Odd Steam PS/3 Upgrade Issue

There's probably a saying somewhere out there about it can't be that, right?

Funny as it is, it can be just about anything. Wouldn't matter what brand it was, it's a defect. Assume nothing! 

A few years back I had a Lionel 8354 than ran awful at slow speeds and even more awful at high speeds.

Turned out one of the armature wires was not connected to the commutator, so at low speeds the thing was running on only two windings, instead of three. Above a certain speed the loose wire would fling outward and contact the field as it rotated putting on a great sparkler display. Was easy enough to diagnose when I ran it with the shell removed. Of course, you can't see inside the can motor.

It was pretty amusing to see the locomotive sitting there and slowly ramping up it's chuff rate while not moving.   I figure that as the smoke motor did whatever noise injection it was doing, the logic thought the motor speed was slightly faster, so the chuff went up slightly.  Of course, that meant motor pulsed more often and it was a continuing ramp up.  I never let it get to warp speed.

I have to admit, I was reluctant to believe my new motor was bad, out of at least a hundred of the MTH motors used, this is the first bad one out of the bag.

John, l have had motors, can motors that have had some metal debris inside, would hang on the magnets and contact the armature. Some times lock up and cause the demise of a motor driver board. Maybe similar to what you ran into, but not enough cause motor to lock up.

Ray

Something has to be happening electrically inside this motor, I'm going to take it apart, just curious what was happening.

Those little suckers are HARD to take apart without totally destroying them!  I didn't find anything that looked out of place inside, looks like a new motor.  Don't know what happened, chalk it up to one of the many electrical oddities.

John, FWIW, I recently saw this issue on a Protosound 2.0 engine too- bad smoke fan and even though the fan would test and be within normal current consumption - it probably has oil/smoke fluid contamination of the brushes and what that does is deposits the carbon conductive brush material into the slots shorting the armature segments. That short every 1/3 of a revolution (most small motors are 3 pole armature) causes a current spike that is not seen in the current averaging your meter is reading. Conversely, that spike in current given the small wiring and shared 5V power for the heater, the speed sensor (which is what is causing the chuff) drops the voltage of all 5V devices in the loco from the small gauge and shared wiring of the tether. Again, we have a situation where you have multiple devices we know to use 5V power and that is shared. Also, DC ground is shared and the speed sensor is often in this shared section of the circuit. When you have one of these motors with carbon brushes- and over time oil or other fluid contaminates- it causes the carbon conductive brush wear material to clump and of course, Murphy at play, it shorts the armature if even just partially. That makes the motor electrically noisier than usual and wreaks havoc on the shared 5V and ground electrical subsystem.

Bet if you put a scope on that motor it would show up.

I first ran across a similar electronics problem with a DC motor in 3D printers. What happened there is these were small DC geared motors, and shipped from China, the manufacturer ensured the entire batch was well lubed- wrapped in plastic. During that long journey- the oil seeped and got into the brushes. Then users would connect these motors and the driver was a small H-bridge with limited current but plenty of margin given the motor's worst case current draw. What happened was a rash of failures of the expensive electronics from the H-bridges blowing up left and right. People would bench test the motor and it looked fine on the bench. They replace the $90 controller board, attach the same now tested motor and a week later blow another $90 controller. What made this problem worse is the motor in the 3D printer often did starts and stops thousands if not millions of times during prints. If it stopped on that semi-shorted segment and attempted to start= instant blown H-Bridge, and that's how the community finally caught onto the problem.

What I would now say is- caution- do not overfill smoke units (not saying you did), and caution, I myself am guilty, blowing down the stack to clear the fluid meniscus that forms after just filling with smoke fluid can then blow smoke fluid back into the fan (I've seen people do it and obvious when smoke comes out the fan inlet hole AKA out from the boiler area when blowing down the stack). That fluid, given enough times and gravity, makes it into the smoke motor and since the brushes are in the bottom end- Being honest, I'm surprised this doesn't happen as a failure more often.

Also, being honest, surprised to not see the typical capacitor suppression of EMI/RFI from the brush terminals. I understand why not to case ground since case is rail ground via the body- not the DC internal ground reference so again, more wires, more complexity- but yes, to not see a small non-polarized cap across these motors- also has me scratching my head. Knowing what that lack of EMI/RFI clamping is wreaking havoc on the electrical system and shared no less with speed sensors- again, amazed this isn't a bigger problem.

Jetguy posted:

John, FWIW, I recently saw this issue on a Protosound 2.0 engine too- bad smoke fan and even though the fan would test and be within normal current consumption - it probably has oil/smoke fluid contamination of the brushes and what that does is deposits the carbon conductive brush material into the slots shorting the armature segments.

The motor was brand new, and I cracked it open and looked to see if anything looked out of place.  Clean as a whistle inside, and clearly no smoke contaminated.

Jetguy posted:

Also, being honest, surprised to not see the typical capacitor suppression of EMI/RFI from the brush terminals. I understand why not to case ground since case is rail ground via the body- not the DC internal ground reference so again, more wires, more complexity- but yes, to not see a small non-polarized cap across these motors- also has me scratching my head. Knowing what that lack of EMI/RFI clamping is wreaking havoc on the electrical system and shared no less with speed sensors- again, amazed this isn't a bigger problem.

 Not hard to explain, they are driven using PWM, and having a cap across there would drive the PWM nuts.

Could have been the screws for mounting too long.  There was a note way back that some motors manufactured with rotor differences and long screws could touch it grounding smoke fan motor.  G

I took the motor apart, couldn't see anything.  The screws used were pretty short, they actually seemed shorter than some I've seen in the fan motor mounts.  In any case, I mounted another new motor (from the same bag) and all was well.  I used the same screws.

I think Jetguy pegged it.  Put the suspect motor on a scope with a current probe.  Average current may be 40 mA, but  I'll bet you'll set an intermittent current spike caused by something shorting something else that drops the 5V supply line that powers the tach circuit.  Of course any glitches in the in the tach supply masquerade as moving tach stripes.  I think you got a bad motor.  I'd bet in production, they apply 5V (or whatever) to the motor and if it spins it gets shipped.

I thing something was contacting the frame of the motor as when I touched it to the chassis, I got a little spark, but only when it was running.  I just couldn't see what it was when I pulled it apart. In any case, a new motor fixed it, but that was an oddball that I hadn't seen before.

Long screws worked on certain motors and not others  But if short, probably not the issue.  Could have saved that motor for a PS-1 repair.  When I have had erratic operation that goes away with smoke off, I have typically found the static resistance of the motor lower then the normal 12 ohms.  Pretty sure it is a noise issue, not a current draw issue.  G

The odd part of this failure was there was no continuity to the shell from either brush, and I even put 12V to the motor and measured the current between the brush terminals and the motor shell, zero current.  However, when the motor was running, something had to be touching the case as I'd get a little spark when I touched the smoke unit body to the frame of the locomotive and the odd chuffing behavior would start.  I can only assume that when the motor was spinning that something was somehow hitting the case, and probably introducing the voltage spikes that caused the behavior.

I wouldn't risk using the motor in anything after seeing this, I ripped it apart, but I found no smoking gun.  A cheap fix, just had be going for a spell.

Thrust of the rotor can hit a long screw G

True, but I did actually look at the pieces, and the screws would not come close to anything in the can, they were pretty short.

I thought maybe a wire on the armature is flaky and as it spun it was arcing and causing the issue.  In truth, it's all guesswork, and a bunch more motors just arrived with my MTH order, so I'm set for a while.

I had the same experience a few years ago, switch the smoke on and the engine would run erratically. I replaced the fan motor and all was well. I have no idea what was wrong with the motor that I replaced.

Eric

TCA, LCCA, MTH Tech