LDraw.org Discussion Forums

Full Version: Existing Part Edit Requests
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3
(2018-03-26, 20:54)Magnus Forsberg Wrote: [ -> ]
(2018-03-26, 18:48)Chris Dee Wrote: [ -> ]OK - I have reversed the changes and made t04ounit a ~Moved to for r04o1000.

I like the logic of this decision.

Thanks. Please review the (F)ixed parts at http://www.ldraw.org/cgi-bin/ptscan.cgi?q=t04ounit so that they don't linger too long on the Parts Tracker.
[url=http://www.ldraw.org/cgi-bin/ptscan.cgi?q=t04ounit][/url]
(2018-03-27, 7:46)Philippe Hurbain Wrote: [ -> ]Chris, could you rename these tori too, for consistency and primitive substitution?
http://www.ldraw.org/cgi-bin/ptdetail.cg...8qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat

That's done.
(2018-03-27, 7:46)Philippe Hurbain Wrote: [ -> ]Chris, could you rename these tori too, for consistency and primitive substitution?
http://www.ldraw.org/cgi-bin/ptdetail.cg...8qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat

They look good, but are unused. Why were they made?
Quote:I like the logic of this decision.

Me too.
(2018-03-27, 19:17)Magnus Forsberg Wrote: [ -> ]
(2018-03-27, 7:46)Philippe Hurbain Wrote: [ -> ]Chris, could you rename these tori too, for consistency and primitive substitution?
http://www.ldraw.org/cgi-bin/ptdetail.cg...8qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat
http://www.ldraw.org/cgi-bin/ptdetail.cg...4qunit.dat

They look good, but are unused. Why were they made?

I couldn't find "t08qunit.dat" at all in my LDraw folders and/or files but "t04qunit.dat" (from Unofficial\p\) is used in "85543c02.dat" (8x), "85543c03.dat" (4x) and "u9228c03.dat" (2x) all from "Unofficial\parts\".
I just came across parts/3660.dat and noticed that the stud rotation is incorrect.
The part needs both studs to be rotated 90 degrees clockwise, so on the following picture ...
[attachment=3165]
... the 'LEGO' logo would be upside down.
just go ahead, correct it and send the corrected file to the Parts Tracker admin so he can upload it to the Parts Tracker Smile
(2018-04-20, 2:18)Steffen Wrote: [ -> ]just go ahead, correct it ...

Done. For the record: I made the change to 3660s01.dat.
(Now I know how to rotate parts too!) Smile
Part 10p04.dat is a little wrong.  The white dots of the house have to be moved 1 stud to the right. As you can see they're current 8 studs from the side of the baseplate, but it should be 9 studs as can be seen here and here.
(2018-09-03, 15:11)Merlijn Wissink Wrote: [ -> ]Part 10p04.dat is a little wrong.  The white dots of the house have to be moved 1 stud to the right. As you can see they're current 8 studs from the side of the baseplate, but it should be 9 studs as can be seen here and here.

Please send the corrected version to [email protected] and I will do a proxy submit under your name.
I was authoring some Minifig accessory when I came across the measurements.

The gap between thumb and fingers with 3820.dat is 4 LDU or 1,6 mm, however measuring a real hand I end up with 2,6 mm or 6.5 LDU

The gap at 93061.dat is 7.4 LDU

Is there another hand somewhere in the Lib with the correct spacing?
(2018-09-14, 16:36)Gerald Lasser Wrote: [ -> ]Is there another hand somewhere in the Lib with the correct spacing?
It doesn't seem. I don't think that correcting the shape of the hand would cause any problem in existing models if we keep exact position connexion points. "Fingers" length and shape could then be improved.
Not sure if this is the right thread, but it seemed a bit silly to make a whole new one for this.
LDraw part 3626bpm0 (and it's counterpart 3626cpm0) and LDraw part 3626bpmb look very much alike. In fact, I think they both represent 3626cpb0719 except one has wrong colored eyes.

Could someone clarify this?
(2018-12-07, 21:28)Merlijn Wissink Wrote: [ -> ]Not sure if this is the right thread, but it seemed a bit silly to make a whole new one for this.
LDraw part 3626bpm0 (and it's counterpart 3626cpm0) and LDraw part 3626bpmb look very much alike. In fact, I think they both represent 3626cpb0719 except one has wrong colored eyes.

Could someone clarify this?

No, the print on the backside is different. It is at Bricklink as 3626cpb0718.
Who is gonna fix this?

w.
(2018-12-18, 20:47)Willy Tschager Wrote: [ -> ]Who is gonna fix this?

w.
what do you think of this?
Rough draft
[attachment=3344]
I personally think that the new one closes too far.
(2018-12-19, 1:42)Travis Cobbs Wrote: [ -> ]I personally think that the new one closes too far.

You think it is still closing too far?
I didn't explain the picture, new one is in the back, so it is wider.

Measuring the hand, the outer diameter could be slightly larger, by approx .5 LDU. However to keep compatibility at max, i.e. the hand can fit into a stud4, the outer diameter shall be kept.

This compatibility issue raises another point:
- The hand can fit into a stud4 only without having a grip on a bar
- If the hand has a grip on a bar, it does not fit into an anti-stud.

However, I would avoid having to use two different hands? What do you think?
(2018-12-19, 7:20)Gerald Lasser Wrote: [ -> ]
(2018-12-19, 1:42)Travis Cobbs Wrote: [ -> ]I personally think that the new one closes too far.

You think it is still closing too far?
I didn't explain the picture, new one is in the back, so it is wider.

Measuring the hand, the outer diameter could be slightly larger, by approx .5 LDU. However to keep compatibility at max, i.e. the hand can fit into a stud4, the outer diameter shall be kept.

This compatibility issue raises another point:
- The hand can fit into a stud4 only without having a grip on a bar
- If the hand has a grip on a bar, it does not fit into an anti-stud.

However, I would avoid having to use two different hands? What do you think?

I like your solution, but I'll need to check against a real part.

As LDraw parts are not flexible as the real parts and therefore several building options will result in collisions anyway, I would go with just one hand model.

/Max
(2018-12-19, 7:51)Max Martin Richter Wrote: [ -> ]As LDraw parts are not flexible as the real parts and therefore several building options will result in collisions anyway, I would go with just one hand model.
Definitely yes, just one hand!!!
Looks good to me. It still doesn't fit a tile without collision, but it's much better. And that's impossible anyway without some elasticity...
The bottom curve of hand could probably be simplified a bit without losing smoothness.
(2018-12-19, 9:10)Philippe Hurbain Wrote: [ -> ]The bottom curve of hand could probably be simplified a bit without losing smoothness.

Yes, it looks good, but keep the curve in normal res. I see no reason to use a hi-res cyli.

How about the difference in height? 
Compared to ldd data, the top of the hand should be lowered.
(2018-12-19, 7:20)Gerald Lasser Wrote: [ -> ]
(2018-12-19, 1:42)Travis Cobbs Wrote: [ -> ]I personally think that the new one closes too far.

You think it is still closing too far?
I didn't explain the picture, new one is in the back, so it is wider.

Measuring the hand, the outer diameter could be slightly larger, by approx .5 LDU. However to keep compatibility at max, i.e. the hand can fit into a stud4, the outer diameter shall be kept.

Sorry, I thought the newer one was in the front. My bad.
@Travis, I shall use colour in future  Smile

@all, Hand outer Diameter measured:
- normal, unloaded: 5 mm
- grip on a standard bar: around 5,1 mm
- grip on a Tile: 5,3 mm !!!

@Magnus, the measured height of the hand is 4,4 mm (11 LDU) the LDraw hand is 12 LDU high

I used the hi-res because the existing hand also uses a higher resolution for the rounded slope.


Questions/Opinions:
- Shall I lower the top of the hand by one (1) LDU?
- Shall I add the rounded corner at the rop of the fingers (r= around 1 LDU)
- Shall I use standard res (16 segments) of some medium resolution (32) for the larger rounded slope.

Also the intersection between arm and hand needs some minor cleaning as well.
(2018-12-19, 21:32)Gerald Lasser Wrote: [ -> ]- Shall I lower the top of the hand by one (1) LDU?
I think it can't hurt.


Quote:- Shall I add the rounded corner at the rop of the fingers (r= around 1 LDU)
a simple 45° bevel at this size would be enough and woudn't cost an arm, triangle wise.

Quote:- Shall I use standard res (16 segments) of some medium resolution (32) for the larger rounded slope.
There's no need to have a precise cylinder structure there.

Attached my take, with a bevel on top and a simplified curve at bottom. 1ldu top move not done. Notice how the triangulation is done to get as good smoothing as possible.


Quote:Also the intersection between arm and hand needs some minor cleaning as well.
Yes!
I like all the changes Philo did on this updated hand, but I would also like to have the top lowered.
(2018-12-20, 18:00)Magnus Forsberg Wrote: [ -> ]I like all the changes Philo did on this updated hand, but I would also like to have the top lowered.

I will incorporate those
When looking through the "HOLD" parts I stumbled across a patterend Dish 4 x 4 with a note that the vertices do not line up.

I remembered that I had a similar issue with the Ghost Face Dish. Where, when zooming really closely in, the pattern was sliced at a multitude of vertices.

Now I looked into the original, non-patterend, Dish and I see that the cond-lines and quads do not share the same vertices. Even some of the Quads do not align, i.e. every second quad doesn't line up with the previous one.

To make a cleaner template for patterned dishes I would rather clean the 3960 one up, to get a clean template for patterned versions.
(2019-05-27, 10:25)Gerald Lasser Wrote: [ -> ]When looking through the "HOLD" parts I stumbled across a patterend Dish 4 x 4 with a note that the vertices do not line up.

I remembered that I had a similar issue with the Ghost Face Dish. Where, when zooming really closely in, the pattern was sliced at a multitude of vertices.

Now I looked into the original, non-patterend, Dish and I see that the cond-lines and quads do not share the same vertices. Even some of the Quads do not align, i.e. every second quad doesn't line up with the previous one.

To make a cleaner template for patterned dishes I would rather clean the 3960 one up, to get a clean template for patterned versions.

Just fix it.

w.
(2019-05-27, 10:25)Gerald Lasser Wrote: [ -> ]Even some of the Quads do not align, i.e. every second quad doesn't line up with the previous one.
To make a cleaner template for patterned dishes I would rather clean the 3960 one up, to get a clean template for patterned versions.
Yes... I already improved this part in the past, I think that the only way to get a better result (considering limited rotation matrix precision) would be to make s02 cover a 45° span and use it in mirrored/90° rotated way.
(2019-05-27, 11:17)Philippe Hurbain Wrote: [ -> ]Yes... I already improved this part in the past, I think that the only way to get a better result (considering limited rotation matrix precision) would be to make s02 cover a 45° span and use it in mirrored/90° rotated way.

I'll do it
I re-did the subfiles of 3960

s02 -> Increased to 45 degrees and removed condlines
s03 -> surface from s02-slices and s05 condlines
s04 NEW condlines 45 degree slice
s05 NEW complete set of condlines, can be used also for patterned dished

Now all gaps are closed and generating patterns should be much easier now!

I will send s02 and s03 to Chris
That was quick. Thanks for the job.

w.
Dear all,

What do you think of moving the Dish 3x3 to 48-segments from the current 16.

See the Orient Pattern as comparison:

[attachment=3708]
Looks much better indeed.
As a rule of thumb, if r>20 ldu, I use high-res. Here r=30, so...
Why can't I see the two posts following this post?
One of them even have a different colour. It is lilac, instead of light blue and lighter blue.

[attachment=3709]
(2019-05-30, 17:41)Magnus Forsberg Wrote: [ -> ]Why can't I see the two posts following this post?
One of them even have a different colour. It is lilac, instead of light blue and lighter blue.

Purple means the post has been soft deleted. This means that Damien probably deleted his post and Philo’s post was a reply to that. I’ve restored both.
(2019-05-30, 9:45)Philippe Hurbain Wrote: [ -> ]As a rule of thumb, if r>20 ldu, I use high-res. Here r=30, so...

So, here we go, I have modified the basic files of43898 to use 48 Segments

43898 -> changed to 48 segment prims and added "arings" where needed
s01 -> changed to 48 segment prims and added "arings" where needed
s05 NEW complete set of condlines, can be used also for patterned dishes

In addition I modified the only existing patterned version on the PT to be on 48 segments and I added five other patterns:

[attachment=3710]
(2019-05-30, 18:04)Orion Pobursky Wrote: [ -> ]Purple means the post has been soft deleted. This means that Damien probably deleted his post and Philo’s post was a reply to that. I’ve restored both.

Hello, you are right, I've answered with my cell phone, and for some reason forgot that when I do this, only the "Quote" part of the post is actually posted.
As I didn't to rewrite what I've written, I choose to delete my post.

I wasn't aware that it would have also deleted Philo's one, sorry for that.
I noticed yesterday that the snowshoe, coming with the current Arctic Sets, is actually another type.

Now comes the dilemma. In our library there is the 30284, the original one and the 11187 which is currently listed as an alias.

At BL and Brickset however, the 11187 is NOT the same design as the 30284. Moreover there is another variant, the 28263, which is an alias to 11187.

The 11187 is official, but in reality it has another origin that 30284. Compared to 30284 it has a short front end and a long backend.

Ideal would be to change 11187 to the correct geometry and add an alias for 28263.

With that approach a complete compatibility cannot be guaranteed due to the different lenghts and existting models may face collision issues (though I think there is only a slim chance...)

[attachment=3787]
(2019-06-26, 11:43)Gerald Lasser Wrote: [ -> ]Ideal would be to change 11187 to the correct geometry and add an alias for 28263.

With that approach a complete compatibility cannot be guaranteed due to the different lenghts and existting models may face collision issues (though I think there is only a slim chance...)
Annoying... I think that we can't simply correct geometry, as the distance between attachment point is wrong (stud and back half bar - if this half bar can be considered as an attachment point). What we do in this case is to create correct 11187a and obsoletize 11184.
(2019-06-26, 15:12)Philippe Hurbain Wrote: [ -> ]Annoying... I think that we can't simply correct geometry, as the distance between attachment point is wrong (stud and back half bar - if this half bar can be considered as an attachment point). What we do in this case is to create correct 11187a and obsoletize 11184.

Thanks for the advice Philo!

I will go ahead and
  • brush up the new 11187a with thenew geometry.
  • Add the alias for the 28263
  • send the obsolete file to Chris
I did my current favourite set, the Quantum Realm Explorer, and I was at first stunned what I got at the engines. Those are made from barrels (2489) and the result did not look nice.

with current official part:
[attachment=3805]


with updated barrel:
[attachment=3806]

I did a quick mock-up and I think I will upload them, which is more or less ready to use...
[attachment=3803]
Another part used in the Quantum Real Explorer that needs a rework, is the Brick 1 x 1 x 2/3 Round with Bar and Clip Vertical (Part 4735)

Currently the grooves on the part are only edge lines, which make it appear smooth during renders. I have reworked it, added proper grooves and a different clip. The clip needs to be a bit narrower though...

[attachment=3824]

[attachment=3825]

with narrower clip
[attachment=3826]
(2019-07-07, 16:01)Gerald Lasser Wrote: [ -> ]Another part used in the Quantum Real Explorer that needs a rework, is the Brick 1 x 1 x 2/3 Round with Bar and Clip Vertical (Part 4735)

Currently the grooves on the part are only edge lines, which make it appear smooth during renders. I have reworked it, added proper grooves and a different clip. The clip needs to be a bit narrower though...





with narrower clip

I distinctly remember fixing this part. Are you sure there isn't a correct one in the library?
(2019-07-07, 21:15)Orion Pobursky Wrote: [ -> ]I distinctly remember fixing this part. Are you sure there isn't a correct one in the library?

This is the screenshot of the part from my last download of the complete.zip after the 2019-01 update:
[attachment=3827]

This corresponds also to the last !HISTORY entry when you download it from the PT.
(2019-07-07, 22:31)Gerald Lasser Wrote: [ -> ]This is the screenshot of the part from my last download of the complete.zip after the 2019-01 update:


This corresponds also to the last !HISTORY entry when you download it from the PT.

I checked that too. It would have been quite a while ago. I might be misremembering.
I checked all the updates, I only find it in its initial release here: LDraw.org Parts Update 2009-02 - Sep 16, 2009

may be still on your HD?
Pages: 1 2 3