It is key to define exactly where the vendorís responsibility ends and begins. Time should be taken to work out an easily understood contract. A ďcustomĒ could be as easy as getting a motor with a custom shaft and your connectors on a specific length of cables. Perhaps you would just like a vendor to assemble the several stock parts into an assembly with only one ordering number to relieve purchasing from buying and tracking a larger number of parts and/or vendors. Of course, a full custom may well be the most desired, as well.
Many machines have evolved, rather than have been designed. As a result, many assemblies are oversized and therefore larger and more expensive than they need to be. Vendor-developed custom assemblies can be a real way to manage risk in development and production. If your vendor starts designing with a known load, speed, settle time, and life calculation, then part count, assembly weight, and settle time all may be reduced while lowering cost. However, the biggest advantage may be that the vendor has said, under the criteria you have laid out, the device will last X amount of time and it will do that on the vendorís warranty. Also, development time may be reduced allowing your (probably very lean) in-house engineering to move on to a less mundane task. Perhaps they will be able to spend more time in your companyís core competency.
If we look at changing an assembly from commercially available to custom, a key question is, will I have to resubmit the new design? In many cases, with the FDA for instance, if the software or firmware doesnít need to change to accommodate the new assembly there is less risk of resubmitting the entire machine for approval. If you need to, add the cost of resubmitting the assembly into your ROI equation so your results are accurate.
The electrical and mechanical footprints can each cause different issues. For mechanical, in many cases, as long as the point of interest doesnít change you should have a very easy conversion. Many times room may be found by making an assembly shorter with a higher natural frequency that will make your production machine weigh less, settle faster, and last longer. Electrically it may be very beneficial to design using the same wiring so that the new custom units may be swapped into the field. If you look at your unique application, will the wiring or mounting need to change, and if so, what does that add to the applied cost of the system? Obviously keeping the same wiring and mounting is optimal.
We should also bring up the difference in support between the standard units and the new custom units. Once the new production units go into the field, is it realistic to just swap units in the field or do you need to actually troubleshoot at the component level in the field? If they actually need to be repaired in the field, do they cause a need for retraining in the field? Will you have to stock both new and older units for service?
You must determine your level of risk to a new approach and the dollars saved by that approach. Say, for instance, you can get rid of the existing PLC or motion controller and go with smart amplifiers on the electrics. This may mean that the calls from your GUI, and to the other systems that the assembly must communicate to, will now be different. This would be a total software overhaul with very high risk. The levels of risk depends on whether you will do this in-house or contract this job out for the real man hours it will take to develop, test, prove, and write a new users guide. You may be very tempted by the savings to proceed, but after an ROI your eyes are fully opened to the risk when doing this.
Of course, no one answer fits every situation. So why would standard product always be the best answer? Take a hard look and you may be surprised by the results.
Steve Reese is the strategic account manager with Kollmorgen.