A. The A12A34 obviously includes a letter in the first and third positions. But there is no meaning implied by the letters. I use them there for two reasons only: The first is to get more numbers than possible from six numbers (999999). The second reason is to force people to say the number in two syllables, reducing the chance that the listener will expect three syllables or some other lingual characteristic. The format complies well with what Frank Watts recommends but it could also skip the tabulation/hybrid format and be used as "A12A34 Rev.A" (a document) or "A00A34" (a part number), complying well with Dave Garwood. Part numbers never have revisions, as all mainstream best practices and Standards dictate.

It is important that people read the books. This is a deep subject and the longer poor systems are used, the harder it is to make the changes that will inevitably be required.