Belts not displaying

[BillM]

Art

I created a timing belt example similar to the one you created in your Ticker Introduction (Part 4, ~time 27:55).

I used the default Timing Pulley object creation for a T20N pulley wheel/pinion pair. I copied the pinion so I could do a 3-pulley system. Selected the pulleys in a CCW and then linked the selected pulleys. No belt was displayed. I verified that the pulleys were connected by running the simulation with cartesians selected.

I have also loaded your example Collections project file and noticed that the belt between the timing pulleys was also missing. I don't know if this is a graphics card issue or a software issue.

My computer is a Dell 7506 2n1 computer running Windows 10.

Bill Michael

[ArtF]

Hi Bill:

Sorry for the trouble. The problem this week that crashed those files I think is now hidden by the
corrections I put in, but the material of the belt is somehow wrong or failing. I will add some more diagnostics
tomorrow and we'll see where the original problem resides. It is likely the material itself. Can you tell me if the belt shows
up if you select "Wire Frame" and try to simulate it? If it does, can you click on it and try to change its material,
( even though you cant see it?).
I suspect we'll find something about that material is bad on your machine, so I will change the material type default
for belts and see what it does, I'll post when I have the corrected file uploaded.

Thx
Art

[ArtF]

Hi Bill :

Version 1.007B is now online, it has additional belt safeties in creation, and fault logs in that area to tell me exactly what is
failing. Try it when you wish and let me see the last 20-30 lines of the logfile, it should point me to the offending line. Its
possible the belt may show up with this change.

Thanks
Art

[BillM]

Art

Before I loaded the new version I did run the simulation that showed the pulley belt failure. When I turned on Wireframe, I was able to see an outline of the belt but could not select the belt or change it's material.

After loading the V107B, I checked out my pulley belt program and could not see the belt unless I selected Wireframe. The following Log output file is for the program with three pulleys.

I also created another 3 pulley program tand included a ball launcher. The balls bounced off the unseen belt which at least confirmed that the simulation knew the location of the belt.

As an interesting observation: The ball launcher would only launch balls when Gravity was turned on. Is this normal?

Bill Michael



Tickerlog.txt excerpt:
er\modeldatabase.cpp(119): Initializing belt data.
er\modeldatabase.cpp(227): Initializing belt motion contour.
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(792): Mesh Stage2: belt
ker\ticker\model.cpp(801): Mesh Stage 3: belt
\tekniquemanager.cpp(171): Texture 1 attempt
\tekniquemanager.cpp(173): Texture 1 complete
\tekniquemanager.cpp(192): Texture 2 complete
\tekniquemanager.cpp(197): LUT complete
ker\ticker\model.cpp(809): Mesh Stage4: belt
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(811): end Mesh Create: belt
er\modeldatabase.cpp(242): Initializing belt Shader. #1
er\modeldatabase.cpp(249): Initializing belt Shader. #2
er\ticker\drawable.h(85): Instanced Object init.
\tekniquemanager.cpp(161): Instanced Binding
\tekniquemanager.cpp(166): Instanced Binding complete
\tekniquemanager.cpp(171): Texture 1 attempt
\tekniquemanager.cpp(173): Texture 1 complete
\tekniquemanager.cpp(192): Texture 2 complete
\tekniquemanager.cpp(197): LUT complete
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\ticker\graphics.cpp(305):
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\tekniquemanager.cpp(227): Shade Failed
er\modeldatabase.cpp(252): Initializing belt Shader. #3
er\modeldatabase.cpp(268): Initializing belt Shader.
er\modeldatabase.cpp(278): Initializing belt constraints.
er\modeldatabase.cpp(288): Initializing belt complete.
icker\tickerview.cpp(177): Remove Objects.
er\ticker\ticker.cpp(262): Exit App.
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))

[ArtF]

Bill:

Thanks for the log, I will analyse it and create some fixes.

>> The ball launcher would only launch balls when Gravity was turned on. Is this normal?

Yes, I decided that without gravity launched articles would sit still, I forgot however that
it can launch things with a speed vector... so I will undo that and make them active even without gravity. Thanks for the comment.

Art

[ArtF]

Bill:

Version 1.008B is online. Your log showed the belts were being initialized twice at creation, once as a default single entity , then properly
as an instanced entity which may be draw multiple times in the shader. Its very possible your system didn't like this, and mine was OK with it.
It also made the log read somewhat psychotically. This version proper only inits the belt once as an instanced object. The log reads more linearly now
so if this doesn't fix the trouble, I can at least see exactly where we're failing.

Thanks for that log, it led to seeing this fairly serious error on my part.

Art

[BillM]

Art:

I loaded v08B. The belts still do not display but are there if I turn on Wireframe during the simulation run.


Bill

Ticker log file excerpt:

er\modeldatabase.cpp(119): Initializing belt data.
er\modeldatabase.cpp(227): Initializing belt motion contour.
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(792): Mesh Stage2: belt
ker\ticker\model.cpp(801): Mesh Stage 3: belt
ker\ticker\model.cpp(809): Mesh Stage4: belt
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(811): end Mesh Create: belt
er\modeldatabase.cpp(242): Initializing belt Shader. #1
er\modeldatabase.cpp(249): Initializing belt Shader. #2
er\ticker\drawable.h(85): Instanced Object init.
\tekniquemanager.cpp(161): Instanced Binding
\tekniquemanager.cpp(166): Instanced Binding complete
\tekniquemanager.cpp(198): LUT complete
\tekniquemanager.cpp(200): Trilinear Sampler complete
\tekniquemanager.cpp(210): Instance Shader Init
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\ticker\graphics.cpp(305):
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\tekniquemanager.cpp(234): Shade Failed
er\modeldatabase.cpp(252): Initializing belt Shader. #3
er\modeldatabase.cpp(268): Initializing belt Shader.
er\modeldatabase.cpp(278): Initializing belt constraints.
er\modeldatabase.cpp(288): Initializing belt complete.
icker\tickerview.cpp(177): Remove Objects.
er\ticker\ticker.cpp(262): Exit App.
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))

[ArtF]

Interesting log. Its pointing to your system refusing to maker the shader from its blob.
I ve added a few debugs so I can see where its failing. I can duplicate your error by deleting the
file PBRVertexInstancedShader.cso in the shaderbins folder. (That files does exist on your system right? )

V1.09b is online to test that theory.
In any event, this log will tell me the size of your shader blob and if its failing on its load or its creation.

Thanks
Art

[BillM]

Art

I've loaded v1.09B. The results are the same. I do have the file PBRVertexInstancedShader.cso on my system.

Bill

Tickerlog excerpt:

er\modeldatabase.cpp(119): Initializing belt data.
er\modeldatabase.cpp(227): Initializing belt motion contour.
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(792): Mesh Stage2: belt
ker\ticker\model.cpp(801): Mesh Stage 3: belt
ker\ticker\model.cpp(809): Mesh Stage4: belt
ker\ticker\model.cpp(769): Mesh Create: belt
ker\ticker\model.cpp(811): end Mesh Create: belt
er\modeldatabase.cpp(242): Initializing belt Shader. #1
er\modeldatabase.cpp(249): Initializing belt Shader. #2
er\ticker\drawable.h(85): Instanced Object init.
\tekniquemanager.cpp(161): Instanced Binding
\tekniquemanager.cpp(166): Instanced Binding complete
\tekniquemanager.cpp(210): Instance Shader Init
ker\vertexshader.cpp(22): Instanced Shader loading
ker\vertexshader.cpp(28): Instanced Shader blob size: 61260
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\ticker\graphics.cpp(305):
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
\tekniquemanager.cpp(234): Shade Failed
er\modeldatabase.cpp(252): Initializing belt Shader. #3
er\modeldatabase.cpp(268): Initializing belt Shader.
er\modeldatabase.cpp(278): Initializing belt constraints.
er\modeldatabase.cpp(288): Initializing belt complete.
icker\tickerview.cpp(177): Remove Objects.
er\ticker\ticker.cpp(262): Exit App.
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))
er\ticker\ticker.cpp(71): Exception code: 0xe06d7363 ((null))

[ArtF]

Bill:

It seems the error is coming from the DX11 subsystem in windows when I attempt to create the
instance shader on your machine. There was a guid error in instance shaders a while back. What version
of windows is this and what is your video card type? It will allow me to investigate why a system might reject a
call to this type of shader.
As we get more feedback I can determine if this is systemic, or some esoteric code call that only some
systems hate or like. I use a rdx2070S as my card and its rather powerful compared to may older cards,
so I'm hoping it isn't a card issue. It may be a driver issue if you haven't updated the video driver in
some time.

Thanks for the tests,

Art