Re: Raster is faster

From: Marshall Burns (Ennex Corporation), André Dolenc (Helsinki University of Technology), Albin Hastbacka (Sanders Design International, Inc.), Dick Newton (Ford Motor Company), Terry Wohlers (Wohlers Associates)
Date: Sunday, December 3, 1995

From: Marshall Burns (Ennex  Corporation), André Dolenc (Helsinki University of Technology), Albin Hastbacka (Sanders  Design International, Inc.), Dick Newton (Ford  Motor Company), Terry Wohlers (Wohlers Associates)
To: RP-ML
Date: Sunday, December 3, 1995
Subject: Re: Raster is faster
Hi Folks,
     Sorry, I must disagree with Ulrich and Terry. Raster is not necessarily faster. It depends on the geometry. And in three dimensions, that dependence is more pronounced than in two, so the history of 2-D plotters is not a good analog here. Some of the relative advantages of the two scanning methods are:
     -- Raster: Easier and less expensive to adapt to multiple jets. This makes it faster for shapes that cover wide swaths of space in each layer.
     -- Vector: Can skip between regions that need processing, without processing all of the empty space in between (except for one tool motion between the regions). This makes it faster for shapes with wide open spaces in many layers. BPM takes advantage of this characteristic by always making only the shell of an object, so that every shape has this characteristic. This is kind-of cheating, and they pay by having dismally fragile models.
     -- Vector: For deposition processes (such as BPM and FDM, but not SLA, SLS, or LOM, which also do vector scanning), easier and less expensive to adapt to aimed (instead of just downward) deposition. The BPM fabricator is the only one that does this so far, with its two rotational freedoms in the jets. The advantages of aimed deposition over downward are (a) possibility of reducing dependence on support structures, which has a whole set of further advantages, and (b) freedom, when the appropriate software is developed, from today's constraint to building in flat layers.
     -- Vector: The resolution of fabrication is limited by the resolution of the motion control systems, not by the resolution of the raster pattern, so it can be better if the raster pattern is a constant of the hardware, as it will usually be for a multi-jet head.
     I wouldn't write off vector deposition processes, like FDM and BPM, just yet! Counterarguments welcome.
     (This is an independent technical opinion. Ennex Fabrication is not engaged by either Stratasys or BPM. Copyright (c) 1995, Ennex Corp.)

Date: Mon, 4 Dec 1995 11:45:28 +0200
From: Andre Dolenc <ado@cs.hut.fi>
M. Burns writes:
 >      Sorry, I must disagree with Ulrich and Terry. Raster is not
 > necessarily faster. It depends on the geometry. And in three dimensions,
     Right. The issue here is *parallelism*. You can make a vector-based system faster by putting in another laser. Besides, if that were the only issue concerning speed, the SGC process from Cubital would offer the greatest potential for being the fastest process on the market, right? But it only works for some special polymers, you say. Yeah, but look at an injection molding device. Doesn't it squeeze a large amout of material "simultaneously" into a cavity? Who said the cavity cannot be a "layer"?
     In computer-related technologies, raster has given excellent results. It may, or may not, apply to RP.
Regards, Andre'

Date: Wed, 06 Dec 1995 09:11:31 EST
From: KHVD07A@prodigy.com (MR ALBIN A HASTBACKA)
     Raster may be faster in some cases, but there are VERY FEW cases where raster is better than vector. When we started the Model Maker development program in 1991, it was done because the raster techniques that we examined DID NOT have the quality of surface finish(all sides-- not just the TOP) that we needed.  The raster techniques that we examined INCLUDED the printhead that is in the MJM.  However, we wanted precision, accuracy, and surface finish.  If you are willing to give up some of those parameters, the raster approach has some merit for certain geometries (e.g., bricks--but then you could use a saw to cut a block of wax to make even a faster prototype).
Regards,
Al Hastbacka

Date: Wed, 6 Dec 95 13:14:24 EST
From: newton@rrm.org (Dick Newton - Ford)
> -- [ From: Al Hastbacka * EMC.Ver #2.10P ] --
> Raster may be faster in some cases, but there are VERY FEW cases where
> raster is better than vector.
     Yes, and there are very few cases were VHS works better than BETA video tape format. It takes more than just superior technology to have a successful product. Things like ease of use, servicability, reliablity etc. all have an impact. 
     At this point, speculating about which is faster/better (raster or vector) is just that, speculation. Only the market will tell which is "Better"
Dick Newton

Date: Thu, 7 Dec 1995 07:32:13 -0800 (PST)
From: "M. Burns (marshall@ennex.com)" <mburns@netcom.com>
On Wed, 6 Dec 1995, MR ALBIN A HASTBACKA wrote:
> Raster may be faster in some cases, but there are VERY FEW cases where
> raster is better than vector. When we started the Model Maker
Al,
     I don't get what you are saying here. It sounds like you are arguing  for the vector technique. But I see your system as a raster device. In my  understanding, the raster-vector dichotomy from the old CRT days refered  to the difference between scanning the beam in a regular pattern of lines  to fill the whole screen with on-or-off pixels (raster) and scanning the  beam along arbitrary curves to form the desired shapes (vector). Doesn't  your technique fall into the first category?
Best regards,
Marshall Burns

Date: 07 Dec 95 19:07:17 EST
From: Terry Wohlers <73417.1465@compuserve.com>
Al Hastbacka writes:
> If you are willing to give up some of those parameters, the raster 
> approach has some merit for certain geometries (e.g., bricks--but 
> then you could use a saw to cut a block of wax to make even a faster 
> prototype).
     I've seen several parts produced by a 3D raster printing device that contain complex surfaces and very fine detailed features.  Plus, the surface finish is excellent.  Al: These parts just might knock your socks off.  It may be a little dangerous to conclude that 3D raster devices can only produce blocks.  As for speed, brace yourself because you may be in for a surprise.  And over time, the speed of the process is likely to improve.  Dick Newton mentioned that the market will decide whether the raster approach has merit.  I believe it does.
Terry Wohlers


Previous message | Next message
Back to 1995 index