Re: Fw: IJK

[Will Shackleford <shackle-at-cme.nist.gov>]

b.eding wrote:

> Will,
> 
> I checked the merge, it is ok.
> I also updated my posemath stuff with the latest
> version and tested it on my machine,
> I do not have the work-around in canon.cpp.
> It is OK.
> 
> Will I noticed the new functionality in PmLine,
> wich handles the orientation as well as the translation.
> 
> I asume this is for the 6DOF trajectory planner functionality, is it?
> Will you do this for circles also? (a circle may also go along with
> a orientaion change).
> 
> Bert
> 

Thanks.


I added code to PmLine functions to handle the orientation. 
Unfortunately there was a problem  with doing this at this level.
Many of the rotation functions assume A=0 means the same thing as 
A=360,720 etc. so it is easy to lose the number of turns. Fred suggested 
trying to keep the rotations in non normalized format, but instead I 
modified the trajectory planner to always pass a 0 rotation to the 
PmLine functions and track the rotations within the trajectory planner. 
I think I now have this working for circles as well as lines in the 
trajectory planner, but I definitely need to test the circles more 
thoroughly. If we keep this style of trajectory planning we won't 
necessarily need to modify PmCircle, although there might still be some 
value in being able to call the functions directly.


-- Will


> ----- Original Message -----
> From: "Will Shackleford" <shackle-at-cme.nist.gov>
> To: "Multiple recipients of list" <emc-at-nist.gov>
> Sent: Thursday, January 04, 2001 5:35 PM
> Subject: Re: Fw: IJK
> 
> 
> 
>> 
>> I have tried to merge changes from Bert's POSEMATH.C into the
>> sourceforge tree.
>> 
>> Bert, Can you check the following URL below to see if I got them right?
>> (I made changes based on a diff with a late 1999 version of _posemath.c)
>> 
>> 
>> 
> 
> http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/rcslib/src/posemath/_posemath.
> c.diff?cvsroot=emc&r1=text&tr1=4.33&r2=text&tr2=4.34&f=h
> 
>> Thanks,
>> 
>> Will
>> 
>> 
>> 
>> b.eding wrote:
>> 
>> 
>>> The problem with full circles is solved with
>>> a workaround in canon.cpp of the EMC latest versions.
>>> 
>>> However the actual problem is inside the posemath library,
>>> I found the problem and solved it, I just tried it on windows-nt,
>>> not yet with the whole emc-sotware on Linux.
>>> 
>>> The modified posemath.c allows full circles without problems.
>>> I have attached the file with this email.
>>> 
>>> Regards,
>>> Bert Eding
>>> 
>>> ----- Original Message -----
>>> From: "Matt Shaver" <mshaver-at-erols.com>
>>> To: "Multiple recipients of list" <emc-at-nist.gov>
>>> Sent: Tuesday, January 02, 2001 6:24 AM
>>> Subject: Re: Fw: IJK
>>> 
>>> 
>>> 
>>> 
>>>> Dan Falck wrote:
>>>> 
>>>> 
>>>>> When EMC hit the code for a
>>>>> full circle it ran away (in G01 mode so it wasn't so spectacular).
>>>> 
>>> Luckily
>>> 
>>> 
>>>>> nothing was ruined, but it surprised me anyway.
>>>> 
>>>> We need to investigate this! It can either execute the block, or flag a
>>>> run time error, nothing else...
>>>> 
>>>> Matt
>>>> 
>>>> POSEMATH.C
>>>> 
>>>> Content-Type:
>>>> 
>>>> application/octet-stream
>>>> Content-Encoding:
>>>> 
>>>> quoted-printable
>>>> 
>>>> 
>>>