The 10th Weekly Power-User Challenge (September 15th, 2017): Measuring Wine Carafe

Trying to reach into the screen but it goes black, just....., before I reach the Carafe handle - or maybe uploading it would be better then I can download it...

Alin Vargatu Looks like the file may be missing...

For question #1. 66.627 US fluid ounces?

I cut off the top that was not level because it would not hold wine.

Question #1. 66.589US fluid ounces

1) 66.629 fl. oz. However, with surface tension it could actually hold a little more. I want bonus points for that. ;-)

only six Lines possible, since the total Volume is less then 70 oz

Hello Alin Vargatu

and welcome back

assuming that the handle is full of liquid "showing only the liquid"the volume is 66.589 US fluid ounces

taking into consideration that the top of the handle is going to be empty "some air is trapped in there" the volume would go down to 66.582 US fluid ounces

If I tip the carafe back 4.05 degrees I can get a little more wine in there. New answer for #1 is 66.745 fl. oz.

I made life simple (because I AM an engineer!) and assumed the handle is solid, so contains no wine. Also makes it easier to wash the carafe later!  I also set the top fill line at 10.75", again for practical reasons that you would never fill it to the brim. (Red wine on a white table cloth = very unhappy wife!) That gives me maximum volume of 65.854.

I then used offset surfaces to create 8 solid bodies (one for each of the fill-line sections). And measured the cumulative volumes.

I used Excel (yes I are an engineer!) to do the math. First I calculated a "volume per vertical inch" for each body. Then decided if the fill line needed to move up or down to get the desired volume to decide if I was subtracting from the current body or adding from the body above.

I saved a copy (-2) of the part, then added (or subtracted) that number in the Equation editor. Then remeasured the volumes. Added these numbers to the spreadsheet.

Also, because the carafe holds less than 70, I edited the top label to make it 65 (I already had the body so why not use it?)

My original volumes, new dimensions and new volumes:

10      7.506 oz    1.201"      9.781oz

20      19.911 oz    2.007"    20.005 oz

30    33.372 oz    2.749"    30.055 oz

40    45.126 oz    3.563"    40.438 oz

50    53.658 oz    4.570"    50.430 oz

60    58.900 oz    6.377'    60.213 oz

65    61.814 oz    9.958"    65.113 oz

I could repeat the above process to get closer dimensions but, as they say, "Good enough for government work".

Files in SW2017 SP4.1

One other practical consideration is that the lines are 0.040" thick, so the vertical dimensions should be to the nearest 0.040".

Here are my findings:

1. Volume = 66.674 US fluid ounces

10 oz. : 1.220"

20 oz. : 2.007"

30 oz. : 2.745"

40 oz. : 3.529"

50 oz. : 4.501"

60 oz. : 6.226"

70 oz. : Not possible

Method I used: (Refer my attached model "mf_20170910_Magic Elixir Carafe" SW2017)

1. Extracted internal surfaces by "Surface Offset=0"
2. Trim that surface at the lowest point of the spout.
3. cap the surface by "Surface Fill" and at the same time knit and made a solid body.
4. Got the answer maximum Volume = 66.674 US fluid ounces
5. Made a cut extrude feature (= H vs V table value finder) to find volumes at different heights by utilizing the above (#3) created body. Entered the values in a table
6. Made a "Curve Through XYZ Points..." By using the above (#5) table values (= V vs. H graph)
7. In the sketch "V vs. H graph" graduated the volume axis 10, 20, 30.... and found their corresponding heights as follows.
8. 

9. Using the method @#5, could verify the volumes.

Hi.

The first part was to get max fluid volume(orange juice).

Internal surfaces offset by 0 and cup top to get solid body. Max volume that i got is 66.555

In the next step i have createt a cut plane and volume sensor then run some design studies to get heights coresponding to given volumes and noted them down.

At last i have edited split sketch to adjust scale on the carafe.

I was golfing on Friday, so I'll have to wait to look at this when I get home tonight (can't hit YouTube here).

Pretty funny though, I already have something similar in my slugme presentation.

Hopefully this might encourage the fine folks at SOLIDWORKS to address SPR#588882

(This is still an issue in 2018 Beta 3)

Alin,

When I compare the max volume numbers from 4 of us, we all have the same vertical dimensions (10.853"), but our volumes are not identical. (66.556, 66.557,66.589 & 66.674).

I know in the certification tests there is a tolerance for the answers (I can't remember the exact number).

Which raises the question: To how many decimals can we rely on the Mass Properties results?

OK, this one is extremely frustrating!

1st of all, I want to point out that I did not look at any of the responses here before I posted this.

I have all the volumes, that is not the hard part.  What I was trying to do was parametrically link the ounce notes to the volume at each level, so that you can just change the height of the line and it would update the volume note.  This means that you need to have the volume calculation for all levels available in 1 configuration.....this is the challenging part.

But I am finding that this is really hard.

Things that I have tried:

1) It turns out that you can't use the volume of a specific body in a calculation, it is all or nothing for that configuration (at least, I can't figure out how to do it).  You can create a configuration that only has that body in it, but that causes other problems (see below).

2) If you link the Volume to a global variable, you can't use volumes from other configurations (sort of...see below).  It is current configuration only.

3) Using custom properties:  If I only use custom properties (not configuration specific), I can link a property to each volume of each configuration (then I can link each property to a global variable if I want).  The problem comes when I have to rebuild the part.  I have to rebuild each configuration every single time.  With Solidworks 2015, this isn't so easy.  An easy hack for SW 2015 is to drag the freeze bar up, then down again (which forces a rebuild of all configurations).  This whole constant rebuilding of configurations is very time consuming and frustrating!  And it still doesn't work as expected.  I can link a property to each configured volume (which includes the Carafe).  Then I can link the Global Variables to the custom properties.  Then I can create an equation subtracting the volume of the empty carafe from the liquid volume.  Then I can link another custom property to each of the variables that show the volumes.  It all seems to work great!

Except that, once I tried to link the custom property to the note, for some reason the note will only display the expression, not the evaluated value!  (I read that this is a major problem if you link a global variable to a custom property) Super frustrating!

So I read something about linking the global variable to a dimension, then linking the dimension to custom property, then linking the note to the custom property....but that doesn't work either.  As soon as I link the dimension to the global variable, everything blows up. Something to do with the update order.  It doesn't work.

Then I tried a Design Table, but that has the same inherent problems as the custom properties, except multiplied...because now you are adding in the update step of updating the design table.  You can't update the Model, all configurations and the design table all in 1 step.  You can get to 2 of the 3, but then you need to update the third, which prompts a cycle of re-updating the other 2, which puts the design table out of sync (which means that the number that I have linked to it gets a little * next to it, turning it into text and not a number, so unusable in equations....*sigh*)

Needless to say, on top of all this, Solidworks died on me about 2,463 times (might be a little exaggerated).

So, what I eventually did was the following:

1) Create my fill configurations (with the carafe full at the different levels) and then add each configuration as a part in an assembly.

2) Added the volume of each configuration (including Carafe) as a separate custom property at the assembly level.

3) Used Global Variables and equations to determine the liquid volume of each configuration (subtract the volume of the empty Carafe).

4) Link Those global variables to dimensions in a sketch at the assembly level (this is because I couldn't figure out how to link assembly global variables directly to part custom properties).

5) Create a sketch at the part level and link the dimensions from the assembly sketch to the part sketch.

6) Use the part sketch dimensions to drive part custom properties.

7) Use the part custom properties to drive the linked notes at the part level.

8) I added a BOM to the assembly with some equations only because I originally thought I might be able to link values from a BOM back to custom properties....can't be done as far as I know.

Notes:  It still doesn't update very well.  In fact, I have a massive circular reference between the part and the assembly and the ounces displayed on the part keep coming out to 0.

I have the calculations in a BOM, so I know that the program knows the numbers are there.

Honestly, I don't think that making the numbers parametric can be done with SW 2015.

AARGH!!  I have to stop...this is taking too long to do.

Attached a picture of the volumes at each level and the Pack and Go assembly that I have so far.  I know it isn't a winning submission, but maybe someone else can make it work.

K, here is my official submission (non parametric).  SW 2015.

I used an assembly to create a BOM of my configurations, because looking at the equation driven BOM was easier than going back to assembly visualization every time.  LOL

I was kind of unsure which way you wanted the Carafe, with the heights as integers, or the volumes as integers...so I did it both ways.

I am only submitting the one with the volumes as integers, though.  If you want the other one, let me know.

1. For eternal glory and 1000 points, be the first to estimate the volume of wine this carafe could hold. Provide the result in US fluid ounces, 3 decimal places, lower accuracy level.

66.630 oz, as shown in the picture of the BOM above.

2. For a bit more glory and 2000 points, be the first to place the measuring lines correctly. Please use 3 decimal places for both volume (US fluid ounces) and dimensions (inch). The height of each line is controlled by a global variable, as explained in the video.

I believe that the measuring lines are placed correctly

3. For maximum glory and 3000 points, be the first to demonstrate that the method you used, in placing the measuring lines correctly, is the most efficient one. You can describe your workflow in writing and/or by recording a video.

I offset the internal surfaces (including the handle) 0" and the used the lowest point of the spout to trim the offset surfaces down the spout level.  Then I added a surface to the top and made it a solid, separate body.

Then I Created planes who's definitions were the same as the definitions of the lines (i.e., I made them linked to the global variables).

I created 7 different "Cut with Surface" commands in each configuration. 

I created 9 different configurations:

1) Carafe empty

2) Carafe full to brim

and 1 configuration for each liquid level.

Then I added the part to an assembly and patterned it 8 more times (total 9 parts).  I made each pattern item a different configuration (as shown in the picture above).

Then I added a BOM to the assembly and added a Volume column.

Then I added another column called "Calculated Liquid Volume".  In this BOM I added an equation in which each config's total volume is subtracted from the volume of the Empty Carafe's volume, leaving the volume of liquid.  This column is parametric (meaning that if you change one of the global variables in the part level that control the line heights, these values will update accordingly).

Then I added a "Height of Line" column.  This column is NOT parametric.

Then I played with the Global variables at the part level until the volumes were what I wanted them to be.

EDIT

Not totally sure that this is the most efficient one.  I discovered the "intersect" tool after I posted this which is a lot more efficient, but I didn't want to redo the study. 

I used the BOM for 2 reasons:  It was an easy and fast way to check my numbers after I updated my heights and it was also an easy way to report out the heights and volumes for the competition.

As for the actual way that I did the volume changes:  All my volumes, including the empty carafe are stored in 1 part and 1 assembly.  Multiple configurations of the part....1 configuration of the assembly.  The only reason that I even added it to the assembly was that it was easiest way to report out the volumes.

And I made the liquid purple, because you said that it was wine, so I made it a nice red wine.  A white wine would have blended with the bottle too much and would have been hard to see.

Hello,

Here is a concept to find the level marks automatically using design checker. I just did the first mark the other can be found accordingly.

Entry just for reference.

Elmar

Here is the skinny –

I thought I had a “great idea”, so I dove in headfirst last night . . .

My original thought was:

1)      Set up planes using the gradations to set the height

2)      Use these planes for an intersect on the carafe to get the internal volumes

3)      Set a sensor for each of the volumes

4)      Drag the planes up & down to get close to the necessary volume

5)      Run a design study on fast quality to get a little closer

6)      Run the study on standard quality to “fine tune” the results

While it appears as though this did work, it ended up being a little more cumbersome than I originally envisioned.

It was still EXTREMELY easy to get very accurate numbers with little effort on my part – which is how I prefer to work: put a lot of thought in up front, carefully set everything up, let the computer get to work while I go out and grab some coffee.

I am not sure if this was a function of SPR#588882, but I set my design study for 10.000 US fluid ounces, but in the results view it would show “is exactly 0.000295735”

As I have seen with this SPR, the sensor is shifted to the next option up in units, just above US fluid ounces is hectoliters – my design study absolutely nailed this to 8 decimal places, but there is such a big difference between the two measurements, that my results are off.

I tried to trick it by shifting the units in document settings and redoing the design study – no dice, it is really hung up on hectoliters now.

To throw a little more gas on this fire, I’ve found that if you edit a sensor, it will change the units until you do a full rebuild.

Now I am starting to get a little bit cranky . . .

Life would be a little better if I could drag to rearrange sensors and bodies in the Feature Manager.

Thanks for the fun & humbling challenge!

todd

Edit: All the gory details . . .

I described my thought process for the challenge above, but here is a little more detail:

1)      I offset the planes from the top plane and used a point in the gradation sketch for the height (This keeps everything linked together quite nicely!)

-          I did use a sketch point at the lip for the last plane (complete volume of carafe.)

2)      Using the intersect command is one of my favorite for getting internal volumes – it is quick, it is clean, and it gives me a separate body that can be easily measured. (This is why I used planes.)

-          This also allowed me to try to hit 10.000 fluid ounces for each individual body.

-          I kept all bodies from the intersect command, and just hid everything but the fluid.

3)      Since each of the levels is a different body, it was easy to set a sensor for each.

(This also makes it easy to see the volume of each body, especially when making updates!)

-          One of the sensors was for two different bodies, because of the hollow handle – no biggie.

4)      Since the planes are linked to the Global Variables, I could do some quick tweaking by dragging dimensions in the gradation sketch and/or using “Manage Equations” to get the volumes to a good “starting point.”

5)      I set up a design study, because, well, as I mentioned before, I would really rather let the computer do the hard work. (Gives me a chance to see what is happening on the forums!)

-          My original plan was to do one level at a time, because when you adjust the dimension for one level, it effects the volume calculations for the neighboring levels.

-          Throwing caution into the wind, I decided to optimize everything at once.

-          Not sure if this was a good idea, but I started by using relatively small steps with the fast solver. This got me really close, so then it was easy to set a range and use the high quality solver.

todd

I thought about doing this, but I can't because I don't have SW premium (maybe someone who does can try it out):

1) Create the internal volume of the Carafe and then save it as a separate part

2) Create a Global variable linked to the volume of the part and call it "VolumeOfLiquid"

3) Create a plane and do a "cut with surface" of the liquid volume at say height = .25" from the origin.

4) Create a global variable linked to the measurement from the origin to the top surface of the cut liquid volume and call it "HeightOfLiquid".

5) Create a sketch on the Right plane and add a single point.  Add 2 dimensions to locate that point, 1 horizontal and 1 vertical from the origin.  Name the horizontal dimension "VolumeVariable" and name the Vertical dimension "HeightVariable"

6) Link the "VolumeVariable" dimension to the "VolumeOfLiquid" Global variable

7) Link the "HeightVariable" dimension to the "HeightOfLiquid" global variable.

8) Now when you drag the slider on the plane height dimension on the plane that is creating the plane cut, you can see the point slide as well.

(Here's the part I can't do because I don't have SW Premium, so this is all theoretical as far as I am concerned  I went so far as to do the rest, though)

9) Now run a motion analysis and use the plane height as your item that is changing and run a trace of the path of the point to create a curve.

10) Then import that curve back into the part and position it correctly and lock it in place.  Now you have a graph of volume vs height of the liquid.

11) Create another sketch and add a point to it and make it coincident to the curve.

12) Add a horizontal dimension to the point, originating from the origin, and give it a value of 10 (placing the point along the graph at 10 fl. oz).

13) Add a vertical dimension to the point and make it driven.  (Now you know the exact height at 10 fl. oz)

14) Link your dimension to 1 of 2 things (not sure which one will work, if any)

          a) a custom property, then link your "TenLevel" global variable to the custom property

          b) Directly to the dimension that locates the "TenLevel" sketch dimension

         EDIT:  c) Just thought of something.   Instead, you could create another global variable linked to the measurement of the point from the origin (Top plane).  Then link the height of the TenLevel Sketch split curve to the global variable.  Then you wouldn't even need step 13.

15) Repeat steps 11-14 for the rest of the volumes.

Not totally sure that it will work (Step 14 might be the game ender).

Edit2: You know what...I also just remembered....going to do the linking at the assembly level.  Can't so this with just one part.  Need an assembly of the liquid and the carafe.

Then have the Global variable mentioned in step 14c drive the height of the split line in the carafe part, through the assembly.

Will someone try it out?

I fully agree with Michael Fernando and Bill Toft - By the rules of the challenge .001" is PLENTY close, and by the rules of drinking, it is either "full", "half empty/time to order another", or "empty/where is the next one?!"

The fun part of setting this up as a design study is that when Alin Vargatu threw down the gauntlet and wanted more accuracy, I spent a minute or two tweaking the ranges for my dimensions, clicked "Run", and went and grabbed another coffee.

If I didn't have other stuff that needed to get done, I probably would have run 2 or 3 more design studies, because I know I can get it even closer . . . then again, I'm pretty sure that a single drop is bigger than .00012 fluid ounces!

Hello All,

I know that time is out I needed to travel back and forth so I didn't have time in the previous days but I really want to share this with you, letting the computer do the hard work getting the volume right up to 8 places after the decimal point and even more:) (you specify the max error you accept) without so much processing time (seconds for level).

I used a macro, the first macro I write to be specific so I'm happy that this challenge encouraged me to learn and try.

In short, I made another configuration for each level and deleted the Carafe itself and cut the fluid with a plane (all of this is done for simplicity, to deal with only one solid body in the part but it could have been done without this)

Then for the first level 10 ounces I tell the macro the needed level 10 and the height of the plane  I used for the cut 1

then the macro measures the volume compare it to the 10 ounces

if it's more, then the height et decreased a bit (0.012m  about half in)

if it's less, then the height et increased a bit

every time this bit is divided by 2 so that the error converge

This will continue to happen until the error (difference between the measured volume and the 10 ounces) gets smaller than a value that I specify at first

every time this bit is divided by 2 so that the error converge

Takes just some seconds to solve then I select the second config change the 10 to 20 and the starting height from 1in to 2in

If I got some people interested I'll make a video discussing that I've done and why.

Have a nice one