New machines vs refurbishment?

At the AIMCAL Fall Technical conference it is intended to have a discussion on the above topic. In my time I have heard passionate arguments for both sides of the question.  Some have bought a second-hand machine and refurbished it without problem and have done so at a cost much lower than they could have bought a new machine.  Others have done the same thing but have been less successful with the costs of refurbishment being much higher than predicted and the time taken to obtain working system much longer that it would have taken to specify and have a new machine built.

Where it becomes easier to argue the case is for a completely new process. Here, unless the process is so similar to the existing process, it usually means that any second-hand system is going to be not quite ideal for the process. Anything that is not quite ideal is likely to be a real problem for production. It usually means that the operators are always fighting to make the process run well which is time consuming and costly.  So the trade off here is between the reduced capital cost with higher running costs versus the higher capital cost and reduced running costs.  The difficulty here is to accurately predict the real running costs. This is where experience tells me that is many of the systems I am aware of that have tried to force fit a process into a machine that was not designed for the process have always been significantly worse and most companies, with the benefit of 20/20 hindsight, given a second chance to chose between anew machine and a refurbished machine would opt for a new machine.

Similarly some companies have chosen not to use one of the regular system manufacturers to build their system and this can add to the build risk.  Any company that is building a system for the first time potentially adds to the risk involved. You may have to pay for their learning experience. This may be found as a longer than normal list of small faults in the system or it can be something more major depending on the experience they have and what mistakes they have made.  This all takes time and money and although the initial price for the system may seem attractive the final cost may end higher that buying from the most expensive established supplier.

If you have a metallizer that is getting old and less productive it can make sense to consider progressive updating of the system. There is still a risk that an old system will still have some problems such as fatigue failure of welds that can cause a vacuum leak but the same is true of any system.  In future this upgrading process may happen more frequently, at least for the control systems, because of obsolescence of the electronics. If the metallizer has been in the same company for many years there will be a history of the failures. This may be used to plot the type and frequency of failures. This can be an indicator of what parts of the system could be upgraded and show the best cost benefit. Whatever fails most frequently or with a high frequency and cost a lot to repair or causes most downtime would be high on the list of items to improve or replace.

So as you can see there are plenty of arguments both for improving existing machines as well as for buying new machines.

But that is just my view, from the outside, as a consultant.  What are your views?  Why would you choose to either buy a new machine or buy second-hand or refurbish?  I would like to hear of your experiences of what went well or what went badly that has coloured you view of which route is best to go down.

I will also report back after the AIMCAL Fall Conference in October with a summary of what was presented in the discussions there.