How do you find out your mistakes ?

[envc77]

Nowaday, often I had many mitakes. missing typing, missing calculation etc.

Do you check the report(or calculate sheet) yourself ?
Or project manager ?

I'm curious your method.

 

[GregLocock]

I look at my calcs to see if there is a 'silly'.

I make small changes and see if the result changes in a sensible way.

I check my result against similar previous results.

I do the same calculation two different ways.

I check the result against a test result.

I check the result against somebody else's calculation or model.

We build a prototype and it doesn't work properly.

We put it into production and it doesn't work properly.

The last method is not recommended.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[ScottyUK]

I find the best way is to work on something else for a while, then come back. Ideally review it as a print - I spot all sorts of problems on paper that I am blind to on-screen.

The best way of finding an error is to issue the drawing to someone else, then all the faults immediately become apparent unless they are huge in which case they become visible only after installation!


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[oldfieldguy]

Standing there in the substation in the dark, listening to the sounds of running feet approaching and the phone ringing off the wall...

old field guy

 

[zdas04]

I've used all of the techniques above (including the Oil & Gas equivalent of OldFieldGuy's technique), I really hate to go back to something I wrote years ago and see a glaring mistake that is now unfixable (I wrote a magazine article in the mid '90's that re-introduced the term "eductor" to the Oil & Gas industry, and defined it improperly, causing no end of confusion to this day--later articles and presentations that define it properly have zero impact).

Like Scotty UK, I really like printing the document and calcs out and marking them up with a red pen. I tend to see errors that were invisible on the screen plus I can stop anywhere I want to and come back to it later without having to re-read the first page 100 times.

I recently read an outstanding Masters Theses that was written by an Asian gentleman (his CV was attached). At one point he said that something was "fairy consistent with the data". I laughed for a long time. It was the only grammatical error I saw in 238 pages, but that one was such a cliche.

David

 

[SomptingGuy]

I work in software. Large projects are done on branches. Branches are merged by someone other than the developer that did the work. This enforces peer review. Then our testers get hold of the new code and find that it doesn't work on some odd machine you'd forgotten we still support.

- Steve

 

[Twoballcane]

Here we pratice haveing peer reviews (sanity check) before we go to print. It could be big (10-20 people including the managers) or small (you and sombody you trust).
However, before you do that, I usally do house cleaning before I show my report, as in redo all my calcs to make sure I did not have a fingure slip on the old calculater. For reports, I usally wait the next morning to read it with a rested brain. With spread sheets, I do a once over with hand calcs for each cell that has an equation to make sure it is doing what I want it to do. For FEMs, I double check with hand calcs to make sure the model is right.

Tobalcane
"If you avoid failure, you also avoid success."

 

[DesignerGuy16]

When I hit print, I tend to always remember what I missed, even before picking it up. I need to check on paper whatever I do on the computer. Just easier to visualize that way.

We also check eachothers drawings as we make releases. Just a quick sanity check, not like going into design intent.

James Spisich
Design Engineer, CSWP

 

[TheTick]

Nothing like a fresh set of eyes to spot the mistakes you've looked at for too long. All healthy, competent engineering organizations mandate independent checks.

 

[ash9144]

The earth shattering kaboom
Just listen for the earth shattering kaboom

Have others look it over. Not just engineers but a variety down to the user/operator. Some of the best corrections/improvements I've received have come from the operators, mechanics, instrument guys that will be working with and on the design.

 

[KENAT]

My job is pretty much to look at other peoples drawings and point out the errors, areas for improvement, omissions etc. – makes me real popular.

On the rare occasion I still do my own original work…

I review it on screen first, I then print it off and take a look. As others say this is when I find a lot of the errors. Ideally I leave it a while between creating it and reviewing it but due to schedules that’s not always possible. I then pass it to someone else, in the case of drawings preferably a suitably qualified ‘checker’ if not then the most senior/skilled drafter/designer around or project manager. Back in Aerospace drawings also got reviewed by the stress guy & technical director.

We also had design reviews etc for significant projects to get the input of other people, and when applicable/possible we’d get user input too.

In Aerospace the entire ECO got carefully checked, here it’s just the drawings (if I have time) the ECO itself has lots of required approvals but I’m not sure how thoroughly they look at it, there is a brief ‘peer review’ when its submitted.

As for calcs, I try and work backwards, or if I set up and excel or similar to to it then I validate it a few times by hand. Back in aerospace the stress guy would sometimes review them for me if they were critical.

Reports, procedures etc. here get several signatures but honestly I’m not sure how thoroughly some of them get looked at, often glaring errors have got through the initial approval process and then I’ve caught them in my reports when I do a final review before release (hence the value of coming back to it some time later).


KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at

http://eng-tips.com/market.cfm?

 

[MotorVib]

When I check a project I start with the assemblies.

Check if all of the parts in the BOM are pick up in the drawing. I normily I highlight each item in the BOM and in the drawing. This also make me look over each page afew times looking for the items numbers. (This pick up most of the small errors) For each item I look to see how it is fix in place (welded, bolted or clamped) and do the notes back it up and does the guy who is going to assembly it have all the informasion he needs.

With all of that done take a look at the loading of the assembly. What is required by the machine and the contract. (Now this step I normily done first but a 2nd check is always good.)

As for Cals I try to do them one day and check them the next and then show them in the peer reviews.

The last step is the peer reviews. You need a group that knowns what you are talking about. Example an EE in a ME peer review might not work to well. In peer review machines and welders are the most importent group of people that need to read a drawing. If you can have the lead welder or machines there for the reviews that is a big step in fewer errors.

Chris

"In this house, we obey the laws of thermodynamics." Homer Simpson

 

[MechEng2005]

As others have said, make a hardcopy and read/review it. This can be done with drawings or emails or whatever. Ideally, I like to wait a day or so and review it. Then, I wait a week or so and review again. Sometimes I will get a piece of wrong information in my head (i.e. the width is 3 inches) and then during an initial review I will accept it because it's what's in my head. A review after a week or so allows me to forget some items, and during the review I will look and verify everything.

For drawings specifically, I have a multiple step process:

1) Verify all components are drawn/dimensioned properly to produce all the pieces.
2) Look at the assembly and verify quantities.
3) Go through the BOM line by line and verify each piece is located properly.
4) Go through the BOM line by line and look for where parts attach to each other. Verify the weld symbols/holes/hardware are the right size and present.
5) Review paint/finish required on assembly (or parts).
6) Check over the title block for errors or omissions. (i.e. "Description: Discharge Shoot" instead of "Discharge Chute")

-- MechEng2005

 

[BillBirch]

If your company is ISO 9000 compliant, it should have a procedure for quality checking. The best design companies in oil and gas that I have worked for would sometimes include typical check-lists in the procedure tailored for the appropriate type of document, e.g drawings or specifications.

I have found that the "self-check" is always mandated and stressed - besides, who wants a checker to discover an error when you could just as easily found and fixed it before anyone else finds it!

With respect to the original post, in my experience formal checks will be carried out within your discipline either by seniors or peers if you are sufficiently senior.

Project managers generally only check for consistency (with other project deliverables) and that contractual requirements have been met, e.g scope. Good project managers should not get involved in the design nuts and bolts, but I'm afraid many can't help themselves and use their rank to enforce preferential engineering.

 

[spider007]

The way I find out I messed up, is when the guys from manufacturing come up rolling their eyes while carrying a print.

I print it, check it (things pop out that you miss on the screen), and then the shop supervisor and eng. manager look at it, before it is released. But lately I’ve been noticing that they talk more about their golf swing, or twists and turns of a particular golf course, then my drawings. On top of that i've been really busy latley and can't spend as much time as i'd like on some jobs......so you can guess the end resoult.

 

[jistre]

Well, hopefully we don't find mistakes when 210,000 cu ft of propane escapes and levels the installation.

To this end, mistakes are caught by constantly having other disciplines looking at the prints. Someone out there is going to see something you haven't, and the more eyes you put on it, the more chance you have to avoid creating a crater.

Even in my previous position designing medical carts, mistakes would invariably be found when someone else got involved, whether it be one of the other disciplines reviewing prints for their suitability, marketing trying to get more feature creep, or manufacturing coming into the office saying the 2000 parts you just bought won't fit.

The short answer is that you will make mistakes. Your job is to do everything in your power to be diligent and make your mistakes minor and far between. To ensure that the mistakes you do make are caught at a manageable point, make everyone look at your work and sign off that they've reviewed it. Nothing makes others more diligent during reviews than having them countersign that they approve the print.

Get it looked at often and get it looked at early.

 

[KENAT]

envc77, while it appears you aren't just talking about checking drawings, there are a number of threads over on GD&t over the last couple of years dealing with drawing check.

If you're interested take a look.

forum1103

Bear in mind the forum is mostly mechanical & related disciplines, not structural/civil or similar.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at

http://eng-tips.com/market.cfm?

 

[cksh]

How do I find mistakes? I usually wait until we spend 100k on tooling and when it doesn't fit plead ignorance Kind of a true story. Thats when I found out no one really checked my designs when they looked at my prints/models. They were just making sure I didn't miss a dimension

I find it difficult to check my own work. I can look at a print a 100 times and think it is perfect. Give it to a coworker and he notices a mistake as soon as he looks at it. I am more of a product manager so all of the checking comes down to me. Not sure I was ready for the responsibility at first. There were a couple of things I had to cross my fingers on and hope they worked. (granted I am talking small parts that are easily fixed not life threatening issues such as a poor bridge design). I have since gone out of my way to recheck my work as many times as needed and with as many people that are willing to take a look.

 

[DKirkham]

Everyone around here pretty much looks at the drawings and it is widely understood not to rail on anyone for a dumb idea (just wait a bit and it will be YOUR dumb idea we are all looking at down the road sometime). Everyone has dumb ideas, that is why pencils have erasers.

Anyway, we ALL look at the part after it is made. Then, we test the car on the track and do a lot of visuals in the pits. Then we congratulate ourselves on how smart we are because everything is working perfectly.

Then someone gets the bright idea to video the suspension in action. After we all suffer a heart attack on seeing the suspension move like a belly dancer on the video, we all go back to the drawing board and see all sorts of new ideas on where we can improve. You would be amazed at what is actually happening in a dynamic situation. I would suggest videoing your products in the actual environment they are actually working in. You will be amazed (and hopefully not scared too much).

David

http://www.kirkhammotorsports.com

 

[Gymmeh]

I normally print it out give it a good going over, fix.

Repeat this process 2 or 3 times if my work is in Inventor, and I may repeat this 10 times if my work is in ACAD.

Then I submit it and hope the purchaser, manufacturing engineer, or any shop guys come to my office.



Official DIPPED Member -
Drank in PP Every Day