Motor behavior and driveline concepts - (a roundtable discussion?)

[PAL_Houston]

... have reached this conclusion: it was idiopathic.
Small things malfunction due to even smaller things.

In other words, it was gremlins!

[TiVoPrince]

When 
stuff acts strangely I usualy have two possible explanations.  Sunspots during the day and moonbeams at night, they both really fluff up technology (based on published reports in The National Midnight Star) and experience tells me that less than 5% of customers will question my explanation...

[up1950s]

Have we reached the conclusion yet that no one really has any clue about this?

I have reached this conclusion: it was idiopathic.

Small things malfunction due to even smaller things.

Repeat the test , but stay awake all night and look at it . It could have been a spider leg that got trapped under the brush . Check your house for a spider walking in circles .http://www.buenascosas.org/?p=3727

[Kisatchie]

When 
stuff acts strangely I usualy have two possible explanations.  Sunspots during the day and moonbeams at night, they both really fluff up technology (based on published reports in The National Midnight Star) and experience tells me that less than 5% of customers will question my explanation...

Hmm... don't forget
cosmic rays...

[mmagliaro]

Have we reached the conclusion yet that no one really has any clue about this?

I have reached this conclusion: it was idiopathic.

Small things malfunction due to even smaller things.

Correct.
I posed the question in case there was something about motor behavior that I had somehow overlooked.
And I posed the whole thread to get ideas on the behavior of the driveline.

Those two things have been very well addressed, and I appreciate it.

[nkalanaga]

Kiz:  Cosmic rays really can mess things up.  With model railroading increasingly high-tech, that could be a factor.

Not railroad related, but in the late 80s I worked for a bank data processing company, as computer operator.  You've probably heard the saying that computers don't make mistakes, it's all the programmer's fault.  Normally that's true.  We had a case where even the maker's (Burroughs) engineers conceded that the computer made a mistake, and they blamed it on a cosmic ray.

Computers use "parity checking" to avoid mistakes.  They count the 1's and 0's in a byte, then add an extra bit to keep the number of ones even or odd.  If a bit is changed, the machine can't tell which one, but it can detect that something is wrong, and give an error.  Newer ones are more sophisticated and can often find and correct the changed bit automatically.

In our case the program added a column of numbers and got the wrong answer.  Rerunning the program gave the right answer, so the program and data were correct.  What had happened was that the "glitch" had flipped two bits in one byte, so the parity remained the same, but the value changed. 

TiVoPrince:  I doubt that moonbeams affect anything, but sunspots can cause real problems, including large-scale power outages.   

[nscaleSPF2]

As a retired Reliability Engineer, I feel compelled to chime in and commend Max on his approach to this issue. 

Max, you were right to stop at the point that you could not duplicate the failure.  If you cannot repeat the failure, then you cannot repair it.  This is common sense, but I can't tell you how many times I've seen an engineer "fix" an issue only to make it worse.

Second, you were on the right track when you were collecting voltage and current data.  I had an esteemed mentor who had seen it all.  For example, he was in the Control Room, helping to get the Apollo 13 crew back to earth.  One of his favorite sayings was "Show me the data."  You use the data to tell you what might be happening, instead of guessing.  The mentor told us that, in his experience, if you guess at a solution, you will be wrong 90 percent of the time.

One of the other posts (from Richie Dost: "Repeat the test , but stay awake all night and look at it .") also got me thinking.  Advances in technology, applied to R/C cars and airplanes, have brought the cost of devices called "data loggers" down into the reasonable level.  For those of you who do a lot of locomotive repair, it is now possible to record voltage, current, motor speed, motor temperature, and maybe even drawbar pull, over minutes or hours.  You can then graph the data and try to make some sense of it.  Someone with the time, a few bucks, and computer skills could do this.  If anyone is interested, here are a few places to start:

http://www.ni.com/data_logger/

http://www.onsetcomp.com/products/data-loggers/U12-data-loggers

http://www.eagletreesystems.com/MicroPower/micro.htm

I don't know if any of this helps (you guys all know more about model locos than I do), but I needed to share the experience.

Jim Hale

[up1950s]

Don't they use a data logger on Mythbusters all the time ?

[TiVoPrince]

Marking  
the tubing (red/black/blue at 120* intervals) to determine if the mechanisim stops at the same point every time.

I'm thinking more like propping it up to run in front of a cheap webcam and going to dinner rather than buying a bunch of data logger hardware/software...

[peteski]

Kiz:  Cosmic rays really can mess things up.  With model railroading increasingly high-tech, that could be a factor.

Not railroad related, but in the late 80s I worked for a bank data processing company, as computer operator.  You've probably heard the saying that computers don't make mistakes, it's all the programmer's fault.  Normally that's true.  We had a case where even the maker's (Burroughs) engineers conceded that the computer made a mistake, and they blamed it on a cosmic ray.

Computers use "parity checking" to avoid mistakes.  They count the 1's and 0's in a byte, then add an extra bit to keep the number of ones even or odd.  If a bit is changed, the machine can't tell which one, but it can detect that something is wrong, and give an error.  Newer ones are more sophisticated and can often find and correct the changed bit automatically.

In our case the program added a column of numbers and got the wrong answer.  Rerunning the program gave the right answer, so the program and data were correct.  What had happened was that the "glitch" had flipped two bits in one byte, so the parity remained the same, but the value changed. 

TiVoPrince:  I doubt that moonbeams affect anything, but sunspots can cause real problems, including large-scale power outages.   

Ah, the old Alpha-particles affecting the DRAM bit cell capacitor charge conundrum.