Re: [CAD_CAM_EDM_DRO] Re: Crashing CNCs: what gives ?

Hi Ballendo;

Adaptative machining - Hmm we never used that term. Interesting
though to hear that here are apparently low cost CNC solutions available
to support this sort of programming now a days. When I was heavily
involved in this sort of manufacturing years ago some of the CNC
controllers barely supported RS232 communicationss so such software
implementations were extremely limited. Some of the later purchased
lathes did have considerable on board inteligence, realtive to the fact
that we where talking 16 mhz 386s. It was a bit strange to realize
that the coputational capability of the PC's driving the lathes were
limiting production rates. Once management realized this it did not
take long for the old PC hardware to disappear.

As far as adaptive machining goes - once the software was validated to
be working correctly the biggest issue with crashes, where the result of
faulty tranducers to measure position to with in a micron on a
production basis. It just goes to show you that while you can
eliminate many problems with solid engineering thee will always be the
potential for a failure someplace.

At the moment I'm looking around for a small lathe to become a home CNC
project. It is a bit amazing to realize that I can do this in an
affordable manner with out a huge outlay of $$$$$$$$$$$.

Thanks
Dave



ballendo wrote:

>David,
>
>This was an excellent post!
>
>Adaptive machining is the nowadays term for what you describe in the
>latter part of this post. Also includes would be in-process
>gauging,probing,etc. Hopefully these "titles" will helpothers find
>more info on what you have posted.
>
>FWIW, two of the current low-cost offerings in CNC controls have
>provision and capability for this type of use; Mach2 by ArtofCNC, and
>TurboCNC by Dak Engineering...
>
>Hope this helps, Thank you again for the post,
>
>Ballendo
>
>--- In CAD_CAM_EDM_DRO@yahoogroups.com, "David A. Frantz"
><wizard@e...> wrote:
>
>
>>Hi Kim
>>
>>
>>I have one suggestion for you based on experience obtained years
>>
>>
>ago.
>
>
>>That is to glue your tool post in place with epoxy. Knocking the
>>
>>
>tool
>
>
>>post of the top slide on a lathe, in the event of a crash, is one
>>
>>
>way to
>
>
>>keep the grief level to a minimum.
>>
>>Now I should mention that this approach was taken due to very
>>
>>
>unreliable
>
>
>>CNC controls that were attempting precision diamond turning.
>>
>>
>Since the
>
>
>>product, in the finshed state was only a few microns thick it
>>
>>
>didn't
>
>
>>take much of crash to damge things. Sometimes the diamond would
>>survive the crash and someitmes it wouldn't. The good thing
>>
>>
>though was
>
>
>>that the rest of the lathe did not suffer.
>>
>>When a upgrade to better CNC equipment took place we moved away
>>
>>
>from
>
>
>>epoxy mounted tool posts. At that point very few crashes where
>>
>>
>due to
>
>
>>anything other than operator error.
>>
>>As far as other aproaches do make use of your hard limit switches
>>
>>
>which
>
>
>>all machines should have. If you have critical tooling or
>>
>>
>product in
>
>
>>the machine and the machine supports it readjust those hard limits
>>
>>
>to
>
>
>>protect those items. Like wise careful use of soft limts on CNC
>>software that supports it will help. Additional switches wire in
>>
>>
>with
>
>
>>you hardstop swtiches may also be of help. All of these take
>>
>>
>setup
>
>
>>time so there has to be some pay off to implementing the
>>
>>
>protection.
>
>
>>Do realize that you have options to the normal hardlimit switch.
>>
>>Though I've seen it attempted many times I've never seen an
>>
>>
>operator
>
>
>>actuated E-stop that really saves tooling. Generally things
>>
>>
>happen to
>
>
>>fast especially with machines running production code, so man in
>>
>>
>the
>
>
>>loop protection is not a good idea.
>>
>>Generally fuses can help but they are rather slow. If the
>>
>>
>amplifiers
>
>
>>support it do use current limiting. This can impact your servo
>>performance and lead to debug issues so be careful and keep the
>>possibilities in the back of your mind. Certain types of crashed
>>involving your spindle can benefit from overload protection of the
>>spindle also, such an overload should also shut down your axis.
>>
>>If you in mass production and the process requires operator tweaks
>>
>>
>do
>
>
>>consider moving the required intelligence to a PC. Have the PC
>>generate the CNC code and have it update the CNC controller.
>>
>>
>Anything
>
>
>>that removes manual computation and data entry should be looked
>>
>>
>into.
>
>
>>So for example if the operator has to monitior the parts for
>>
>>
>certian
>
>
>>dimensions that are key to the process and make program parameter
>>changes based on those samplings then it is a very good idea to
>>
>>
>have the
>
>
>>entire process computerized. To get further into this example if
>>
>>
>one
>
>
>>or more of the parameters that the operator is monitoring is
>>
>>
>directly
>
>
>>related to thermal growth of the spindle or growth of the
>>
>>
>leadscrews,
>
>
>>develope a computer program to generate the compensations. It is
>>amazing how many unexplained crashes go away once simple validation
>>
>>
>for
>
>
>>rational data is done. While I'm not currently in a situation
>>
>>
>where
>
>
>>I'm up to date on the latest CNC software available I do know that
>>
>>
>some
>
>
>>CNC packages can support such coftware developement within a
>>
>>
>modified
>
>
>>form of CNC code. With modern PC's running CNC interpeters,
>>
>>
>there is
>
>
>>a very real option of running your SPC software right beside the
>>
>>
>CNC
>
>
>>code. It should be noted though that you will have to figure out
>>
>>
>how
>
>
>>to get the CNC side of the machine to re read the CNC code
>>automatically. While I realize that these are techniques of mass
>>production where you want the cheapest operator possible running
>>
>>
>your
>
>
>>process, it can be very usefull for projects of lesser scale. It
>>
>>
>is
>
>
>>also surprising how many mistakes even a skilled and reasonably
>>intelligent operator can make when under pressure or simply
>>distracted. The trick is to use accumulated intelligence,
>>
>>
>experience,
>
>
>>a bit of engineering math to remove the human mind from the chain
>>
>>
>as
>
>
>>much as possible. This will save you time and materials due to
>>
>>
>mistakes.
>
>
>>Another thing that can be very useful if you go the software route
>>
>>
>is
>
>
>>the concept of traceability. It is amazing how usefull a change
>>
>>
>log
>
>
>>automatically generated can be. What I'm getting at here is to
>>
>>
>have
>
>
>>all paramters touched by an operator saved in a log file. At
>>
>>
>one
>
>
>>point we even had the lathes computing compensations on the fly and
>>
>>
>also
>
>
>>recording those parameters in the log files. So when thinking
>>automation software do not forget LOG FILES!!!!. A little bit of
>>information can go a long way to correcting or identifiying
>>
>>
>problems.
>
>
>>So things like time and date, before and after values of the
>>
>>
>parameter
>
>
>>in question and what ever other parameters that may be of
>>
>>
>importance is
>
>
>>recorded.
>>
>>Log files are very usefull beyond the software maintenance issues
>>also. If you have a lathe making automatic tweaks to parameters,
>>
>>
>and
>
>
>>that lathe has issues, a log file can be very useful in determining
>>
>>
>what
>
>
>>went wrong. How the lathe was behaving when no body was around
>>
>>
>can be
>
>
>>very useful to the electromechanical maintenance people.
>>
>>I hope this isn't long winded.
>>
>>Thanks
>>Dave
>>
>>
>>
>>Kim Lux wrote:
>>
>>
>>
>>>I'd like to know what is supposed to "give" when one crashes a CNC
>>>tool/slide into a chuck, tailstock, etc ?
>>>
>>>Lets say we've got a servo/ballscrew set up that generates 1000 to
>>>
>>>
>1500
>
>
>>>pounds of force when called on. Next, we accidentally command
>>>
>>>
>that axis
>
>
>>>to run into something immovable: a chuck, tailstock, etc. On a
>>>commercial machine, how is this handled without wrecking
>>>
>>>
>anything ?
>
>
>>>Right now we've got a quick blow fuse that pops in our equipment.
>>>
>>>
>It is
>
>
>>>the only way that I can see to limit the damage. What is everyone
>>>
>>>
>else
>
>
>>>doing ?
>>>
>>>BTW: now that we are experienced, we rarely crash our equipment.
>>>
>>>
>I'm
>
>
>>>just curious/wondering how others and commercial equipment handle
>>>
>>>
>this.
>
>
>>>
>>>
>>>
>>>
>
>
>
>
>
>
>