LEGO Set Restoration -- Creating an 'Original Parts' resource

[Fugazi]

When faced with the prospect of restoring a classic LEGO set, rebuilding a childhood set from a mix of parts, or evaluating the authenticity of a set bought second hand, the same old issues crop up for the perfectionist in me. I like to make sure that my restored set is made of correct vintage parts, with vintage colour, vintage clutch and vintage moulding characteristics. Sometimes it's easy to tell apart a 1980s part from a modern one, using mould variations, colour changes, mould markings. Buying a 'vintage' part off Bricklink often proves more difficult, especially in the case of common parts -- for instance standard bricks and plates. More often than not a part will be described as New or Used, without further details. The seller most likely won't know the approximate age of the part, and certainly not the original set it came from.

Which brings me to this idea I'm bringing up for discussion. Yesterday I opened a mint polybag, set 30210 Frodo With Cooking Corner (Brickset). Could I write down an inventory of this set that would help me to rebuild or restore it in the future, or make it possible to determine if other (opened) sets like it are made of the correct 'vintage' parts? This would involve describing the 'age' of the parts this set is made of. And the best and (to my knowledge) only consistent way to do this would be to inventory the mould numbers used for each part.

What are the mould numbers? Most parts nowadays have engraved somewhere both the part number and the mould number. In the example below, a picture borrowed from member tomdobs55, 3710 would be the part number and 58 the mould number. '2' is this brick's position within the mould, and very often this information would be moulded as 2-58 or 2-58. The mould number is a rough estimate of the part's age, mould numbers increasing over time as moulds wear down and are replaced. Very common parts and/or parts introduced a long time ago generally have higher mould numbers, recently introduced parts start at mould number 01 and work their way up over the years (if ever).

In my previous example, the 'Vintage Inventory' of set 30210 would look like this:

part number - mould serial number

2343 - 03

2489 - 09

3004 - 183

3020 - 77

3021 - 45

3040 - 27

3069 - 62

3626 - 41

3741 - 11 and 12

3742 - 07

3794 - 32

4032 - 16

4073 - 41

4528 - 03

4599 - 07

4865 - 15

10048 - 01

52107 - 05

64647 - 02

unreadable mould number: minifig torso, 3062, 41879, 95228

no mould number: 33078

Note that I don't keep track of part colours or the number of each part in a set. Such information is possibly superfluous for this purpose.

If a few more people were to do the same with their copy of this polybag, we would obtain a range of mould numbers for most parts in the set. This range could be very small (a single mould number for some parts in every 30210 polybag produced) or much wider. It wouldn't be the last word on the 'age' of the parts in this set, especially in this case where production may still be ongoing. But it would still constitute a useful resource for:

- finding appropriate vintage parts for rebuilding/restoring this set in the future, and

- verifying the authenticity of a second-hand/opened set.

There are limitations to this concept, including:

- Not all parts have readable mould numbers

- Parts older than circa 197x? have no mould numbers

- Inventorying a set is time-consuming, and

- Requires either a mint set or a set kept separate from other parts since its opening.

But I still think this approach deserves consideration. Such a resource would grow more valuable over time, and as older sets are added to it. It might not matter today whether my Frodo polybag has the right parts, but in 50 years time when mint 30210 sets might not be available any longer, collectors would want to have a reasonable guarantee that their Frodo set is made of 2012 parts rather than much later ones. This would help the obsessive-compulsive collectors but also the more casual ones who are only trying to make sure that their classic sets are reasonably accurate and not a hodge-podge of parts from all eras. Selling sloppily restored sets as 'authentic original parts' for a premium would become more difficult. But this would also help people Bricklink the right parts for their needs. For instance, if I need 40 1980s-vintage yellow 1x2 bricks to restore my Yellow Castle and the seller agrees to pick those that have a mould number of (fictitious number) '70' and lower, then I can at least be reasonably confident that the bricks I'm paying for are 80s bricks and not more recent ones with different clutch, colour or moulding properties.

Mould number ranges can also be matched with other descriptions that are currently used on Bricklink, such as moulding pip positions or 'Pat pend removed' for a more complete picture.

So while I'm personally convinced that this sort of (long-term) endeavour is worthwhile, there are probably issues that I haven't though of and this is why I'm opening this discussion. One question that I have considered without reaching a conclusion is the format such a resource should take. Ideally this information would need to be matched/merged with an existing online database such as Bricklink or Brickset. How much information would be enough? Every part of every set, or only the most common (and thus more difficult to filter by age) parts of each set, or even only the most common parts of the most collectible sets? How many individual copies of each set would have to be inventoried in order to achieve an representative range of mould numbers, and how should those set copies be selected -- ideally from different countries? Who should be allowed to contribute to such a database? Has anything similar been attempted before?

Any input (comments, suggestions, mockery, potatoes) welcome!

[pinioncorp]

Seems like a ridiculous amount of work for something so frivolous. Obviously it's my opinion, but I can't see the value of this. Does it matter that much whether a brick was from the set or not? The whole point of Lego is that it's a system of the same bricks.

That said, a less accurate but considerably less time consuming method would be to filter brick moulding numbers by year, rather than by set. That way you can know that the bricks included in a used set are from the correct generation.

[Fugazi]

For some people it matters. If I'm MOCing I don't care about the age/source of the parts I'm using. But if I want to display or preserve a classic set, or rebuild a set I had in my childhood, I will try to find parts that are contemporary with the set. It can also become more important when one is trying to restore a set for selling.

I agree with most of what you said though, and the idea is not to find if a brick belongs to a set but whether it's within the correct age range for that set. I think it can prove easier and more accurate to check a mould number than trying to compare clutch, colour variations or other subjective characteristics when looking for a good match.

Your suggestion of determining the appropriate mould numbers for a given year is very good, though for many sets that were available for a number of years the year of manufacture may not match the year of first release. And in order to build a record of mould/year progression, the method is roughly the same -- mould number inventories of a number of sets from each year.

Edit: What you're suggesting would look roughly like this (fictitious numbers), only much bigger:

[davee123]

Honestly, I think the biggest problem is coordinating with people, and the volatility of the data.

Most people aren't interested in being that anal. And a lot of people may volunteer their information without actually being very careful-- they think they're helping, when in actuality, they may be screwing up data. Suppose (for instance) that someone constantly transposed the cavity # and mold #, where both were relatively low and indistinguishable? You've got to make sure that you only attract persnickity LEGO experts, and you've got to make sure they keep at it.

It's a hugely difficult task-- at the lest, you'll want representatives in the USA, Europe, and the far east, because sets made for different markets may have been packed with totally different types of inventories. If you open up to letting anyone submit entries, you'll have to get a way of vetting their submissions-- you don't want someone that's sloppy submitting an inventory. And if you only rely on a list of pre-approved submitters, you'll probably have to be very pro-active in identifying quality candidates.

That aside, the data is pretty ... difficult. LEGO will do some pretty strange things. They may (I don't know) keep a bin of 1x6 bricks in limbo for a few years until they're ready to go into a set. Suppose (for instance) if one of their packing facilities had an extra 20 bins of 1x6 bricks, which weren't used because that facility never packed any sets containing that piece. So they sit there for a few years and eventually get "forgotten about" until the facility closes down or does some house-cleaning, at which point the elements resurface and get shipped out somewhere else to get used. Similar things have happened, so it's difficult to know "for sure" when the life cycle for a particular mold was used.

Also, some sets get "refreshed" with new inventories. The worst offender is Dacta/LEGO Education, who will take a set like 9320 ("Journey into Space", AKA "Voyage into Space"), and totally re-structure the inventory. I bought a copy in 2003-ish, and it had a HUGE amount of chrome. Then I bought another copy in about 2005-ish, and it had a TOTALLY different inventory with lots of trans-neon-green parts. I think that's the reason that BrickLink says the new 1x1 cone element (4589b) came out in 1996. It obviously DIDN'T, but I believe that set 9287 has a re-vamped inventory (including the new cones).

However, assuming that you get around those issues, the data itself wouldn't be overly hard to compile. IDEALLY, I think I'd go with essentially 2 tables for inventories:

Inventory Row Records:

- Submission ID (ties to the set number, submitter, etc)

- Quantity (number contained in this submission)

- Extras (number deemed "extra" in this submission)

- Bag number (the ID of the plastic baggie that this element came in)

- Design ID (the 4/5 digit design ID of the element, like 3710 for a 1x4 plate)

- Material ID (ABS plastic, CA plastic, metal, rubber, etc)

- Color ID (could be combined with Material ID)

- Mold variant ID (the type of mold-- IE hollow tubes vs. solid tubes, etc)

- Mold number (the one that LEGO imprinted)

- Cavity number (the one that LEGO imprinted)

== OR ==

- Sub-Submission ID (a multi-part element like a Ninjago spinner)

Submission ID:

- Type of inventory (Set, part, book, etc)

- Inventory type (from MISB, derived, "best guess", etc)

- Submitter's ID

- Time of submission

- Set/Part ID

- Country/region of packaging

- Year (most recent listed year)

- SKU

- Notes about the submission

I ... think ... that should cover pretty much everything, although I admit it would be a PAIN to enter all that data. Imagine going through your copy of the Taj Mahal and checking the cavity and mold number on EVERY 1x1 plate that came with the set, just to make sure that you included ALL the possibilities! Yikes!

DaveE

[Fugazi]

Thank you Dave for your input. You certainly raise points that I hadn't considered.

Most people aren't interested in being that anal. And a lot of people may volunteer their information without actually being very careful-- they think they're helping, when in actuality, they may be screwing up data. Suppose (for instance) that someone constantly transposed the cavity # and mold #, where both were relatively low and indistinguishable? You've got to make sure that you only attract persnickity LEGO experts, and you've got to make sure they keep at it.

I understand this problem. On more recent parts the mould vs cavity (thanks for pointing the correct term!) confusion is lessened as the mould number is often underscored. On older parts it may become problematic, and sometimes only one number is present (the cavity number?). I would probably make mistakes myself. I guess that by picking dedicated collaborators and keeping records so that any unusual reports can be double-checked, good quality data can be obtained.

It's a hugely difficult task-- at the lest, you'll want representatives in the USA, Europe, and the far east, because sets made for different markets may have been packed with totally different types of inventories. If you open up to letting anyone submit entries, you'll have to get a way of vetting their submissions-- you don't want someone that's sloppy submitting an inventory. And if you only rely on a list of pre-approved submitters, you'll probably have to be very pro-active in identifying quality candidates.

I agree, and I think that the second option is more appealing even though it means the project progresses more slowly. I would rather collect a little amount of good data rather than a huge database of dubious information. There's no time frame or performance objectives involved here -- it's a long term project that I think should be tackled carefully.

That aside, the data is pretty ... difficult. LEGO will do some pretty strange things. They may (I don't know) keep a bin of 1x6 bricks in limbo for a few years until they're ready to go into a set. Suppose (for instance) if one of their packing facilities had an extra 20 bins of 1x6 bricks, which weren't used because that facility never packed any sets containing that piece. So they sit there for a few years and eventually get "forgotten about" until the facility closes down or does some house-cleaning, at which point the elements resurface and get shipped out somewhere else to get used. Similar things have happened, so it's difficult to know "for sure" when the life cycle for a particular mold was used.

True, but the purpose of such a repository would be to determine which parts are likely to belong in a given set, not the life cycle of a particular mould. So if there are odd findings, such as a batch of 1999 parts resurfacing in 2005, then it would validate the use of those parts in sets released in 2005. Considering this, instead of giving a mould number range for a part every year I realise that the enumeration of every mould number used that year would be more informative.

Also, some sets get "refreshed" with new inventories. The worst offender is Dacta/LEGO Education, who will take a set like 9320 ("Journey into Space", AKA "Voyage into Space"), and totally re-structure the inventory. I bought a copy in 2003-ish, and it had a HUGE amount of chrome. Then I bought another copy in about 2005-ish, and it had a TOTALLY different inventory with lots of trans-neon-green parts. I think that's the reason that BrickLink says the new 1x1 cone element (4589b) came out in 1996. It obviously DIDN'T, but I believe that set 9287 has a re-vamped inventory (including the new cones).

Sets with availability spanning many years are a headache -- their inventory will obviously not only reflect the moulds used for their parts in the year of release, but also in all the subsequent years they were also manufactured. Not to mention the instances of refreshed inventories you describe. Which is why using information from only the sets that had a short life span is tempting. In which case the information generated would be aggregate -- not the specific moulds used for one part in one set, but for that part in the sets manufactured that year. It's less precise, but I could live with using a replacement part that I know was manufactured the right year though maybe not in that specific set. It's a highly debatable point of view, of course.

However, assuming that you get around those issues, the data itself wouldn't be overly hard to compile. IDEALLY, I think I'd go with essentially 2 tables for inventories:

Inventory Row Records:

- Submission ID (ties to the set number, submitter, etc)

- Quantity (number contained in this submission)

- Extras (number deemed "extra" in this submission)

- Bag number (the ID of the plastic baggie that this element came in)

- Design ID (the 4/5 digit design ID of the element, like 3710 for a 1x4 plate)

- Material ID (ABS plastic, CA plastic, metal, rubber, etc)

- Color ID (could be combined with Material ID)

- Mold variant ID (the type of mold-- IE hollow tubes vs. solid tubes, etc)

- Mold number (the one that LEGO imprinted)

- Cavity number (the one that LEGO imprinted)

== OR ==

- Sub-Submission ID (a multi-part element like a Ninjago spinner)

Submission ID:

- Type of inventory (Set, part, book, etc)

- Inventory type (from MISB, derived, "best guess", etc)

- Submitter's ID

- Time of submission

- Set/Part ID

- Country/region of packaging

- Year (most recent listed year)

- SKU

- Notes about the submission

I ... think ... that should cover pretty much everything, although I admit it would be a PAIN to enter all that data. Imagine going through your copy of the Taj Mahal and checking the cavity and mold number on EVERY 1x1 plate that came with the set, just to make sure that you included ALL the possibilities! Yikes!

This kind of thorough inventory could be used in different ways, but it may go beyond what would be required for this specific purpose. Depending on how many assumptions and approximations we can make and still remain relevant, would it be possible (and desirable) to limit the workload? Is it important to know how many of each part come from a specific mould, or is the list of mould IDs for each part in a given set sufficient? Can we assume that the same part in different colors will come from roughly the same moulds during a given year, or will there be a huge disparity? Do we need the cavity number to double-check for eventual mistakes (swaps), or do we trust the people doing the inventory? Do the end-uses of this data require exhaustive and precise records, or is a decent approximation of reality good enough?

And perhaps most importantly, would anyone care?

[SNIPE]

I like what your doing, bricks with an aging look or that have a different mould, put back to make an original set.

New parts might not look aged enough and you learn a little from lego history and how they 'inproved' brick manufacture and design ,and it gives you more knowledge of more parts that are rare because of their mouldings.

It will probably be a little hard to simply pick up or view these old parts in detail to verify if they are suitable or rare but if thats what you want to do then thats that.

Edited November 6, 2012 by SNIPE

[davee123]

True, but the purpose of such a repository would be to determine which parts are likely to belong in a given set, not the life cycle of a particular mould.

Isn't it the point, though? IE, you want to know whether or not a particular mold's life cycle is "reasonably" within a certain range? Essentially, my point was more along the lines of: if you determine that a particular mold's lifespan was from (say) 1982-1991, and then, when you try to assemble a set from 1994, and find that mold included, would you assume that it's an erroneous data point? What if the set is from 1999, or from 2006? Personally, I would guess that the answer is that you can't prove an *end* of a particular mold's life-cycle, but you probably COULD prove a beginning.

Is it important to know how many of each part come from a specific mould, or is the list of mould IDs for each part in a given set sufficient?

Probably not much more important than identifying mold lifespans :) Ultimately, I'd say if you're looking to spearhead such a project, then it's your call. But if you want to be as anal as possible, such that nobody could ever demand anything further, then that's probably what you'd have to include.

Can we assume that the same part in different colors will come from roughly the same moulds during a given year, or will there be a huge disparity?

I would expect some degree of disparity, but how much would be ... interesting to see. I wouldn't assume anything until you had a pretty large chunk of data!

Do we need the cavity number to double-check for eventual mistakes (swaps), or do we trust the people doing the inventory?

I'd probably get a good chunk of data for each element first, and try to make sure that you know which is the cavity number and which is the mold ID. And ideally, ask someone at LEGO!

Do the end-uses of this data require exhaustive and precise records, or is a decent approximation of reality good enough?

And perhaps most importantly, would anyone care?

That'll be the big question, I think... My guess is that a lot of people would care, but only from a point of interest perspective. I don't think you'll get a lot of people for whom this is a really critical piece of information, or who would be willing to spend sufficient effort to get an effective data set together.

Honestly, if you want to make it happen, my experience is that you need to make it easy for contributors. Visual interfaces for inputting data, along with providing easy access to similar data so they can double-check themselves as they add new data. Having people submit things like text files or spreadsheets, and having to manually cross-reference things is a big pain, and can quickly discourage would-be contributors.

DaveE

[LEGO Historian]

OK... first of all I want to apologize for my absence (and lack of shipping)... just went thru a very rough period where I lost one of my best friends (died of a heart attack at 49)... and nearly lost one of the most treasured people in my life (my 18 month old grandniece)....

Let me just chime in quickly....Fugazi... this is a very very difficult thing to accomplish... only because I have tried it with some of the parts of the 10 6075 Yellow Castle sets I bought back in 1984 on clearance...

I put together all the 462 yellow parts (new condition) from that set... and I found quite a few serious issues... mold number aside.

Here's some of the parts, ALL YELLOW, and the issues with them (USA 6075 Yellow Castle):

---------------------------------------------------------------------------

(71) 1x1 and (183) 1x2 bricks.... all had pip's on the side of the brick... with a bolder font on the studs.

(62) 1x3 and (30) 1x4 bricks.... about 60-80% of these had the pips on the studs, (instead of the sides) although it was a bolder font than later bricks.

(29) 1x6 and (15) 1x8 bricks.... about 50+% of these had the pips on the studs (instead of the sides) and again it was a bolder font than later bricks.

(3) 1x4 arch bricks - all of these had the pips on the normal font studs.

(8) 1x6 arch bricks - all of these had the pips on the sides, with normal font studs.

(6) 1x8 arch bricks - all of these had the pips on the sides, with normal font studs.

(6) 2x2 bricks - all had the pips on the sides, and bolder font on the studs.

(6) 2x3 bricks - w/o cross supports, all had the pips on the sides, with bolder font studs.

(12) 1x2 inverse 45 degree slopes - all had pips on the side, with bold font studs.

(8) 2x2 inverse 45 degree slopes - all had pips on the side, with bold font studs.

(6) 1x1 round bricks - early versions had solid studs, later set versions had hollow studs.

___________________________________________________

TLG often throws in a "monkey wrench" into the best laid plans for trying to recreate original sets. Here's another modern example... the 10152 Maesk Sealand Ship....

The 2004 edition 10152 has Maersk blue 2x4 bricks WITH cross supports.

The 2005 edition 10152 has Maersk blue 2x4 bricks WITH cross supports.

The 2006 edition 10152 has some sets with Maersk blue 2x4 bricks WITH cross supports... but then when TLG ran out of Maersk blue 2x4 bricks in their inventory... rather than create another batch of Maersk 2x4 bricks in that color... they scouted the LEGO model shops for Maersk blue 2x4 bricks. As it turned out, at least one of the shops had some in inventory, and they were sent to TLG to be put into the last of the production run of the 10152. Well as it turns out... these last batch of 2x4 Maersk blue model shop 2x4 bricks were of a very old vintage... without cross supports... dating to the early 1980s... The irony here is that the 10152 was the first set to contain Maersk blue 2x4 bricks... so one would think that all should have bricks with cross supports... and this is not the case in this instance. No other LEGO set has 2x4 Maersk blue bricks without cross supports... except this 2006 set.

So there are a LOT of variables that TLG throws into the equation of what to expect in recreating old LEGO sets... thus making it very difficult.

And as for real old 1960s USA/Canada Samsonite LEGO sets... putting them together becomes nearly impossible... since there are 4 distinct fonts used in many of the sets (no USA Samsonite set should contain less than 3 different fonts).

P.S. This is my little princess that almost didn't make it....

http://www.flickr.com/photos/82930629@N08/8184257095/in/photostream

Edited November 14, 2012 by LEGO Historian