Re: Tolerances

From: Marshall Burns (marshall@ennex.com)
Date: Mon Jun 08 1998 - 12:34:23 EEST


Phil,

    Your numbers below seem to suggest that the SLA is much more accurate than
the Sanders and that the Actua is almost as accurate as the Sanders. Do you agree
with this interpretation of your data?

    There has been little if any discussion on this list of the difference
between precision and accuracy. Precision relates to the ability to make fine
details; accuracy relates to the correctness of dimensions. They are related but
they are not the same thing. It may be that the forte of the Sanders machine is
not accuracy, but precision.

    Yet I wonder why this would be the case. Are there internal stresses induced
during drop-on-drop deposition that are resolved through creeping distortions?
Internal stresses used to cause terrible inaccuracy in SLA models but improved
chemistry and process dynamics have drastically reduced those stresses. Is there
similar work needing to be done on drop-on-drop techniques such as in the Sanders
machine? But still, why is the Actua coming out ahead?

Marshall Burns

Professor P M Dickens wrote:

> Marshall,
> It might be worth looking at a paper that we published at the Rapid
> Tooling & Manufacturing conference in Denmark last year. - "The accuracy
> of RP models for Investment Casting of Injection mould Tool Inserts",
> Heather Almond et. al.
> I give a few results of accuracy tests:
> Sanders
> epsilon 90 = 0.35mm, 39% of error values within +/- 0.1mm, max deviation
> = 0.7mm
> Actua No.1
> epsilon 90 = 1.0mm, 38% of error values within +/- 0.1mm, max deviation
> = 1.6 mm
> Actua No.2
> epsilon 90 = 1.2mm, 26% of error values within +/- 0.1mm, max deviation
> = 1.7mm
> SLA QuickCast 1.1 No.1
> epsilon 90 = 0.05mm, 82% of error values within +/- 0.1mm, max deviation
> = 1.1mm
> SLA QuickCast 1.1 No.2
> epsilon 90 = 0.05mm, 86% of error values within +/- 0.1mm, max deviation
> = 0.02mm
> SLA QuickCast 2.0 No.1
> epsilon 90 = 0.2mm, 44% of error values within +/- 0.1mm, max deviation
> = 0.4mm
> SLA QuickCast 2.0 No.2
> epsilon 90 = 0.1mm, 70% of error values within +/- 0.1mm, max deviation
> = 0.3mm
>
> Phill
>
> Professor Phill Dickens
> De Montfort University
> Leicester
> England
>
> Marshall Burns wrote:
> >
> > Dear Dan,
> >
> > Okay, but it would be interesting to know what are the tolerances that
> > you guys who are actually using the various machines are accustomed to
> > getting from them? Some people say that an important reason for choosing an
> > SLA today is that the machine is mature and gives good accuracy. How good?
> > And can it be as good as Sanders?
> >
> > Marshall
> >
> > DANIEL CHARLES DAVIS wrote:
> >
> > > Marshall,
> > >
> > > That's what I get for answering before my morning coffee. Two
> > > thousands......twenty thousandths.......what's a few thou between
> > > friends. Sorry for any confusion caused by my brain failure.
> > >
> > > This thread is timely for me as I was just going over tolerances here
> > > for some plastic parts. I'm sure you realize that in the auto industry,
> > > especially Asia, our tooling is held to 0.1-0.2mm (or even looser) for
> > > 90+% of what we make. That leads me to wonder: how many consumer
> > > product industries REALLY need to hold accuracy to 0.05mm overall? Of
> > > course that can get us into a whole debate regarding engineering
> > > practices ala Taguchi methods and "robustness." This is the wrong forum
> > > for that although I would be happy to discuss in private.
> > >
> > > Regards,
> > > Dan Davis
> > > Head, RP&M Division
> > > PROTON Berhad
> > > Hicom Industrial Estate
> > > Batu Tiga, PO Box 7100
> > > Shah Alam, 40918 Selangor MALAYSIA
> > > +60 3 515-2380
> > >
> > > > ----------
> > > > From: Marshall Burns[SMTP:marshall@ennex.com]
> > > > Sent: Friday, June 05, 1998 10:36 AM
> > > > To: DANIEL CHARLES DAVIS
> > > > Cc: 'James Theoharris'; 'RP-ML'
> > > > Subject: Re: Too many SB's & a little late
> > > >
> > > > Dan,
> > > >
> > > > Would you please clarify? You seem at first to be implying that
> > > > machines
> > > > other than the Sanders can hold 0.002 inch tolerance. But then your
> > > > data
> > > > talks about holding 0.010 to 0.014 inch. Accuracy has always been a
> > > > complex
> > > > issue. What point are you actually making here?
> > > >
> > > > Marshall Burns
> > > >
> > > >
> > > >
> > > > DANIEL CHARLES DAVIS wrote:
> > > >
> > > > > Hmmmm, two thousandths of an inch. Are you saying that all RP
> > > > equipment
> > > > > except Sanders is only accurate to that or is this only the Actua?
> > > > >
> > > > > Of course the Actua (and Z-Corp and other concept modelers) will not
> > > > > hold tooling tolerances. If we are talking about SLA & LOM, we
> > > > > consistantly hold 0.25-0.35mm (0.010-0.014 inch) tolerance evben on
> > > > full
> > > > > envelope parts. I have many CMM printouts to support that. Maybe
> > > > the
> > > > > folks in question just don't know how to dial in their machines?
> > > >
> > > >
> > > >

--
Marshall Burns
Marshall@Ennex.com

***************************************************************** ***** ENNEX(TM) CORPORATION ***** Fabbing the Future(TM) ***************************************************************** ***** 10911 Weyburn Avenue, Suite 332, Los Angeles, U.S.A. 90024 ***** Phone: +1 (310) 824-8700. Fax: +1 (310) 824-5185 ***** E-mail: Fabbers@Ennex.com. Web site: http://www.Ennex.com *****************************************************************

For more information about the rp-ml, see http://ltk.hut.fi/rp-ml/



This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:45:51 EEST