Parts Numbering Scheme

Boy, this is one of my pet peeves, the typical part number is incomprehensible to a human. Untold hours are wasted looking up part nmbers.

I suggest you invent a part number that is descriptive such as

RES 1000 Ohms 1/8 W 5% 0402 SMD Rev 3

This ASCII string can BE THE PART NUMBER. You can add as many fields as needed to completely describe the part

Computers don't need random digits to be the part number. They work fine with ASCII strings.

I would be glad to work with you to invent this new part number system. We can both become millionaires.

Mark

There is also copy and paste.

Regards,

Boris Mohar

Got Knock? - see: Viatrack Printed Circuit Designs (among other things)

void _-void-_ in the obvious place

Agree, if it is smart enough to know that

RES 1.0 KOhms 1/8 W 5% 0402 SMD Rev 3

can be the same part.

Not to mention 1K0

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Dan,

You have some good insights, but let me add a couple of things:

1) One of the biggest frustrations I've had as a working engineer at a small handful of companies is that many systems used for entering BOMs are some horribly old archaic pieces of software that have rotten search facilities. A smart design would let something like "res 2.2k" find any of "resistor 002.2kilohms" "r 2.2 kohms" or "res 2.2kilohms", yet most have limited exact search functionality much less the "fuzzier" search needed to find "r 2.2 kohms".

The problem with this is that you can never foresee all the fields you need. In a reasonably sophisticated manufacturing operator, something like 99% of your resistors will be something like "10k, 5%, 0805," but there's always that design where things like temperature coefficients, the build style (inductive, non-inductive, etc.) matters. This gets even worse with capacitors.

I've had more than one product fail during production due to someone deciding that they could go and substitute some "run of the mill" part for something that had tight specs. To avoid that I sometimes will go to the effort to find a part where the "standard" description (e.g., 2.2uF 0805 25V) specifically _isn't_ "in the system!"

...and even with the exceptions to the rules, it's still a lot easier than not making the attempt.

---Joel Kolstad

Since part numbers are entered by hand over and over again into computer systems the key is to use the smallest number of digits possible.

Part numbers should not attempt to define the part.

I would suggest a 6 digit system xxyyyy where xx is a gross identifier and yyyy is just a sequential number within the xx class.

Do not use dashes or periods in the part number - that is just one more keystroke to enter.

Do not use Alpha characters - you should be able to enter the part number with a key pad for speed.

--

Dan Hollands

1120 S Creek Dr Webster NY 14580 585-872-2606 snipped-for-privacy@USSailing.net

Please explain why:

4387645237-045893-001

is less error prone compared to :

RES 1000 Ohms 1/8 W 5% 0402 SMD Rev 3

There's the old story about an army guy that ordered a small audio transformer to repair a radio and he slipped up on one digit in the part number.

The transformer finally arrived.. on a flat bed trailer truck. and was big enough to power a small city.

As far as unforseen variations, you can alway append a small sequence number at the end as I did using Rev 3 as an example.

Mark

I would suggest that you don't even try to organize the storeroom by part number. If you do then you will have to dedicate enough space for each part to accept the maximum amount you may ever have on hand -- and when you add one miserable little part you'll have to rearrage the whole room.

Any 1/2-way decent stock tracking program will allow you to track the locations of any parts, and to print pick tags with the location for the stock boys. Use one, put the parts where it makes sense, ask the computer when you're looking for parts and tell the computer when you move them.

--
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Bingo!

Yup. . .

Bang-on. BTDTGTTS

Back in the dark ages, our department had 2 copies of the Parts Book:

1 sorted by P#; 1 sorted by description.

Trying to find parts by Description, it was obvious that the Co. had started with a system but (over time) the gatekeeper function had been abandoned. BAD MANAGEMENT DECISION.

(The Co. had also started with a P#-is-indicative system, but abandonded that--perhaps at the same time.)

Yes, that is just how I reasoned when setting up our database. Our part numbers start from 1. Maybe I took the principle too far...

We are up to 1654 now. (BFS17W, a NPN RF transistor. Nice part.)

--

John Devereux

Hello John,

And dirt cheap. My impression is that many young engineers don't even know about these anymore.

Regards, Joerg

typical

So what's a 10% 100k resistor? RB100K ???

I tend to waffle about it, but usually I figure that arbitrary numbers are best -- it doesn't seem there are that many cases where "100k" is the _only_ piece of information you'd like to know about the part, in which case you have to look up the rest anyway.

Engineers often like descriptive part numbers because they rarely have to enter the part number. They may enter it into a bill of material but thats about it.

Manufacturing and Purchasing are the groups that spend lots of time entering part numbers over and over again. When parts are ordered, received, removed from stock etc. Long alph numeric part numbers not only take more time to enter but the increase the probability of error.

It is important to get it right from the start. Changing the system later is a major pain.

The ideal system uses sequential numeric part numbers linked to a standardized description field. Then reports can be generated for the engineers that have similar parts grouped together. This makes it easy to know what parts are already stocked as new products are designed reducing the rate at which new parts are brought into the system.

A standardized description field is designed for each type of component such

RES 5W 104 SM

It takes considerable work to standardize the descriptions and enforce their use but is well worth the trouble.

Dan

--

Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
QuickScore@USSailing.net
www.QuickScoreRace.com


"Mark"  wrote in message 
news:1118156165.514339.259050@f14g2000cwb.googlegroups.com...
> Boy, this is one of my pet peeves, the typical part number is
> incomprehensible to a human.  Untold hours are wasted looking up part
> nmbers.
>
> I suggest you invent a part number that is descriptive such as
>
> RES 1000 Ohms  1/8 W 5% 0402 SMD  Rev 3
>
> This ASCII string can BE THE PART NUMBER.
> You can add as many fields as needed to completely describe the part
>
> Computers don\'t need random digits to be the part number.  They work
> fine with ASCII strings.
>
> I would be glad to work with you to invent this new part number system.
> We can both become millionaires.
>
>
> Mark
>

Yes, that too :) Good job too since I just ordered a reel of 3k. A lifetime supply at the rate we will use them!

Well, it would be if not for RoHS (as discussed in another thread...)

--

John Devereux

I agree completely. I once thought "engineer readable" part numbers were necessary and designed my own system where the part number encoded all sorts of useful (to me) info. This ended up being a nightmare because the person assigning the numbers had to keep coming to me to check that new numbers were correct. Then there were the situations that I hadn't thought of, or were ambiguous. In the end, it generated more work for many people and didn't really improve anything.

If I had to do it again, I would choose a system based on purchasing, inventory, and manufacturing requirements. I'd probably use 2-3 digits to define broad categories (mechanical, electrical, etc.), 1-2 digits for special handling requirements (ESD sensitivity, fragility, humidity, temperature, etc.) and 5-7 digits for a sequential stock number. I'd also include a check digit or two to help detect entry errors (particularly transpositions [430 instead of 403] and phone/calculator substitutions [143 instead of 749]).

I would use numbers with fields separated by dashes. This is both human (numeric keypad) and "barcode" friendly. Dashes reduce human error by breaking numbers into subunits that fit into our short-term memory. Numeric+dashes allows use of the less error prone (and shorter) numeric-only barcodes. Check digits are also required to detect read errors for most barcodes, so including it everywhere keeps the number consistent.

Initial entry of the standardized descriptions should be software-verified, or menu-driven. If the description doesn't match an established template, it should require someone with management-level authority to enter an override code. Systems that allow entry without checks are a recipe for enormous headaches. I would also include a free-form notes field to help cover those special cases nobody thought of.

--
Tim Hubberstey, P.Eng. . . . . . Hardware/Software Consulting Engineer
Marmot Engineering . . . . . . .  VHDL, ASICs, FPGAs, embedded systems
Vancouver, BC, Canada  . . . . . . . . . . . http://www.marmot-eng.com

Hello Dan,

That is exactly the point. You need to be able to enter the PN with just a numeric keypad. Also, in most companies it needs to be suitable for bar coding.

In a controlled environment you can't change them later. This can easily require an ECO re-release of everything that had been designed to date.

Regards, Joerg

We are a small company, and we would like to implement a better components numbering scheme.

The parts numbers need to go into our accounting system and on parts lists, build lists and some schematics, so we don't want them to be too long.

Letters and digits are OK, but each must start with a letter. Sort order is a consideration, because computers sort strings from left to right, often in ascii or similar order. That means each character position has significance and it might be best if every part had the same number of characters.

The storeroom must be organised by part number, so that we can work systematically to find the correct shelf, container, then envelope or reel inside the container. We need to store similar parts together in the storeroom, to make the most of the containers. For example, SMD resistor reels are stored together, with a number of different resistor values in each container. This means SMD resistor part numbers must form a sequence, though not necesarily in order of ohms value.

A part number should at least tell roughly what kind of part it is - SMD resistor, leaded electrolytic, etc, so that means the leftmost characters should carry that info so the sorted list is by part type.

Now, does anyone have a system like that ? Or a different or better system ?

thanks Roger Lascelles

Hello Joel,

At the end it all boils down to how well the search functions on a parts database work. For example if you start a new design you may want to know what kinds of SOT23 transistors are already released and which ones are used on current products. This ensures that you get to share in all the candy that Purchasing receives from the vendors before Christams ;-)

Regards, Joerg

One of the best I've ever seen, used at a pro-audio manufacturer used a typical prefix and and sub-descriptor suffix follwed by the value.

E.g. RA100K0 was a 2% 1/4W metal film resistor with value = 100k RA033R0 was 33R in the same type. Note the use of 4 numerals in both cases to make listings consistent.

RB, RC etc were different families of resistor type / tolerance / rating etc.

Simialar method used for other parts.

I like the text string idea too. But you need to keep to the format - and that may be rather less easy to remember.

Graham