Antonius, I get the same on regular dimensioning, but I'm talking about within a GD&T block..
The Primary is set to 3 places but the secondary is just cutting off the digits past two places, so instead of rounding up to .08mm from .0762, it just displays [0.07]
Thanks for the clarification. Indeed, this is still an issue in Creo 2.0 M010. Thanks for the heads up.
I will report this to CS and see what they have to say. Yet another reason -not- to use GTOL! 
"BRIAN!"
No problem, Brian. I too was taken out of this community for about 3 weeks due to "obligations" and happy to see your smiling avatar back at it.
In this case, yes, PTC did due diligence but only half way. It appears, and I do not have the reference, but it appears that there are more than one way to skin this cat. I have seen many references since that simply round appropriately and do not round down to the minimal tolerance.
You will probably need to be armed with SAE J390 to make a case. Somehow I think you just might have access to that 
of course we could do well to avoid the issue all together and covert the entire planet over to metric as it should be. 
Only if you're using a system which relates the tolerance to the number of displayed decimal places... otherwise machining to 0.995/0.795 is no harder than machining to 1.0/0.8. 0.9 satisfies either of them!
And in the case of your GD&T, machining to 0.076 is definitely (slightly) easier than machining to 0.07.
John, again, thank you for bringing this up as this will certainly be a problem I will run into. I have been very diligent in learning how to make model annotation work for me in my contract design business. This is yet another stumbling block.
I have reported this to CS and created an Idea (bug report!) that needs a flood of votes if you can access the Ideas section. Below is a cross-post of the idea for those who cannot access it:
This really shouldn't be an Idea, but rather a "Quality Oversight Bulletin!".
This error should never have existed outside the halls of PTC, much less considered for release into production.
Why is it that dual dimensions round properly, yet GTOL dual dimensions truncate the secondary dimension?
I will let the image speak for itself... ref - .0030 [0,07620]
model annotation shown.
detail config options:
dual_digits_diff -1
dual_dimensioning primary[secondary]
dual_secondary_units mm
This makes for yet another obstacle in 3D annotation not to mention using relational GTOL throughout development.
CS confirmed that this is expected behavior only with regard to tolerances. I have requested ASME and ISO references to confirm this complies to standards. Jonathan's explanation is what seems to be the guiding factor.
Title
Tolerance values of Secondary Units are tighter than Primary Units in Creo Parametric
Description
- With Dual Dimensioning, Tolerance values of Secondary Units are rounded incorrectly.
- The Secondary Dimension tolerance values are Truncated instead of rounding
- Why are the secondary units rounded differently than the primary?
Applies To
- All releases of Pro/ENGINEER and Creo Elements/Pro
- Creo Parametric 1.0, 2.0
Cause
- Works to Spec
Resolution
- Secondary values are rounded to fit within the range indicated by primary units
- See also Help files at #Help# Help Center > <Product> Functional Areas > Detailed Drawings > Detailed Drawings > Annotating the Drawing > Dimensioning the Model > Formatting Dimension Display > To Format Existing Dual Dimensions in a Drawing.
- Product Help Extract:
Note
"When you are using dual dimensioning with dimensional tolerances, the system rounds the secondary values so that they always fit within the range indicated by primary units (that is, so that they are tighter than primary values), and retain the desired design intent".
Of course, this opens another can of worms... what dimension is the "valid" dimension? I always considered the secondary to be a "reference". Regardless, the primary must be met. Am I wrong in the words of ASME and ISO? If I thought about it hard enough, can't you make a dimension contradict the limits by rounding the secondary (nominal secondary + tolerance range compared to primary range)? Calling on you mathmatiocians here
Regardless, this seems odd until I find out more. Clients tend to have rules that don't always conform to standards. At the very least, this should a config option.
Idea edited.
A little more research shows that dual dimensioning is governed by SAE J390 (1999).
Does anyone have access to this standard? Please weigh in as I need more evidence to support the "idea" as a viable config option.
Anyway, I found the following information on All-Seals Inc web site. Please note they refer to "method A" per the standard. Below is their statement regarding dual dimension references and an example of a .003" tolerance.
This document seems to guide ASME more than ISO. But according to these findings, there is a legitimate call for having the option to properly round the value vs always rounding to minimal.
Additional comment added to the Idea.
