Post By Tonytn36
Post By Miguels244
How do you ensure you get truly reliable equipment?
I work for a small company that buys a fair amount of semi-automated, custom equipment per year. We have a tremendous amount of problems with getting reliable equipment. It seems like everything we get in - "turn-key", we end up having major downtime, multiple days of downtime every 3-4 months. Most equipment ends up taking 9-12 months before it is reliable enough that it doesn't require constant babysitting. We lose almost any ROI by buying huge inventories of spare parts, outside service, and missed shipments when we get the inevitable "big screwup"
So I ask this, as simply as possible.
How do you ensure that you get the most reliable equipment possible?
Do you have the expectation that a piece of equipment should be able to run at least a year with no unplanned downtime?
Do you extend the construction timelines, have strict QC checks early on in the process? Such as serious plan reviews, machine runoffs?
Do you change the payment terms so you are holding a larger percentage on the back end? Like , 20%/50%,30% instead of 40%/50%/10%? So the integrator has a very vested interest in getting that final acceptance payment?
My guess is that the answer for a Fortune 500 company would be "Tons of lawyers". For a Fortune 1000, "We do it all in house, what's the problem?"
I know a lot of integrators follow this board so I want to hear their opinion too.
First off, you need a set of standards and procedures for machine purchases.
You should have an up to date specification list of preferred parts vendors for the machines you are purchasing. This list should be very detailed. The last thing you want is a machine coming in with parts you can't get or parts from companies not well established and possibly not going to be around long.
This should be supplied to the vendor, along with a checklist. The checklist is a simple "does comply" or "doesn't comply" check box for each item in the standard and should have an area by each item to explain exceptions to the requirements. You must review and approve these exceptions.
The vendor should be supplied with a document that contains the detailed specifications and requirements for the process. This will include the required cycle time, Cpk's, desired modes of operation, expected operating time per year, operator interface requirements, safety requirements, environmental requirements, etc.
You want a run-off at the vendor's facility. The run off should include a strip-chart monitored 20 to 36-hour continuous run. The purpose of this run is to find any quick-failing items, potential manufacturing defects in components and any design issues. Any stoppages of the machine during this run must have a notation as to cause. In our case, the machine must run the entire time period with zero stoppages.
The second part of the run off would be a quality run where all quality functions of the machine are verified. Be sure to use a significant sample size for this initial quality run.
Doing this at the vendor allows them to fix issues that arise in-house at less expense than trying to do it on your floor.
After the machine is on your floor and ready to run, a second 20-36 hour strip-chart recorded continuous run should be required along with a second quality run. Things can go bad during shipping / installation.
You can require a 2 year parts and labor warranty. You can also go with performance penalty clauses in the contract.
You should have weekly or bi-weekly design and project reviews with the vendor, and keep a running, itemized "hit list" of any issues, responsible parties, and dates.
None of this will guarantee you a problem-free machine, however it will go a long way to eliminate a lot of issues that cost time and money once the machine is in production.
Very good post.
What do you do about a run off when their automation is build for your machine? We has a conveying and boxing system that cycled at their place, but did not show all the flaws until set up at our plant.
Cold parts hand feed into the system worked fine, but parts, literally " hot off the press" failed.
They ended up walking away from the final payment and we had to redesign inhouse at our expence. We had to put a servo drive on the conveyer to control part movement.
The above practices are good policy. A good machine builder will already be asking for your spares list as they build the quote, they will have a meeting schedule and a communication plan in place, and will provide runoff dates in the contract. The NAAMS(?) book Toyota puts out is a good example of the kind of documentation and the process.
It DOES increase the buyer cost though...keeping a handle on your end of the process is time consuming. Flying out to the builder for runoff is expensive as well.
As an integrator I can tell you one of our biggest issues is change traffic. The product is not stable and we aren't in the development loop. I can't tell you how many science projects I have had to "automate" by the ship date some sales guy sold to the venture capitalists and CEO.
The other thing is that equipment is the LAST thing people think about. The time lines are hyper compressed. This is a big issue, it means there is less time to revise designs as the project comes together. Compromises have to be made based on parts availability.
Building a good machine takes time...budget for it in your own timeline.
A lights out 24/7 machine is a totally different thing than an 8/5 machine, keep that in mind as you write your requirements.
If feasible, involve your machine vendor in the product design.
It's good to have a partner in the automation rather than just a vendor.
I have often thought there was a good living to be made in consulting and managing automation for companies that need machinery. Many don't have the experience in the industry to effectively manage the process, and never will because they are only buying a machine occasionally and cant afford the time to develop the processes and knowledge to do it well.
This happens a lot. The parts are different at the builder than at the plant. This where the partnership comes in...a good vendor will put an engineer at the interface point to figure out what's different.
Originally Posted by Toolroomguy
The odds are that the builder took a loss on the project because they under bid it.
Did you take the low bid or look at the technical merits first.
A good quote will give you enough detail so you can make decisions on both.
Yes, low bidder.
It was funny, well not at the time, but when we reported the problems, they sent out a kid to fix them. A young man, barely out of high school, who was supposed to bolt on some parts that would 'fix it".
When I asked him what machine building experence he had, he said none.
Ouch!! I bet that kid was humiliated by being put in that position by his boss.
Originally Posted by Toolroomguy
This is part of the procurement process and is part of the problem getting good machinery. Automation is unlike other purchases. It's custom, unproven, stuff. It impacts the profit margin in more ways than the price.
Often the low bid is low because they don't entirely understand the problems involved. By the same token the buyer does not understand the problems either. So the bean counters look at it and only see dollar signs. How many people quoted it? What was the price/delivery spread?
Hey, just throw a cheap conveyor with a sensor on it...how hard can that be right?
Yea, that works EVERY TIME!......hehe
Originally Posted by Miguels244
Like Miguels, I do this stuff for a living (automation, process development, machine design/building). He has some very good points also.
To answer one of your earlier questions, we keep 20% until it runs to our satisfaction. This is usually more than the profit margin on the machine, so it gives them a bit more incentive to make it work. They aren't just losing profits, it's going to cost them out of pocket if it doesn't work.
The vendor MUST understand the *real* process they are dealing with, otherwise it gets really ugly, really fast as you are aware.
As per your description of the packaging system, future considerations for your specifications should include the temperature the parts are at when they are presented to the machine. We deal with AL parts fresh out of the mold (actually the robot takes them out of the mold.... and robots pour the molten AL in the mold....) these things are still not even solidified completely yet, so I understand the issues regarding "hot parts" to the process.
You must clearly define what is expected as much as possible. The more clearly it is defined, the less chance there is of a screw up, miscommunication, or misunderstanding between the customer and vendor.
Clear specifications will also reduce your cost because they know exactly what they are bidding.
Tonytn36, you mentioned
"The run off should include a strip-chart monitored 20 to 36-hour continuous run"
I am curious about this one. Are you actually connecting a chart recorder to an input that is indicative of the machine running? Such as a NC E-stop, 'gate closed', or "Spindle Enable" circuit?
Or are you saying that you are manually monitoring the process, setting an engineer there with pencil and paper to record the minute details?
Tonytn36 and Miguels, I sincerely appreciate your responses. You've always given me good responses, and you both have put some real thought and effort into your responses for this thread.
Toolroomguy, I'm sorry you have to be in the same boat as I am. Redesigning things as soon as they hit the door sucks. Maybe one day- things are going to get better.
Muffler, Yes you tie the strip chart into the PLC and tie to a 1 time per cycle bit that is handy. We try to tie to something that can't be fooled. Our requirement is 20 hours and we do go monitor it. We don't spend 20 hours right at the machine, we may start it at like 3:30 in the morning, stay with it a few hours, go eat, check on it....get some training.....etc....then go take a nap and come back in late that night to see it finish up.
We do carefully review the strip chart afterward also.
I have seen 20 hour buy offs where the buyer plugs his laptop in to the net and goes to the strip club...no shit. It works well, the lappy logs all I/O and times.
The 20 hour only tests mechanical failures, not actual function.
Sounds like a job we did several years ago
Originally Posted by Toolroomguy
Seriously - a big part of how we grow our business is cleaning up after the "low ball" bidder.
I fault the buyer in many cases when a machine fails to perform.
We added 3-axes of servo drives to a packaging machine in Wisconsin after the OEM (Superior Cartoners) went bankrupt trying to make the production line work. It was a fairly significant redesign, eliminating portions of a line shaft and then coming up with a servo drive system that would smoothly introduce product into open boxes and then glue the flaps and send them on. We have done smart conveyors for singulating product at speeds of 300 units per minute . . . sounds easy, but with uneven inflow . . . and varying amounts of back pressure on the infeed conveyor - not a walk in the park.
I agree with the comments above by Tony - the buyer should have standards for what they expect in the plant. AND - they should have standards for selecting the machine builder that include a variety of elements not mentioned here. Engineering / Experience should be given high weight in the decision, a proven track record of supporting machines of similar duty cycle in their region, proven experience with the technology that the plant has standardized on (typically Rockwell, Siemens, GE, Mitsubishi, Koyo, whatever . . .), agreed upon milestones in the design that are tied to payment terms, a well documented change order process, there are many other things that should be brought to mind when selecting an integrator / machine builder who you are not familiar with.
We typically have our final payment tied to a performance test that measures actual throughput of the machine while logging all sources of downtime over a 48 hour production period. No more than 3% of downtime can be attributed to machine events during that time - if/when this performance test fails - we must evaluate what the root cause of the downtime is and take corrective action. We have never failed a performance test except in one case listed below. We have built over $50million in control systems / retrofits and machines over the last 15 years.
About 5 years ago we were bidding a contract well in excess of $1million to retrofit a large converting line in Savannah Georgia. The customer specified a specific technology and the vendor had recently released this technology to the market place. We had done roughly 2 dozen machines like this with an alternate technology with excellent results and I was skeptical as to whether this newly released product was ready for show time. Both the customer and the vendor insisted that this would be fine - in my opinion, neither of them had the engineering experience or understanding of the process to make that judgment - I argued against it and ultimately was told that if I didn't have the confidence that it could be done that they would go with another vendor. They went with that vendor (a large well known automation vendor with an in-house integration group) . . . the machine ran, but it ran well below pre-retrofit performance levels. We recently were asked to provide a bid to rip and replace that system.
18 months ago - I was in the same position with the same two players - (my largest customer with a corporate sponsored project) . . . this time we went with the vendor's technology which we had gained considerable experience with and oddly enough had just gone through another major revision that brought a lot of nice features related to safety which were a big part of the justification for the retrofit.
Long story short - compared to every single retrofit we had ever done before - it was a miserable arduous effort to get it to work. Our start up ramp for this class of machine is typically two weeks - this one took 4 months. We have had nearly 30 firmware revisions - 1 of which was a complete rewrite. We worked with the technology vendor at every single junction to outline the problems, assist with the testing, driving toward a robust solution. We are today - almost there with the first machine and I have been sitting on over $500k of product turned into a complete machine center on my floor for a year that the customer does not want until the first machine is done. We have not yet satisfied the performance guarantee and we have identified one additional mechanical change that will be needed before we are confident that it will run at the peak performance levels required by the contract. This mechanical component was supplied by the machine OEM - but we are taking responsibility for it because it is being controlled by our servo / gearbox / software.
This is a situation where the customer set a specification that was unrealistic - heavily influenced by the efforts of a corporate sales team fielded by the vendor - decisions made at the purchasing level that don't allow for evaluation of risk or appropriateness of technology to the task.
The customer lost millions in opportunity cost - we lost hundreds of thousands on this job. The vendor has worked hard to fix all of the problems and also has gone a long way toward helping us with credits toward future purchases. In the end - I believe we will be roughly $200k in the hole on this project even after significant credits awarded to us by the technology vendor.
Who is at fault here? I think there is shared responsibility all around . . .
In retrospect - we should have walked away from this as the new products from the vendor specified by the contract were not ready and they have a history of launching products that need a year or two of shaking all the bugs out. That would have been tough to do - 18 months ago we were very slow and it would be tough to give the news to my sales guy and employees that we were not going to accept this order when the shop was on 4 day work weeks with a 20% cut in pay.
The customer at this plant and at the corporate level should not have insisted so stridently that they get the latest / greatest technology without acknowledging that there was significant risk that they were willing to accommodate in the name of "development". We don't shy away from using a product just because we don't like the color . . . we shy away from stuff that hasn't been proven to work well.
The vendor took way too long to take us seriously - that is partially because we won this contract in competition with their integration group. They don't have as strong of engineering team as they would like to think they have and it took several months for them to begin to investigate the specific portions of their firmware / hardware that we had identified as being problematic.
We build a lot of custom machinery - about 90% of our machines work in 24x7x365 applications in the Glass container and Paper industries along with quite a bit of machinery in the veneer / engineered wood products industry. Down time is in the thousands of dollars an hour - if we sense that the customer is looking for the lowest bid and they see our cautionary statements as "a lack of confidence" . . . while seeing the competing machine builders "high level of confidence" when they say "piece of cake" . . . well, we typically decline to bid those machines.
In this case - we are out of the ditch now, but it took a lot of effort and expense to get out of the ditch. We fulfill our commitments and would never consider walking away from a job - even if it cost us a lot of money. . . and in this case - our customer recognizes now the big picture and we collaborate on a much higher level on development work to ensure that all the stake holders understand the risk / reward merits.
Custom machine building is a tough gig - you often don't get several machines to amortize your engineering effort over - I have never known a one-off machine to ever be profitable for the machine builder unless you have a lot of very smart people involved who resist the urge to be penny wise and pound foolish.
Last edited by motion guru; 10-19-2011 at 11:35 AM.
Reason: additional information
Good points motion......it's like I tell my boss.....It cost this much $$$....to do it right...and that $$ figure is what it is. As soon as you start cheaping out and cutting corners....the phone will start ringing at 2 AM for you to come fix the POS because it won't run.
+1 for motion guru's tale.
My first industrial employer made a thriving business out of "second surgery". We were large among mid-sized suppliers, despite ambitions definitely not one of the larger suppliers. Prospective customers would often buy a complex system from another vendor either due to a lower bid or a larger reputation, then 9 to 18 months later come looking for a system that actually worked for them and they'd frequently get it from us.
After a couple of dozen rescues, our sales guys started leading with that point.