What's new
What's new

Creating a rough estimate of machining time without CNC-preparation

When my friend Steve built that cost estimating spreadsheet 15 years ago it was a workbook of a couple of dozen sheets and some very complex macros. Took him a lot more than 80 hours. He was a career manufacturing engineer with a degree in Mathematics from the University of Chicago. The spreadsheet was designed for someone experienced in machine shop processes and mechanical design to use. Because solid modeling was in its infancy at the time one was required to have a detailed drawing to study and visualize the process as well as make rough calculations of machined features. The key here is that a mature business needing machined parts will employ good mechanical designers who know what expensive design features to avoid where possible. The estimator seeing one such issue on a drawing will immediately flag that element for special estimating effort. So a few minutes spent by the estimator counting and measuring features, roughing out the CNC process variables like tool changes, tool travel distances in both cutting and rapid traverse and cycle stops will give quite usable estimates of time. A good estimator will have at hand the current capabilities of machine tools in use so those numbers can be entered in the time calculation. Another very important factor is the relative size of parts. If you can hold it in the palm of your hand without straining ordinary materials like aluminum and steel will have low costs and might be made out of mill shapes. On larger parts materials costs start to strongly favor near net shapes on production parts. This greatly simplifies the cost estimating because such parts have much simpler material removal features and will generally be better optimized designs for efficient machining. Add to that solid modelers that can easily give the estimator part volumes and weights, part areas and easily produce a visual solid model of the material removed by machining. (providing there are models for both the near net shape and the finish machined part.)
Ed Weldon
 
Might work in my favour, may blow up in my face like a Peter North film. Could be good or bad, depending on how you look at it. LOL



It's too early for this. I'm freaking dying over here. Aside from the anodizing post, this might be the best thing I've ever seen you post here. :bowdown:
 
About 15 years ago a fellow I worked with developed a fairly sophisticated Excel spreadsheet for our company to do just what you are looking at. I worked with it a lot myself and felt it was a pretty powerful way to estimate part fabrication costs. The purchasing department quickly lost enthusiasm for our work because they lacked the most basic understanding of it and managed to accomplish little other than infuriating the supplier base of fabrication shops. Another division of the company bought proprietary software to do the same thing, paid well into the 5 figures for it and used it with some success. At the time (2001) our company was pretty close to Autodesk so we shared the method with them. At the time of my retirement late 2002 they hadn't done anything with it although the method seems a natural for 3d CAD systems to give designers a real time cost estimate of their design as it develops. Maybe that's actually happening now. But when I was working the design engineers still threw the designs "over the wall" and the purchasing people in the manufacturing department looked like heroes when they compared their production volume contract costs with the prototype costs even if the company's competitors were buying similar parts for half the price.
The foundation of our spreadsheet approach was the estimating method outlined by Ostwald in the 1984 edition of the AM Cost Estimator. A rare book even in the USA. The basic idea was to use the retail cost of hours on a specific machine or work center and estimate minutes of time the job is being processed there. Add the material cost and a scrap and rework factor. The idea is that all the overhead and profit stuff the company accountants deal with to satisfy the tax collectors are combined into one hourly rate on each process. That's the way job shops in the USA quote their jobs. The additional effort on the part of the company planning to buy a job is to maintain credible local cost information on various machining processes as well as materials. Ostwald's book had that info for 30 years ago. A lot has changed in fabrication technology. But a lot hasn't. You still have to set up machines, load and unload parts, change tools, cut metal. Technology changes the time constants; but the operations are still there. And note here we are counting work center time here. Labor hours are just part of that single work center price whether it's a CNC machine in a lights out line or a simple assembly bench.
Ed Weldon

AutoDesk may not be doing anything along these lines (I wouldn't know, I'm not as up to date with their stuff) but Dessault Systems (the company who makes SolidWorks) has an integrated system called 'Design for Cost' - Design for Cost: Manufacturing Cost Estimation and Quoting | SOLIDWORKS

We looked into it briefly as a company, and we have also used other estimating software, but all of these systems are geared towards work in your own shop, and of pretty limited use in a lot of cases.

Has anyone else used this program? Experiences with it?
 
A possibility would be to exclude the tolerances and the finishing from the model, since its a comparison between the methods that is the most crucial part of the project we could simply state that the amount of work required on a casted or 3d-printed part is equal to the amount of work required on a rough-machined part (with an added factor for setting up the rough material in the machine). And from that assumption work out a simpler formula for machining time. What do you think of this approach?
 
I'm starting to think I should quote someone 200 bucks... just to quote them. Because I use my CAM to estimate machining time, I have to CAD/CAM it first, then put it into my spreadsheet for the final quote which includes the cost of the CAD CAM I've just done. Maybe its a really bad way of doing business or a really good way of loosing money if they don't go ahead with it. Would be interested to see other ways. At least if I say "It will be 200 to quote you". Pressures on us both to go ahead with work. Might work in my favour, may blow up in my face like a Peter North film. Could be good or bad, depending on how you look at it. LOL


I am glad I did not read that first thing in the morning,I would of spat coffee all over my computer screen.
 
Just putting in a little draft from the report here. Anything that you consider wrong or would add? This is not a finalized version and the language is a bit off here and there.


"If you want a mathematical model that could give you beautiful graphs which tells you when a method gets more economical than another, you need to get the answers in absolute numbers. To make such a model, you need a different approach.
The idea was to measure the time for the rough machining process, using one of the manufactures NC-program. This should be done for a couple of different parts and then a system of linear equations should be made. One equation for each part that has its rough machine time measured. The variables in the equations should be the area of non-developable surfaces on the model, material removed and the distance that needs machining with a tool that has to process more than four times its diameter in height, and a variable of the number of machining steps, which variables that will be chosen can be determined later.
The idea is that the equation system should give you how much of each of these variables affect the machine time.
It was realised that it is very important that the chosen factors match the machine time. Some test was made and an error of 10% in the machining time could give a as much as an error of 50% in one of the factors. To overcome this problem, one solution could be to use a lot of machine times and then use the averages the you get on the factors, if the chosen factors are the factors that affect the time most, it will probably give you a good estimate on the machine time."
 
Complexity in terms of features that require non standard tools would add to the cost? This is an older thread but still interesting. Any new thoughts on how to rough cost machining from solid (especially as the original authors were looking to compare with metal printing/additive manufacture). They did publish a paper, and used quotes from their machine shops for specific parts to try and anchor their comparisons and extrapolations.
 








 
Back
Top