BreakPoint

cancel
Showing results for 
Search instead for 
Did you mean: 

This is Hooovahh


@Hooovahh wrote:

Pressure Bomb

(Long one sorry)....

 

  I checked the heater tank and the temperature railed at 200C. 

...

 

...

 

After that we installed a hardware temperature limiter which ensured that wouldn't happen again.


This would have been handy and is not subject to crashing...

 

Spoiler

 

Mercury.JPG

 

Unless some drops it and the we have a mercury clean-up on our hands.

 

 

That would have taken a week or more.

 

Ben

Retired Senior Automation Systems Architect with Data Science Automation LabVIEW Champion Knight of NI and Prepper LinkedIn Profile YouTube Channel
0 Kudos
Message 221 of 460
(10,751 Views)

Yeah I absolutely should have known better.  But we were replacing an existing non-RT system and whoever budgeted it said it would be a direct replacement of what essentially was a Windows 3.x PC.  The system locking up any other IO forever would have no negative effect, but that one heater control was an issue.  I was young and somewhat inexperienced.  I'd hope that if a similar thing came up today I would put my DFMEA hat on and insist it be done right.  Well lesson learned.

0 Kudos
Message 222 of 460
(10,741 Views)

@Hooovahh wrote:

... but that one heater control was an issue.  I was young and somewhat inexperienced. ...

 

 Well lesson learned.


I will spare you the mistakes I made over 43 years like down all of the computer users across the U of Pitt campus and many others but will admit I never liked it when I heard a customer say...

 

"If you want to make an omlet, you have to break a couple of eggs."

 

Like hearing fingernails on the chalk board.

 

Ben

Retired Senior Automation Systems Architect with Data Science Automation LabVIEW Champion Knight of NI and Prepper LinkedIn Profile YouTube Channel
0 Kudos
Message 223 of 460
(10,739 Views)

I felt like my dad was yelling at me

 

This story doesn’t involve me but was told to me by my then manager just before they hired me.

 

This “Company A” made some widgets that needed testing.  So they got a 2 page summary of the things that the test system needed to do, along with the basics of how the widget works.  Part of the problem was that the widget hadn’t been made, and a production intent part wouldn’t be available until the program was almost over.  So the requirements document was pretty thin in details.  The expectation was that the company making this system would have to adapt and change as more information about the widget was known.  In the past “Company A” would just do all the hardware and software themselves, but as the program scope grew they knew they needed to hire another company to help.

 

This requirements document was sent to a couple system integrators to get a feel for the cost and time that would be needed to make the systems.

 

One company called back and said they wanted to have a meeting to understand in more detail what the systems would be like.  This “Company X” had made test systems but usually for a more regulated industry like aerospace and biomedical.

 

As soon as the meeting started, employees from Company X started asking questions about the system they would make.  How big would it be?  What types of sensors would be needed?  How would IT be involved?  How does the widget communicate?  What hardware was needed?  What steps were needed in the sequence?  Down to what file format the system should log to?  Company A didn’t have an answer for most of these questions.  After the meeting ended the manager from Company A said he felt like his dad was just yelling at him the whole meeting.  They weren’t used to working with outside companies for this part of the program and they should have tried to understand how little information they were giving them.  But at the same time Company X probably should have known working in this industry is a bit more fast and loose.

 

Later Company X asked for $200,000 for the program which Company A thought was crazy.  I assume it was because so few details were known they put in a lot of padding in for potential risk.  Another Company Y out of India said they would do the software only portion for $8,000.  The company that actually got the program was around $50,000 with an hourly price after a specific date.  They told me that they thought about hiring Company Y in addition to the real one just to see what they would come up with never intending on actually using it.

Message 224 of 460
(10,658 Views)

@Hooovahh wrote:

They told me that they thought about hiring Company Y in addition to the real one just to see what they would come up with never intending on actually using it.


I'm sure we lost some projects because we where too cheap. ~20k, where all 4 other parties where in the x00k range.

 

20k vs x00k happens more often, and sometimes the customer is smart and picks the 20k. The additional costs for other companies where (at least in some cases) for 2 project managers, and their (personal) secretaries.

0 Kudos
Message 225 of 460
(10,644 Views)

Customer: Hey so we're using your software here and at the end of every part it says it failed, its supposed to say that right?

Hooovahh: Uh...no that means it failed.

Customer: Oh well why would it fail?

Hooovahh: There's a whole boat load of reasons it might fail, there is a report that you can open that will tell you why it failed.

Customer: So you're telling me that when it says it failed, that part shouldn't be used?  I thought it says failed for all parts even good ones?

Hooovahh: Nope I wouldn't make the software do that.

Customer: Well, that sucks I have a couple dozen parts in a row and it says they all failed.

Message 226 of 460
(10,383 Views)

I feel your pain.  I have test programs run fine for years and every now and then I will get a call saying something is wrong with the program and it needs to be fixed.  When I ask what happened they say a part failed and I need to check the program for errors. 

 

The latest was a sensor causing a test to fail.  It turns out that just before the Covid shutdown there was a change in the field that required a pin on a controller to be changed and the wiring to be modified.  When production/testing started back up the change was not passed to manufacturing so the test station failed the unit for a missing signal.  After a wiring change the program worked fine 🤔

0 Kudos
Message 227 of 460
(10,377 Views)

@Hooovahh wrote:

Customer: Hey so we're using your software here and at the end of every part it says it failed, its supposed to say that right?

Hooovahh: Uh...no that means it failed.

Customer: Oh well why would it fail?

Hooovahh: There's a whole boat load of reasons it might fail, there is a report that you can open that will tell you why it failed.

Customer: So you're telling me that when it says it failed, that part shouldn't be used?  I thought it says failed for all parts even good ones?

Hooovahh: Nope I wouldn't make the software do that.

Customer: Well, that sucks I have a couple dozen parts in a row and it says they all failed.


Obviously, you need to change the software so it doesn't reject all those parts.

0 Kudos
Message 228 of 460
(10,360 Views)

wiebe@CARYA wrote:


Obviously, you need to change the software so it doesn't reject all those parts.


Here you go!

 

 

Message 229 of 460
(10,356 Views)

@RavensFan wrote:

wiebe@CARYA wrote:


Obviously, you need to change the software so it doesn't reject all those parts.


Here you go!

 

 


Thanks! I think I've seen that before... Customers seem to want this all over the globe.

 

Sometimes they get suspicious though... No fails at all? Is this software doing anything? That's where fails can be filtered with a random function.

 

Software can fix everything!

Message 230 of 460
(10,304 Views)