The gripper is the project
Buyers compare robot arms. Integrators worry about grippers. That gap explains most of the surprises in a first automation quote.
An arm is a well understood, mass produced, largely interchangeable thing. Several manufacturers will sell you a competent arm in the payload and reach class you need, and the differences between them matter far less than most sales conversations suggest. The part that decides whether your cell works is the thing on the end of it, and that part is specific to your product in a way the arm never is.
Start with the arithmetic nobody enjoys. The arm's payload rating has to carry the part and the gripper. A vacuum tool for cases runs a couple of kilograms. A palletizing gripper with fingers and vacuum can be five or more. A screwdriving tool is heavier than most people expect. If you size an arm to your part weight and forget the tool, you have undersized the cell, and this is the single most common error we see in specifications written by enthusiastic first time buyers.
Then there is the harder problem, which is whether the tool can hold your product at all. Vacuum needs a surface that seals, so vented, dusty, flimsy or heavily printed cartons fail in ways a clean sample never reveals. Fingers need a graspable geometry and enough clearance to get around the part. Magnets need ferrous material and derate badly when the part is not flat against the face. Every one of these is obvious in hindsight and invisible in a specification document.
This is why we push for a picking trial on your actual material rather than on the clean samples that come out of a drawer. Test the dustiest carton, the oiliest sheet, the thinnest gauge, the damaged one at the bottom of the stack. If a tool handles your worst material, it will handle your best. If it only handles your best, you will discover that on a Thursday when a bad lot arrives.
There is also a published specification trap. Many gripper makers publish two payload ratings: a force fit figure where friction holds the part, and a form fit figure where the part is shape locked. The form fit number is often forty to a hundred percent higher, and it is the one that appears in marketing. Unless your part is genuinely captured by the finger geometry, size on the friction number.
None of this makes gripping mysterious. It makes it the part of the project that deserves the most attention, the most testing, and the most honesty in a quote. If a proposal has a paragraph about the arm and a line item about the tooling, ask why round.
Size the arm for part plus tool, test the tool on your worst material rather than your best, and treat the published grip rating as two different numbers, because it usually is.
Put it to the test on your own job
Tell Joe what you are running and he will apply this to your numbers, including telling you when a robot is the wrong answer.
More notes