I'd just like to compare notes with a few machine designers about how they figure time on machine design tasks. I personally find that about
1 1/4 hours per fabricated part is minimum to estimate for design (3D) and detailing (2D), and that's really a very conservative minimum. 1
3/4 hours to 2 hours is probably more realistic unless the task is very straightforward. Then one has to figure in time for changes and time to do assembly drawings. It's certainly not nearly as simple as that, and there are a lot of variables that can drive an estimate drastically one way or another, but that's where I begin. How about you? I'm just looking for general approach "rules of thumb" as contrast to how I start off.
Don't know if this helps you, but we do an automotive wheel (family, to be fair) in two days. We aren't very good and we don't work very hard, so that oughta give you an outside number for that kind of part.
My rule of thumb is 2.5 hours per part. My field is custom automation, so the following provisos apply:
The equipment designed is usually a one-off deal, not something that needs to be extensively refined for mass production.
However, it needs to work the first time. Redesigning the machine which has been built and is not working because its underlying concept is wrong is NOT an option.
A significant amount of on-the-shop-floor tweaking is required to bring the machine to its customer-accepted, shipping condition. The results of this tweaking need to be reflected in the as-built documentation.
Since this is a one-off piece, not all minor details are reflected in the model and/or documentation. For example, although all the fastener holes are present and properly dimensioned in the model and on the drawings, the fasteners themselves are usually not inserted into the model and therefore are not reflected in the Bill of Materials.
When my ex-employer was still designing these machines in 2D using good ol' AutoCAD, our Mechanical Engineering Department manager did a time study. Results were pretty consistent across a team of 6 engineers. When the full project, from the date PO was awarded to the shipping date, was taken into account, the per-part time spent on it by the engineer was about 2.5 hours (concept design, creation of assembly, detailing, documenting changes to create as-built documentation). So to break even in the engineering part of the project we needed to have these 2.5 hours per part.
Since I switched to Pro/E in 1998 I find that when I estimate the number of parts in the machine within 10-15% of the actual final number and plug in these 2.5 hours, I am actually left with a pretty sizable safety factor. So when possible I still use the 2.5 hours rule: it is nice to know you can certainly beat your own estimate.
A very useful post, Alex, and tends to echo my own experience pretty much. It's nice to get that kind of "feel good" confirmation. Am using SolidWorks which is probably pretty much the same as Pro/E as long as one is competent with it (and we are). Thanks for the time you took!!