: | So this is how its done?

Ok I work for a systems Integrator and I have been given a few projects to work on nothing major. Well I like to write my code then simulate it and get most of the bugs out before I go to the field for startup so I dont have to be worried with writting a ton of code in the field. My boss who has quite abit more experience than I do says he just writes the manual stuff then waits to get in the field to write the automatic parts. He said if he writes the automatic operational stuff in the office its normally wrong in the field. To me this sounds completely nuts! and scares the hell out of me. But he has been doing it longer than i have.

Is writing and simulating code for the entire project in office the way to go or his way"flying by the seat of your pants" ?????

Thanks

The more I can test and simulate in the shop the happier I am. Field time is the most expensive, and if you can minimize field time your are ahead of the game.

Gabsdad said:

Ok I work for a systems Integrator and I have been given a few projects to work on nothing major. Well I like to write my code then simulate it and get most of the bugs out before I go to the field for startup so I dont have to be worried with writting a ton of code in the field. My boss who has quite abit more experience than I do says he just writes the manual stuff then waits to get in the field to write the automatic parts. He said if he writes the automatic operational stuff in the office its normally wrong in the field. To me this sounds completely nuts! and scares the hell out of me. But he has been doing it longer than i have.

Is writing and simulating code for the entire project in office the way to go or his way"flying by the seat of your pants" ?????

Thanks

Click to expand...

That scares the hell out of me too. That is how you hurt people.

Any good integrator will do as you first indicated. Lots of time in the office writing the functional spec, coding and fully simulating the plant equipment in code before going though a comprehensive FAT procedure.

Of course there are always going to be changes (sometimes substantial) during commissioning, but that is no excuse not to do the work upfront. Anyone who says otherwise is a cowboy and has no business programming PLC's.

I suppose maybe it depends on the industries you work for/in. I can tell you coming from a Pharmaceutical/Biotech SI background that I don't know of ANY firm who would contract an integrator who doesn't include extensive testing in their proposal.

At the least, not properly testing even the smallest project opens you up to liability in the event of an accident (personal or property) and of course extended time in the field.

I'd say develop your habits now, test as much as you can, and I can guarantee you that it will pay off in the long run.

zankorel said:

That scares the hell out of me too. That is how you hurt people.

Any good integrator will do as you first indicated. Lots of time in the office writing the functional spec, coding and fully simulating the plant equipment in code before going though a comprehensive FAT procedure.

Of course there are always going to be changes (sometimes substantial) during commissioning, but that is no excuse not to do the work upfront. Anyone who says otherwise is a cowboy and has no business programming PLC's.

Click to expand...

Well, I'm an in house programmer. So I do about half of my program writing at the machine, and it doesn't cost me any more than doing it 100 feet away in my office. It also helps me visualize the machine better when I'm actually at the machine.

I'll have the skeleton of the program ready, which is just the basic sequence, maybe an HMI with I/O descriptions.

I write the manual and auto at the desk. I create a dry mode using timers for motion. I always have the general time I expect the motion to happen in sometimes I have to calculate it. I have very little or no tweeking to do on the real machine. I have been successfully doing this for 15 years. I also create a single step mode to walk through the sequence.
It would take for ever to get a new system up if I had to wait for the construction to be finished before program development.
I have never done any high speed electronics machines. When I see those machines move I wonder if I could use my system.

As with most people here I would say that getting it right in the office is the best way of doing it.

Simulate everything you can and prove it. As zankorel says a FAT is the best way to go. Whilst not everything I do relates to PLCs, it does involve software control of one form or another.

I have just completed 2 days of testing a CCTV interface with a client and several things came up such as graphics, a faulty relay on a PCB and some tamper issues.

These are easy enough to sort in the factory, but if we had to do this on site it would be completely different. We'd probably kill several trees getting permits to do the work.

Generally if it takes an hour to do something off site, it will take a whole day on site.

Jon.

Typically, I design and simulate the entire program before installing it in the feild. Sometimes I have to go to great lengths to simulate all of the required functions and actions, such as having a 480V, Three Phase circuit installed in my test lab so that I can run frequency drives and servos, but it is well worth the effort. I spend less time on startup and have to correct fewer bugs than I do if I am forced to do a majority of the coding in the field.

Recently, I had to rebuild the controls on a machine that was currently being used and we had a one week window to remove the machine from production, rip out the old controls and install the new controls. We were reusing a lot of components from the existing controls, so I could not simulate the system as well as I had wanted to, so I spent the next week finishing code and debugging with the production guys standing over my shoulder asking "When are you going to be finished?". Fortunately, we finished the machine and everyone is happy now.

I say always simulate as much as possible to reduce code errors and field programming.

Regards

An important factor in getting your mistakes done in the office and not the field: you make your field mistakes while the client might be watching. Easy way to come off as incompetent, IMO.

Tharon said:

Well, I'm an in house programmer. So I do about half of my program writing at the machine, and it doesn't cost me any more than doing it 100 feet away in my office. It also helps me visualize the machine better when I'm actually at the machine.

Click to expand...

You are in a fortunate position if you can do this. I have worked "in-house" for a number of special purpose machine manufacturers and without exception they would all prefer that the software works as soon as you load it into the machine. The sooner the kit can be shipped, the sooner your employers can recoup the investment they have made in building it. A proper specification and extensive simulation and testing before the machine is built is surprisingly cost-effective.

you must have a good machine sequence and understanding of how the machine will operate and do everything you can at your desk. i worked with a guy that said he was about 50% done when he first downloaded and we had a 2nd guy that we agreed this guy was about 90% finished when he first downloaded (he is good). also what hurts is when the Machine Owner starts deciding what he wonts after you are about finished.( Can it do this? can we do that? and so on)

I'll bet that your boss has modular code like
Pawl in, Pawl out, Spray on, Spray off etc.

He probably also knows the sequence.
It makes sense to make sure that all of the manual operations work, before you tie it all together.
But you have to have a plan.

klittle said:

you must have a good machine sequence and understanding of how the machine will operate and do everything you can at your desk. i worked with a guy that said he was about 50% done when he first downloaded and we had a 2nd guy that we agreed this guy was about 90% finished when he first downloaded (he is good). also what hurts is when the Machine Owner starts deciding what he wonts after you are about finished.( Can it do this? can we do that? and so on)

Click to expand...

First job I did with 100% specification, simulation and testing was a replacement control system for an irradiation cell (the sort of thing that kills all the bugs in your toothpaste after its made and packaged). This was better than 99% complete and correct at the start of live testing (only a few timers to tweak). OK, I know this is exceptional (can't fool around with those cobalt radiation sources) but if you don't understand the machine before you start the code - why don't you find out first?
1. Get the user requirement specification right.
2. Get the machine specification right from that.
3. Don't even start your code specification without the 1 & 2.

We all have had customers who don't know what they want until they see what they've got - and the two are different. If all the specifications have been signed off in advance, everyone's covered and there's extra to pay for afterthoughts. If your customer won't sign off the spec before you start then you need a good slice of contingency time and money built into the project. In my book customers of that ilk are just plain unprofessional.

Gabsdad said:

Ok I work for a systems Integrator and I have been given a few projects to work on nothing major. Well I like to write my code then simulate it and get most of the bugs out before I go to the field for startup so I dont have to be worried with writting a ton of code in the field. My boss who has quite abit more experience than I do says he just writes the manual stuff then waits to get in the field to write the automatic parts. He said if he writes the automatic operational stuff in the office its normally wrong in the field. To me this sounds completely nuts! and scares the hell out of me. But he has been doing it longer than i have.

Is writing and simulating code for the entire project in office the way to go or his way"flying by the seat of your pants" ?????

Thanks

Click to expand...

Sorry, but sounds unprofessional to me.

Actually, your boss is better than the last 2 integrator I hired (and will never hire again). The first one printed all the drawings I sent as 11x17 on 8.5x11 but didn't reduce. When he got to the machine, he said wow, now that makes better sense. He then opened his laptop to bang out the program in just a couple of hours. 3 days later - and $5K worth of damage and the machine could do some basic functions. I wound up finishing it myself. The worst is he tried to charge me extra work for all the time he put in.

The second guy - sent by our corporate group - had absolutely nothing when he got in the plant. 2 weeks and still no working machine. That's when he said his time was up. I called the owner and stated that he had a weird way of telling me he never wanted to work for us again.