ADVERTISEMENT

Off Topic...desired engineering or physics brain that works...at least once.

tjreese

tjreese

All-American
Gold Member
Sep 27, 2008
25,813
24,151
113
Is there a conversion relationship between "input settings" with a known force in newtons and an angular speed of so many radians/sec without a radial distance or radius and an output measured in newton-meters and seconds? Is there a relationship that ties those variables together...power ??? with only what I listed?

Been thinking about this as long as Crean does his recruits...sorry for the off topic, just thought some youthful brain might steer this old man down a good path. Hope someone has the rudder.
 
tjreese said:
Is there a conversion relationship between "input settings" with a known force in newtons and an angular speed of so many radians/sec without a radial distance or radius and an output measured in newton-meters and seconds? Is there a relationship that ties those variables together...power ??? with only what I listed?

Been thinking about this as long as Crean does his recruits...sorry for the off topic, just thought some youthful brain might steer this old man down a good path. Hope someone has the rudder.
Click to expand...
Take it the genius forum. :D How long did it take you to think that up?
 
  • Like
Reactions: boiler_89_
tjreese said:
Is there a conversion relationship between "input settings" with a known force in newtons and an angular speed of so many radians/sec without a radial distance or radius and an output measured in newton-meters and seconds? Is there a relationship that ties those variables together...power ??? with only what I listed?

Been thinking about this as long as Crean does his recruits...sorry for the off topic, just thought some youthful brain might steer this old man down a good path. Hope someone has the rudder.
Click to expand...
P=T*w
P= power in watts which is equal to torque in Newton meters
T= terminal velocity
w=radians per second

You may be able to use...power in watts = newton meters/secs.

I am not sure if that's what you are looking for and I am assuming you probably already know these formulas.
 
TwinDegrees2 said:
Take it the genius forum. :D How long did it take you to think that up?
Click to expand...
It is a fricking problem that I need to make equal to plot a 45 degree line for efficiency of design ignoring friction and other losses by analyzing 6 variables in at least three settings each for main effects and interactions. Once I can equate a dynamic input with an output and its values I can then study the variables to decrease the output difference between initial results and after usage. I can do the variable studying, but it is equating input and output to similar metrics that is making my head hurt. I'm just hopeful that some person that knows his or her physics has an answer...but I think those people are probably studying instead of being on a basektball forum.

It is my last gasp...before I start making some assumption... :(
 
proudopete said:
P=T*w
P= power in watts which is equal to torque in Newton meters
T= terminal velocity
w=radians per second

You may be able to use...power in watts = newton meters/secs.

I am not sure if that's what you are looking for and I am assuming you probably already know these formulas.
Click to expand...

Thanks...yes that would be power in n-m./sec for the ouput. My input is in newtons and angular velocity (radians/sec). if I knew the radial distance for my angular verlocity I could get there as well. What I would like to do is plot an input power on the X axis and an output power on the Y axis. For this example I would neglect various losses and plot a 45 degree slope from the origin and compare my improvements to "perfection"
 
tjreese said:
It is a fricking problem that I need to make equal to plot a 45 degree line for efficiency of design ignoring friction and other losses by analyzing 6 variables in at least three settings each for main effects and interactions. Once I can equate a dynamic input with an output and its values I can then study the variables to decrease the output difference between initial results and after usage. I can do the variable studying, but it is equating input and output to similar metrics that is making my head hurt. I'm just hopeful that some person that knows his or her physics has an answer...but I think those people are probably studying instead of being on a basektball forum.

It is my last gasp...before I start making some assumption... :(
Click to expand...
Have you tried Bayes or related theroms to predict the actions of each variable?
 
proudopete said:
Have you tried Bayes or related theroms to predict the actions of each variable?
Click to expand...
The statistics is no problem. I will be using orthogonal arrays L27 with two dynamic signals to test for main effects and any interactions is my current thought. I have a baseline design and will have no problem beating that (everything is done on a simulator...class work). When it is all said and done I would like to compare my new variable settings to a theoretical perfect situation where "input" equals "output"...which would be those straight lines.

I'm familiar with improving through statistics (taguchi) as I was doing that back in the 80s. I am not a designer. This class is to take the optimization of variables and various settings and compare them to the "physics" of the situation. It is to make designs robust over noise or environmental conditions you can't control. This problem is to improve a chainsaw. My noise is pine &oak wood and synthetic & organic oil that my setting must be robust. My control variables are tooth hardness, tooth spacing, height, tooth angle, another angle of something??? and track clearance between chain and bar. I chose the initial settings at three levels to check for linear realtionships...of which I think most will be as well as potential savings if a less costly variable setting will essentially have the same effect...

I volunteered to take this class because, I was a glutton for punishment. I'm as knowledgable as anyone in the class and showed the teacher a few things relative to this...but the design part is burying me. I have many more things I would like to do and paired up with the head of engineering to work on this. he thinks we cdan't get there and will just have to compare the improvement to the basic design, but I wasn't ready to quit just yet....it is just that I'm about there as well.... :) thought maybe some youthful person might know...as a last gasp...
 
I am no physicist so I may not be the best person to ask. I am also not young, so there may be a limit there as well.

Poincaré could possibly be used to track the variables and then you may be able to plot out where they equal what you are looking to prove. I would assume if 45 degrees is your as your control, finding them, if they line up as you hope, shouldn't be an issue.

I wish you luck. You are right in the fact that those who would chuckle and give you the answer are more than likely somewhere eith their heads in a book.
 
Can you not find some reference value (or set of values) for the missing information as a sort of constraint on your model? If the problem is under-specified assumptions will be needed at some point. As you noted you are limited by lack of data regarding 'a radial distance or radius'
 
proudopete said:
I am no physicist so I may not be the best person to ask. I am also not young, so there may be a limit there as well.

Poincaré could possibly be used to track the variables and then you may be able to plot out where they equal what you are looking to prove. I would assume if 45 degrees is your as your control, finding them, if they line up as you hope, shouldn't be an issue.

I wish you luck. You are right in the fact that those who would chuckle and give you the answer are more than likely somewhere eith their heads in a book.
Click to expand...
Thanks...I would rather talk basketball too :)
 
i may be misunderstanding you but ill throw this out. you can use rpm instead of radians/sec. plot your torque curve vs rpm like a dynamometer would read? force torque in this case is instantaneous and power has a time component so you need to rectify that somehow. good luck!
 
cgnops said:
Can you not find some reference value (or set of values) for the missing information as a sort of constraint on your model? If the problem is under-specified assumptions will be needed at some point. As you noted you are limited by lack of data regarding 'a radial distance or radius'
Click to expand...

yeah, I may have to go there...just making sure I'm not missing something..and getting tired in the process as there are more enjoyable things at my age.... :)
 
tjreese said:
The statistics is no problem. I will be using orthogonal arrays L27 with two dynamic signals to test for main effects and any interactions is my current thought. I have a baseline design and will have no problem beating that (everything is done on a simulator...class work). When it is all said and done I would like to compare my new variable settings to a theoretical perfect situation where "input" equals "output"...which would be those straight lines.

I'm familiar with improving through statistics (taguchi) as I was doing that back in the 80s. I am not a designer. This class is to take the optimization of variables and various settings and compare them to the "physics" of the situation. It is to make designs robust over noise or environmental conditions you can't control. This problem is to improve a chainsaw. My noise is pine &oak wood and synthetic & organic oil that my setting must be robust. My control variables are tooth hardness, tooth spacing, height, tooth angle, another angle of something??? and track clearance between chain and bar. I chose the initial settings at three levels to check for linear realtionships...of which I think most will be as well as potential savings if a less costly variable setting will essentially have the same effect...

I volunteered to take this class because, I was a glutton for punishment. I'm as knowledgable as anyone in the class and showed the teacher a few things relative to this...but the design part is burying me. I have many more things I would like to do and paired up with the head of engineering to work on this. he thinks we cdan't get there and will just have to compare the improvement to the basic design, but I wasn't ready to quit just yet....it is just that I'm about there as well.... :) thought maybe some youthful person might know...as a last gasp...
Click to expand...

Don't forget to address the angle of the grain of wood with a factor leading to the semi-dampening of any non-linear vibration of the cutter.
 
tjreese said:
yeah, I may have to go there...just making sure I'm not missing something..and getting tired in the process as there are more enjoyable things at my age.... :)
Click to expand...
Who doesn't enjoy doing calculus? I think I get what you are trying to find, most desirable output/most economic input. Once again, physics isn't my thing. But I do enjoy it somewhat...at times.

Top secret may be right. The key may be in the conversion.
 
TopSecretBoiler said:
i may be misunderstanding you but ill throw this out. you can use rpm instead of radians/sec. plot your torque curve vs rpm like a dynamometer would read? force torque in this case is instantaneous and power has a time component so you need to rectify that somehow. good luck!
Click to expand...
yeah, I can get the rpm by converting radians. and would have force and RPM for inputs. thought about that. What I need to do is equate input data into some power input that I measure some power output when I am supplied n-m (torque) and seconds. If I don't get my inputs and output readings on the same scale I won't have my slope for perfection. My input should equal my output ignoring losses...hence the 45 degree angle between them. I have an idea...just not sure how put it all together yet. The output is the torque (N-M) and seconds to cut through a 4"x4" piece of wood...and I think THAT is my clue somehow. Somehow the 4"X"4" block of wood is the key. You see I select within a range a force and an angular speed for my input and the simulator depending on my variable selections outputs an initial torque (not force) and seconds to cut through the block of wood and does the same after a defined amount of cuts to simulate usuage. The delared problem is to reduce teh delta in cuts initally and after usage, but naturally not increase the overall time. The design must be teh best design whether cutting pine, oak, synthetic oil or organic oil by manipulating my control variables at three settings all at a couple of throttle or input settings. Somehow, I have to grasp the 4"X4" wood as I think that is a clue... :)
 
Inspector100 said:
Don't forget to address the angle of the grain of wood with a factor leading to the semi-dampening of any non-linear vibration of the cutter.
Click to expand...
Oh this is a classroom problem...that is not one of the variables as that might push me over teh ledge... :)
 
proudopete said:
Who doesn't enjoy doing calculus? I think I get what you are trying to find, most desirable output/most economic input. Once again, physics isn't my thing. But I do enjoy it somewhat...at times.

Top secret may be right. The key may be in the conversion.
Click to expand...
He took me where I was and it may be the right track...once I figure out the 4"X4" block as that defines a distance...maybe coffee in the morning?
 
tjreese said:
well, right now I'm in the spaced out zone...just a shell of my former self! :)
Click to expand...

No, no. Analyse the piece of wood and determine the angle of the grain to the cutter. Make the factor a variable based on the delta of that angle within the piece of wood and integrate the required power to make the cut using another integration of the wear on the cutter and it's relationship to the forces required to cut through each grain. As Jake Queery(sp) say, that is just the voice of reason.
 
  • Like
Reactions: tjreese and proudopete
Inspector100 said:
No, no. Analyse the piece of wood and determine the angle of the grain to the cutter. Make the factor a variable based on the delta of that angle within the piece of wood and integrate the required power to make the cut using another integration of the wear on the cutter and it's relationship to the forces required to cut through each grain. As Jake Queery(sp) say, that is just the voice of reason.
Click to expand...
I'm tired...calling it a night. Thanks everyone...I'll catch a little tube.
 
Last edited:
tjreese said:
The statistics is no problem. I will be using orthogonal arrays L27 with two dynamic signals to test for main effects and any interactions is my current thought. I have a baseline design and will have no problem beating that (everything is done on a simulator...class work). When it is all said and done I would like to compare my new variable settings to a theoretical perfect situation where "input" equals "output"...which would be those straight lines.

I'm familiar with improving through statistics (taguchi) as I was doing that back in the 80s. I am not a designer. This class is to take the optimization of variables and various settings and compare them to the "physics" of the situation. It is to make designs robust over noise or environmental conditions you can't control. This problem is to improve a chainsaw. My noise is pine &oak wood and synthetic & organic oil that my setting must be robust. My control variables are tooth hardness, tooth spacing, height, tooth angle, another angle of something??? and track clearance between chain and bar. I chose the initial settings at three levels to check for linear realtionships...of which I think most will be as well as potential savings if a less costly variable setting will essentially have the same effect...

I volunteered to take this class because, I was a glutton for punishment. I'm as knowledgable as anyone in the class and showed the teacher a few things relative to this...but the design part is burying me. I have many more things I would like to do and paired up with the head of engineering to work on this. he thinks we cdan't get there and will just have to compare the improvement to the basic design, but I wasn't ready to quit just yet....it is just that I'm about there as well.... :) thought maybe some youthful person might know...as a last gasp...
Click to expand...
Okay, now I'm interested....one question for you or two. Wouldn't tooth thickness be a necessary variable/equation along with various material of the bar if you're to "improve" the current situation to maximum conditions? What about including the various bearings available to reduce drag?
 
  • Like
Reactions: Inspector100
also remember newton breaks down to 1kg per meter per second. it may help to break it down. then your newton meter in the other side of the equation is the torque
 
tjreese said:
Is there a conversion relationship between "input settings" with a known force in newtons and an angular speed of so many radians/sec without a radial distance or radius and an output measured in newton-meters and seconds? Is there a relationship that ties those variables together...power ??? with only what I listed?

Been thinking about this as long as Crean does his recruits...sorry for the off topic, just thought some youthful brain might steer this old man down a good path. Hope someone has the rudder.
Click to expand...
This instructional video should help... one way or the other.

 
TwinDegrees2 said:
Okay, now I'm interested....one question for you or two. Wouldn't tooth thickness be a necessary variable/equation along with various material of the bar if you're to "improve" the current situation to maximum conditions? What about including the various bearings available to reduce drag?
Click to expand...
This is a "classroom" exercise where I have a range of inputs and a calculated output based upon the variables I must work with (6) at whatever level within the range allowed. Where I am stuck is equating the input numbers (F, radians/sec) with the ouput numbers (n-m, sec). My problem is not in determining which variables affect the answer within the ranges allowed, it will not be a problem in determining which variables have the greatest effect on the ouput with reduced noise or the interaction between those 6 variables if they exist...all across an environmental noise of two types of wood and two types of oils.

My output simulator provides 4 pieces of data Initial torque (N-M) and initial seconds to cut through the 4"X4" piece of wood and the same data after usuage. The goal is to reduce the effect of usage differences in time so that time required to cut the wood with a given input has a small delta in time between initial cut and time after it has made several cuts. Reducing that delta of time between initial and usage results should not provided an answer that overall makes teh difference small, but increases teh time overall. I'm approaching this as a fractional factorial experiment...doing a fraction of the total combinations which would be 3**6 or 729 possible settings--6 variables at three settings. In any fractional factorial (reduced study) there are many aliases. Since there are 729 possibilities and I intend to use 27 (L 27 Orthoganal Array) combinations for this study I must give up something. Essentially, there are 26 aliases buried in the results. 27*27=729 and I'm using 1/27 (the fraction studied) or 27 different scenarios to extract main effects (ALL 6 VARIABLE SETTING EFFECTS) and three interactions. If the slopes of the lines of various main effects plotted against each other are NOT parallel, then that "out of parallel condition indicates some strength of interaction between the variables in question.

In "real life" main effects and only first order interactions typically matter. Suppose I have variables A,B,C,D,E,F I may find that Variable A and B are important in affecting the outcome and AXB (first order of interaction) is also important. C may also be important and AXC interaction may be important, but usually AXBXC interaction is not and certainly AXBXCXDXEXF is not. Careful selection of the various combinations allow us to reduce the study combinations to a fraction of the possible outcomes and get the important data in a fraction of the possibilities. So, when I determine the effects of a "column" which is a "main effect", that outcome could be 26 other aliases and so it is important to understand the right column in an orthoganal array to not confound main effects and first order interactions, but rather confound the higher order interactions. The L27 orthogal array allows me to study 6 variables at three levels and get 3 different interactions.
Another study may be needed based upon the initial results for more clarification, but together...ALL of these will significantly be less than 729 studies. Here is an orthognal L8 for a visual understanding. An L8 is a study for only 2 levels and when you look at this you will see letters (main effects) and the combinations of those settings at one of the 2 levels. If you replace a 1 with your chosen low setting and 2 as your chosen level high setting...the table (saves teh manual calculation) for determining your settings. Replacing a 1 with a - and the 2 with a + (two settingss) is the classical layout whihc is the same. Now if I put main effects in A,B and C and study their effects Column C (main effect C) may be significant...but it could be the interaction between A and B as well. If you replaced the 1 wiht a - and the 2 with a + in all columns, you can now notice the settings are the same, but column 3 (main effect C) is confounded wiht the potential interactions of A and B. Multiply the replaced numbers with - adn + in all columns and see if Column C is the result of A and B multiplied together and you will see it is. Consequently, I do NOT want to use that column for a main effect if I can move it. https://www.google.com/search?q=l27+orthogonal+array&espv=2&biw=977&bih=500&tbm=isch&imgil=oycffqNdO4ElhM%3A%3BsnnkD3y5uz-Q8M%3Bhttp%253A%252F%252Freliawiki.org%252Findex.php%252FTaguchi_Orthogonal_Arrays&source=iu&pf=m&fir=oycffqNdO4ElhM%3A%2CsnnkD3y5uz-Q8M%2C_&usg=__N1UlL1nCUzOSECjhPJ3PwHBhSC0=&ved=0ahUKEwi-vsnmrrrSAhUh1oMKHei2DLcQyjcIMw&ei=fGW5WL6QLKGsjwTo7bK4Cw#tbm=isch&q=l8+orthogonal+array&*&imgrc=ajoktuzNxzxkfM:
 
Last edited:
  • Like
Reactions: Inspector100
TwinDegrees2 said:
Thanks....I really am enjoying reading this. Told someone today I'm bored and need a new challenge.

But I was thinking about another coaching job.;)
Click to expand...
Well, if you care about the kids in being fair to them and yet trying to drive them to heights they may not achieve on their own (pushing them) coaching can really be difficult...and consume your life thinking about what if we did this or that and so forth. If this is interesting to you, rather than getting bogged down in classical statistic Design of Experiements I would suggest getting a book on Taguchi understanding. His Linear graphs allow a visual for determining where interactions are located and save the manual calculations. Here is a pic of what I'm saying and going back to the L8 orthoganal I posted in this thread. Teh numbers represent variables for consideration. Teh lines represent interactions between those numbers. Going back to a previous post this morning I said that columns 1,2 and 3 or A,B and C...C is the math result of the A and B values (remember the - adn + signs). Here in the lineaer graph is a visual for that calculation. The line is the interaction between the dots and so if I think teh interaction is important between A and B I should NOT place a main effect in that column but chose another column for confounding data or aliasing. That way I can study column 3 and determine if the interaction between A and B is in effect.
https://www.google.com/search?q=l8+...i=UWy5WICMDcS0jwSs4IaIAw#imgrc=o-CNwx2sH-YtaM:
 
Inspector100 said:
IU not making NCAA tournament = $20. Iowa beating Wisky = $1M. tjreese discussing molecular motion without loses = PRICELESS.
Click to expand...
You ought to see me when I start discussing pros and cons on zones. Seriously I'm much more interested in basketball.... :)
 
TwinDegrees2 said:
I know I've never had one of these discussions in the transmission plants.:rolleyes:
Click to expand...
Someday I may have a few stories to tell about that. I dropped in a few years later to see if anyone was there and many years later they still had my pic and were telling stories about me. I almost dropped a car off a rack, caught one on fire and had people pull off the road and watch me do various tests...not to mention snowballs I might sneak in. Someday perhaps...
 
You must log in or register to reply here.
ADVERTISEMENT

Latest posts

ADVERTISEMENT
Top Bottom
Back