Q: My company builds large power control field units for the oil industry. We had been outsourcing much of the fabrication work, but a while back we decided to bring more manufacturing in-house, including sheet metal bending. Although we’ve done machining in-house for years, we had never done any sheet metal manufacturing before. Much of it is straightforward, except for bending. No one in our shop had done any actual bending on a press brake before. We have worked through many of the forming issues on our own, except for one involving panels.
Each product we produce uses between 10 and 24 panels, and we’re having trouble assembling them. Even though we form them within tolerances specified in the print, they do not match up when we get to the last panel. Ultimately, we have to manufacture a single custom panel to make it all fit correctly. Why does this happen consistently, and what can we do to correct it?
A: I get similar questions from time to time about cumulative error when placing multiple “like” parts together to build a completed unit. The solution can be a bit tricky. It is all in how you are gauging the parts.
Errors and Gauging
Some minor errors can be attributed to the press brake, but not many. Most errors come from the sheet metal itself, though they’re not really “errors” so much as they are variations within the material aspects and their related tolerances. For example, sheet metal has a tolerance zone for thickness. In 16-ga. sheet, that tolerance runs from 0.053 to 0.067 in. Material has tolerance zones around hardness, yield strength, ultimate tensile strength, and more. The material is considered “good” if its characteristics remain within the tolerance limits.
Nonetheless, all of those minor variations within the tolerance zone can affect forming dramatically. For example, a change in thickness changes the final formed dimension. It also changes the bend deduction and the bend angle. Each of the variations adds up to a lot of differences from part to part. Some forming problems can come from machine errors, but again, these are minor.
Calculating Cumulative Error
You’re probably forming something similar to the panel in Figure 1. Let’s say your overall panel dimension is consistently off by 0.015 in. It’s within your overall tolerance of ±0.020 in., but by the time you multiply 0.015 in. by 20 panels, how much cumulative error do you have? That’s right—you have a total of 0.300 in. of error when the unit is assembled. No wonder you need to build a single custom panel to complete the project, wasting resources, time, and talent.
If we gauge from the cut edge to the first bend, we hold that dimension; that is, we maintain the dimension between the edge and first bend and push the error inward. Now, what happens if you gauge off the first flange to make the second bend? You’ll hold the dimension between the first and second bend. So, where does the error end up? It goes to the overall dimension. Add up that error from panel to panel, and you’ll find the assembly won’t fit and you’ll need to make that final custom panel.
Generally speaking, you need to gauge the first bend from the cut edge to hold that dimension—especially if there are called dimensions between features on opposing flanges, as shown in Figure 1.
Our error must go between flange one and two, and we need to hold the overall dimension. But how do we do this? The answer is all about where you gauge the part.
Again, whenever you gauge from a cut edge to the first bend, all the error moves inward, and you hold the dimension from the edge to the first bend. Next, the stops move, and you make the second bend off the first bend, holding the dimension from flange to flange, between the first and second bends—but the error hasn’t vanished. Big or small, any error moves inward and becomes part of the overall dimension. In this case, when you have 10 to 20 panels stacked side by side, the overall dimension is the last place you want the error to appear!
Read more: The art of gauging on the press brake