Attached are some photos of a handle for a speargun. Would it be possible to model this? What would be the best approach?
I've had a few does at this. The first step is to obviously draw and outline of the handle. 
I had a modeling challenge on my blog several months ago where the theme was handles. Check out the links. There were a lot of good submissions, some with models you can download.
The handle will be used as part of the overall model of the gun. The purpose is manufacture of the barrel, not the handle.
To give some idea, I started off with a few sketches based on handle measurements. The first sketch being vertical profile in section. I'd like to be within a millimetre (0.039 or 1/25 inch) on the overall contours. The parts at the handle interface need to be say within 0.2 of a millimetre (8/1000 of an inch), but these are typically flat surfaces.
There are different techniques for modeling a finger grip on a handle. I prefer to use a boundary surface by sketching the path (direction 1) for the finger grips with profile or cross-section curves (direction 2) to control the height / elevation (see image below).
This could also be done using an indent feature, or deform (curve-to-curve) feature. However, I would use the boundary surface feature first and try the indent feature, or a combination of both second. I rarely use the deform for anything that is to be manufacturered by injection-molding.
If you need help on this project, I could use the cash as I am in-between paid projects and cannot attend SW World this year.
Chris Thompson
Derek,
I do not mind answering questions, and I ask some myself on this forum from time-to-time. The purpose of these forums are for asking questions, so everyone should feel free to ask for advice.
Only when people start asking for work, which requires more than one hour maximum of my time, do I ask to be paid as a independent contractor / consultant. I have a few start-up companies not be able to pay for my services as they were trying to attact seed-money from VCs.
Concerning your question, I think Matt Lombard has an example of a finger grip on his website as well. I do not recall where it is located, or what technique he used.
Find someone who can 3D scan it for you, then create your surface model over top of this. With all the Next Engine scanners out there, it should be pretty easy to find someone who could get you .020 accuracy for cheap. I'd offer some suggestions, but you're on the wrong continent for any help I can provide.
If I were to create this I would probably do it with 1 boundary surface for the three bottom fingers, then trim away for the trigger. Then create a boundary for the index finger, and another for the top part (shown on the bottom in your image). Then I would blend both of those and build the trigger guard. All the surfaces would be boundaries, exept a few filled surfaces for some of the blends at the end.
That's just a rough guess though, looks like a day or two of work, to me.
Well Derek.
Working from the assumption that 3D scanning is not an option.
Firstly I would take measurements in the FRONT, TOP and RIGHT view. Normally as accurately as possible. This will always give you clues as to how the original was modeled. I would spend most of my time getting the measurements right. This is very important.
Secondly I would start modeling. At this point I believe any approach is a good approach. Keep on trying different methods and you'll get closer and closer to the original model with every attempt.
This is the nature of reverse engineering.
However you'll find that moment when everything falls into place very rewarding. And the result will be a model that can be used to reproduce the part if anything goes wrong with changes or modifications to the original.
The most common mistake I see in this sort of model is that the user assumes the surfaces are somehow defined by their borders. This is a trap that usually leads to unmanageably wrinkled surfaces.
For instance, imagine what the finger indents would like like if they were extended in every direction and not cut by other intersecting surfaces. Then each groove becomes a simple sweep, perhaps with an additional guide curve. The groove surfaces then take their final shape by where they intersect the larger outer surface (and are thus trimmed).
Hey Derek,
Fun project, it's ambidextrous so that saves you a ton of work. Yeah, you need to sketch the profile first...even if you're gonna cheat (indents and squishy solids). This picture isn't much help, you need something with less perspective. You can try a scanner, or cut it in half and then scan it!
The ridges of the finger indents are actually fairly complex so you can't just drop a plane in there, a 3d curve driving a multiple section loft will do it. The valleys of the finger indents are normal to the "main" sketch. You can drop a a plane in there with a 2d sketch to provide an anchor point to keep everything curvature correct. The idea is to get a skeleton that you can draw the loft sections onto so you have control of tangency and curvature, things you can dimension and control. Anyways, my opinion...let me know if that's completely greek I'll make you some notes if you want.
-Mike
Progress to date is posted below.
Using the boundary surface I notice it is critical to get the number of curves and straignts the same on both boundaries. Can anyone advise how to join two curves that are converted from the same sketch that has a break. This means that the two converted curves are coincident with the same line and just need to be joined. Thanks.
Feel free to coment on how best to proceed.
Derek wrote:
Can anyone advise how to join two curves that are converted from the same sketch that has a break.
Try using Fit Spline (Tools> Spline Tools)
Derek,
I don't have time to look at your actual model, but I'm impressed by how well you did, especially given your lack of experience. The only thing I noticed from your pictures is that the corners of the front edges of the finger grips and the back edge of the trigger opening look much too sharp, so that they would create high pressure points on your fingers as you held it. Without looking at your part I can't offer much advice on how to smooth those areas out.
Jerry Steiger
Thanks Jerry.The handle is made by another company. I am more focussed on how to model the handle rather than handle design.
The main issue I'm having here is getting continuity between boundary surfaces. Much of the trouble relates to getting open edges or boudaries ie. those not connected to another boundary when formed, to have the right curvature. This creates problems when another boundary surface connects to the open boundary. If you specify an equal curvature relation to this original open boundary any misalignment gets accentuated. You can see this in some of the patching. Not so bad if you are freeforming but a problem if the rest of the model is following sketches. I suspect that the only answer is lots of guide curve sketches.
Here are a c ouple of screen shots to show the problem. The bottom surface was ceated using surfaces 1 and 2 with an open edge of surface 1 being used as one of the boundaries of surface 2. There is a severe change of curvature in region 10. Surface 3 shows a surface created from one boundary surface. The surface is smooth and continuous in region 20.
I guess the ideal is to minimise the number of surface boundaries and joins that are used. That is not always feasible.
Derek,
I wish I had the time to look at your part and give you the feedback you want. I feel guilty enough just taking the time to write this!
It's hard to come up with hard and fast rules for working with this type of shape. As you said, it is often best to minmize the number of surfaces and boundary curves. (Many people fall into the trap of adding more guide curves to try and get the geometry to behave, when the added curves just make it worse.) On the other hand, sometimes you need to split a surface up into two or three surfaces when the "character" of the surface changes too much for one surface to work well. Sometimes a Fill Surface will work better than a Boundary Surface. Sometimes a Loft is the best choice. I end up doing a lot of trial and error.
For your region 10, you might want to try breaking the surface up into four surfaces or split it into two surfaces running the other direction. Or you might try adding a guide curve running right to left. You might want to make it one surface, with a left to right guide curve and an up to down guide curve. You might want to make the center surface that you haven't added yet wider, so that the broad flat area is all in that one surface and the more abrupt changes are limited to the top and bottom surfaces. You can spend any number of hours trying out the possibilities! With any luck you will be much better at this type of modeling when you are done. You've made a heck of a good start!
Jerry Steiger
Jerry Steiger
Sorry Jerry. Didn't want to pressure anyone. Just keen to get feedback from people who are able. I might keep bumping this thread from time to time. I understand this week is a busy one in the world of SolidWorks. I'm also getting feedback from my VAR on the matter.
The VAR has suggested using the intersection curve feature in the sketch tools to lay part of a spline, used for a mating surface, over the edge of the already formed surface. I'm really looking for a way of controlling the curvature near the open edge. I played around with free forming. It was too difficult to control.
I remodelled the handle with some fairly major changes to the sketches. Was able to do a cut out on the trigger and use a variable fillet to give a more gradual radius. I also used a projected 3D sketch from two 2D sketches to get the ridge that runs along the base of the trigger and onto the handle. Both worked well. Set at the start of the featue tree this ridge formed the basic topography on which other 2D profiles could be contructed.
In the end the approach that worked best was to create base sketches that give the characteristic shape of the model and whose profiles could be easily measured and constructed. That included front, top and side views, ridges and trigger guard profiles. A matrix of 2D section sketches using splines was created over that basic outline. More 3D spline sketches were added following the contours of the surface and in convenient locations for building the surfaces. Sketches were converted off existing sketches to get a mesh of lines for the region to be surfaced.The other sketches were hidden using display states and only the converted sketch shown. It was relatively easy to then select the required sketches. Selection manager tools were used extensively in creating and selecting sketches and surface edges.
I ended up with a surface on either edge of the handle running from the handle base to the top of the barrel, one at the base of the handle, two over the handle and back end of the trigger, one along the top of the barrel, two over the trigger and then a number of smaller surfaces for the trigger and region where handle mates to barrel.
The whole process is relatively tedious and time consuming though no doubt that would be reduced with experience. A 3D scanner may have been quicker and cheaper. I've never used one and am not sure of the limitations of models formed in this way.
The later attempt gave a better surface but I'm still having problems controlling the curvature of free edges of surfaces.
Derek,
No need to apologize; this kind of exchange is what makes this forum worth visiting. Which is one reason why I want to respond to you, even though I don't have the time to do a very good job of it. On your last question, controlling the curvature at free edges, there are several methods that I have used. One I use quite a bit is to build a mating surface that is just a tool to help you get your real surface the way you want it. I use this when I want my surface to have a particular draft angle at the parting line, for instance. The other method I use a lot, which you are already using, is to use guide curves perpendicular to the free edge.
If you are going to be doing this type of work fairly often, you might want to buy Matt Lombard's Surfacing Bible. Also check out Ed Eaton's tutorials:
http://www.dimontegroup.com/Tutorials/SolidWorks_Tutorials.htm
At the rate you are picking this up, you'll be giving presentations yourself soon.
Jerry Steiger
Generally have to say that I'm pretty impressed with the SolidWorks surfacing capability. Not that I have tried any other surfacing packages. I've played around with it for a while and am pretty much a complete novice in this aspect of design. Given the tools, the hard bit is working out strateges for using the tools to achieve the required end. The rest is just trial and error and plenty of patience.
The software performed pretty well in general. I had one bug that caused the computer to crash a number of times. This seemed to right itself in the end and I could carry out the same sequence without failure. Not sure what the problem was. Given the way spline sketches go haywire I was surprsied there were not more crashes.
The renderings took a matter of seconds in Photoview 360. The only gripe there would be the difficulty in rotating the model.
This is a surface model. Any hints on how to convert it to a solid would be helpful.
Glad you finally made one and happy to know that it was a good learning curve for you as well
Any chance or possibilities of sharing the final part.
Try knit command under surfaces to combine all the surfaces and convert to solid.
