cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Relative vs Absolute accuracy?

Participant

Relative vs Absolute accuracy?

What are the circumstance where one would want to use Absolute accuracy?


Is it so that you can set all parts to be equal to one accuracy by changing one part?


Why is it any easier than manually changing all the parts to be the same number?


11 REPLIES 11

Relative vs Absolute accuracy?

To Try to get Pro-Mold to work. Which can be trail and error at times.

Seems like the software should have one accuracy that worked across all applications. So far been to much to ask of PTC.

Best regards,
James Starkey
Starkey Tech. Serv., L.L.C.

Relative vs Absolute accuracy?

I always use and advise absolute accuracy.

That way the part accuracy won't change as the model changes.

Obviously the value will have something to do with the part/assembly. I
won't use 0.01mm for an agricultural machine that will mainly get parts
cut using a blowtorch, but for a mold of a plastic part, I would, for
example.

So far, I have seen problems using things like Mold and even Inheritance
features, that fail using relative accuracy, but absolute works. You
may need to tweak the value depending on the origin geometry (imported,
modeled better or worse, geometry checks, etc), but one way or another,
absolute has worked.

I don't see a reason for using relative accuracy, TBH.

Rui





On 12.12.

RE: Relative vs Absolute accuracy?

Michael,


You will use it when dealing with imported geometry or parts and assemblies that are dissimilar in size that must relate to one another. By size I mean the bounding box (or model size). Essentially, with Relative accuracy, the model regens with only as much "accuracy" to work and is based on the ratio of smallest to largest features within the model. Think of it as related to processor crunch time. too much accuracy and math gets intense. too little and the model is junk. And it's a bit ofBlack Magic to pin down where the break even point is for users. The standard .0012 Relative is the mathematicians Happy Place. In 25 years, I have had to change accuracy maybe, five times within any givenpart. Not so when dealing with assemblies and/imported data. In those cases, it can be common-place to need Absolute when two dissimilar data sets have to play nice together.


Think of a vehicle unibodythat is 11 feet long,6 feet wide, 4.5feet talland has a .118" radius on a surface somewhere. The software can have a very dificult time regenerating situations like that. It really is immaterial whether you model it natively or import a data set for reference. It has to do with the model calculations based on the ratio I spoke of above. of course, EVERY unique situation has it's unique math chanllenge. Your mileage will vary.



Hope it helps...

Highlighted

Relative vs Absolute accuracy?

Relative accuracy was, I believe, a way to balance geometry and computer
power. Building to an absolute accuracy of X thousandths or X ten
thousandths of an inch took a lot of computing and was frequently more
than needed. On a large part, you may only need to go to the hundredths
of an inch, on a smaller part you'd need more. So accuracy relative to
model size was born.



Of course, that presented problems when parts had to interact with each
other, like doing a mold split for example. It also gets problematic
when you have a large part that requires fine detail, like an injection
molded part with small radii or snap details. In those cases, the large
size of the part produced a large accuracy value up front, and then when
the fine details were attempted, they'd fail due to the accuracy being
too 'loose'.



With newer PCs, the use of fine absolute accuracy values became more
practical, and PTC switched the default from relative to absolute with
WF4, I believe. Of course, if your start part dates to before then, it
may have relative accuracy so your models will continue to have relative
accuracy unless you change the start part.



Personally, I started using Pro/E in 1996 and was doing mold design at
the time, so I always used absolute accuracy and have ever since. I've
not run into many times where it was a problem, on the contrary, I have
run into issues with building on customer start parts with relative
accuracy and not being able to create the fine detail required.



So for me, the time to use absolute is always. I'd ask "When is a good
time to use relative?"



Doug Schaefer
--
Doug Schaefer | Engineering Manager
Crow Works

Relative vs Absolute accuracy?

At one time, I think .0012 relative accuracy WAS the standard. There
was some discussion about this on the forum several years ago.



I use absolute accuracy for forging part and tool design. Because of
the process I go through, relative accuracy ends up being a royal pain.
You really need to create company wide start parts, etc... that enforce
this policy for all designers



We set all of our products to have .0001" absolute accuracy. Tooling
for parts is approx 36" dia x 18" high. No problems at our end,
especially with tool paths.

Having .0001" absolute accuracy may make models take longer to regen,
and file sizes larger, but storage is cheap, and computers are too.

At one time, dragging and dropping .STEP file parts into the pro-e
window would create a part that did not use the default start part, and
came in at .0012 relative. I think WF 5 fixed that now and you get a
dialog asking you to use the default start part.



I find that quite a bit of the large file problem and regen accuracy vs.
regen time does not matter as much if I use ext-copy-geom. vs. merge /
cutout for my dies.



As I remember though, the scripts in Mold design want to use Merge /
cutout and they prompt you to match accuracies.



I use TDO, and write my own scripts using ext copy geom function.



Watch out also with merge / cutout cavities if you make changes. I may
be wrong, but I seem to remember that if you edited the part you were
using in the cavity, the cavity in your tooling would not update unless
you have both models in session. I think that there may be a config
option to force opening of related parts and force regen.







Christopher F. Gosnell



FPD Company

124 Hidden Valley Road

McMurray, PA 15317

Relative vs Absolute accuracy?

All,
This came from an old PTC document.

Relative vs. Absolute Accuracy

Description

This document describes the definition and use of model accuracy, which is set using #Setup, #Accuracy. This information expands on what is found in the PTC Help System. The main sections of this document are:

Accuracy Overview<">http://www.hub.slb.com/Docs/ofs/tech_centers/src/Relative_vs_Absolute.htm#Accuracy_Overview>
Relative Accuracy<">http://www.hub.slb.com/Docs/ofs/tech_centers/src/Relative_vs_Absolute.htm#Relative_accuracy>
Absolute Accuracy<">http://www.hub.slb.com/Docs/ofs/tech_centers/src/Relative_vs_Absolute.htm#Absolute_accuracy>
Modifying Accuracy: When is it necessary?<">http://www.hub.slb.com/Docs/ofs/tech_centers/src/Relative_vs_Absolute.htm#Modifing_accuracy>

RE: Relative vs Absolute accuracy?

I often find myself using absolute accuracy. My company produces large sheet metal parts. Since the thickness of the material relative to the size of the part is quite small, as I am creating my part features will "mysteriously fail." Changing from relative to absolute accuracy fixes the problem.

Relative vs Absolute accuracy?

Thanks Michael,

"Accuracy" in a succinct and useful explanation - which I've copied for
other users. Without wishing to initiate another "help" thread, why is
this old stuff so much easier to get to back then?



Richard A. Black

Lead Design Engineer

Eaton Corporation

440 Murray Hill Road

Southern Pines

NC 28387 USA



tel: 910 695 2905

fax: 910 695 2901

-

www.eaton.com

RE: Relative vs Absolute accuracy?

I think I have been conviced about Absolute Accuracy's benefits.


I set my config.pro option enable_absolute_accuracy to yes.


When I input .0001 I get:


Enter absolute part accuracy or <esc> to return to Accuracy menu [2.0555e-04 inch]: .0001
Range is 0.0003 to 0.0205. Please re-enter: 0.0003


Another user is gettingthe range 0.0073 to 0.7217 only?