Our Internal Process for Software

Our first attempt at simplifying our engineering efforts was to create a custom takeoff program.  We were already creating our own shop drawings, what if we could use the geometric data from Autocad to drive our material takeoffs and fabrication tickets?

We started with an article out of Cadence magazine that showed how you could attach xdata to entities inside Autocad to create BOM information.  We created an Access database that could hold the information and using the Open Design Alliance libraries, we were able to extract the xdata from the Autocad drawings and create a table of information inside Access.

The tag lines inside Autocad were system agnostic, allowing you to apply different systems to each elevation or change the system information without editing the data inside Autocad.  Routines were created to understand where and how each part would intersect other parts so shear block clips and notching could be automatically applied.  Once the data was inside Access, we could create custom reports or write code to fill out our existing material forms in Excel.  The Excel forms were then customized to be able to feed directly into our accounting system so we could eliminate almost all of the data entry duplications.

The tag lines would represent a part (lets say a horizontal).  One line would have a tag called “Horizontal” and would know what elevation it belonged to.  In the database, you could assign your horizontal body, body stop, face stop and shear block clips to that tag line.  You could then create adjustments to the line length.  maybe your face stop needs to stop 1/16″ short or the shear block clips need to be a fixed length.  This data could be arranged into systems and then applied to any elevation.  Finish colors could be assigned at the system or elevation level.  Overrides could be created inside Autocad for any individual part that needed to be adjusted.  Vertical mullions in an elevation would look for horizontal intersections and automatically assign machinings at those locations.

We have been using this system for about 6 years now and have tied the data output to our machines and can create text files to feed the Emmegi software and also some of the other machines in our shop.  We have increased accuracy and by creating a very specific method for takeoffs and fabrication tickets we have been able to hire entry level people and make them successful within 2 to 3 months.

But it still isn’t easy.  It still requires database and Autocad skills.  We have had limited success getting sales to use the program because of this.  When you get a multi-million dollar project with 6 custom systems, it still takes 3 weeks to create and check the takeoffs and another 6 weeks to create and check the fabrication tickets.  That is only one person which makes if really cost effective, but because we are using Access, we can not have more than one person working on one project and that also is a limitation that hampers our ability to get things done faster.

On the plus side, the shop gets really consistent tickets and paperwork no matter who does the work.  The purchasing department and project management departments have very little data entry to do because most of our data is digital (shop work orders, purchase orders, etc…)

Software

In our company and others that I have worked for, the typical critical path through the office is engineering.  From the time I kick a project off, the critical path is profile drawings, shop drawings, material takeoffs and fabrication drawings.  Even the material lead times (metal 12 weeks or high performance glass) don’t seem to be the issue.  By the time we are done with the material takeoffs, the fabrication tickets can take as long as the materials.

We have done a number of things to try and simplify this process.  I thought I would spend the next couple of posts documenting our ideas and achievements to help others.

So, first, let’s define the problem.  I find in most process improvement the biggest issue is getting the question right.  We all can come up with good answers, but will it really fix the underlying problem?

To start with, we are a manufacturing company.  95% of all of the projects we do, we design, engineer, extrude and fabricate our own systems.  We don’t buy systems from outside vendors very often and all of the engineering is done in-house by our own employees.  This is probably different from most companies that either do not do their own systems or have some type of mix of outside and internal systems.

By doing this, we are able to control our own destiny.  We are not at the mercy of outside resources.  We are a customer focused company, so our ability to meet the schedule needs drives this business decision.  But it creates its own problems also.  When the market gets hot and there is lots of work, our ability to scale is limited by this decision.  When times get tough and there isn’t much work in the marketplace, you also have to make some tough decisions.

For me, the issue comes down to people.  In our business, there is a huge amount of specialized knowledge.  When I came to this company 8 years ago, the people doing fabrication tickets were the most experienced people in the company.  They knew how the systems worked, what they could do and what they couldn’t and how to represent it to a less experienced shop worker.  Finding more people in our industry with this knowledge is difficult if not impossible.

The second issue we face is time.  There never seems to be enough of it.  In our process, everything was still being done by hand.  The estimating was done with some type of spreadsheet and manual takeoff.  The shop drawings were digital, but then pushed into an analog (paper) format.  Takeoffs were done from the paper and put back into excel for processing.  Fabrication tickets were done from paper and then pushed out into another analog format for the shop.

We also needed to make the flow of information to the machinery easier.  I already have an engineering bottleneck, now the new machinery actually adds a process to the steps.  How do we make this easy?

So, the question we were trying to answer was:

  • How do you create a seamless “once generated, constantly refined” data set that can flow from sales through operations to the shop?
  • How do you make this system scale?  You want to have a process that requires less people so it can survive the down times, but be able to scale it up with the same people during the really busy times.
  • Can the system encorporate enough Knowledge Management (KB) so you can have new bright less experienced people able to do the work?
  • Can the system provide a more consistent output?  Our shop is our least experienced staff and most do not have a good grasp of english.  If we can standardize the fabrication tickets, cutting lists and packaging it makes it easier to avoid “interpretation” errors.
  • Can the system speak directly to the machinery?  If we can make it mostly digital, can the data flow to the machines without additional steps?
  • Can it accomodate existing processes?  We had already developed specific Excel formats that worked for our vendors.  Can we keep these forms as the output with a new system?

Follow Up on the Emmegi Planet

We have now had the equipment up and running since July of 2007. The machine is working excellent with minimal issues. We typically run 2 to 3 shifts on the machine with about 1 hour of maintenance per day. We are on track with our ROI calculations and this purchase was definitely a success. Some of our lessons learned:It will take much longer than you think to get the machine up and running the way you want. I would allow 6 months for you to be able to run continuously with few problems. You will have to train people, they will make mistakes and they will cause issues that will require service. You will need to train service people and they will take a while to figure out how to keep the machine running or to debug issues. The machine itself takes a while to work out the kinks. Don’t figure on less. I have a few friends at other companies and for a 5 axis machine like this, they all agree.Software will be an issue. The CAM Plus software from Emmegi is excellent. We have found very few issues with this software. It can do almost anything you want and very minimal time for setup. Your parts libraries need to be in DXF files and saved in metric units. You can just write a script for this in Autocad and create a library on the fly. That being said, there is a learning curve in the office for people to understand what they can and can’t do with the machine. They will need to learn three pieces of software: CAM Plus, Job and Drill. They will need to understand the limits of the tools in the tool changers. They will need to learn which way to cut the metal with the flying blade. All of this takes time and people will make mistakes. We tried setting up the clamp options in the office but it was actually easier in the shop. Most things work with the default clamp locations but it is hard for the office to know when the clamps need to be closer to the machining. In the shop, you can hear it.

The Emmegi training wasn’t much help here. The people from Italy were able to get us up and running, but because of the language differences, it was mostly trial and error on our part.

You will also need other software besides the Emmegi stuff mentioned above. You will want to be able to import information into job and not have to enter it manually. We have created our own software to do our takeoffs and fabrication tickets. We then modified the program to be able to output the “unified standard format” files to the job software. This saves a ton of time and makes it possible to program as needed and not just high quantity repetitive parts. If you are purchasing a machine, make sure you get multiple copies of the software and the USB Key that enables the file imports. You will need one of these keys for every machine that needs to import files.

We have been reviewing the software from SoftTech and Orgadata. These are programs that allow you to create elevations inside the program and get BOM data and fabrication information. I’ll talk more about this in a future blog.

Service is another key aspect. We had many issues when we first started using the machine. Emmegi was very responsive. They flew people out from Italy and we got through all of our issues. I think that Elumatic has a pretty good service crew in The States, but I’m not sure about Tekna, Fom or Mubea. This is a critical component of the decision. I guess the best way to tell is to talk to other people who have the machines. I usually try to find people that have the machines to see what their experience has been. You can usually talk to the machine sales people to find out who their customers are, but you want to talk to the competing company to see who might be having trouble with a competitor’s machine. For instance, we were seriously considering a machine from Mubea, but I had heard from the Emmegi sales rep. that U.S. Aluminum was having issues with the machine. I contacted him and got some great feedback on the Mubea machine. He wasn’t as happy as Mubea made it sound, but he also wasn’t as unhappy as Emmegi made it sound. Most people I’ve talked to seem happy to help.

Installation of the Emmegi Planet

So our machine shipped on time.  The machine arrived in the port of Oakland and we had Bigge Crane Company transport it to our shop.  We removed it from the truck and moved it into our shop using one of the machines we would typically move our precast panels with.  We removed the case and everything appeard to be in great shape.  We had our electrical and networking connections already completed.  All we need now is the Emmegi personel to do the installation and training.

We had a few problems here.  The communication between Emmegi in the states, in Italy and our office could have been a bit better.  We had a hard time getting a straight story on when the installation would take place.  When the installer finally showed up at our office, he was very proficient in the hardware and software.  He spent the first week getting the machine installed and ready for use.  A lot of that time was actually the installation of the fence around the machine.  The machine itself had already been tested and run in Italy.

The biggest issue was the training.  Software training was minimal and there was no game plan for what they would train on.  What we needed was to try and program our first project for the machine while Emmegi was at our office.  We got a few test pieces programmed, but not a full order.  This allowed maybe one day of testing and then they were gone and we were on our own.  We then spent the next week getting the progams for our first order ready to run.  We hit several walls during this process.  Emmegi eventually sent out a second representative and we worked through our issues.  We were now ready to begin production.

The trip to Italy

One of the big factors in our decision was timing.  Every machine manufacturer was very busy.  Some companies had told us they waited 12 – 18 months for their machines.  We wanted to have something in place by late July.  Emmegi assured us we could have our machine by early June.Several weeks prior to the shipment we received a call from our sales rep. indicating that we needed to go to Italy to sign off on the machine.  This was never discussed in any of our meetings and was a compete surprise to us.  Chris (our shop manager) and I planned a last minute trip to go to Italy to see our machine in action.

We had been given copies of the CamPlus, Job and Drill software from Emmegi and we decided to try to program some of the parts from our first targeted job for the testing.  Chris in our shop had several extrusion shipped over to Italy so we could use these when we were there.

As we tried to program our parts for the test, we had many issues.  Phone calls back to Emmegi in New Jersey did not answer the questions we had.  Emails to Italy were also not very helpful.  At this point, I’m getting worried.  If you can’t program your parts, or if you are getting tons of errors and no one knows why, it scares the hell out of you.  We did manage to program several parts and sent these to Italy ahead of our trip.

We arrived in Italy and were schedule to go to Carpi first (this is where the main Emmegi plant is and their software division) and then to Piacenza where they manufactur the Planet machines.

The trip could not have gone better.  We spent an entire day with their software group and they answered all of our questions.  We were able to see all of the equipment they make and it was quite amazing.  We were able to visit some glass companies there and see what they do and how the utilize the Emmegi equipment.  But the best part is that we were able to see our machine manufacture our parts.  We even had them try some things on the fly to see how easy it would be to adjust the machine and there were no issues.  We signed off on the machine and they loaded it for shipment the next day.

I’d like to say after having been in this industry for over 25 years that it is refreshing to see a company exceed your expectations.  The trip to Italy was really amazing.  Everything was handled very professionally and all of our questions and concerns were answered easily.  Despite all of the people who told us it wasn’t possible, they produced our machine and shipped it right on time.  Any limitations we had found with the software were reallyu mis-understandings on our part.  A few of the bugs we did find were corrected the same day and our machine was updated prior to shipment.  Most of those revolved around the metric to english units.  Some parts of the software were still displaying dimensions in metric.  These small things were corrected and the machine performed perfectly.

Making the decision

So we received quotes from Emmegi, Mubea and Elumatic for machining centers.  We spent several weeks creating an outline of the positive and negative aspects of each.  We also contacted several customers regarding the machines they had purchased to see what they thought of them.The Mubea machine was the least expensive of the group.  They had excellent sales representation and we really liked the software.  The fenestration company that we visited in Hayward that had their machine was very happy.  We also contacted U.S. Aluminum because they had purchased a machine and were using it for door fabrication.

Emmegi had the most expensive machine.  We spoke to several of their customers also.  Everyone was very happy both with the machines and the service.  We weren’t as impressed with the software.

In the end, we made our pitch to Randy regarding our findings.  The Emmegi machine was really built to do curtainwall.  It had the profile capabilities that we needed, it could fabricate from the underside of the profile (Mubea could also, but it had to be lubricated in the spindle where Emmegi is lubricated at the tool.  This was a limitation we discovered very late in the game), and we could find very few limitations.  It was very expensive.  You could almost buy two Mubea machines for the price.

One thing we found in speaking with the other companies that had gone through this process was that if they were to do it over, they would have gone with the better machine.  We didn’t want to spend less money now only to spend a lot more money in the long run.  As my brother says: “Buy nice or buy twice”.

Despite all the homework and all of the information, we weren’t really prepared to speak to Randy’s concerns.  I learned a new lesson in this process.  When selling things to the boss, you better know the return on investment numbers back and forth.  Features and benefits are great, but when someone is going to invest 3/4 of a million dollars in a machine, you better know how long it will take for it to pay for itself.  After meeting with Randy the first time, I went back and did all of the math.  We took the upcoming backlog and spread the numbers in fabrication improvements and created the back-up we needed to really speak  to Randy’s concerns.  In the end, using very conservative figures, we felt the machine could pay for itself in 3 years.

With those numbers and a bit more discussion, we got released to make the purchase.